Proceedings of the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst

Citation
Proceedings of the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst

Material Information

Title:
Proceedings of the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst
Series Title:
NCKRI Symposia
Added title page title:
Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst
Alternate title:
NCKRI Symposium 5
Creator:
Land, Lewis
Doctor, Daniel H.
Stephenson, J. Brad
National Cave and Karst Research Institute
Publisher:
University of South Florida
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Regional Speleology ( local )
Genre:
Conference Proceeding
serial ( sobekcm )
Location:
United States
Vietnam
China
United Kingdom
Belgium
Georgia
Romania
Italy
Coordinates:
46 x 25

Notes

Abstract:
These proceedings represent the talks, posters, and symposia presented at the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, which took place in Rochester, Minnesota, October 5-9, 2015. This international conference series creates a better understanding of environmental issues and geohazards associated with karst environments. This 14th conference was jointly organized by NCKRI and the Minnesota Ground Water Association. The 611 page proceedings volume contains 69 peer-reviewed papers organized under the following headings: a) Upper Mississippi Valley Karst Aquifers, b) Karst Hydrology, c) Karst Geology, d) GIS Databases and Mapping of Karst Regions, e) Contamination of Karst Aquifers, f) Geophysical Exploration of Karst, g) Karst Management, Regulation, and Education, and h) Modeling of Karst Systems. Also available is the Program with Abstracts, which includes five abstracts not in the proceedings, and the two field trip guidebooks: Ordovician Karst of Southeast Minnesota and Living with Karst in Rochester, Minnesota.
General Note:
National Cave and Karst Research Institute Symposium 5 Sinkholes and the Engineering and Environmental Impacts of Karst PROCEEDINGS OF THE FOURTEENTH MULTIDISCIPLINARY CONFERENCE October 5 through 9, 2015 Rochester, Minnesota EDITORS:​ Daniel H. Doctor, United States Geological Survey Reston, Virginia, USA Lewis Land, National Cave and Karst Research Institute Carlsbad, New Mexico, USA J. Brad Stephenson, CBI Federal Services Knoxville, Tennessee, USA
Restriction:
Open Access - Permission by Publisher
General Note:
See Extended description for more information.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-03300 ( USFLDC DOI )
k26.3300 ( USFLDC Handle )
21858 ( karstportal - original NodeID )

USFLDC Membership

Aggregations:
Added automatically
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item has the following downloads:

The 14th Sinkhole Conference Program with Abstracts ( .pdf )

A Comparative Study Between the Karst of Hoa Quang, Cao Bang Province, Vietnam and Tuscumbia, Alabama, USA / Gheorghe M. Ponta, Nguyen Xuan Nam, Ferenc L Forray, Florentin Stoiciu, Viorel Badalita, Lenuta J. Enache and Ioan A. Tudor ( .pdf )

A Method of Mapping Sinkhole Susceptibility Using a Geographic Information System: A Case Study for Interstates in the Karst Counties of Virginia / Alexandra L. Todd and Lindsay Ivey-Burden ( .pdf )

A Proposed Hypogenic Origin of Iron Ore Deposits in Southeast Minnesota Karst / E. Calvin Alexander, Jr. and Betty J. Wheeler ( .pdf )

A Semi-Automated Tool for Reducing the Creation of False Closed Depressions from a Filled LiDAR-Derived Digital Elevation Model / John Wall, Daniel H. Doctor and Silvia Terziotti ( .pdf )

Accounting for Anomalous Hydraulic Respones During Constant-Rate Pumping Tests in the Prairie Du Chien-Jordan Aquifer System - Towards a More Accurate Assessment of Leakage / Justin Blum ( .pdf )

Building Codes to Minimize Cover-Collapses in Sinkhole-Prone Areas / George Veni, Connie Campbell Brashear and Andrew Glasbrenner ( .pdf )

Cars and Karst: Investigating the National Corvette Museum Sinkhole / Jason S. Polk, Leslie A. North, Ric Federico, Brian Ham, Dan Nedvidek, Kegan McClanahan, Pat Kambesis, Michael J. Marasa and Hayward Baker ( .pdf )

Case Studies of Animal Feedlots on Karst in Olmsted County, Minnesota / Martin Larsen ( .pdf )

Characterization of Karst Terrain Using Geophysical Methods Based on Sinkhole Analysis: A Case Study of the Anina Karstic Region (Banat Mountains, Romania) / Laurentiu Artugyan, Adrian C. Ardelean and Petru Urdea ( .pdf )

Concepts for Geotechnical Investigation in Karst / Joseph A. Fischer and Joseph J. Fischer ( .pdf )

Conduit Flow in the Cambrian Lone Rock Formation, Southeast Minnesota, USA / John D. Barry, Jeffrey A. Green and Julia R. Steenberg ( .pdf )

Creation of a Map of Paleozoic Bedrock Springsheds in Southeast Minnesota / Jeffrey A. Green and E. Calvin Alexander, Jr ( .pdf )

Extended description ( .htm )

Detection of Voids in Karst Terrain with Full Waveform Tomography / Khiem T. Tran, Michael McVay and Trung Dung Nguyen ( .pdf )

Determination of the Relationship of Nitrate to Discharge and Flow Systems in North Florida Springs / Sam B. Upchurch ( .pdf )

Development of Cavity Probability Map For Abu Dhabi Municipality Using GIS and Decision Tree Modeling / Yongli Gao, Raghav Ramanathan, Bulent Hatigoplu, M. Melih Demirkan, Mazen Elias Adib, Juan J. Gutierrez, Hesham El Ganainy and Daniel Barton Jr ( .pdf )

Down the Rabbit Hole: Identifying Physical Causes of Sinkhole Formation in the UK / Tamsin Green ( .pdf )

Driftless Area Karst of Northwestern Illinois and its Effects on Groundwater Quality / Samuel V. Panno and Walton R. Kelly ( .pdf )

Dye Tracing Through the Vadose Zone Above Wind Cave, Custer County, South Dakota / James Nepstad ( .pdf )

Evaluation of Cavity Distribution Using Point-Pattern Analysis / Raghav Ramanathan, Yongli Gao, M. Melih Demirkan, Bulent Hatipoglu, Mazen Elias Adib, Michael Rosenmeier, Juan J. Gutierrez, and Hesham El Ganainy ( .pdf )

Evaluation of First Order Error Induced by Conservative-Tracer Temperature Approximation for Mixing in Karstic Flow / Philippe Machetel and David A. Yuen ( .pdf )

Evaluation of Veterinary Pharmaceuticals and Iodine for Use as a Groundwater Tracer in Hydrologic Investigation of Contamination Related to Dairy Cattle Operations ( .pdf )

Evaporite Geo-Hazard in the Sauris Area (Friuli Venezia Giulia Region - NE Italy) / Chiara Calligaris, Stefano Devoto, Luca Zini and Franco Cucchi ( .pdf )

Experimental and Numerical Investigation of Sinkhole Development and Collapse in Central Florida / Xiaohu Tao, Ming Ye, Xiaoming Wang, Dangliang Wang, Roger Pacheco Castro, JIan Zhao ( .pdf )

Ordovician Karst of Southeast Minnesota Field Trip Guidebook / Calvin Alexander, Jeff Green, Tony Runkel and John Barry ( .pdf )

Living with Karst in Rochester, MN / Jeffrey S. Broberg ( .pdf )

Finding Springs in the File Cabinet / Mason Johnson and Ashley Ignatius ( .pdf )

Using Nitrate, Chloride, Sodium, and Sulfate to Calculate Groundwater Age / Kimm Crawford and Terry Lee ( .pdf )

GOLIATH’S CAVE, MINNESOTA: EPIGENIC MODIFICATION AND EXTENSION OF PREEXISTING HYPOGENIC CONDUITS ( .pdf )

HAZARD OF SINKHOLE FLOODING TO A CAVE HOMININ SITE AND ITS CONTROL COUNTERMEASURES IN A TOWER KARST AREA, SOUTH CHINA ( .pdf )

History and Future of the Minnesota Karst Feature Database ( .pdf )

Human_Impacts on Water Quality in Coldwater Spring Minneapolis ( .pdf )

Hydrochemical Characteristics and Formation Mechanism of Groundwa ( .pdf )

Hydrocompaction Considerations in Sinkhole Investigations ( .pdf )

Hydrogeological dynamic variability in the Lomme Karst System (Be ( .pdf )

Hydrologic and geochemical dynamics_of vadose zone recharge in a ( .pdf )

Hydrological and Hydrogeological Characteristics of the Platform ( .pdf )

Integration and Delivery of Interferometric Synthetic Aperture Ra ( .pdf )

Introduction and Table of Contents ( .pdf )

INVESTIGATION OF A SINKHOLE IN OGLE COUNTY NORTHWESTERN ILLINOIS ( .pdf )

Karst Hydrogeologic Investigation of Trout Brook ( .pdf )

Karst Influence in the Creation of a PFC Megaplume ( .pdf )

Karst paleo-collapses and their impacts on mining and the environ ( .pdf )

KARST SPRING CUTOFFS CAVE TIERS AND SINKING STREAM BASINS_CORRE ( .pdf )

LEGACY DATA IN THE MINNESOTA SPRING INVENTORY ( .pdf )

LEPT a simplified approach for karst assessing vulnerability in ( .pdf )

Media Sinkholes and the UK National Karst Database ( .pdf )

Monitoring the Threat of Sinkhole Formation Under a Portion of US ( .pdf )

New_Methodologies and Approaches for Mapping Forested Karst Lands ( .pdf )

Numerical Simulation of Karst Soil Cave Evolution ( .pdf )

Numerical Simulation of Spring Hydrograph Recession Curves for Ea ( .pdf )

Pre-Contruction Rock Treatment and Soil Modification Program ( .pdf )

Predicting Compaction Grout Quantities in Sinkhole Remediation ( .pdf )

Pre-event and post-formation ground movement associated with the ( .pdf )

Recharge Area of Selected Large Springs in the Ozarks ( .pdf )

RELAY RAMP STRUCTURES AND THEIR INFLUENCE ON GROUNDWATER FLOW IN ( .pdf )

Rollalong resistivity surveys reveal karstic paleotopography deve ( .pdf )

Seeps and Springs at a Platteville-Observatory-on the River Blu ( .pdf )

SHALLOW DEPRESSIONS IN THE FLORIDA COASTAL PLAIN-KARST AND PSEUD ( .pdf )

Sinkhole Physical Models to Simulate and Investigate Sinkhole Col ( .pdf )

SINKHOLE VULNERABILITY MAPPING-RESULTS FROM A PILOT STUDY IN NOR ( .pdf )

Spatiotemporal Response of CVOC Contamination and Remedial Action ( .pdf )

STUDY ON MONITORING AND EARLY WARNING OF KARST COLLAPSE BASED ON ( .pdf )

Study on the critical velocity of groundwater ( .pdf )

Successful Foundation Preparations in Karst Bedrock of the Masonr ( .pdf )

The application of passive seismic techniques to the detection of ( .pdf )

THE COST OF KARST SUBSIDENCE AND SINKHOLE COLLAPSE IN THE UNITED ( .pdf )

THE MILLION DOLLAR QUESTION-WHICH GEOPHYSICAL METHODS LOCATE CAV ( .pdf )

The Sandstone Karst of Pine County Minnesota ( .pdf )

TRACER STUDIES CONDUCTED NEARLY TWO DECADES APART ELUCIDATE GROUN ( .pdf )

Tracking of karst contamination using digital mapping technologie ( .pdf )

Using Electrical Resistivity Imaging to Characterize Karst Hazard ( .pdf )

Proceedings of the 2015 Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst (complete issue) ( .pdf )


Full Text

PAGE 1

The 14th Sinkhole Conference Program with AbstractsOctober 5-9, 2015 Rochester Minnesota, USA Integrating Science and Engineering to Solve Karst Problems THE SINKHOLE C ONFERENCEM u l tidiscip l i n a r y C o n f e r e n c e o n S i n k holes a n d t h e E ngineering and E n vir onmen tal Impac ts of K arst O dessa Spr ing Fillmore County, MN. Photo by Andrew Peters. P assagew a y of N iagar a C a v e C our t esy of M ar k Bishop Channel R ock C a v er n, under east er n M inneapolis MN. P hot o b y Johns L o v aas Cover-collapse sinkhole in the pool area of a water-retention structure, Goodhue County, MN. Photo by Stephen Fay. B edr ock G eology of M innesota, M innesota G eolog ical Sur v ey Rochester

PAGE 2

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 1 O ct ober 5 9, 2015 Rochester, Minnesota, USA Table of ContentsSponsors .................................................................................................................................................................................................... 2 Welcome Letter s ................................ ...................................................................................................................................................... 4 Organizing Committee ................................ ............................................................................................................................................. 6 History of Conference ................................ .............................................................................................................................................. 8 Beck Scholarship ................................ ....................................................................................................................................................... 9 Host City Ro chester, Minnesota ........................................................................................................................................................... 10 Field Trip s ................................ ............................................................................................................................................................... 13 Short Courses ................................ .......................................................................................................................................................... 15 Program at -a -Glance .............................................................................................................................................................................. 17 Detailed Program ................................ ................................................................................................................................................... 18 Invited and Keynote Speakers ................................ ............................................................................................................................... 25 Presentation a nd Poster Abstracts ........................................................................................................................................................ 27 Cr edits: Program with Abstract s prepared by Brian A. Smith, Brian B. Hunt and Justin P. Camp, Barton Springs/Edwards Aquifer Conservation District, Austin, Texas. Cover photograph: Stagecoach Spring, the entrance to the Stagecoach Caver n s cave system and the headwater of Watson Creek, a state de signated trout stream. Photo by Jeff Green Release: By submitting the registration form, you hereby release any photographs that may be incidentally taken of you during these events by Sinkhole Conference 2015 to be used for any purpose. Wa iver: By registering, you agree and acknowledge that you are participating in the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst (Sinkhole Conference 2015) and its activities intentionally and of your own free will, and you are fully aware that possible physical injury might occur to you as a result of your participation. You give this acknowledgement freely and knowingly that you are, as a result, able to participate in Sinkhole Conference 2015, and you hereby assume responsibili ty for your own well being. Rec ording of Presentations : The recording of any oral or poster presentation is prohibited without the prior approval of the author.

PAGE 3

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 2 Sponsors Rochester Public Utilities (RPU) has been the municipal utility of Rochester, Mi nnesota f or more than 120 years. RPU serves over 50,000 electric customers and over 37,000 water customers in Rochester. With a focus on customer service and efficiency in operations, we continue to strive for excellence through effective water quality pro grams, robust infrastructure, and customer education efforts that will ensure a safe and reliable drinking water supply for years to come. Rochesters water supply comes from the Jordan aquifer, in which RPU has 32 deep wells, with another well planned to go in service by the end of 2016. Additionally, RPU works with customers to lower usage of electricity and water by offering incentives for qualifying conservation and efficiency equipment purchases through RPUs Conserve & Save rebate program. In 2014 RPU helped customers save over 9 million gallons of water through these efforts. Additional Water Highlights : There are 20 water storage facilities in Rochester. RPU was voted by the AWWA Southeast Minnesota Chapter Best Tasting Drinking Water in 2014. RPU developed and installed a number of water conservation and water quality interactive exhibits at St. Marys UniversityCascade Meadow Wetlands & Environmental Science Center. RPU is also a significant employer in Rochester, employing over 160 f ull time employees. For more information on RPU visit RPU at www.rpu.org Braun Intertec assists public and private organizations and property owners with site evaluations, site preparation recommendations, and construction support services. With more than 800 employees corporate wide across 16 offices, Braun Intertec staff represents multipl e technical disciplines including karst investigations and engineering, environmental consulting, geotechnical engineering, testing, special inspections, geothermal consulting and facilities evaluations. Advanced Geosciences is the manufacturer of t he SuperSting, geophysical resistivity imaging system. It is used for investigations in geotechnical applications such as; detecting cavities and sinkholes, mapping of bedrock topography, mapping of soil profile, moisture content is clay, locating leakage through dams, environmental work and checking soil resistance for corrosion studies.

PAGE 4

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 3 GeoTDR applies Time Domain Reflectometry (TDR) technology throughout the world for automated remote monitoring of subsidence over active and abandoned mines as well as monitoring sinkhole subsidence in karst areas. GeoTDR is a subsidiary of Geotechnical Consulta nts Inc, which provides geotechnical engineering, environmental services, and consultation materials, engineering and testing. The Edwards Aquifer Authority is a groundwater district, mandated by the 1993 Edwards Aquifer Authority Act. The Act gran ts all of the powers, rights, and privileges necessary to manage, conserve, preserve, and protect the aquifer. The EAA regulates the portion of the Balcones Fault Zone Edwards Aquifer a jurisdictional area that provides water to over 2 million people, an d covers more than 8,000 square miles across eight Texas counties. The Barton Springs/Edwards Aquifer Conservation District is a groundwater district in central Texas. The agency has the mission to conserve, protect, recharge, and prevent the waste of groundwater and preserve all aquifers in the District. The District man ages well drilling and groundwater production from the Barton Springs segment of the Edwards Aquifer and the underlying Trinity Aquifer two major karst aquifers in Texas. Cover collapse sinkhole in the Edwards Formation within a stormwater retention pond, Austin Texas. Left photo taken soon after collapse, right photo showing rock fill and concrete plug. Photo by BSEACD.

PAGE 5

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 4 5 October 2015 Welcome Karst Engineers and Scientists! We are delighted to be your hosts for the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst. For the past 31 years, this series of meetings has been among the most important in developing a better understanding of karst processes that result in envir onmental problems, and in creating effective measures that identify those problems before they occur, remediate them when they occur, and prevent them from occurring in the first place. This years conference is the first in the series that is jointly org anized with another conference and organization. We are happy to be working with the Minnesota Ground Water Association (MGWA ) and to combine our convention with the MGWA Fall Conference. We hope this merging of conferences will make this the biggest Sinkh ole Conference and greatest exchange of knowledge to dat e. We are also delighted to award a record four Barry Beck Sinkhole Conference Student Scholars this year each from a different country. We thank the generous sponsors who made that possible. This y ears Sinkhole Conference offers an excellent series of papers, thought provoking keynote addresses, two fascinating and fun field trip s and ample time for you to meet new and old friends to discuss and collaborate on karst engineering and environmental research projects. Dont forget to visit the booths of our generous exhibitors and sponsors and support them for supporting the Sinkhole Conference! If you have any questions or concerns about the meeting, please tell us directly or leave a message at t he registration desk and we will address them as soon as possible. We look forward to visiting with you soon. Sincerely, George Veni James W. LaMoreaux Conference Co chairman Conference Co chairman Executive Director President National Cave & Karst Research Institute PELA GeoEnvironmental Kelton Barr Conference Co chairman Past President Minnesota Ground Water Association

PAGE 6

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 5 5 October 2015 Welcome to Minnesota! On behalf of the Minnesota Ground Water Association, I welcome you to Minnesotas karst country. We are proud to cohost this important conference in a part of the world where karst is a bit secretive, but often at the root of things (literally). Our karst country is mostly obscured beneath a thick layer of plant roots, sometimes with a nice layer of glacial drift to keeps things interesting. This conference uncovers the deep inner workings of karst landscapes, and it helps to hone the many skills n eeded to research and manage them. The purpose of our Association is to promote public policy and scientif ic education about groundwater This 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst serves our objectives very well: Promotes and encourages efforts that address the scientific and public policy aspects of groundwater ; Establishes a forum for scientists, engineers, planners, educators, attorneys, and others concerned with groundwater Educates the general public regarding groundwater ; Shares information on groundwater through meetings of our membership. This is the first time that the Minnesota Ground Water Association has partnered with the National Cave and Karst Research Institute, and this conference is the satisfying result. The outstanding presenters, short courses, field trips, and social events are a testament to the immense dedication and cooperation of many people working together across the globe. If you find yourself with the opport unity, I encourage you to offer up a hearty thank you to our Organizing Committee members and volunteers who generously support this work. I encourage you to enjoy yourselves, share your expertise across disciplines, and learn something new. And have fun! Sincerely, Lanya Ross President Minnesota Grou nd Water Association

PAGE 7

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 6 Organizing Committee General Conference Co Chairs George Veni, Ph.D., P.G., National Cave and Karst Research Institute (NCKRI), Carlsbad, NM Kelton Barr, P.G., Braun Intertec, Minneapolis, MN Jim LaMoreaux, Ph.D., P.E., PELA GeoEnvironmental Tuscaloosa, AL Program Co Chairs Lynn B. Yuhr, P.G., Technos, Inc., Miami, FL Michael Alfieri, P.G., Water Resource Associates, Tampa, FL Audrey Van Cleve, Hydrologist, Minnesota Pollution Control Agency, retired, Minneapolis, MN Proceedings Managing, Assistant and Copy Editors Daniel H. Doctor, Ph.D., U.S. Geological Survey, Eastern Geology & Paleoclimate Science Center, Reston, VA Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and National C ave and Karst Research Institute, Carlsbad, NM J. Brad Stephenson, P.G., L.R.S., CB&I Federal Services Knoxville, TN Rebel Cummings Sauls, Kansas State University Julie Fielding, U niversity of Michigan Proceedings Associate Editors Gregory Brick, Ph.D., Minnesota Department of Natural Resources, St. Paul, MN James Kaufmann, U.S. Geological Survey, Rolla, MO Mustafa Saribudak, Ph.D., P.G., Environmental Geophysics Associates, Austin, TX Samuel V. Panno, CGWP, Illinois State Geological Survey, Prairie Res earch Institute, University of Illinois, Champaign, IL Jason Polk, Ph.D., Western Kentucky University, Bowling Green, KY David J. Weary, U.S. Geological Survey, Reston, VA Ming Ye, Ph.D., Florida State University, Tallahassee, FL Lynn B. Yuhr, P.G., Technos, Inc. Miami, FL Field Trips E. Calvin Alexander, Jr., Ph.D., Department of Earth Sciences, University of Minnesota, Minneapolis, MN Jeff Green, LPG, Minnesota Department of Natural Resources, Rochester, MN Jeffrey Broberg. LPG, WSB Associates, Inc ., Rochester, MN Short Courses Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and National Cave and Karst Research Institute, Carlsbad, New Mexico Joe Fischer, Ph.D., P.E., Geoscience Services, Clinton, NJ Invited Speakers Yongli Gao Ph.D., University of Texas San Antonio, San Antonio, TX Logo Samuel V. Panno, CGWP, Illinois State Geological Survey, Prairie Research Institute, University of Illinois, Champaign, IL Public Relations La n ya Ross, President, Minnesota Ground Water Association George Veni, Ph.D., P.G., National Cave and Karst Research Institute, Carlsbad, NM

PAGE 8

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 7 Beck Scholarship E. Calvin Alexander, Jr., Ph.D., Department of Earth Sciences, University of Minnesota, Minneapolis, MN Ira D. Sasowsky, Ph.D., P.G., Dept of Geosciences University of Akron, Akron, OH Beck Scholarship Silent Auction John Barry, Minnesota Dept. of Natural Resources, St. Paul Minnesota Circulars and Publicity Samuel V. Panno, CGWP, Illinois State Geological Survey, Prairie Research Inst itute, University of Illinois, Champaign, IL Program with Abstracts Brian A. Smith, Ph.D., P.G., Barton Springs/Edwards Aquifer Conservation District, Austin, TX Brian B. Hunt, P.G., Barton Springs/Edwards Aquifer Conservation District, Austin, TX Justin P. Camp, Barton Springs/Edwards Aquifer Conservation District, Austin, TX Website Gheorghe Ponta, P.G., P.E. Geological Survey of Alabama, Tuscaloosa, AL Sean Hunt, Hydrologist, Minnesota Dept. of Natural Resources, St. Paul MN Registration Sean Hunt, Hydrologist, Minnesota Dept. of Natural Resources, St. Paul MN Audrey Van Cleve, Hydrologist, Minnesota Pollution Control Agency, retired, Minneapolis, MN Hotel/ Conference facilities Audrey Van Cleve, Hydrologist, Minnesota Pollution Control Agency, retired, Minneapolis, MN Kelton Barr, P.G., Braun Intertec, Minneapolis, MN Mindy L. Erickson, Ph.D., P.E., Hydrologist, U.S. Geological Survey, Minnesota Water Science Center Mounds View, MN Treasurer Jeanette Leete, Minnesota Department of Natural Resources, St. Paul MN Professional Organization Liaisons Lanya Ross Minnesota G round W ater A ssociation St. Paul, MN Wanfang Zhou, Ph.D., P.G. ERT Inc. International contacts Members at Large Scott Alexander Universi ty of Minnesota, Minneapolis, MN Phil Carpenter Dept. of Geology and Environmental Geosciences, Northern Illinois University Ralph Ewers Ewers Water Consultants Inc., Richmond, KY Bashir Memon PELA GeoEnvironmental Tuscaloosa, AL Deana Sneyd, Golder Associates, Atlanta, GA

PAGE 9

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 8 History of C onference The roots of the Sinkhole Conference can probably be traced back to when Barry Beck became interested in caves as a teenager where he lived in Rochester, New York. His interest in caves and karst studies eventually led him to Texas, Mexico, Puerto Rico, G eorgia, and Florida for extended periods, plus visits to many other karst areas around the world. Barry organized the First Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst which was held in Orlando, Florida Oct ober 15 17, 1984. Subsequently, second and third conference s were held under the sponsorship of the Florida Sinkhole Research Institute, a division of the University of Central Flori da in Orlando, in 1987 and 1989, which he directed. These conferences were established to meet a critical need for applied research information on the very complex hydrogeological environment of karst areas of the world. In 1992, following the closing of the Florida Sinkhole Research Institute, Barry joined the staff of P. E. LaMor eaux & Associates, Inc. (PELA) and opened the company's Oak Ridge, Tennessee office. Beginning with the Fourth Multidisciplinary Conference in 1993, PELA sponsored t he continuation of this important series of conferences along with many other disti nguished organizations. The Geo Institute of the American Society of Civil Engineers took the lead in sponsoring the conference in 2003, 2005, and 2008 after which P ELA took over the sponsor ship again for the 2011 conference. Since the 2011 conference, thi s conference series has come under the management of the National Cave and Karst Research Institute (NCKRI). As a government established non profit organization NCKRI is focused on karst phenomena and organized in part to conduct and support such conferen ces. PE LaMoreaux & Associate s, Inc., as cosponsor, and the Organizing C ommittee remain an integral part of the conference. This years Sinkhole Conference is co organized with the Minnesota Ground Water Association. The proceedings of these conferences h ave been valuable additions to karst libraries around the world. Bel ow is a list of the proceedings from the beginning conference to 2013 Previous Conferences and Proceedings No. Proceedings Year Location 1st Sinkholes: Their Geology, Engineering and Environmental Impact 1984 Orlando, FL 2nd Karst Hydrogeology: Engineering and Environmental Applications 1987 Orlando, FL 3rd Engineering and Environmental Impacts of Sinkholes and Karst 1989 St. Petersburg Beach, FL 4th Applied Karst Geology 1993 Panama City, FL 5th Karst GeoHazards: Engineering and Environmental Problems in Karst Terranes 1995 Gatlinburg, TN 6th The Engineering Geology and Hydrogeology of Karst Terranes 1997 Springfield, MO 7th Hydrogeology and Engineering Geology of Sinkholes and Karst 1999 Harrisburg/Hershey, PA 8th Geotechnical and Environmental Applications of Karst Geology and Hydrology 2001 Louisville, KY 9th ASCE Geotechnical Special Publication No. 122 2003 Huntsville, AL 10th ASCE Geotechnical Special Publication No. 144 2005 San Antonio, TX 11th ASCE Geotechnical Special Publication No. 183 2008 Tallahassee, FL 12th Carbonates and Evaporites volume 27, nos. 2 3 2011 St. Louis, MO 13th NCKRI Symposium 2: 13th Multidisciplinary Conference on Sinkholes and Karst 2013 Carlsbad, NM

PAGE 10

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 9 Barry F. Beck Sinkhole Conference Student Scholarship The Barry F. Beck Sinkhole Student Scholarship (Beck Scholarship) is a competitive grant awarded to one or more students who presents the results of their research at the bi annual Sinkhole Conference. The award is being inaugurated at this years conference in memory of the late Dr. Barry Beck, a pioneer in the scientific study of sinkholes who founded the Sinkhole Conference. At least one Beck Scholarship will be awarded for each c onference. Additional scholarships may be awarded if funded from donors for the conference. Beck Scholars receive: 1. One free Sinkhole C onference registration. 2. One free registration to a field trip and short course (pending space availability). 3. An award ce rtificate. 4. Recognition through name badge ribbon, mention in the Sinkhole Conference program and website, and announcements at the opening ceremony and banquet. 5. Reimbursement for up to $1000 of person al individual travel, food, and lodging expenses associa ted with attending the Sinkhole Conference. For more information about the Beck Scholarship and to fund future scholars, watch for future Sinkhole Conference websites or visit the National Cave and Karst Research Institute website at https://support.nckri.org/barrybeck scholarship and contact info@nckri.org or by calling 5758875518. B arry F. Beck 2015 Sinkhole Conference Student Sch olarship Recipients Laurentiu Artugyan West University of Timisoara Romania John Wall North Carolina State University USA Tamsin Brittany Green University of Leeds UK Caren Raedts University of Western Ontario Canada

PAGE 11

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 10 Host City : Ro chester, MN Roches ter welcomes the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst! Rochester combines the best aspects of an innovative and cosmopolitan city with the charm of a small town environment. Named "Best Small City" in America by nationally recognized magazines, Rochester merges a multicultural atmosphere with midwestern hospitality, giving convention attendees a most memorable experience. Rochester will pleasantly surprise you with the new developments and amenities downtown and throug hout and around the city. Start experiencing Rochester by strolling along the dynamic Peace Plaza, around Peace Fountain and into The Shops at University Square. We offer traditional and new restaurants, fine shops and food courts all connected b y our convenient climate controlled pedestrian skyway/subway system downtown. Conference Location: Mayo Civic Center 30 Civic Center Drive SE, Rochester, MN 55904 Mayo Civic Center is Southern Minnesota's premier destination for local, regional, national and international conventions, entertainment, social an d sporting opportunities. Designed as a dual venue complex to provide much needed entertainment for the citizens of a growing community, the Mayo Civic Center first opened its doors in 1938. The 1700 seat arena was designed to promote athletic events and included a rink for skating and ice hockey. The 1340 seat theatre was intended as a showcase for fine arts and was designed to accommodate midsized theatrical and musical productions (Source: mayociviccenter.com) During this years Sinkhole Conference, the Civic Center will undergo another major renovation and expansion. We appreciate your patience and understanding if some details in this program or announced during the conference must change due to construction activities. Map of downtown Rochester with the conference location indicated by the star.

PAGE 12

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 11 Mayo Civic Center Map

PAGE 13

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 12 Map of conference location and venues. Image courtesy of Google Earth.

PAGE 14

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 13 Field Trip s Main Field Trip: Karst Features in the Ordovician Rocks of Southeast Minnesota Tuesday, 6 October, 7:30 am 6:00 pm Trip Co Leaders: Jeff Green E. Calvin Alexander Jr., Tony Runkel and John Barry Assemble and board buses at the Rochester Civic Center at 7:00 to 7:20. Buses will depart promptly at 7:30 AM On Tuesday we offer a full day tour of sinkholes, stream sinks, springs, blind valleys, caves and other karst features developed in the glacio fluvial landscape of Olmst ed and Fillmore counties in sout heast Minnesota. Although often referred to as the Driftless Area of southeast Minnesota, the area has been glaciated several times during the Pleistocene but has not been covered with ice in the last few hundred thousand years. The P leistocene process es are relatively recent perturbation on the karst processes, which have been shaping the surface and subsurface of this area ever since the rocks were deposited 400 to 500 million years ago. The ma jor human activity in southeast Minnesota is modern intens ive row crop agriculture (corn and soybeans) and confined animal operations. These human activities directly affect the karst landscapes and the underlying karst aquifers. The karst processes in turn di rectly affect and limit modern agricultural practices We will visit on of two commercial cave s Niagara Cave, near Harmony, Minnesota or Mystery Cave in Forestville/Mystery Cave State Park near Spring Valley Minnesota. The cave temperature is about 8 C and a light jacket is recommended. We will also be hiki ng to severa l stops across uneven terrain. Hats, sturdy shoes, field clothing and cameras are recommended. Post Conference Field Trip: Karst in our Built Environment: the Rochester Area Friday, 9 October, 12:00 noon 4:00 pm Trip Leader: Jeff Broberg As semble and board buses at the Rochester Civic Center at 11:45 am to 12:00 noon. Buses will depart promptly at 12:00 noon Lunch will be provided. This field trip will present the karst features and sinkhole challenges in and around Rochester and will orient participants to the karst geology and hazards, land use controls, and engineering solutions in this rapidly developing community. A number of sites will be visited, including: A sinkhole plain developed as a high value residential subdivision in a township with a sinkhole ordinance where developers are required to identify sinkholes on official maps, plats and site plans and where some owners have preserved sinkholes, while others have sealed them. A sinkhole plain with dozens of sinkholes surroundi ng a large beef cattle feedlot An exposed sandstone breccia pipe next to a theater complex where subterranean karst on an Ordovician conformity has resulted in the collapse of overlying sandstone The restored site and bedrock cores of a major sinkhole collapse in a blocked drainage way that threatened a major highway and commercial property The site and viewing of bedrock cores where a Pleistocene age sinkhole at the base of the St. Peter Sandstone was mitigated to build a 9 story hospital building A bui lding stone quarry where saw cut rock faces dissect solution enlarged joint planes and paleo karst surfaces.

PAGE 15

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 14 Caving Trips at the 14th Sinkhole Conference Mr. John Ackerman is allowing wild caving trips in his Minnesota Cave Preserve by 14th Sinkhole Conference participants. Preferential access will be given to out of state and international participants. See http://www.cavepreserve.com/ for information about the individual caves and the Minnesota Cave Preserve. Spring Valley Caverns, Goliaths Cave, Holy Grail Cave and Tysons Spring Cave will be available for potential trips. These trips are being coordinated by Calvin Alexander independent of the 14th Sinkhole Conference. The White n ose Syndrom e (WNS) that is devastating bat populations in the eastern and central U.S. has not yet reached Minnesota. Our sincere hope is that WNS does not reach Minnesota. All cavers will be required to follow the USFS decontamination guideli ne for all caving gear see https://www.whitenosesyndrome.org/topics/decontamination for the most up to date protocols. Each trip will be restricted to 5 (five) cavers plus a guide. None of the trips req uire vertical gear but most cave s are entered through 30 inch (0.76 m) diameter vertical shafts with permanently installed ladders. Trips to Spring Valley Caverns, Goliaths Cave and Holy Grail are relatively dry. Trips to Tyson Spring Cave will require thick wet suits. Trips to Tyson Spring Cave do not require SCUBA gear, i.e. do not involve diving, but Tysons is a river cave that involves wading, crawling and some swimming in water that will b e about 8 C (47 F) and require thick wet suits to avoid hypot hermia. Depending on demand, trips may be organized on Sunday 4 October 2015, Monday evening 5 October 2015, Friday afternoon 9 October 2015 and or Saturday 10 October 2015. There will be a $25 per trip charge mainly to cover transportation from the Ro chester Civic Center to each cave and return. The caves are about 25 to 45 miles south of Rochester, Minnesota. Kings and Queens Bluff overlooking Mississippi River Valley, Great River Bluffs State Park, MN. Photo by MNDNR.

PAGE 16

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 15 Short Courses (Monday, October 5, 2015) F our courses are offered to expand your knowledge of practical applications of karst science. Space is limited and registrations are processed on a first come, first served basis. Onsite registrations can only be accepted on a space available basis. Short Course 1: Geologic Site Characterization with Emphasis on Karst Instructor: Lynn Yuhr, Technos, Inc. Location: McDonnell Suite Mayo Civic Center A geologic site characterization is the technical foundation for all geotechnical and environmental projects. The objective o f a geologic site characterization is to gain an understanding of subsurface conditions that will impact the engineering or environmental decisions made at a site. This effort can be fairly straightforward in a uniform geologic setting. However, karst often pr ovides some of the most variable and diverse geologic settings to deal with. A wide range of topics will be covered to include a discussion of the problem, a strategy using an integrated systems approac h of appropriate methods, using adequate levels of s ite coverage and considering the impact of scale. Case histories will be presented to illustrate the process. If the geologic site characterization is done right, subsequent geotechnical and environmental decisi ons can be made with a high degree of confidence and be supported by reliable technical data. This course is based upon many decades of experience, which is presented in the book Site Characterization in Karst and PseudoKarst Terrains written by Richard C. Benson and Lynn B. Yuhr being published by Springer in 2015. Lynn Yuhr has specialized in site characterization with an emphasis on karst for more than three decades. Short Course 2: Groundwater Tracing as a Hydrogeologic Tool in Karst and Other Landscapes Instructors: Shiloh Beeman, RG, Sr. Hydrogeologist, Ozark Underground Laboratory, Inc. Tom Aley, PG, PHG, President and Sr. Hydrogeologist, Ozark Underground Laboratory, Inc. Jeff Green, Springshed Mapping Hydrologist, Minnesota Department of Natural Resources Location: Legion Suite, Ma yo Civic Center Over the last 50 years, groundwater tracing has been developed as an important tool for the assessment of the hydrogeology of karst environments. In more recent years, it has also been applied successfully in many other subsurface environm ents for both environmental and engineering applications. However, groundwater tracing is still an often overlooked and underutilized tool for understanding subsurface hydrology. This short course focuses on the application of groundwater tracing through the examination of a variety of case histories. Case histories will include natural resource evaluations, environmental contaminant transport, highway construction, and leaking reservoirs. In addition, participants will receive an overview of dye tracing m ethodologies, commonly used tracer dyes, and sampling and analysis techniques.

PAGE 17

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 16 Short Course 3: Grouting in Karst Instructors: Joseph A. Fischer, P.E., Geoscience Services Michael J. Miluski, P.E., Compaction Grouting Services Location: McDonnell Suite, Mayo Civic Center There will be an overview of the differing karst environments and how they influence the most effective grouting procedures. Types of grout, procedures, equipment descriptions and examples will be discussed. Handouts of the course content will be provided. Also discussed will be the design and preparation of a grouting program. Instructors Joseph A. Fischer, P.E. of Geoscience Service s and Michael J. Miluski, P.E. of Compaction Grouting Services, Inc. have extensive experience in grouting in karst environme nts from both the geotechnical engineers and contractors viewpoints. Short Course 4: Minnesota Environmental Management Rules, Regulations and Permits for Southeast Minnesota Karst Landscapes and Karst Aquifers Presenters: George Schwint, Principal Engineer Feed Lots, Minnesota Pollution Control Agency Scot Johnson, MS, PG, Hydrologist 3, Minnesota Department of Natural Resources Daniel Stoddard, Assistant Director, Pesticide & Fertilizer Management Division, Minnesota Department of Agriculture Mode rator: E. Calvin Alexander Jr., University of Minnesota Location: Legion Suite, Mayo Civic Center This sh ort course will present an over view of the various rules, regulations and permits that are in place to help environmental managers deal with the challenges associated with southeast Minnesotas karst landscapes and karst aquifers. Rules pertaining to feed lots, spill sites, silica sand mining, landfills, pesticides, nutrients, etc. will be presented and discussed. Handouts will present URLs and con tact people for the various Minnesota karst rules, regulations and permit requirements and the basics of each will be discuss ed. The short course is meant to give Minnesota environmental managers an overview of the karst related regulations and to presen t enough information that environmental managers from other states and countries can compare to their own management tools.

PAGE 18

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 17 Program at a Glance

PAGE 19

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 18 Detailed Program Monday, October 5 7:30 AM Registration opens Continental breakfast in Civic Center Suites Lobby 8:30 AM Short C ourses 1 and 2 (McDonnell and Legion Suites) SC 1: Geologic Site Characterization with Emphasis in Karst (4 hrs); Instructor: Lynn B. Yuhr SC 2: Groundwater Tracing as a Hydrogeologic Tool in Karst and Other Landscapes (4 hrs); Instructors: Shiloh Beeman, Thomas Aley, and Jeff Green, 12:30 PM Lunch on your own 1:30 PM Short courses 3 and 4 (McDonnell and Legion Suites) SC 3: Grouting in Karst (4 hrs); Instructors: Joseph A. Fischer and Michael J. Miluski SC 4: Minnesota Environmental Management Rules, Regulations and Permits for Southeast Minnesota Karst Landscapes and Karst Aquifers (4 hrs) Tuesday, October 6 7:00 AM Registration opens (Civic Center Suites Lobby) 7:30 AM 5:30 PM Field Trip: Karst of Southeast Minnesota; Trip Co Leaders: Jeff Green, E. Calvin Alexander, and Tony Runkel Note: Board Buses on Center Street (north side of civic center ). Lunch will be provided. 6:30 PM Welcome reception Legion Suite Wednesday, October 7 7:30 AM Registration opens North Lobby 7:40 AM Continental breakfast in Exhibit Hall IIII 8:00 AM Introductory Remarks from George Veni (NCKRI) and La n y a Ross ( MGWA ) 8:10 AM Remarks from Mayor of Rochester, Ardell F. Brede 8:20 AM Remarks from Harvey Thorleifson, Minnesota Geologic Survey Plenary Session: Upper Mississippi Valley Karst Aquifers Greg Brick (chair); Exhibit Hall IV 9:10 AM Joel T Groten, US Geological Survey; E. Calvin Alexander, Jr., University of Minnesota Karst Hydrogeologic Investigation of Trout Brook 9:30 AM Sophie M Kasahara, Scott C Alexander and E Calvin Alexander, Jr., University of Minnesota Human Impacts on Water Quality in Coldwater Spring, Minneapolis, Minnesota 9:50 AM Daniel H Doctor, US Geological Survey; E. Calvin Alexander, Jr., University of Minnesota; Roy Jameson; Scott Alexander, University of Minnesota Hydrologic and Geochemical Dynamics of Vadose Zone Recharge in a Mantled Karst Aquifer: Results of Monitoring Drip Waters in Mystery Cave, Minnesota 10:15 AM Break in E xhibit H all IIII

PAGE 20

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 19 10:30 AM John D. Barry and Jeffrey A. Green, Minnesota Dept. of Natural Resources; Julia R Steenberg, Minnesota Geological Survey Conduit Flow in the Cambrian Lone Rock Formation, Southeast Minnesota, U.S.A. 10:50 AM Kimm Crawford, Crawford Environmental Services; Terry Lee, Olmsted County Environmental Resources Department Using Nitrate, Chloride, Sodium, and Sulfate to Calculate Groundwater Age 11:10 AM Garre A Conner, Pangea Geoservices Karst Spring Cutoffs, Cave Tiers, and Sinking Stream Basins Correlated to Fluvial Base Level Decline in South Central Indiana 11:30 AM Samuel V Panno, Illinois State Geological Survey and Dr. Walton R Kelly, Illinois State Water Survey Driftless Area Karst of Northwestern Illinois and its Effects on Groundwater Quality Plenary Session: Karst Hydrology : Sam Panno (chair) ; Exhibit Hall IV 11:50 AM Amal Poulain, Gatan Rochez and Vincent Hallet, University of Namur, Belgium Hydrogeological Dynamic Variability in the Lomme Karst System (Belgium) as Evidenced by Tracer Tests Results (KARAG project) 12:10 PM Catered Lunch in Exhibit Hall IIII 1:30 PM James W Duley Cecil Boswell and Jerry Prewett, Missouri Geological Survey Recharge Area of Selected Large Springs in the Ozarks 1:50 PM Zaza Lezhava, Nana Bolashvili, Kukuri Tsikarishvili, Lasha Asanidze and Nino Chikhradze, Javakhishvili Tbilisi State University Hydrological and Hydrogeological Characteristics of the Platform Karst (Zemo Imereti Plateau, Georgia) 2:10 PM Keith A White, ARCADIS; Thomas J Aley, Ozark Underground Laboratory; Michael K Cobb, ARCADIS; Ethan O Weikel, US Army Corps of Engineers; Shiloh L Beeman, Ozark Underground Laboratory Tracer Studies Conducted Nearly Two Decades Apart Elucidate Groundwater Movement Through a Karst Aquifer in the Frederick Valley of Maryland 2:30 PM James Nepstad, National Park Service Dye Tracing Through the Vadose Zone Above Wind Cave, Custer County, South Dakota 2:50 PM Gheorghe M. Ponta, Geological Survey of Alabama; Nguyen Xuan Nam, Vietnam Institute of Geoscience and Natural Resources; Ferenc L Forray, Babes Bolyai University; Florentin Stoiciu, Viorel Badalita, Lenuta J Enache and Ioan A Tudor, R&D National Institute for Nonferrous and Rare Metals A Comparative Study Between the Karst of Hoa Quang, Cao Bang Province, Vietnam and Tuscumbia, Alabama, USA 3:10 PM Break in E xhibit H all IIII Plenary Session: Karst Geology John Barry (Chair) ; Exhibit Hall IV 3:30 PM Gongyu Li, Xian Research Institute of China Coal Technology & Engineering Group Corp.; Wanfang Zhou, Ph.D., P.G. ERT, Inc. Karst Paleo Collapses and Their Impacts on Mining and the Environment in Northern China 3:50 PM Beverley Lynn Shade, University of Texas; E Calvin Alexander, Jr. and Scott C Alexander, University of Minnesota The Sandstone Karst of Pine County, Minnesota 4:10 PM E Calvin Alexander, Jr. and Betty J. Wheeler, University of Minnesota A Proposed Hypogenic Origin of Iron Ore Deposits in Southeast Minnesota Karst 4:30 PM Tamsin Green, University of Leeds Down the Rabbit Hole: Identifying Physical Causes of Sinkhole Formation in the UK

PAGE 21

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 20 4:50 PM Brian B Hunt, Brian A. Smith and Alan Andrews, Barton Springs/Edwards Aquifer Conservation District; Douglas Wierman, The Meadows Center for Water and the Environment; Alex S Broun, Hays Trinity Groundwater Conservation District; Marcus O Gary, Edwards Aquifer Authority Relay Ramp Structures and Their Influence on Groundwater Flow in the Edwards and Trinity Aquifers, Hays and Travis Counties, Central Texas 5:30 PM Poster Presentations ; Wine and Cheese Reception ; Exhibit Hall I III E. Calvin Alexander, Jr., University of Minnesota; Scott C. Alexander, University of Minnesota; Kelton Barr, Braun Intertec; Andrew Luhmann, University of Minnesota; Cale Anger, University of Minnesota (deceased) Goliaths Cave, Minnesota: Epigenic Modification and Extension of Pr eexisting Hypogenic Conduits Nozad Hasan Azeez, Dr. Landis Jared West, Professor Simon H Bottrell, University of Leeds Numerical Simulation of Spring Hydrograph Recession Curves for Evaluating Behavior of the East Yorkshire Chalk Aquifer BJ Bonin, Greg Brick Minnesota Department of Natural Resources and Julia Steenberg, Minnesota Geological Survey Seeps and Springs at a Platteville Observatory on the River Bluffs Toby Dogwiler and Blake Lea, Missouri State University Using Electrical Resistivity Imaging to Characterize Karst Hazards in Southeastern Minnesota Agricultural Settings Zhanfei Gu, Qi Liu, Yaoru Lu, and Zhenming Shi, Tongji University; Gaoyu Su, Guangdong Freeway Co, Ltd. Analysis and Prevention of Sinkhole Collapses During the Reconstruction and Extension of the Guang Qing Freeway, China Fuwei Jiang, Guizhou Institute of Technology; Mingtang Lei and Dai Jian ling, Institite of Karst Geology Study on the Critical Velocity of Groundwater to Form Subsidence Sinkholes in Karst Area Mason Johnson and Ashley Ignatius, Minnesota Pollution Control Agency Finding Springs in the File Cabinet Philippe Machetel, Geosciences Montpellier; David A. Yuen, University of Minnesota and China University of Geosciences Evaluation of First Order Error Induced by ConservativeTracer Temperature Approximation for Mixing in Karstic Flow Michael G Raines, Dr. Vanessa J Banks, and Jonathan E Chambers, British Geological Survey; Philip E Collins, Brunel University London; Peter F Jones, University of Derby; Dave J Morgan, James B Riding and Katherine Royse, British Geological Survey The Application of Passive Seismic Techniques to the Detection of Buried Hollows Tim Stokes, Vancouver Island University; Paul Griffiths, Consultant; Carol Ramsey, Vancouver Island University New Methodologies and Approaches for Mapping Forested Karst Landscapes, Vancouver Island, British Columbia, Canada. Kamal Taheri, Kermanshah Regional Water Authority; Milad Taheri, BuAli Sina University;Fathollah Mohsenipour, Kermanshah Regional Water Authority LEPT, A Simplified Approach for Assessing Karst Vulnerability in Regions by Sparse Data; A Case in Kermanshah Province, Iran Sam B Upchurch, SDII Global Corporation Determination of the Relationship of Nitrate to Discharge and Flow Systems in North Florida Springs Min Yang, Feng'e Zhang, Sheng Zhang, Miying Yin and Guoqing Wu, Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences Hydrochemical Characteristics and Formation Mechanism of Groundwater in the Liulin Karst System, Northwestern China Wei Zhang, Wang Guiling and Liu Feng, Institute of Hydrogeology and Environmental Geology, CAGS Environmental effects of rational utilization of karst geothermal resources in the North China Plain

PAGE 22

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 21 Feng'e Zhang, Sheng Zhang, Miying Yin, Guoqing Wu, Institute of Hydrogeology & Environmental Geology, Chinese Academy of Geological Science Chemical Evidence for Biokarst Development in the Ordos Basin (Northwest China) in Laboratory Batch Experiments Edward D Zisman, Cardno ATC; Stephen West, BTL Engineering Services, Inc. Hydrocompaction Considerations in Sinkhole Investigations 7:30 PM Planning Meeting for 15th Sinkhole Conference ; Board Room off the North Lobby Thursday, October 8 7:30 AM Registration opens North Lobby Continental breakfast in Exhibit Hall IIII 8:20 AM Announcements/Reminders in Exhibit Hall IV Plenary Session: GIS Data bases and Map ping of Karst Regions : Jason Polk (Chair) ; Exhibit Hall IV 8:30 AM Jeffrey A Green, Minnesota Dept. of Natural Resources, E Calvin Alexander, Jr., University of Minnesota Creation of a Map of Paleozoic Bedrock Springsheds in Southeast Minnesota 8:50 AM Vanessa J Banks, H J Reeves, E K. Ward, E R Raycraft, H V Gow, D J R Morgan and D G Cameron, British Geological Survey Media, Sinkholes and the UK National Karst Database 9:10 AM Sam B Upchurch and Thomas M Scott, SDII Global Corporation; Michael C Alfieri, Water Resource Associates; Thomas L Dobecki, SDII Global Corporation Shallow Depressions in the Florida Coastal Plain: Karst and Pseudokarst 9:30 AM Clint Kromhout and Alan E Baker, Florida Geological Survey Sinkhole Vulnerability Mapping: Results from a Pilot Study in North Central Florida 9:50 AM John Wall, North Carolina State University; Daniel H Doctor and Silvia Terziotti, US Geological Survey A Semi Automated Tool for Reducing the Creation of False Closed Depressions from a Filled LiDAR Derived Digital Elevation Model 10:10 AM Break in E xhibit H all IIII 10:30 AM Robert G Tipping, Mathew Rantala, E Calvin Alexander, Jr., University of Minnesota; Yongli Gao, University of Texas; Jeffrey A Green, Minnesota Department of Natural Resources History and Future of the Minnesota Karst Feature Database 10:50 AM Gregory Brick, Minnesota Dept. of Natural Resources Legacy Data in the Minnesota Spring Inventory 11:10 AM Yongli Gao, University of Texas; Raghav Ramanathan, Bulent Hatigoplu and M. Melih Demirkan, Rizzo Associates; Mazen Elias Adib, Abu Dhabi City Municipality; Juan J Gutierrez, Hesham El Ganainy and Daniel Barton Jr., Rizzo Associates Development of Cavity Probability Map For Abu Dhabi Municipality Using GIS and Decision Tree Modeling

PAGE 23

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 22 11:30 AM Raghav Ramanathan Rizzo Associates; Yongli Gao, University of Texas; M. Melih Demirkan; Bulent Hatipoglu, Rizzo Associates; Mazen Elias Adib, Abu Dhabi City Municipality; Michael Rosenmeier, Juan Gutierrez, and Hesham El Ganainy, Rizzo Associates Evaluation of Cavity Dist ribution Using Point Pattern Analysis 11:50 AM Alexandra L Todd and Lindsay Ivey Burden, University of Virginia A Method of Mapping Sinkhole Susceptibility Using a Geographic Information System: A Case Study for Interstates in the Karst Counties of Virginia 12:10 PM Catered Lunch in E xhibit H all IIII Track A -Session: Contamination of Karst Aquifers : Ming Ye (Chair) ; Mc Donnell Suite 1:30 PM Larry Boot Pierce, Missouri Geological Survey; Honglin Shi, Missouri University of Science and Technology Evaluation of Veterinary Pharmaceuticals and Iodine for Use as a Groundwater Tracer in Hydrologic Investigation of Contamination Related to Dairy Cattle Operations 1:50 PM Virginia Yingling, Minnesota De partment of Health Karst Influence in the Creation of a PFC Megaplume 2:10 PM Caren Raedts and Christopher Smart, Western University of Ontario Tracking of Karst Contamination Using Alternative Monitoring Strategies: Hidden River Cave Kentucky 2:30 PM Ingrid Y Padilla, Vilda L Rivera and Celys Irizarry, University of Puerto Rico Spatiotemporal Response of CVOC Contamination and Remedial Actions in Eogenetic Karst Aquifers Track A --Session: Geophysic al Exploration of Karst: Mustafa Saribudak (Chair) ; Legion Suite 1:30 PM Mustafa Saribudak, Environmental Geophysics Associates The Million Dollar Question: Which Geophysical Methods Locate Caves Best Over the Edwards Aquifer? A Potpourri of Case Studies from San Antonio and Austin, Texas, USA 1:50 PM Lewis Land, National Cave and karst Research Institute; Lasha Asanidze, Tbilisi State University Rollalong Resistivity Surveys Reveal Karstic Paleotopography Developed on Near Surface Gypsum Bedrock 2:10 PM Brian Bruckno Virginia Department of Transportation; Andrea Vaccari, University of Virginia; Edward Hoppe, Virginia Center for Transportation Innovation and Research; Scott T Acton, University of Virginia; Elizabeth Campbell, Virginia Department of Transportation Int egration and Delivery of Interferometric Synthetic Aperture Radar (InSAR) Data Into Stormwater Planning Within Karst Terranes 2:30 PM Khiem T Tran, Clarkson University; Michael McVay, University of Florida; Trung Dung Nguyen, Clarkson University Detection of Voids in Karst Terrain with Full Waveform Tomography 2:50 PM Laurentiu Artugyan, Adrian C. Ardelean and Petru Urdea, West University of Timisoara Characterization of Karst Terrain Using Geophysical Methods Based on Sinkhole Analysis: A Case Study of the Anina Karstic Region (Banat Mountains, Romania) 3:10 PM Break in E xhibit H all IIII Track A --Session: Karst Management, Regulation and Education: David Weary (Chair) ; McDonnell Suite 3:30 PM David Weary, US Geological Survey The Cost of Karst Subsidence and Sinkhole Collapse in the United States Compared with Other Natural Hazards 3:50 PM Fang Guo and Guanghui Jiang, Institute of Karst Geology, Chinese Academy of Geological Sciences; Kwong Fai Andrew Lo, Chinese Culture University; Qingjia Tang, Yongli Guo and Shaohua Liu, Institute of Karst Geology, Chinese Academy of Geological Sciences Hazard of Sin khole Flo oding to a Cave Hominin Site and its Control Countermeasures in a Tower Karst Area, South China

PAGE 24

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 23 4:10 PM Martin Larsen, Olmsted County Soil and Water Conservation District Case Studies of Animal Feedlots on Karst in Olmsted County, Minnesota 4:30 PM Chiara Calligaris, Stefano Devoto, Luca Zini, and Franco Cucchi, Trieste University Evaporite Geo Hazard in the Sauris Area (Friuli Venezia Giulia Region NE Italy) 4:50 PM George Veni, National Cave and Karst Research Institute; Connie Campbell Brashear and Andrew Glasbrenner, Bracken Engineering, Inc. Building Codes to Minimize Cover Collapses in Sinkhole Prone Areas 5:10 PM Jason S Polk and Leslie A North, Western Kentucky University; Ric Federico, Brian Ham and Dan Nedvidek, EnSafe; Kegan McClanahan and Pat Kambesis, Western Kentucky University; Michael J Marasa, Hayward Baker Cars and Karst: Investigating the National Corvette Museum Sinkhole Track B --Session: Geophysical Exploration of Karst, Mustafa Saribudak (Chair) ; Legion Suite 3:30 PM Philip J Carpenter and Lauren M Schroeder; Northern Illinois University Investigation of a Sinkhole in Ogle Cou n ty, Northwestern Illinois, Using Near Surface Geophysical Techniques 3:50 PM Zhende Guan, X Z Jiang, Y B Wu, Z Y Pang, Institute of Karst Geology, China University of Geosciences Study on Monitoring and Early Warning of Karst Collapse Based on BOTDR Tech nique 4:10 PM Cathleen E Jones and Ronald G Blom, Jet Propulsion Laboratory, California Institute of Technology Pre Event and Post Formation Ground Movement Associated with the Bayou Corne Sinkhole Track B --Plenary Session: Modeling of Karst Systems : Ming Ye (Chair) ; Legion Suite 4:30 PM Long Jia, Yan Meng and Zhende Guan, Institute of Karst Geology; Li peng Liu, China Institute of Water Resources and Hydropower Research Numerical Simulation of Karst Soil Cave Evolution 4:50 PM Xiaohu Tao, Ming Ye, Dangliang Wang, Roger Pacheco Castro and Xiaoming Wang, Florida State University; Jian Zhao, Hohai University Experimental and Numerical Investigation of Cover Collapse Sinkhole Development and Collapse in Central Florida 5:10 PM Justin L Blum, Minnesota Department of Health Accounting for Anomalous Hydraulic Responses During Constant Rate Pumping Tests in the Prairie Du ChienJordan Aquifer System Towards a More Accurate Assessment of Leakage 6:30pm Banquet and Guest Speaker: David Rogers, Missouri University of Science and Technology ; Location: The Kahler Grand Hotel, Heritage III, 20 SW Second Ave Rochester Hales Bar and the Pitfalls of Constructing Dams on Karst Friday, October 9 7:30 AM Registration opens North Lobby Continental breakfast in Exhibit Hall IIII 8:20 AM Announcements/Reminders in Exhibit Hall IV Plenary Session: Engineering and Geotechnical Investigations in Karst, Lynn Yuhr (Chair) ; Exhibit Hall IV 8:30 AM Joseph A Fischer and Joseph J Fischer, Geoscience Services Concepts for Geotechnical Investigation in Karst

PAGE 25

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 24 8:50 AM Mohamed Alrowaimi, Hae Bum Yun and Manoj Chopra, University of Central Florida Sinkhole Physical Models to Simulate and Investigate Sinkhole Collapses 9:10 AM Kevin M O'Connor, GeoTDR, Inc; Matthew Trainum, Iowa Department of Transportation Monitoring the Threat of Sinkhole Formation Under a Portion of US 18 in Cerro Gordo County, Iowa Using TDR Measurements 9:30 AM Edward D Zisman, Cardno ATC Predicting Compaction Grout Quantities in Sinkhole Remediation 9:50 AM Chase L Kicker, New Mexico Institute of Mining and Technology A Feasibility Study of the Implementation of a Flowable Fill Material to Prevent Sinkhole Occurrence at the I& W Brine Well Site in Carlsbad, New Mexico 10:10 AM Break in E xhibit H all IIII 10:30 AM Steven W Shifflett, US Army Corps of Engineers Pre Construction Rock Treatment and Soil Modification Program Using Low Mobility Grout to Mitigate Future Sinkhole Development in a 2,787.1 Square Meter (30,000 SF) Maintenance Facility 10:50 AM David M Robison, U.S. Army Corps of Engineers Successful Foundation Preparations in Karst Bedrock of the Masonry Section of Wolf Creek Dam 12:00 AM Post conference field trip: Karst in Our Built Environment Trip Leader: Jeff Broberg Note: Box lunches will be provided. Moth Spring, Fillmore County, MN. Photo by Jeff Green

PAGE 26

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 25 Keynote Spea ker: Michael T. Osterholm, PhD, MPH The Midwest Karst and Historic Tall Grass Prairie: An Overlooked International Treasure Dr. Osterholm is the McKnight Presidential Endowed Chair in Public Health at the University of Minnesota and director of the Center for Infectious Disease Research and Policy (CIDRAP), a professor in the Divis ion of Environmental Health Sciences, School of Public Health, a professor in the Technological Leadership Institute, College of Science and Engineering, and an adjunct professor in the Medical School, University of Minnesota. He is also a member of the In stitute of Medicine (IOM) of the National Academy of Sciences and the Council of Foreign Relations. In June 2005 Dr. Osterholm was appointed by Michael Leavitt, Secretary of the Department of Health and Human Services (HHS), to the newly established Nation al Science Advisory Board on Biosecurity. In July 2008, he was named to the University of Minnesota Academic Health Centers Academy of Excellence in Health Research. In October 2008, he was appointed to the World Economic Forum Working Group on Pandemics. In the late 1960s Mike was part of the cave diving teams that dove the Coldwater Spring sump, discovered and mapped the initial exploration of Coldwater Cave which is now the longest cave (currently at17 miles) in the Upper Mississippi Valley Karst. In the early 1970s Mikes PhD thesis at the University of Minnesota concerned the impact of bacterial contamination of groundwater on childrens health in southeast Minnesotas karst lands Dr. Osterholm served for 24 years (1975 1999) in various roles at the Minnesota Department of Health (MDH), the last 15 as state epidemiologist and chief of the Acute Disease Epidemiology Section. While at the MDH, Osterholm and his team were leaders in the area of inf ectious disease epidemiology. Dr. Osterholm has been an international leader on the critical concern regarding our preparedness for an influenza pandemic. His invited papers in the journals Foreign Affairs the New England Journal of Medicine, and Nature detail the threat of an influenza pandemic before the recent pandemic and the steps we must take to better prepare for such events. Dr. Osterholm has also been an international leader on the growing concern regarding the use of biological agents as catast rophic weapons targeting civilian populations. In that role, he served as a personal advisor to the late King Hussein of Jordan. Dr. Osterholm provides a compr ehensive and pointed review of America's current state of preparedness for a bioterrorism attack in his New York Times best selling book, Living Terrors: What America Needs to Know to Survive the Coming Bioterrorist Catastrophe. One of Mikes passions is the Prairie Song Farm in northeastern Iowa karst. (Google Prairie Song Farm). With 380 feet of vertical relief the farm stratigraphically spans from the Galena Limestone down to the Prairie du Chien Group. Mike notes that the Upper Miss issippi Valley is the only place on Earth where Tall Grass Prairie on loess soils occur on top of karst. This unique superposition creates a unusual landscape the Mike is working to restore.

PAGE 27

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 26 Banquet Speaker : David J. Rogers, PhD Hales Bar and the Pitfalls of Constructing Dams on Karst Hales Bar Dam was built on the Tennessee River 33 miles downstream of Chattanooga by a private company to generate power in 1905 1913. The dam site was selected because it was the narrowest reach in the downstream end of the Walden Ridge Gorge. The site i s underlain by Mississippian Bangor Limestone on the southeast flank of the Sequatchie Anticline. Three different contracts failed to complete the dam because of difficult foundation conditions. From 19101913 diamond drill core holes were used to explore the site and a series of reinforced concrete caissons 40x45 ft on upst ream side and 30x32 ft on the downstream side were installed. Excessive leakage soon appeared near the eastern abutment, and gradually increased. Soundings were made in 1914 to ascer tain the areas of gross leakage thereafter rags were placed over suction h oles on the river bed and concrete pumped over these. Once a leak was stemmed, leakage would resume at other, adjacent locations. The owners tried to stem the leaks by inserting hay bales, old mattresses, chicken wire, and even corsets! In 1919 the owners began drilling grout holes from the inspection gallery within the dam and pumping hot asphalt into the voids. This was followed by the injection of 78,324 cubic feet of hot asphalt grout into the dam foundation, using 6,266 lineal feet of boreholes with a verage hole depth of 92 ft. By 1922 the problem appeared solved, but le akage gradually resumed between 192219 29, rising to the same level as had been observed previously. In 19301931 a new program of exploration was undertaken, using dyes and oils to i dentify conduits under the dam. Leakage was found to vary between 100 and 1200 cubic feet per second (cfs). When the dam was acquired by the Tennessee Valley Authority (TVA) in 1939 they employed fluorescein dyes to track the under seepage. Dye tests rev ealed that the leakage varied between 1720 and 1650 cfs; about 10% of the rivers normal flow. They also noted seepage boils forming in the gravel bars, which increased eac h year. The TVA began constructing the most expensive cutoff wall ever built, drilling 750 18inch diameter holes along the dams centerline and backfilling this with concrete to a maximum depth of 163 feet, extending 25 to 103 feet below the river bed. In April 1963 the TVA announced it was abandoning Hales Bar Dam, due to increasing leakage. Biography: J. David Rogers holds the Karl F. Hasselmann Chair in Geological Engineering at the Missouri University of Science & Technology in Rolla, Missouri. He is presently representing the geological and geotechnical engineering professions on the National Academies panel that has been charged with examining Levees and the National Flood Insurance Program: Improving policies and practices, being funded by FEMA. Dr. Rogers has been fascinated by dam and levee failures, evaluating the stability of natural slopes, embankments, stream channels, highways, and hydraulic structures. He has served as principal investigator for research funded by the NSF, U.S. Geological S urvey, Na tional Geospatial Intelligence Agency, Federal Highway Administration, Department of Defense, and several state departments o f transportation. He has served on numerous panels, including the Mississippi Delta Science & Engineering Special Team, the Coastal Louisiana Recovery Panel, the NSF Independent Levee Investigation Team and USGS Investigation Teams evaluating the impacts of Hurricane s Katrina and Rita, the NSF team evaluating the 2008 and 2011 Mississippi River fl oods, and the Resilient and Sustainable Infrastructure Networks team funded by NSF to make a five year examination of the California Bay Delta flood protection systems. He recently completed a book on the Engineers Who Built the Panama Canal and delivere d one of principal history and heritage presentations at the 100th anniversary of the completion of the Panama Canal, down in Panama for the ASCE annual meeting. He is a frequent guest on newscasts and television documentaries on geologic and mancaused ha zards. Dr. Rogers received his B.S. degree in geology from Cal ifornia State Poly technic University at Pomona, his M.S. degree in civil engineering from the University of California, Berkeley, and his Ph.D. in geological and geotechnical engineering at th e University of California, Berkeley. He served on the Berkeley faculty in civil engineering for seven years prior to accepting his current position in 2001.

PAGE 28

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 27 Presentation Abstracts (in order of presentation) Wednesday, October 7th 9:10 am 1 1:5 0pm Upper Mississippi Valley Karst Aquifers Karst Hydrogeologic Investigation of Trout Brook Joel T Groten, US Geological Survey E. Calvin Alexander, Jr., University of Minnesota Trout Brook in the Miesville Ravine County Park of Dakota County Minnesota is the trout stream with the highest nitrate concentration in the karst region of southeastern Mi nnesota. Water quality data from 1985 and 1995 (Spong, 1995) and from 2001, 2002, 2006, 2010, and 2014, collected by the Dakota County Soil and Water Conservation District (Dakota SWCD, 2014) document an increasing level of nitrate in Trout Brook. A karst hydrogeologic investigation was designed to measure nitrate levels at sampling points along the stream and to increase our understanding of the source and movement of nitrates throughout the length of Trout Brook. Eighteen springs and seeps have been locat ed in the Main Branch and tributaries of Trout Brook. A previously unreported flowing section and stream sieve, Weber Sieve, were found above what had been thought to be the head of perennial flow in the East Branch of Trout Brook. Two new sinkholes developed after the 14 15 June 2012 flood in a field northeast of the East Branch of Trout Brook. This investigation included regular monitoring of major anions in the streams and springs, synoptic stream flow measurements, and a dye trace of a sinking stream in the Trout Brook drainage. The initial assumption was that the majority of the baseflow of Trout Brook was from discrete springs. However, synoptic baseflow and nitrate measurements show that only 30 40 percent of the total flow in Trout Brook is from discrete springs, and the rest appears to be from distributed groundwater discharge directly into the stream. Both the discrete springs and the distributed recharge occur along reaches of Trout Brook that drain the significant high transmissivity zone nea r the bottom of the regionally important Shakopee aquifer. Dye traces have confirmed flow paths from Weber Sieve to LeDuc and Bridgestone Springs and have begun to define springsheds for these head water springs. Nitrate concentrations and chloride/bromide ratios decreased systematically from the upstream springs to the downstream springs. The nitrate concentrations have been increasing at four springs from 1985 to 2014 and at two surface sampling points from 2001 to 2014. The nitrate concentration of a nother surface sampling point increased from 2001 to 2006, decreased from 2006 to 2012, and increased from 2012 to 2014. Snowmelt and rainfall runoff was sampled on 2 March 2012 and showed no detectable nitrate in the runoff from a watershed with no row cr op agriculture, but elevated nitrate was detected in an adjacent watershed with row crop agriculture. All of these trends illustrate the dominance of agricultural sources of nitrate in Trout Brook. Human Impacts on Coldwater Spring Water Quality, Minneapolis, Minnesota Sophie M Kasahara, University of Minnesota Scott C Alexander University of Minnesota E Calvin Alexander, Jr., University of Minnesota Coldwater Spring in Minneapolis, Minnesota was the water supply for Fort Snelling from the 18 40s to 1920. The spring site has been declared a sacred site by some federally recognized Native American tribes. The site is managed by the National Park Service. This project has monitored the water chemistry of Coldwater Spring to document human impacts on the springs water quality. Temperature, dissolved oxygen, conductivity, pH and anions were monitored weekly and cations and alkalinity monitored monthly at Coldwater Spring and the adjacent Wetland A from 15 February 2013 through 18 January 2015. Cold water Springs water flows through fractures in Platteville Limestone of Ordovician age. The basic chemistry of Coldwater Spring should be the calcium magnesium bicarbonate water typical of carbonate springs. However, on an equivalent basis, Coldwater Springs water currently contains almost as much sodium as calcium + magnesium and more chloride than bicarbonate. The chloride concentrations are about 100 times the levels from 1880. Maguire (1880) reported the chloride levels of Coldwater Spring were about 4.5 ppm. During the current study the chloride content in the spring increased from about 320 ppm from March 2013 to about 410 ppm in December 2014. In April, May and June of 2013 and 2014, the chloride rose about 100 ppm in three month long pulses. The ch loride concentration of the water in Wetland A ranges from about 400 ppm to over 600 ppm with a pattern that is a mirror image of the Coldwater Spring pattern. This major anthropogenic chloride component has a chloride to bromide ratio of 2,500 300, well within the range of chloride to bromide ratios of road salt, 1,000 to 10,000. Road salt is applied to two major multi lane highways close to the spring and is used extensively in this heavily urbanized area throughout the

PAGE 29

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 28 winter The temperature of the sp ring is variable and higher than its pre settlement temperature. Nicollet (1841) recorded the temperature of Coldwater Spring multiple times in summer of 1836 as 46 F (7.8 C) and multiple times in winter of 1837 as 45.5 F (7.5 C). More recently the tem perature of Coldwater Spring fluctuates smoothly between 10.7 and 13.1 C. The higher temperature of the springs discharge also indicates an anthropogenic source of heat within the spring shed or spring recharge area. The spring water is coldest in May an d June and warmest in October and November. The temperature of the water in Wetland A fluctuates from 6.4 to 13.8 C in a pattern that is opposite of that in Coldwater Spring. Coldwater Spring also contained a significant, increasing nitratenitrogen co mponent which ranged from 2.5 to 5.2 ppm with dips at the same times as the chloride pulses. Wetland As nitrate nitrogen level varied between 0.2 to almost 6 ppm with large pulses at the same time as Coldwater Springs dips. A 2014 study performed by th e U.S. Geological Survey came to the conclusion that increasing chloride levels in lakes and streams are likely driven by increasing road salt application, rising baseline concentrations, as well as an increase in snowfall in the Midwestern area of the U.S during the time of the study (Corsi 2014). The significant chloride, temperature and nitrate levels are likely to be driven by anthropogenic sources. Hydrologic and Geochemical Dynamics of Vadose Zone Recharge in a Mantled Karst Aquifer: Results of Mo nitoring Drip Waters in Mystery Cave, Minnesota Daniel H Doctor, US Geological Survey E. Calvin Alexander Jr., University of Minnesota Roy Jameson Scott Alexander, University of Minnesota Caves provide direct access to flows through the vadose zone that recharge karst aquifers. Although many recent studies have documented the highly dynamic processes associated with vadose zone flows in karst settings, few have been conducted in mantled kar st settings, such as that of southeastern Minnesota. Here we present some results of a long term program of cave drip monitoring conducted within Mystery Cave, Minnesota. In this study, two perennial ceiling drip sites were monitored between 1997 and 2001. The sites were located about 90 m (300 ft) apart along the same cave passage approximately 18 m (60 ft) below the surface; 7 to 9 m (20 to 30 ft) of loess and 12 m (40 ft) of flat lying carbonate bedrock strata overlie the cave. Records of drip rate, electrical conductivity, and water temperature were obtained at 15 minute intervals, and supplemented with periodic sampling for major ion chemistry and water stable isotopes. Patterns in flow and geochemistry emerged at each of the two drip sites that were re peated year after year. Although one site responded relatively quickly (within 2 7 hours) to surface recharge events while the other responded more slowly (within 25 days), thresholds of antecedent moisture needed to be overcome in order to produce a disc harge response at both sites. The greatest amount of flow was observed at both sites during the spring snowmelt period. Rainfall events less than 10 mm (0.4 in) during the summer months generally did not produce a drip discharge response, yet rapid drip re sponses were observed following intense storm events after periods of prolonged rainfall. The chemical data from both sites indicate that reservoirs of vadose zone water with distinct chemical signatures mixed during recharge events, and drip chemistry ret urned to a baseline composition during low flow periods. A reservoir with elevated chloride and sulfate concentrations impacts the slow response drip site with each recharge event, but does not similarly affect the fast response drip site. Nitrate concentr ations in drip waters were generally less than 4.0 mg/L as NO3 (or less than 1 mg/L as N). Nitrate was either stable or slightly increased with drip rate at the fast response drip site; in contrast, nitrate concentrations decreased with drip rate at the s low response drip site. Conduit Flow in the Cambrian Lone Rock Formation, Southeast Minnesota, U.S.A. John D. Barry Minnesota Dept. of Natural Resources Jeffrey A. Green, Minnes ota Dept. of Natural Resources Julia R. Steenberg, Minnesota Geological Survey The karst lands of southeast Minnesota contain more than one hundred trout streams that receive perennial discharge from Paleozoic bedrock springs. Several of the Paleozoic bedrock units that provide discharge are karst aquifers. Field investigatio ns into the flow characteristics of these formations have been conducted using fluorescent dyes to map groundwater springsheds and characterize groundwater flow velocities for use in water resource protection. Recent field work has focused on the Camb rian Lone Rock Formation, a siliciclastic unit consisting of very fine grained sandstone and siltstone with minor beds of shale and dolostone. The formation is mapped within tributary valleys of the Mississippi River throughout southeastern Minnesota and s outhwestern Wisconsin. Overlying the Lone Rock is the Cambrian St. Lawrence Formation. Over a dozen streams have been observed to disappear into stream sinks where the upper St. Lawrence is the bedrock unit closest to the land surface. At three of these si nking stream locations, dye was recovered emanating from springs located in the basal St. Lawrence or from springs located in two distinct zones in the Lone Rock. Dye breakthrough velocities calculated using passive charcoal detectors ranged between 21 214 meters/day at one location and 88153 meters/day at another. At a third site, automatic

PAGE 30

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 29 water samplers were placed at a spring that had been previously demonstrated to be connected to a St. Lawrence stream sink through dye tracing. In that trace, an eight hour sampling frequency determined the dye breakthrough velocity was 314 meters/day. Based on outcrop and borehole observations in Minnesota, secondary pore networks in siliciclastic dominated units generally have bedding parallel and vertically orie nted apertures less than a few centimeters. The process by which the bedding parallel secondary pore networks form remains obscure; some appear to be mechanically developed. However, interstitial carbonate cement within these units leads to the possibility of dissolution being a minor factor in the formations groundwater flow characteristics. These dye traces were conducted at three different sites across a twenty three kilometer distance and are evidence that the siliciclastic Lone Rock Formation has a co nduit flow component similar to that fou nd in carbonate karst aquifers. Using Nitrate, Chloride, Sodium, and Sulfate to Calculate Groundwater Age Kimm Crawford, C rawford Environmental Services Terry Lee, Olmsted County Environmental Resources Dept. Regression analysis is used to identify monotonic trends to assign water age using ion data from two large well water databases from southeast Minnesota (SE MN). Nitrate (NO3N), chloride (Cl), sodium (Na), and sulfate (SO4) ions in the commonly used a quifers in SE MN can be used as groundwater tracers since they are either entirely or partly anthropogenic in their sources, their loading occurs on a regional scale, and they are almost entirely conserved. Ion concentrations over time are used to establish six trend patterns. Two patterns are unchanging (background and stable above background), and four are changing (linear up, exponential up, peaking, and down). These patterns are then used to assign specific age values or age ranges to the well b ased upon that ion. For ions with linear upward trends, specific ages are derived from the Intercept Year representing a time when water extracted from the well first infiltrated from the land surface containing the ion at detectable concentrations and the Marker Year representing the beginning of large scale trend changes for that ion source. Karst Spring Cutoffs, Cave Tiers, and Sinking Stream Basins Correlated to Fluvial Base Level Decline in SouthCentral Indiana Garre A Conner, Pangea Geos ervices The Mitchell Aquifer averages 80 m in thickness and underdrains a karst region in the Crawford Upland and Mitchell Plateau region in south central Indiana (110,000 km2). The Springville Escarpment is a transitional boundary between the upland and plateau. Cave stream linking between cave tiers in the aquifer and correlation of cave tier inception horizons to a base level decline surface is interpreted for the Kirby Wate rshed, encompassing the prekarst headland of Indian Creek (42 km2). The watershed was severed from lower Indian Creek at Eller Col by limestone cavern drainage on the ridge between White River and East Fork. Correlation of recharge basin topography and cav e tiers is possible owing to the observation of 55 karst springs confined to lithostratigraphic contacts at three spring stratigraphic levels. Karst Spring Cutoffs are a specific type of vadose canyon diverting cave streams, bypassing around springs and pa ssing into the laterally offset cave streams in the next lower cave tier. Cutoffs connect upper to middle tier cave streams and middle to lower tier cave streams as they enlarge below sinking stream basins and tributary spurs. Three speleogenic enlargeme nt cycles characterize the eastern Leonard Springs Area, but only two cycles have enlarged in the western Garrison Chapel Area. Driftless Area Karst of Northwestern Illinois and its Effects on Groundwater Quality Samuel V Panno, Illinois State Geological Survey Walton R Kelly, Illinois State Water Survey The bedrock aquifer of the Driftless Area of northwestern Illinois is Ordovician age Galena Dolomite. Previous work by the authors and others showed that the geology and hydrogeology of this a rea fall well within the definition of karst. Bedrock in the study area has been shown to be extensively fractured and creviced; karst features in the county are dominated by solution enlarged crevices from 0.4 inches to 3 feet wide within most road cuts a nd quarries examined. Other karst features include cover collapse sinkholes ranging from 3 to 25 feet in diameter overlying Galena Dolomite, karst springs and crevice caves. A preliminary evaluation of the groundwater quality Jo Daviess County in the D riftless Area of northwestern Illinois was conducted to assess the susceptibility of the Galena Dolomite aquifer to surface borne contaminants. This was done by evaluating available groundwater quality data from published sources and the Illinois State Water Survey

PAGE 31

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 30 Groundwater Quality Database (i.e., wells and springs), and also by sampling 11 private wells across the county and analyzing for inorganic chemistry, dissolved organic carbon, stable isotopes and tritium. We found that groundwater in the study a rea is of a CaMg HCO3 type as would be expected for an aquifer dominated by dolomite. In parts of the county, the upper part of the carbonatehosted aquifer contains elevated concentrations of chloride, nitrate and potassium. Likely contamination sources are anthropogenic and include road salt, potash and nitrogen fertilizers, and livestock/human waste. The presence of these contaminants suggests movement of surfaceborne contaminants into the aquifer and into wells screened at depths ranging from 65 to 150 feet. Historic data reveal stratification of surface borne contaminants (greatest concentrations nearest the surface) within the fractured and creviced aquifer to a depth of about 300 feet. Nitrate N (NO3N) concentrations in karst springs are typic ally between 10 and 13 mg/L, but can exceed 30 mg/L. Because the predominant land use in the study area is row crop agriculture, it is likely that much of the NO3N is derived from N fertilizer. In the 11 water well samples, NO3N concentrations ranged fro m < 0.04 (detection limit) to 5.4 mg/L and concentrations were clearly related to tritium. Specifically, NO3N concentrations in groundwater containing less than 3 TU were below detection (0.04 mg/L), and above 3 TU, NO3N and tritium were positively corre lated. This relationship suggests a nonpoint source of N and denitrification within the aquifer. Chloride (Cl-) concentrations in karst springs were between 15 and 25 mg/L above background concentrations (1 to 15 mg/L). Water wells samples had lower Clco ncentrations with 7 of 11 wells below background (ca. 15 mg/L), although the concentration in the shallowest well was 110 mg/L and was probably derived from road salt. Overall, the groundwater quality of the Galena Dolomite aquifer in Jo Daviess County is what would be expected in an open, dolomite dominated karst aquifer. Wednesday, October 7th 11:50am 3: 1 0 pm Karst Hydrology Hydrogeological Dynamic Variability in the Lomme Karst System (Belgium) as Evidenced by Tracer Tests Results (KARAG project) Amal Poulain, University of Namur, Belgium Gatan Rochez U niversity of Namur, Belgium Vincent Hallet, University of Namur, Belgium Paleozoic carbonate aquifers represent major groundwater resources in Belgium. Karstification processes affect most of them and Belgium counts many hydrologically active karst networks. Given the intrinsic vulnerability of such geological objects, comprehensive studies are required in order to improve their protection and management. The KARAG project (Karst Aquifer ReseArch by Geophysic, 20132017) aims to identify the specific dynamic of karst aquifers by using geophysical and hydrogeological tools. This research is funded by the Belgium Nation al Fund for Scientific Research (FNRS) and conducted by the University of Namur, University of Mons and the Royal Observatory of Belgium. The LKS Lomme Karst System (Rochefort, southern Belgium) was chosen as it is a major Belgian karst system (10 k m long) in the Givetian carbonate aquifer (Middle Devonian). The system is formed by two parallel components: the surface system (the Lomme River) and a complex underground system (multiple sinkholes with one main resurgence). Based on this layout, it is p ossible to study the aquifer dynamic and its relat ionship with the surface river. A high resolution monitoring has been installed since July 2013 in order to follow the system dynamic during several hydrogeological cycles. Multitracing experiment s with different injections and monitoring points highlight the complexity of underground flow dynamics. Investigations enlightened the connectivity between monitoring points and how dependent of the hydrological conditions were these connections. The breakthrough curves analysis allows to characterize the hydrodynamic behavior of the underground flows within the aquifer. Modeling will be necessary to link hydrological and tracer tests data in order to build a detailed conceptual model for this karst system. This model will also be used to interpret geophysical data (ERT, gravimetry) collected in order to study the unsaturated and epikarst zones.

PAGE 32

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 31 Recharge Area of Selected Large Springs in the Ozarks James W Duley, Missouri Geological Survey Cecil Boswell Missouri Geological Survey Jerry Prewett, Missouri Geological Survey Ongoing work by the Missouri Geological Survey (MGS) is refining the known recharge areas of a number of major springs in the Ozarks. Among the springs being investigated are: M ammoth Spring (Fulton County, Arkansas), and the following Missouri springs: Greer Spring (Oregon County), Blue Spring (Ozark County), Blue/Morgan Spring Complex (Oregon County), Boze Mill Spring (Oregon County), two different Big Springs (Carter and Dougl as County) and Rainbow/North Fork/Hodgson Mill Spring Complex (Ozark County). Previously unpublished findings of the MGS and United States Geological Survey (USGS) are also being used to better define recharge areas of Greer Spring, Big Spring (Carter Coun ty), Blue/Morgan Spring Complex, Rainbow/ North Fork/Hodgson Mill Complex, Wilder Spring (Ozark County) and Althea Spring (Ozark County). MGS is applying a graphical method of data analysis using spectrofluorometric scan results. Comparing the dye peak intensity to the intensity of the valley preceding the peak yields a ratio that can be used to standardize and quantify water traces. This method can be applied to current and some legacy traces with comparable results. In some cases, past tracer inj ection sites were utilized in attempts to replicate older traces from those locations. The data clearly show that there is value in applying spectrofluorometric dye detection techniques in attempting to replicate older traces. Some repeat injections and su bsequent monitoring confirmed earlier traces. Other replication efforts revealed multiple recovery points that were undetected by the legacy traces, thus expanding known recharge areas. Still other replication efforts indicate that some older traces are no t repeatable. The effect of the replication efforts significantly changes the logical interpretation of a number of recharge area boundaries. Among the findings of the overall study to date: Mammoth Spring and Greer Spring share a portion of their rec harge, with the majority of Greer Springs flow apparently passing under a gaining segment of the 11 Point River, ultimately emerging more than four kilometers to the southeast. This and other findings raise questions about how hydrology in the study area may be controlled by deep seated mechanisms such as basal faulting and jointing. Research and understanding would be improved by 1:24,000 scale geologic mapping and increased geophysical study of the entire area. Hydrological and Hydrogeological Characte ristics of the Platform Karst (Zemo Imereti Plateau, Georgia) Zaza Lezhava, Javakhishvili Tbilisi State Univ ., Georgia Nana Bolashvili, Javakhishvili Tbilisi State Univ ., Georgia Kukuri Tsikarishvili, Javakhishvili Tbilisi State Univ ., Georgia Lasha Asani dze, Javakhishvili Tbilisi State Univ ., Georgia Chikhradze, Javakhishvili Tbilisi State Univ ., Georgia The article discusses the hydrological and hydrogeological characteristics of the platform karst of Zemo Imereti plateau. The structural plateau of Zemo Imereti is the part of the intermountain plain karst zone of Georgia and one of the interesting parts of the karst relief development. The above mentioned karst region includes the easternmost part of western Georgia, which is characterized by peculiar natural conditions (relief, tectonics, climate, surface and underground streams) and represents one of the significant platform karst regions in the Caucasus. On the basis of the cartographic materials (GIS) analysis and borehole data the general scheme of h ydrogeological setting of the Zemo Imereti structural plateau (two hydrogeological basins were defined) was created as confirmed by experiments. In addition, it was identified that underground karst water flowing from the periphery to the center determines sedimentation together with the broken dislocations within the frame of the structural plateau. The study found that within the Chiatura structural plateau the joint karst hydrogeological system (with enough dynamic water resources) has been established, which mainly is unloaded in sources of Ghrudo vaucluse and the surrounding area (local erosion basis). Ghrudo hydrogeological system and Chiatura structural plateau are characterized by the systems of isolated karst fissure waters with different hypsometri c location and orientation. Therefore, based on these studies, it could be said that in karst areas the structural features can define the characteristic of groundwater circulation, but karst age can also make a significant adjustment. Tracer Studies Con ducted Nearly Two Decades Apart Elucidate Groundwater Movement Through a Karst Aquifer in the Frederick Valley of Maryland Keith A. White, ARCADIS Thomas J Aley, Ozark Underground Laboratory Michael K Cobb, ARCADIS Ethan O Weikel, US Army Corps of Engineers Shiloh L Beeman, Ozark Underground Laboratory A pair of groundwater tracer studies at a single karst test site were completed 18 years apart. The results of these studies have

PAGE 33

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 32 provided evidence of both relatively rapid advective transport via conduits and an extreme capacity for dye storage and retardation. The tracer results, coupled with other subsurface investigation data, are used to develop a conceptual model for groundwater movement through this karst aquifer in the Frederick Valley of Maryland, as well as identify implications for remediation. Three fluorescent tracer dyes used in the initial study were detected in several background monitoring locations established for the second study conducted 18 years later, demonstrating the pers istence of these dyes in the aquifer. One of these dyes was not detected during the original study, providing useful information regarding flow and transport in the aquifer. At some of these sampling locations, at least one of the dyes was degraded, and wo uld have gone undetected without the use of activated carbon samplers. Lastly, even though relatively rapid first detections occurred during both studies (as compared to non karst groundwater systems) the majority of injected dye mass remained in the aquif er after the studies were completed. This suggests that the aquifer has a large capacity to store contaminants and that low levels of contaminants can be expected to persist in groundwater discharged from springs for a long period of time. Dye Tracing Th rough the Vadose Zone Above Wind Cave, Custer County, South Dakota James Nepstad, National Park Service During the 1990s, in an attempt to better understand threats posed by surface developments overlying the cave, National Park Service staff at Wind Cav e National Park in Custer County, South Dakota carried out a series of dye traces through portions of the vadose zone overlying the cave. Wind Cave is located within the 100m thick Madison formation (limestone and dolomite), which in most locations is capp ed by varying thicknesses of the basal units of the Minnelusa formation (intermingled beds of sandstone, limestone, and shale). A variety of cave locations with dripping or pooled water were monitored for up to five years following dye injection. Transit t imes to the cave varied from less than six hours to as much as 4.8 years. Despite a variety of positive results, there appears to be little correlation between transit time and lateral or vertical distance from the injection site. Data analysis produced tr aditionalshaped dye recovery curves in some locations, albeit stretched out over hundreds and possibly even thousands of days beyond dye injection. The results strongly suggest that chemical or sewage spills in the vicinity of the dye injection sites woul d quickly enter multiple sites in the cave system, and could persist for years. A Comparative Study Between the Karst of Hoa Quang, Cao Bang Province, Vietnam and Tuscumbia, Alabama, USA Gheorghe M. Ponta Geological Survey of Alabama Nguyen Xuan Nam, Vietnam Institute of Geos and Nat Res Ferenc L F orray, Babes Bolyai University Florentin Stoiciu, R&D Nat. Inst. for Nonferrous & Rare Metals Viorel Badalita, R&D Nat. Inst. for Nonferrous & Rare Metals Lenuta J Enache R&D Nat. Inst. for Nonferrous & Rare Metals I oan A Tudor, R&D Nat. Inst. for Nonferrous & Rare Metals Some of the most beautiful karst features created by the dissolution of limestones are residual hills with steep or vertical sides rising from a flat plain, known as tower karst. Tower karst to be developed requires a mean annual temperature of minimum 170C to 180C and 1,000 to 1,200 mm/m of annual rainfall (Jakucs, 1977). Two sites matching this criteria were selected: the karst of Hoa Quang District, Cao Bang Province, Vi etnam, and Tuscumbia, Colbert County, Alabama, U.S.A. Preliminary observations regarding similarities and differences between these two sites are presented in this paper. The Hoa Quang karst area is located in the northern Vietnamese Province o f Cao Bang. In 2014, a large number of karst springs, caves, sinking streams, and karst landforms were identified. Eighteen water samples were collected and analyzed for anions, cations, oxygen and hydrogen stable isotope ratios. The pH values are typ ical for karst waters and ranged from 7.23 to 7.97. Specific conductance values ranged from 153.2 to 421.6 S/cm, the total alkalinity as CaCO3 varies from 125 to 207 mg/L, carbon dioxide varies between 40.8 and 123.4 mg/L, whereas the values for the total hardness (as CaCO3) are between 143 and 220 mg/L. The local meteoric water line, based on our measurements is d by Craig (1961) and revised by Rozanski, et al. (1993). The intercept value differs very slightly from both local and global water lines. Due to the short sampling period, the information provided by the water stable isotopic composition is limited. Carbonate rocks underlie many areas of north Alabama. Karst features can be found around Tuscumbia, in northwestern Alabama, which is part of the Tennessee Alabama Georgia karst area that is called TAG. TAG has the highest concentration of caves in United States, and home for a few large springs. Tuscumbia Spring is a municipal water supply with a base flow of 1,500 L/s. The field parameters measured in January 2014 were: pH 6.81, specific conductance 292 uS/cm, and temperature 5.310 Celsius. In 19891 990, the Geological Survey of Alabama conducted an extensive investigation in the area, performing dye studies in storm water drainage wells (SDW 1 through SDW 20) to

PAGE 34

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 33 define the recharge area of Tuscumbia Spring. The storm water drainage wells can be a pot ential source of contamination for the springs. Two rock samples from Vietnam and one from Tuscumbia, Alabama (U.S.A.) were collected and examined using the X ray diffraction (XRD) analysis, microscopic analysis in polarized light and Differential Scanning Calorimetry Thermogravimetry (DSC TG) analysis. The quality of limestones in Vietnam and Tuscumbia (38.7 percent and 39.6 percent versus 31.10 percent calcium concentration) and the amount of pre cipitation (1,500 to 2,000 mm/m in Vietnam versus 947 m m/m to 1,960 mm/m per year in Tuscumbia) are comparable. Thick limestone beds, massively jointed, combined with frequent tectonic uplifts and a complex geologic pattern result in the tower karst landscape in Vietnam versus a leveled landscape in Tuscumbi a, Alabama. Tectonics is the primary driver for the formation of tower karst landscape in Cao Bang Province, Vietnam. Wednesday, October 7th 3 :30pm 5:15 pm Karst Geology Karst Paleo Collapses and Their Impacts on Mining and the Environment in Northern China Gongyu Li, Xian Research Inst. of China Coal Tech. & Eng. Wanfang Zhou, ERT, Inc. Karst paleo collapses are unique collapse structures widely found in the coal measures of northern China. Their geometric dimensions and internal properties indicate that a compound dissolution of carbonate and gypsum rocks may contribute to their formation. When these collapses are permeable to groundwater flow, they hydraulically connect the coal seams and the karst aquifers, which is a prerequisite for water inru shes during coal mining. Over the last 40 years, water inrushes through these collapses have caused fatalities, economic losses, and degradation in the environment in northern China. Determination of locations and hydrogeological characteristics of the kar st paleo collapses are essential in preventing water inrush incidents through them. Advanced geophysical prospecting, aquifer testing and accompanied dye tracing are effective approaches to investigating these structures. The Sandstone Karst of Pine County, Minnesota Beverley Lynn Shade, University of Texas E Calvin Alexander, Jr., University of Minnesota Scott C Alexander, University of Minnesota The glaciated, forested landscape of central Pine County in east central Minnesota contains a series of sinkholes, stream sinks, springs and caves. The features are formed in Precambrian Hinckley Sandstone and overlying unconsolidated glacial deposits. This is a sandstone karst. The features serve the same function as in carbonate karst terrains: sinkhol es and caves focus recharge into a heterogeneous subterranean flow system that discharges into springs. The Hinckley Sandstone is a quartz arenite. No carbonate grains or cements have been found in sandstone samples from the sinkhole area, nor is there evi dence that calcite solution controls bedrock permeability. Three parameters appear to control the distribution of sinkholes: depth to bedrock, type of underlying bedrock, and meter scale heterogeneity of surface sediments. The permeability structure of the Hinckley Sandstone appears to be controlled by fractures and depositional features at centimeter to meter scale. Field mapping in the area has revealed 309 karst features: 237 sinkholes, 25 stream sinks, 32 springs and 15 caves. Recent LiDAR coverage indi cates that there are many more sinkholes and other karst features than the original mapping was able to locate. Interpretation of the LiDAR images is challenging because karst processes, glacial processes and human activity have all produced natural and an thropogenic closed depressions of a variety of sizes and shapes in this landscape. A Proposed Hypogenic Origin of Iron Ore Deposits in Southeast Minnesota Karst E Calvin Alexander, Jr. University of Minnesota Betty J Wheeler, University of Minnesota From 1942 through 1968 there was an active iron ore mining industry in western Fillmore, eastern Mower and southern Olmsted Counties of Minnesota. This iron mining district was 250 miles south of, and the ores were a billion years younger than, the ores of the classic iron mining districts in northern Minnesota. The high grade iron ore was mostly goethite and hematite and occurred as near surface relatively small pods which unconformably filled paleokarst depressions in the Devonian Spillville Formation and the Ordovician Stewartville Formation. The source of the iron has long been cryptic. The available field and textural evidence is consistent with a hypogenic origin of these iron deposits. Before the current Mississippi River drainage system was incis ed, regional ground water flow

PAGE 35

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 34 systems could have emerged through the karst conduits in the Paleozoic carbonates. The waters in the deeply buried aquifers underlying this area currently are anoxic and enriched in dissolved ferrous iron and would have been more so before the entrenchment of the Mississippi River reorganized the regional ground water flow system. When that water emerged into the atmosphere the ferrous iron would have quickly been oxidized by a combination of biotic and abiotic processes produ cing the ferric oxide ore at the spring orifices. Numerous springs and seeps in Minnesota are currently building iron oxide deposits at their orifices. Down the Rabbit Hole: Identifying Physical Causes of Sinkhole Formation in the UK Tamsin Green, University of Leeds Heavy precipitation in the UK in February 2014 induced ground subsidence and consequently a rapid increase in the frequency of sinkhole occurrences. These new sinkhole collapses emph asize the need to further analyze the causes of the increased occurrence by investigating the relative importance of various surficial factors. Malham and the Mendips are two areas of particular interest, since both are underlain by limestone bedrock and are susceptible to subsidence. This is due to limestone being primarily permeable in joints, and so it dissolves to form an extensive network of karstic caves. It was therefore useful to compare two sites of similar geology, both from the Triassic and Jurassic periods, as this controlled the amount of presentl y exposed limestone from past glacial retreat, for accurate comparison of susceptibility. Susceptibility maps of the two areas were created by integrating GIS application and statistical methods to develop algorithms to address the issue of dissolution The maps aim to identify the physical surficial conditions, in addition to heavy precipitation that exacerbates subsidence development. Statistical testing of the GIS data indicated that in Malham, slope is the most significant parameter (Kruskal Wal lis, formation; while in the Mendips altitude is the most significant parameter (Kruskal p<0.001 respectively). Curvature appeared less statistically signific ant with fewer values reported from post hoc Mann Whitney U tests. This integrated geological mapping and statistical approach will prove useful in delineating susceptibility zones in areas within the UK. Relay Ramp Structures and Their Influence on Groundwater Flow in the Edwards and Trinity Aquifers, Hays and Travis Counties, Central Texas Brian B. Hunt, Barton Springs/Edwards Aquifer Cons. Dist. Brian Smith, Barton Springs/Edwards Aquifer Cons. Dist. Alan Andrews, Barton Springs/Edwards Aquifer Cons. Dist. Douglas Wierman, Meadows Center for Water & Environ Alex S Broun, Hays Trinity Groundwater C onservation Dist. Marcus O Gary, Edwards Aquifer Authority The Cretaceous Edwards and Middle Trinity Aquifers of central Texas are critical ground water resources for human and ecological needs. These two major karst aquifers are stratigraphically stacked (Edwards over Trinity) and structurally juxtaposed (normal faulting) in the Balcones Fault Zone (BFZ). Studies have long recognized the importance of faulting on the development of the karstic Edwards Aquifer. However, the influence of these structures on groundwater flow is unclear as groundwater flow appears to cross some faults, but not others. This study combines structural and hydrological data to help characterize the potential influence of faults and relay ramps on groundwater flow within the karstic Edwards and Middle Trinity Aquifers. Detailed structure contour maps of the top of Walnut Formation in the study area were created from a geologic and drillers logs. The data were then contoured in Surfer (Kriging) with no faults. Structure contour surfaces revealed detailed structural geometries including linear zones of steep gradients (interpr eted as faults) with northeast dipping zones of low gradients (interpreted to be ramps) between faults. Hydrologic data (heads, dye trace, geochemistry) were overlaid onto the structure contour maps in GIS. Results for the Middle Trinity Aquifer suggest r elay ramps provide a mechanism for lateral continuity of geologic units and therefore groundwater flow from the Hill Country (recharge area) eastward into the BFZ. Faults with significant displacement (>100 m) can provide a barrier to groundwater flow by t he juxtaposition of contrasting permeabilities, yet flow continues across fault zones where ramps exist, or where permeable units are juxtaposed with other permeable units. In the Barton Springs segment of the Edwards Aquifer the primary flow path defined by dye tracing and heads is coincident with the Onion Creek relay ramp dipping to the northeast. This work addresses the lateral continuity (intra aquifer flow) of the Edwards and Trinity Aquifer systems, which has importance for conceptual models and ulti mately resource management.

PAGE 36

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 35 Wednesday, October 7th 5 :30pm 6:30 pm Poster Presentation (alphabetical) Goliaths Cave, Minnesota: Epigenic Modification and Extension of Preexisting Hypogenic Conduits E. Calvin Alexander, Jr., University of Minnesota Scott C. Alexander, University of Minnesota Kelton Barr, Braun Intertec Andrew Luhmann, University of Minnesota Cale Anger, University of Minnesota (deceased) Goliaths Cave is developed in the Ordovician Dubuque and Stewartville Formations of the Galena Group in Fillmore County, MN. The cave currently functions as an epigenic karst system with allogenic surface water sinking into the cave and a vadose stream running through the cave and resurging at springs a few kilometers away. Passages in the cave ar e locally controlled by vertical joints in the nearly flat lying carbonate bedrock, but the water flow directions often do not correspond to the systematic joint directions. The cave contains straight, joint controlled passages that appear to predate the current epigenic drainage systems. These old passages contain hypogenic features and are connected and modified by distinct, younger epigenic passages often with very sharp transitions back and forth between the two passage types. The epigenic flow in cises vadose canyons into the hypogenic passages. The hypogenic passages represent ancient, deep, compartmentalized flow systems that predate the present topography. The concept ancient is poorly constrained, however. These ancient cave passages are b eing reactivated by epigenic processes, while undergoing destruction by general erosion of the landscape. Numerical Simulation of Spring Hydrograph Recession Curves for Evaluating Behavior of the East Yorkshire Chalk Aquifer Nozad Hasan Azeez, University of Leeds Landis Jared West, University of Leeds Simon H Bottrell, University of Leeds The Cretaceous Chalk aquifer is the most important in the UK for the provision of water to public supply and agriculture. The Chalk has both matrix and fract ure porosity and is thus best considered as a dual porosity aquifer system. Although the matrix porosity is large, typically around 0.35 in the study area of East Yorkshire, UK (ESI, 2010), pore diameters are typically very small, and the water contained i n them is virtually immobile. The high permeability fracture network is responsible for the ability of water to drain; spatial variations in fracture network properties mean conventional approaches to aquifer characterization such as borehole pumping tests are of limited utility. Hence this study attempts to better understand the flow system and characteriz e aquifer properties from the recession response seen at springs during the spring/summer period when recharge is minimal. This approach has the advantage that spring hydrographs represent the sum of the response from entire catchments. This paper reports numerical modeling for simulating aquifer and spring responses during hydrological recession. Firstly, available geological and hydrogeological info rmation for the study area was used to develop hydrogeological conceptual models. Four different numerical models have been constructed representing four possible scenarios that could represent the aquifer in the selected area. These are: single reservoir aquifer, double reservoir aquifer, single reservoir aquifer with highly permeable vertical zone intersecting the spring location and single reservoir aquifer containing tunnel shaped highly permeable zone at the spring elevation respectively. The sensitiv ity of spring recession response to various external and internal parameter values was investigated, to understand relations between spring recession, hydrological inputs (recharge) and aquifer structure. Spring hydrographs from the real aquifer were compared with the hydrographs generated from models, in order to estimate aquifer properties. The work aims to identify the utility of spring hydrographs in eliciting aquifer permeability structure, as well as identifying the conceptual scenario which best represents the Chalk Aquifer in East Yorkshire, UK. Seeps and Springs at a Platteville Observatory on the River Bluffs BJ Bonin, Minnesota Department of Natural Resources Greg Brick, Minnesota Department of Natural Resources Julia Steenberg, Minnesota Geological Survey Residential building construction along the Mississippi River bluffs in the 1970s created a unique enclosed outcrop of the Late Ordovician Platteville Limestone at Lilydale, Minnesota. This outcrop was examined in early 2013 after a newl y formed spring flooded an elevator shaft the previous year, drawing attention to the foundation conditions. The Lexington Riverside property is a six story condominium complex constructed within the top of the bluff. A two level underground parking ga rage was built into the bluff. Bedrock was mechanically excavated to accommodate

PAGE 37

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 36 the construction of the building, creating an unweathered rock surface. The space between the structure and the excavated rock face, running for 150 meters, was roofed over, a nd is used as a utility space. At least three dominantly carbonate members of the Platteville Formation are visible: Mifflin, Hidden Falls, and Magnolia, in ascending order. The foundation of the structure was constructed on the lowermost Platteville limes tone and Glenwood shale and is tile drained to the nearby river gorge. Most of the seeps and springs on the property, both inside and on the grounds, belong to the three Platteville spring lines identified for the Twin Cities Metropolitan (TCM) area by Brick (1997). Groundwater emanates from both vertical joints and horizontal bedding plane partings within the Platteville Limestone and at the Platteville Glenwood Shale contact. Overall, the hydrostratigraphic attributes of this site are consistent wi th how the Platteville has been recently characterized in the TCM area in a fractured bluff edge setting (Anderson et al., 2011). The enclosed outcrop features many seep and spring related mineral deposits. Most notable were the iron stained flowstone and microgours near the seeps and springs along fractures in the limestone, and calcite rafts on the surfaces of the pools. At some damp locations a fungal ecosystem has developed. Gypsum beards have grown in dry portions of the cavern. This man made cavern, and others nearby, present unique opportunities to research groundwater flow in fractured bedrock settings. Studying the spring locations relative to joints and bedding, changes in spring flow rate over time, and mineral deposition rates, are possi ble in this accessible location without the complication of surface water inputs or instrumental interference from the general public. Using Electrical Resistivity Imaging to Characterize Karst Hazards in Southeastern Minnesota Agricultural Settings Toby Dogwiler, Missouri State University Blake Lea, Missouri State University Much of the Driftless Area of southeastern Minnesota is underlain by karstified carbonate bedrock. Land use in this karst terrain is dominated by agriculture, including row cro p and dairy operations. The karst in this region is often mantled with up to 15 m of soil and unconsolidated sediments. As a result, underlying karst hazards such as incipient sinkholes are often hidden until they are suddenly revealed by the collapse of subsurface voids. Regionally, the economics of the dairy industry is causing a trend toward the consolidation and expansion of existing operations. As concentrated animal feeding operations (CAFO) or feedlots expand, state and local agencies are charged with enforcing regu lations designed to protect environmental and water resources in agricultural areas. One of the key challenges in reviewing and siting expanded dairies is identifying potential karst hazards, particularly where they might undermine manure storage facilities or where they occur on croplands where manure is applied. Uncertainty about the location of karst hazards relative to proposed feedlot facilities is one of the reasons that feedlot expansions in the Driftless Area are often controversial. Recently, the Minnesota Pollution Control Agency enacted strict guidelines that severely limit bedrock removal in order to facilitate the construction of manure storage containments. The purpose of this rule is to ensure that a minimum separation is maintained between intact bedrock and the containment liner to provide the opportunity for attenuation within the soil of contaminants that could potentially leak if the containment structure is compromised. Electrical Resistivity Imaging (ERI) techniques have been em ployed to screen for karst hazards during the planning phase of feedlot expansions, and where present, to more accurately characterize the nature of the karst hazard. Because depth to bedrock is highly variable in the karst terrain of southeastern Minneso ta, ERI has also been a useful tool to characterize this spatial variation under proposed manure containment sites. In this study, ERI was performed using a 56channel AGI Supersting system with post processing of the data in EarthImager software. Dipo le Dipole and Wenner electrical resistivity arrays have been the most useful for identifying karst hazards. Electrode spacing of 3 to 5 m has provided a good balance between depth of image and the spatial resolution necessary to locate and identify karst hazards. Soil boring data, which is typically collected during pre construction site investigations, is critical to the interpretation of ERI data. Although individual sites vary, most surface materials in southeastern Minnesota have resistivities that f all within predictable ranges: 20 80 ohm m for soils, 80100 ohm m for epikarst and weathered residuum, and >100 ohm m for bedrock. Karst voids in the subsurface typically display resistivities greater than 1000 ohm m, providing good contrast with the r esistivities of the surrounding bedrock. ERI has been an effective tool in identifying karst hazards in agricultural settings of southeastern Minnesota. In addition to improving pre construction site assessments, ERI has also helped to reduce poten tial controversy surrounding the karst hazards of proposed projects by providing more certainty about the underlying geology.

PAGE 38

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 37 Analysis and Prevention of Sinkhole Collapses During the Reconstruction and Extension of the Guang Qing Freeway, China Zh anfei Gu, Tongji University; Qi Liu, Tongji University; Y aoru Lu, Tongji University; Zhenming Shi, Tongji University; Gaoyu Su, Guangdong Freeway Co, Ltd. The GuangQing freeway reconstruction and extension project is located in Guangdong province, China extending from Guangzhou city to Qingyuan city. The total project length is 57.6 kilometers (km), 33 km of which is in a karst area where sinkhole collapses are common. Therefore, there were concerns about the safety and security of the roadbed, bridge piles, and other structures associated with the original roadway. The study addresses karst development, covering layer (overburden), and external factors that contribute to catastrophic sinkhole collapse along the project. The study found that the formatio n of sinkhole collapses is affected significantly by the degree of karst development and soil characteristics, including thickness. Collapses may be triggered directly by external factors that disturb the natural equilibrium. Such factors include heavy rai nfall that impacts soil and groundwater conditions, as well as vibration and groundwater flow disruptions that are caused by pile construction and other engineering activities. Prevention measures were adapted to local conditions and optimized based upon s afety, reliability and costeffectiveness. Study on the Critical Velocity of Groundwater to Form Subsidence Sinkholes in Karst Area Fuwei Jiang, G uizhou Institute of Technology Mingtang Lei, Institute of Karst Geology Dai Jian ling, Institute of Karst Geology Subsidence sinkholes in karst area are a common geological hazard causing disaster accidents. However, the critical hydraulic conditions of forming subsidence sinkholes have not been well understood. Based on the theories of pipe flow, this paper derives expressions of a critical hydraulic condition for assessing whether it leads to subsidence sinkholes. A case study with samples of the cohesive soils taken from the Wuxuan county, Guangxi province, China, was conducted to evaluate the derived critical value. Combining the derived critical value and the monitoring data of hydraulic conditions in the study field, our results indicate that subsidence sinkholes are not forming under the current hydraulic conditions. Finding Springs in the File Cabinet Mason Johnson, Minnesota Pollution Control Agency Ashley Ignatius, Minnesota Pollution Control Agency The Minnesota Pollution Control Agency (MPCA), in partnership with other agencies, is currently undertaking comprehensive sub basin assessments statewi de over a ten year period. Southeast Minnesota has over 17,500 kilometers of perennial and intermittent streams, making the task of comprehensive sub basin assessment challenging; the task is further complicated by karst geology. In the summer of 2014, a pilot project began between the MPCA and Minnesota Department of Natural Resources (DNR) to digitally preserve paper documents which capture qualitative and quantitative data about the hydrology, water chemistry, geomorphology, biology, land use and karst features of southeast Minnesota streams. The paper documents in file cabinets were not in an accessible or easy to use format; as such, they were in a data silo. The task was to preserve the documents so as to make the data usable by converting the docu ments into a digital format (Adobe PDF, GeoTIFF, ESRI Feature Class). To date, more than 4,000 documents (of an estimated more than 12,000) have been converted, made text searchable, prepared for storage in a document management system, and made more accessible through a geographic information system (GIS). This previously inaccessible data is an important piece in understanding the karst region of southeast Minnesota. Within the documents scanned thus far, over 400 springs and other karst features have been identified, which are not currently recorded in Minnesotas Karst Feature GIS Database. Evaluation of First Order Error Induced by ConservativeTracer Temperature Approximation for Mixing in Karstic Flow Philippe Machetel, Geosciences Montpellier David A. Yuen, Univ. of Minn. and China Univ. of Geosciences Fluid dynamics in karst systems is complex due to the heterogeneity of hydraulic networks that combine the Porous Fractured Matrix (PFM) and the interconnected drains (CS). These complex dynamic systems often need to be treated as black boxes in which only input and output properties are known. In this work, we propose to assess the first order error induced by considering the temperature as a conservative tracer for flows mixing in karst (fluv io karst). The fluvio karstic system is treated as an open thermodynamic system (OTS), which exchanges water and heat with its surrounding. We propose to use a cylindrical PFM drained by a water saturated cylindrical CS, connected on one side to a sinkhole zone and, on the other side, to a resurgence flowing at the base level of

PAGE 39

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 38 the karstic system. This framework allows us to develop the equations of energy and mass conservation for the different parts of the OTS. Two numerical models have been written to s olve these equations: 1) the so called AW (for Adiabatic Wall) configuration that assumes a conservative tracer behavior for temperature with no conductive heat transfer, neither in the liquid, nor in the PFM or even through the wall separating the CS from the PFM; and, 2), the CW (for Conductive Wall) configuration that takes into account the heat and mass transfers in water and from water to aquifer rocks both in the CS and in the PFM. Looking at the large variability of karstic system morphologic propert ies, dimensionless forms of the equations have been written for both AW and CW configurations. This approach allows us to gather the physical, hydrological and morphological properties of karstic systems into four dimensionless numbers: the Peclet, Reynold s, Prandtl and dimensionless diffusivity numbers. This formalism has been used to conduct a parametric exploration across several orders of magnitude based on the Peclet and the Reynolds numbers. The final errors, between the AW and CW configurations, rema in less than 1% across the entire parametric range. The combination of error curves bounds a closed volume in error space that gives a first upper bound of the error made by considering the temperature as a conservative tracer. Applying the method to an il lustrative example of karst allows us to reach a first order error within a few degrees C. The Application of Passive Seismic Techniques to the Detection of Buried Hollows Michael G Raines, British Geological Survey Vanessa J Banks, British Geologica l Survey Jonathan E Chamb ers, British Geological Survey Philip E Collins, Brunel University London Peter F Jones, University of Derby Dave J. Morgan, British Geological Survey James B Riding, British Geological Survey Katherine Royse, British Geologica l Survey Pilot studies involving the use of passive seismic techniques in a range of geological settings and applications (e.g., mapping bedrock, studies of soil erosion and Quaternary surficial mapping) have shown that it is a versatile, noninvasive and economic technique. This paper presents the findings of three case studies that used passive seismic techniques for the detection and characterization of buried hollows in carbonate rocks: 1) a buried hollow in the Cretaceous chalk at Ashford Hill in the Kennet Valley, a tributary of the River Thames, UK; 2) buried karst in the foundation excavations for wind turbines in Carboniferous limestone at Brassington, Wirksworth, Derbyshire, UK; 3) defining the extent of solution hollows that host terrestrial Mio Pliocene deposits, near Friden, Newhaven, Derbyshire, UK. Whilst case studies 2) and 3) are focused on areas of buried dolines, the geological context of the Ashford Hill site is more complex, including a deformation hollow with an uplifted pinnacle of chalk bedrock at the centre. The data were collected using Tromino, a threecomponent, broadband seismometer, to measure background ambient noise (microtremors induced by wind, ocean waves, industrial machinery, road and rail traffic etc.). The Tromino is a small, portable device with an operating range of 0.1 Hz to 30 kHz and interpreted using proprietary software (Grilla), which subjects the data to Fourier transformation and smoothing. Where possible, shear wave velocities have been calibrated using borehole data or parallel geophysical techniques. In each case, the karst features were defined by Nakamuras horizontal to vertical (H/V) spectral ratio technique, where microtremors are converted to show impedance contrasts (velocity x density), or a pseud o layered seismic stratigraphy of the near surface along each profile. An additional benefit of the use of this technique is its depth of penetration and potential for defining the structural and lithological context of the hollows, thereby contributing to the process understanding associated with their formation. To this end the technique has helped define the structural discontinuities (fault, joint or bedding) that guide formation of the hollows. New Methodologies and Approaches for Mapping Forested Ka rst Landscapes, Vancouver Island, British Columbia, Canada Tim Stoke s, Vancouver Island University Paul Griffiths, Consultant Carol Ramsey, Vancouver Island University Mapping is an essential tool for land management and is typically used to assess the nature and characteristics of a land surface, along with its resource features and values. Mapping of karst landscapes is of particular importance for the temperate rainfo rests Vancouver Island on the west coast of British Columbia (BC), where both forestry and natural resource development activities occur. A set of BC Government standards for mapping karst have been developed at varying scales (reconnaissance, planning lev el and detailed), and incorporate procedures to assess the potential and vulnerability of karst, applying qualitative analysis of various surface and subsurface karst attributes. Previously, karst maps have been compiled by GIS mapping software and generat ed as static graphic images. However, it is now possible to make maps more accessible and interactive by uploading them as overlays within Google Earth (as KML files) or other imagery platforms. It is also possible for these maps to be taken into the field using tablets, phones or iPads, allowing for on site data collection and resource evaluation. A key focus of this research is the

PAGE 40

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 39 development of a `Karst Map of Vancouver Island that outlines the known extent of the karst and the likely contributing nonkarst catchments, and also identifies regions of varying karst potential/vulnerability and notable karst areas/features. Trials have been completed to see how detailed imagery of karst features and areas collected by UAV technology can be used for karst ev aluation. Karst areas have also been mapped using LiDAR, which has the potential for both detecting surface karst features beneath the forested canopy and for assessing the overall sensitivity of forested karst areas. LEPT, a S implified A pproach for A sse ssing K arst V ulnerability in R egions with S parse Data; a C ase S tudy from Kermanshah P rovince, Iran Kamal Taheri, Kerma nshah Regional Water Authority Milad Taheri, Bu Ali Sina University Fathollah Mohsenipour, Kermanshah Reg. Water Authority There are a v ariety of widely used methods for porous aquifer protection to assess the vulnerability of groundwater resources, such as DRASTIC; Depth to water, net Recharge, Aquifer media, Soil media, Topography, Impact of vadose zone, and hydraulic Conductivity, SINTA CS; depth to ground water (S), effective infiltration (I), unsaturated zone attenuation capacity (N), soil attenuation capacity (T), hydrogeologic aquifer characteristics (A), hydraulic conductivity range (C) and hydrological role of the topographic slope (S). And GOD; Groundwater occurrence, Overlying lithology and Depth of groundwater However, some more limited methods (including EPIK; Epikarst development, Protective cover, Infiltration conditions and Karst network development, PaPRIKa; Protection of karst Aquifers based on their Protection, Reservoir, Infiltration and Karstification type and COP; Concentration of flow, Overlying layers and Precipitation regime) are also suggested for karstic aquifer vulnerability analysis. The latter methods are app lied using different parameters such as karst network development, depth of karstification, and protective cover. Due to the nature of the data, these methods are highly affected by local and regional climate conditions. Data gathering of these methods is difficult, time consuming and needs a full understanding of karst system. Data shortages, especially those related to karst formations in some parts of the world including the west part of Iran, and crucial demands for utilizing water resources demonstrat e a great appeal to find a representative method for evaluation of these regions. Conventional methods of karst aquifer evaluations cannot be properly applied in the absence of a required karst data base; therefore, there is a need for a method that could be applied with the least amount of available data. The LEPT method introduced in this paper is a simple approach which provides rough evaluation of the general information gathered from karst areas of the west of Iran combined with field experiments. This method, which utilizes four parameters to assess the vulnerability of karst aquifers, was applied to the karst areas of Kermanshah (a province in the west of Iran) for the first time. Results of this approach categorize karst plains into four zones with very high, high, low and very low sensitivity in terms of their vulnerability to environmental impact; these classes positively correlated with fiel d information. Determination of the Relationship of Nitrate to Discharge and Flow Systems in North Florida Springs Sam B Upchurch, SDII Global Corporation The Suwannee River Water Management District has collected quarterly discharge and water quality data from 30 1st and 2nd magnitude springs in the Suwannee River Basin since 1998. These data were collected quarterly well into the late 2000s and constitute a valuable database for characterizing spring discharge behavior. Trend and correlation analyses were used to compare the relationships of NO3 + NO2 (nitrate in this paper), specific conductance, and spring discharge. Trends were considered Data from 50% of the springs show that nitrate concentrations increase as discharge from the spri ng increases. Forty five percent of the remaining springs showed no correlation between discharge and nitrate, and only 5% (2 springs with poor data) have relationships where high discharge was related to lower nitrate concentrations. Twenty percent of the springs had positive correlations of specific conductance with discharge, 37% showed no correlation, and 43% had negative correlations between specific conductance and discharge. Most important in terms of understanding the plumbing of the conduit systems, 40% of the springs had positive correlations between nitrate and specific conductance, 48% showed no correlation, and 12% had negative correlations.

PAGE 41

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 40 Hydrochemical Characteristics and Formation Mechanism of Groundwater in the Liulin Kars t System, Northwestern China Min Yang, Inst. of Hydrogeology and Env. Geology, CAGS Feng'e Zhang, Inst. of Hydrogeology and Env. Geology, CAGS Sheng Zhang, Inst. of Hydrogeology and Env. Geology, CAGS Miying Yin Inst. of Hydrogeology and Env. Geology, CA GS Guoqing Wu, Inst. of Hydrogeology and Env. Geology, CAGS The Liulin karst system is typical of hydrogeological systems in northwestern China, with a group of springs as the dominant mechanism for regional groundwater discharge. To reveal the hydrochemical formation mechanism of the Liulin karst groundwater system, we studied the hydrogeochemical processes of karst groundwater in aquifers at the base of the hydrogeological investigation. Then starting from the chemical composition of karst groundwater together with the recharge runoff discharge process of groundwater systems, we analyzed the solutes origin and the dissolved mineral facies of the groundwater chemical composition. The results showed that the anionic and cationic compositions of karst water were different in recharge, runoff and discharge areas, with the main anions of HCO3 and SO4 2in recharge areas, and HCO3 and Clin runoff and discharge areas, as well as the main cationic for Ca2+ and Na+, of which the molar concentrations of Ca2+ was greater than Na+ in recharge areas and contrary to the runoff and discharge areas. Karst water was influenced by carbonate and evaporite dissolution while flowing through the aquifers, of which carbonate rock dissolution dominated in recharge areas, and evaporite rock dissolution increased to be the dominate lithology in runoff and discharge areas. Based on analysis of water rock interaction, the main dissolved mineral facies included dolomite, calcite, gypsum and halite. Dolomite is the most important dissolved mineral, followed by calcite and gypsum in recharge area, as well as calcite, gypsum and halite in runoff and discharge areas. Environmental E ffects of R ational U tilization of K arst G eothermal R esources in the North China Plain Wei Zhang, Inst. of Hydrogeology and Env. Geology, CAGS Wang Guiling, Inst. of Hydrogeology and Env. Geology, CAGS Liu Feng, Inst. of Hydrogeology and Env. Geology, CAGS As a clean energy, geothermal resources play an important role in protecting and improving the environment. This paper uses the well known North China Plain, with a carbonate karst reservoir, to demonstrate how significant the geothermal resource is, and t o highlight essential environmental differences between karst geothermal resources and coal resources. After analyzing the different geothermal areas, the maximum allowable drawdown method and mining coefficient method are applied as tools for evaluating g eothermal recoverable reserves. Evaluation procedures include geological characteristics research; select an assessment method, and, a final comprehensive evaluation. The evaluation results show that the karst geothermal resource for the North China Plain is 5.281017J. The exploration and utilization of the karst geothermal resource will beneficially result in energy savings of coal resources, of about 307t and emission reduction of carbon dioxide about 7.17107 t. Chemical Evidence for Biokarst Develo pment in the Ordos Basin (Northwest China) in Laboratory Batch Experiments Feng'e Zhang, Inst. of Hydrogeology and Env. Geology, CAGS Sheng Zhang, Inst. of Hydrogeology and Env. Geology, CAGS Miying Yin, Inst. of Hydrogeology and Env. Geology, CAGS Guoqing Wu, Inst. of Hydrogeology and Env. Geology, CAGS The present work is designed to simulate the dissolution of sulfate minerals (primarily gypsum) under various conditions of different bacterial cell numbers, temperatures and reaction times both in water rock and water rock bacteria systems by laboratory experiment. The amount and rate of dissolved sulfate rock and sulfate reduction rates were estimated using the experimental data. The results suggest that sulfatereducing bacteria promote gypsum dis solution and temperature plays a more important role on sulfate reduction rates than the number of bacteria. The dissolution of gypsum driven by bacterial sulfate reduction results in the formation of karst features. The research is an insight into biokars t, which provides a new perspective for the field of petroleum geology. Hydrocompaction Considerations in Sinkhole Investigations Edward D Zisman, Cardno ATC Stephen West, BTL Engineering Services, Inc. The cause of ground settlement is a significant concern in sinkhole investigations where the potential for shallow and deep seated instability in the subsurface is a major focus of the investigation. Complicating the investigation is the occurrence of hydr ocompaction of surficial soils caused by introduction of large amounts of surface water particularly from improper maintenance of rainfall runoff. This condition is usually followed by the subsequent loss of soil moisture during dry

PAGE 42

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 41 periods. This manuscript will discuss how hydrocompaction plays a role in the analysis of settlement in the investigation of sinkhole loss and how one can distinguish between hydrocompaction settlement and deepseated settlement (note, that hydrocompaction is one of many factors that can account for settlement of structures). It will consider the effects of soil density as it impacts hydrocompaction in the investigation of building distress. Also discussed are the results of laboratory tests of simulated hydrocompaction on fine s and samples in loose and dense states. In one of the tests, the formation of a collapse sinkhole occurred at the end of the test. Photographs depicting the sequence of soil failure are attached at the end of this paper. Thursday October 8th 8:30 am 12:10pm GIS Databases and Map ping of Karst Regions Creation of a Map of Paleozoic Bedrock Springsheds in Southeast Minnesota Jeffrey A Green, Minnes ota Dept. of Natural Resources E Calvin Alexander, Jr., University of Minnesota Springs are groundwater discharge points that serve as vital coldwater sources for streams in southeast Minnesota. The springs generally emanate from Paleozoic carbonate and siliciclastic bedrock aquifers. Use of systematic dye tracing began in the 1970s a nd continues through the present as a standard method for investigating karst hydrology and to map springsheds,. The work was accelerated in 2007 because of increased funding from the State of Minnesotas Environment and Natural Resources Trust Fund. A com pilation springshed map of dye traces conducted over the last several decades has been assembled for the region. In southeast Minnesota, the springs are the outlets of conduit flow systems in both carbonate and siliciclastic bedrock aquifers. Conduit flow dominates groundwater transport in carbonate aquifers and is an important component of groundwater flow in siliciclastic aquifers. Conduit flow in aquifers occurs independently of the presence or absence of surface karst features. The springsheds of t hese springs have three interacting components: Groundwater Springsheds (analogous to classic karst autogenic recharge areas), Surface Water Springsheds (analogous to classic karst allogenic surface runoff areas), and Regional Groundwater Springsheds. Surface Water Springsheds can be up to several orders of magnitude larger than the Groundwater Springsheds to which they contribute water. Surface Water Springsheds can feed surface flow into one or several stream sinks. Those multiple stream sinks may be in one or more Groundwater Springsheds. The leading edges of dye tracing breakthrough curves typically show groundwater flow velocities in the hundreds of meters to kilometers per day range in all of the bedrock aquifers tested. The width and duration of tails of breakthrough curves in these conduit flow systems vary with the bedrock aquifers. The Galena Group has Full Widths at Half Maximums (FWHMs) of a few hours and tails that are down to background in a few days. The Prairie du Chien Group also has FWHMs of hours but has tails that continue for weeks. The St. Lawrence and Lone Rock Formations have FWHMs of months to years. Media, Sinkholes and the UK National Karst Database Vanessa J Banks, H. J Reeves, E K Ward, E R Raycraft, H V Gow, D J R Morgan and D G Cameron, British Geological Survey The British Geolo gical Survey (BGS) maintains a number of databases that feed into hazard susceptibility assessments, including karst, landslide and mining susceptibility. The winter period from D ecember 2013 to January 2014 was one of, if not the most, exceptional periods of winter rainfall in the last 248 years for England and Wales. During this period the Jet Stream diverted easterly tracking cyclones along a more southerly route than is usual a cross the United Kingdom (UK). This resulted in south east and central southern England experiencing total rainfall values of 372.2 mm for this period, which was the wettest two month period since 1910. This period was associated with extensive flooding an d increased numbers of slope failures, landslides and sinkholes, which affected transport routes into and out of London, thereby generating considerable media attention. In addition to government and stakeholder requirements, the BGS experienced an unusual ly high level of enquiries from the public and the media pertaining to sinkholes, which put an additional strain on res ources, but is an acknowledged component of the BGS remit. During February alone, the BGS received reports of 19 sinkholes. The majority of these occurred in the Cretaceous chalk of southern England. Approximately half were not naturally occurring sinkholes, but were due to the collapse of anthropogenic features. Typically, the anthropogenic subsidence collapse features included: the collap se of chalk shafts associated wit h historic extraction of chalk for brickworks; the collapse of deneholes (medieval chalk workings for chalk for ground improvement), and chalk mine

PAGE 43

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 42 shaft collapses. This paper describes the National Karst Database, stakehol der requirements and how the BGS has responded with new and improved mechanisms for data collection, storage and dissemination. Shallow Depressions in the Florida Coastal Plain: Karst and Pseudokarst Sam B. Upchurch, SDII Global Corporation Thomas M Scott, SDII Global Corporation Michael C Alfieri, Water Resource Associates Thomas L Dobecki, SDII Global Corporation In Florida, shallow depressions (i.e., depressions <1 2 m in depth) on the land surface are often attributed to sinkhole development. However, it has become evident that there are at least six different mechanisms through which these depressions can form in geologically young cover sediments. These mechanisms include: 1. Cover subsidence sinkholes over shallow limestone; 2. Suffosion sin kholes over shallow limestone; 3. Cover settlement over shallow shell beds; 4. Large, aeolian deflation areas that resemble Carolina bays; 5. Depressions that mimic landforms developed on a shallow paleosol; and 6. Depressions created by pedodiagenesis ( i.e., conversion of smectite to kaolinite) in a soil forming environment. Of these, only the first two appear to represent traditional mechanisms for sinkhole development in eogenetic karst. Cover settlement over shell beds is poorly understood and inc orrectly attributed to sinkhole development processes. This type of depression has serious limitations in terms of cover thickness and shell content of the substrate. The last three mechanisms are pseudokarst created by aeolian and soil forming processes. In this paper we present examples of each and discuss their constraints and evidence. Sinkhole Vulnerability Mapping: Results from a Pilot Study in North Central Florida Clint Kromhout Florida Geological Survey Alan E Baker, Florida Geological Survey At the end of June in 2012, Tropical Storm Debby dropped a record amount of rainfall across Florida which triggered hundreds, if not thousands, of sinkholes to form which resulted in tremendous damage to property. The Florida Division of Emergency Management contracted with the Florida Department of Environmental Protections Florida Geological Survey to produce a map depicting the states vulnerability to sinkhole formation. The three year project began with a pilot study in three northern Florida counties: Columbia, Hamilton and Suwannee. Utilizing the statistical modeling method Weights of Evidence, results from the pilot study yielded a 93 percent success rate of predicting areas where the geology is conducive to sinkho le formation. Lessons learned and field mapping techniques developed during the pilot study are now being applied to map the entire States vulnerability to sinkhole formation. A Semi Automated Tool for Reducing the Creation of False Closed Depressions fr om a Filled LiDARDerived Digital Elevation Model John Wall, N orth Carolina State University Daniel H Doctor, US Geological Survey Silvia Terziotti, US Geological Survey Closed depressions on the land surface can be identified by filling a digital elevation model (DEM) and subtracting the filled model from the original DEM. However, automated methods suffer from artificial dams where surface streams cross under bridges and through culverts. Removal of these false depressions from an elevation model is difficult due to the lack of bridge and culvert inventories; thus, another method is needed to breach these artificial dams. Here, we present a semi automated workflow and toolbox to remove falsely detected closed depressions created by artificial dams in a DEM. The approach finds the intersections between transportation routes (e.g., roads) and streams, and then lowers the elevation surface across the roads to stream level allowing flow to be routed under the road. Once the surface is corrected to match the approximate location of the National Hydrologic Dataset stream lines, the procedure is repeated with sequentially smaller flow accumulation thresholds in order to generate stream lines with less contributing area within the watershed. Through mu ltiple iterations, artificial depressions that may arise due to ephemeral flow paths can also be removed. Preliminary results reveal that this new technique provides significant improvements for flow routing across a DEM and minimizes artifacts within the elevation surface. Slight changes in the stream flow lines generally improve the quality of flow routes; however some artificial dams may persist. Problematic areas include extensive road ditches, particularly along divided highways, and where surface flow crosses beneath road intersections. Limitations do exist, and the results partially depend on the quality of data being input. Of 166 manually identified culverts, 125 are within 25 m of culverts identified by this tool. After three iterations, 1,735 culv erts were identified and cataloged. The result is a reconditioned elevation dataset,

PAGE 44

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 43 which retains the karst topography for further analysis, and a culvert catalog. History and Future of the Minnesota Karst Feature Database Robert G Tipping, University of Minnesota Mathew Rantala, University of Minnesota E Calvin Alexander Jr., University of Minnesota Yongli Gao, University of Texas Jeffrey A Green, Minnesota Department of Natural Resources Since the 1990s, the University of Minnesota and the Minnes ota Department of Natural Resources have maintained a karst features database that is used to conduct research on karst processes and inventory karst features. Originally designed as a tabular database only, the karst features database developed into a sp atial database in 2002, with tabular data stored in Microsoft Access and a spatial component managed in ESRI ArcView. In 2012, the database was converted to a single, relational database platform, PostgreSQL, with both tabular and spatial components edite d in ESRI ArcMap. Custom editing forms are written in Visual Basic and are accessed in ArcMap sessions by ESRI add ins. The current database infrastructure allows for remote editing. Read only versions of the data are available in GIS/spatial format for public use via web services. Future development plans include links to water chemistry data, water level measurements and other ancillary data, along with the addition of vectors to represent dye traces and polygons for larger karst features. Legacy Dat a in the Minnesota Spring Inventory Gregory Brick, Minnesota Dept. of Natural Resources Past spring inventories have covered certain parts of Minnesota reasonably well, notably, the springs of the Minneapolis St. Paul metropolitan area and the southeastern Minnesota karst. But hitherto, there has not been a systematic effort to create a uniform statewide inventory. The first step, before hunting down new springs, was to compile existing data and the most fruitful source of hydrological legacy data for the Minnesota spring inventory was the DNR Fisheries files. Once entered into a GIS capable database, these spring locations can help seed the ground so that when crews finally do take to the field to map more springs, they will have known examples to work from. Good baseline and time series data should also help evaluate the impact of climate change and land use changes on Minnesotas springs over time. Development of Cavity Probability Map For Abu Dhabi Municipality Using GIS and Decision Tree Modeling Yo ngli Gao, University of Texas Raghav Ramanathan, Rizzo Associates Bulent Hatigoplu Rizzo Associates M. Melih Demirkan, Rizzo Associates Mazen Elias Adib, Abu Dhabi City Municipality Juan J Gutierrez, Rizzo Associates Hesham El Ganainy Rizzo Associates Daniel Barton Jr., Rizzo Associates Cavity collapse and settlement due to the presence of shallow solution cavities cause significant geotechnical and other engineering problems in certain areas within the Abu Dhabi City Municipality (ADM). A c avity probability map helps to identify regions that are more susceptible to the formation of cavities by identifying and analyzing influential factors contributing to its formation. Information relating to cavities was cataloged and reviewed based on avai lable data from the Geotechnical Information Management System (GIMS), which is a consolidated geotechnical database developed by the ADM. Geological and geotechnical subsurface conditions are obtained from previous site investigation campaigns performed i n the ADM region. All geotechnical, geological and cavity related information are stored in a GIS geodatabase system. Based on detailed literature review, primary factors influencing formations of cavities are identified: presence of soluble bedrock, depth to Gascharan Formation, cavity density, cavity thickness and distance to nearest neighbor. Another important factor in the formation of cavities is fluctuations in the groundwater table. Fluctuations are mostly caused by dewatering projects during constru ctions of buildings or infrastructure. The effect of variation in groundwater is not considered in this study due to lack of high quality data. A decision tree model based on cavity distribution was developed for cavity hazard assessment. The primary contr ols on cavity development are lithostratigraphic position or bedrock geology and depth to the soluble Gachsaran Formation. Most cavities tend to form in highly concentrated zones. Implementation of the decision tree model in ArcGIS resulted in a cavity pro bability map. This cavity probability map is mainly based on existing borehole data. Areas not fully mapped by boreholes must be re evaluated for cavity risk when new borehole data is available. Low Probability, Low to Moderate Probability, Moderate to Hig h Probability, High Probability, and Very High Probability areas were delineated in the probability map.

PAGE 45

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 44 Evaluation of Cavity Distribution Using PointPattern Analysis Ragh av Ramanathan, Rizzo Associates Yongli Gao, University of Texas M. Melih Demirka n, Rizzo Associates Bulent Hatipoglu, Rizzo Associates Mazen Elias Adib, Abu Dhabi City Municipality Michael Rosenmeier, Rizzo Associates Juan Gutierrez, Rizzo Associates Hesham El Ganainy, Rizzo Associates Presence of solution cavities of different sizes poses major engineering problems in some areas of Abu Dhabi City Municipality (ADM) underlain by soluble rocks such as gypsum, calcarenite, or mudstone. This is especially critical if they are located at a relatively shallow level and are likely to cause settlement or sudden soil collapse. The Gachsaran Formation, which is composed of interlayered mudstone and gypsum, underlies all of the ADM and is known to be vulnerable to cavity formation in the area. The mudstone and gypsum beds within the upper part of the Gachsaran Formation are prone to dissolution. Numerous sinkholes have been reported, particularly in the zone between Abu Dhabi International Airport and Mafraq. Many geotechnical and geological reports and surveys were reviewed for extracting inform ation related to geological and hydrogeological conditions causing cavities in the ADM. Information associated with cavities was cataloged and reviewed based on available data from an existing geotechnical borehole database maintained by the ADM. Cavity data obtained from borehole information was analyzed to examine cavity distributions based on the following factors: lithology, geographic clusters, cavity density, cavity size, depth to cavity, and depth to bedrock. All cavities were grouped into geographic clusters and lithological clusters for pointpattern analysis. Most cavities (87 percent) occur in mudstone or gypsum, or at an interface between these two rock types, which compose part of the Gachsaran Formation. The dissolution of carbonate and evapori tes within this formation causes the formation of subsurface voids in the ADM area. Geographically, majority of cavities occurred in the Shakhbout City area, hence pattern analysis including average nearest neighbor analysis, Morans I for measuring spatial autocorrelation, and G statistics for measuring high/low clustering were conducted in this area using spatial statistics tools in ArcGIS. Average nearest neighbor analysis and Morans I show that cavities are strongly clustered in this area with a high confidence level (>99 percent). General G statistics identified a high clustering (hot spot) of cavities with relatively high values of depth to cavity, depth to bedrock, and number of cavities per borehole. No highly clustered large cavities were detected by the General G statistics. Additionally, distances to the first through the 9th nearest neighbors were determined for cavities in different lithological materials and geographical clusters. Outcome of these spatial correlation and statistical analysis c an be used to conduct risk assessment and the probability of occurrences of future cavities. A Method of Mapping Sinkhole Susceptibility Using a Geographic Information System: A Case Study for Interstates in the Karst Counties of Virginia Alexandra L Todd University of Virginia Lindsay IveyBurden, University of Virginia Karst is a landscape underlain chiefly by limestone that has been chemically dissolved by acidic groundwater, producing subsurface voids that may lead to sinkholes at the surface if the overlaying soils can no longer support their own weight and collapse. The western counties of Virginia have a high concentration of karst areas due to widespread occurrence of carbonate rock exposures, and their geomorphic development within the Appal achian Mountains As a result, the Commonwealth of Virginia Hazard Mitigation Plan recommends that the Virginia Department of Transportation (VDOT) develop a method to determine the roadways and regions most susceptible to experiencing sinkholes, in an eff ort to reduce the possibility of reported sinkhole damage to property. While many noninvasive methods exist to detect subsurface voids, such as electric resistivity imaging, microgravity, ground penetrating radar, and seismic surveys, these methods are tim e consuming and costly. This study proposes the use of a geographic information system (GIS) to create a susceptibility map of regions in the karst counties of Virginia, and in particular along interstate highways, that are most susceptible to future sink hole development. Five factors that have previously been shown to play a role in the acceleration of sinkhole formation in Virginia include: bedrock type, proximity to fault lines, drainage class, slope of incised river banks, and minimum soil depth to bed rock. The analysis compares 1:24,000 scale maps of existing sinkholes developed by Virginia Department of Mines Minerals and Energy (DMME) with a series of maps representing differing combinations of each of the five factors to determine which weighted com bination is most appropriate to use for a final representative sinkhole susceptibility map. The layers representing each factor are created using publicly available tabular and spatial data taken from the United States Department of Agriculture (USDA) Soil Survey Geographic (SSURGO) Database, the United States Geological Survey (USGS) National Map, the USGS Mineral Resources Online Data, and the National Weather Service. The methodology used to gather information specifically from the SSURGO database is highlighted within this paper. Data from the SSURGO database is used to create the bedrock type, drainage class, and minimum soil depth to bedrock layers. A substantial benefit to

PAGE 46

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 45 this methodology is that the new technique can be adjusted to accommodate for sinkhole susceptibility in other karst regions, by simply adjusting the input layers to consider the specific geology of a particular region. Thursday October 8th 1:30p m 3:10pm Track A: Contamination of Karst Aquifers Evaluation of Veterinary Pharmac euticals and Iodine for Use as a Groundwater Tracer in Hydrologic Investigation of Contamination Related to Dairy Cattle Operations Larry Boot Pierce, Missouri Geological Survey Honglin Shi, Missouri University of Science and Technology Standard groundwater tracers such as Rhodamine WT, Fluorescein, Eosin and Tinopal CBX effectively provide a snapshot of hydrological conditions over a brief period of time and in a tightly controlled setting. However, in complex environmental situations wi th multiple potential sources, groundwater hydrologists are often seeking groundwater tracers that have extended longevity in the natural environment and the ability to directly pinpoint source locations. After reviewing operations of the nearby dairy it was determined that emerging contaminants, specifically two bovine veterinary pharmaceuticals (antibiotics), cephapirin sodium (CEPNa) and cephapirin benzathine (CEPB), and a sanitation agent, elemental Iodine (I) may have potential as extended longevi ty groundwater tracers if analytical methodology could be established. Initially, sample analysis indicated that cephapirin is undetectable in unconcentrated samples of lagoon wastewater at parts per billion (ppb) concentration; pre concentrated sample s which utilized solid phase extraction allowed for better detection at part per trillion level. Concentrated samples from one of the two lagoon cells sampled (cell #3), detected cephapirin at 13.14 ppt level, while cell #1 failed to detect any cephapirin present. Controlled laboratory testing later indicated that in a wastewater environment cephapirin degrades to approximately 20% of initial concentrations within 4 days, with complete degradation within 6 days. Degradation patterns in surface water and gr oundwater samples were less dramatic and at slower rates. Degradation curves of the surface and groundwater samples indicate that concentrations of cephapirin are still detectable for approximately 25 days. Unconcentrated Iodine samples collected in lagoon cells ranged from 50.896 ppb and 1,704.55 ppb with variations determined to be a result of the primary inflow of the lagoon. Cephapirins use as a long term groundwater tracer does not seem to be an immediate option. Further research may reveal that its degradation products are potentially useful as a tracer. In some instances, such as catastrophic discharges of large volumes of milk when samples can be collected and analyzed quickly, the use of cephapirin as an environmental tracer may prove possible The validity of pharmaceutical iodine as a groundwater tracer appears to be much greater than that of cephapirin. Iodine was detected in all of the environmental samples including the highly organic and anaerobic environment of the dairy wastewater lagoo n. This study concludes that iodine is capable of surviving the hostile wastewater environment. If sufficient data is collected to determine natural background levels, iodine may prove useful in determining hydrological connections between iodine laden dai ry effluent and the underlying groundwater. Karst Influence in the Creation of a PFC Megaplume Virginia Yingling, Minnesota Department of Health Perfluorochemicals (PFCs) are fully fluorinated organic chemicals used to produce a wide range of industrial and commercial products. Their extreme persistence and mobility in the environment and nearly ubiquitous presence in humans and wildlife has raised serious concerns regarding their environmental and human health effects. In the 1940s to 1970s, PFC bearing wastes were disposed of in three unlined landfills in Washington County, Minnesota. The resulting co mingled PFC plumes created a megaplume that contaminated over 250 km2 of groundwater and four major drinking water aquifers; affecting eight municipal water supply systems and thousands of private wells. Site investigations revealed that karst features, particularly in the Prairie du Chien Group (OPDC), and groundwa ter surface water interactions played a critical role in contaminant migration. Tracking of Karst Contamination Using Alternative Monitoring Strategies: Hidden River Cave Kentucky Caren Raedts Western University of Ontario Christopher Smart, Western Uni versity of Ontario Karst groundwater contamination presents great challenges for efficient monitoring because of rapid, discrete transport and the

PAGE 47

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 46 diversity of contaminants. Here a low cost approach is described and applied to Hidden River Cave, Kentucky where a long history of contamination has been experienced. Local knowledge was acquired through informal interviews and coupled with observations of contaminant residues, faunal distributions and fluorescence spectra in the cave. The resulting pattern s were interpreted using Google Earth and Street View to identify specific contaminant sources in the affected sub catchment of the cave. Despite success in matching contaminant sources with the contamination history and pattern, the informal nature of th e investigation renders it unacceptable as the basis for any intervention. But such low cost studies will be needed for the majority of contamination occurrences where financial resources are very limited. A radical revision of our adversarial approach to environmental management will be required for such a change to occur. Spatiotemporal Response of CVOC Contamination and Remedial Actions in Eogenetic Karst Aquifers I ngrid Y. Padilla, University of Puerto Rico Vilda L. Rivera, University of Puerto Rico Celys Irizarry, University of Puerto Rico The northern karst region of Puerto Rico has a long and extensive history of toxic spills, chemical waste, and industrial solvent release into the subsurface. High potential for exposure in the region has prompted aggressive remediation measures, which have extended for over 40 years. Of particular concern is contamination with chlorinated volatile organic compounds (CVOCs) because of their ubiquitous presence and potential health impacts. This work evaluates histo rical groundwater quality data to assess the spatiotemporal distribution of CVOC contamination in the karst aquifer system of northern Puerto Rico, and its response to remedial action in two superfund sites contaminated with CVOCs. Historical data collecte d from different information sources with different monitoring objectives is evaluated spatially and temporally using Geographic Information System (GIS) and statistical analysis. The analysis shows a significant extent of contamination that comes from mul tiple sources and spreads beyond the demarked sources of pollution. CVOCs are detected in 65% of all samples and 78% of all sampled wells. Groundwater shows continued level of contamination over long periods of time, demonstrating a strong capacity of the karst groundwater system to store and slowly release contamination. Trichloroethene and Tetrachloroethene are the most frequently found, although other CVOCs (e.g., Trichloromethane, Dichloromethane, Carbon Tetrachloride) are detected as well. The spatial and temporal distributions of CVOCs seem to be highly dependent upon the monitoring scheme and objectives, indicating that the data does not adequately capture the contamination plumes. Targeted remedial action using pump and treat (air stripping) and soil vapor extraction in two superfund sites has reduced concentrations over time, but the spatial and temporal extent of the contamination reflect inability to completely ca pture the heterogeneous plumes. Thursday October 8th 1:30p m 4:30pm Track A : Geophysical Exploration of Karst The Million Dollar Question: Which Geophysical Methods Locate Caves Best Over the Edwards Aquifer? A Potpourri of Case Studies from San Antonio and Austin, Texas, USA Mustafa Saribudak, Environmental Geophysics Associates The existence of caves represents a hazard for urban areas. Therefore it is important to know the size, position and depth of caves before building or reconstruction. Cavity imaging using geophysical surveys has become common in the San Antonio area since early 2000 although their use has been going on in other parts of country for the last 25 years. It appears from these studies that the resistivity imaging method has been the primary technique among others, such as gravity, ground penetrating radar (GPR ), magnetic, conductivity, and self p otential(natural potential). This study mainly describes resistivity imaging and natural potential data (NP), and some other geophysical data collected over several known and unknown caves during the study between the y ears of 2000 and 2015. All caves but one was encountered through drilling and/or excavation for building and utility lines or power pole reconstructions. The study area falls into the part of the Recharge Zone of the Edwards Aquifer region and it represent s a well developed karstified and faulted limestone in the Austin and San Antonio areas. 2D resistivity imaging data is presented as a colored 2 D electrical image of subsurface (i.e. a vertical cross section of the distribution of subsurface resistivity) Such a display section indicates low, medium, and high resistivity areas and the structural configuration of the subsurface geology. Based on our 15 years of experience working with resistivity data, we like to indicate that all range of resistivity val ues (low to high) can cause caves and voids. However, purely air filled cavities cause high resistivity anomalies whereas clay filled caves are the source of low resistivity anomalies. But it is rarely that caves are purely filled with air. A variety of se diments accumulates in caves and can be preserved more or less intact for long periods of time.

PAGE 48

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 47 Presence of sand and gravel and clay deposits, mineralization, faults and fractures, perched water in caves are the rules rather than the exception. It is clear from this body of work over the 15 years that the resistivity method does not always successfully delineate the location of caves and other karstic features. However, it provides significant information on the near surface geology and geological structure. The NP data, on the other hand, notably defines the location of karstic features. This study thus demonstrates that the resistivity method is not always a reliable predictive technique, but is useful in karst terrains to cover large areas quickly and, th e merits of integrating natural potential method, in order to reduce the ambiguity in the interpretation, are quite evident from the case studies. Rollalong Resistivity Surveys Reveal Karstic Paleotopography Developed on Near Surface Gypsum Bedrock Lewis Land, National Cave and karst Research Institute; Lasha Asanidze, Tbilisi State University Following flooding in September 2013, several areas in northern Eddy County, New Mexico were damaged by multiple sinkhole collapses. We conducted rollalong electrical resistivity (ER) surveys for subsurface cavities parallel to roads within and near the community of Lakewood, NM to guide road repairs. The rollalong method allowed us to generate exceptionally long, continuous ER profiles of the survey area. E R surveys attained a maximum exploration depth of 55 to 62 meters over a lateral extent of ~1000 meters, revealing an unconformable surface developed on gypsum bedrock, punctuated by shallow depressions. Subsurface stratigraphy, including clay rich valley fill sediments, and mudstone and gypsum of the underlying Seven Rivers Formation, can be identified by vertical and lateral variations in electrical resistivity. The irregular bedrock surface of the Seven Rivers Formation reflects paleotopography developed on that surface prior to its burial by floodplain sediment. Some of the negative paleotopographic features are probably filled sinkholes, which may be associated with shallower karstic features not imaged on the profiles. Integration and Delivery o f Interferometric Synthetic Aperture Radar (InSAR) Data Into Stormwater Planning Within Karst Terranes Brian Bruckno, Virgini a Department of Transportation Andrea Vaccari, University of Virginia Edward Hoppe, Virginia Cntr for Trans. Innovation & Research Scott T Acton, University of Virginia Elizabeth Campbell, Virginia Department of Transportation As part of a USDOT funded study focused on the implementation of satellite based Interferometric Synthetic Aperture Radar (InSAR) technology, the researchers integrated InSAR derived point cloud data into the transportation design process within karst terrain. The groups workflow included initial validation of InSAR data by acquiring over 1.67 million InSAR data points (various scatterers) which were then brought into a GIS dataframe and georeferenced to locations of mapped sinkholes. The technique was then applied to the evaluation of karst hazard of within a 40x40 km data frame located in the Valley and Ridge Province of Virginia. The group identified s ystematic kinematic differences in scatterer behavior with respect to their proximity to mapped karst geohazards, and used this method to screen for and identify unknown karst features, revealing numerous previously unidentified sinkholes. After validatin g the data with quantitative field correlations, the group then integrated the InSAR data into a traditional, CADD developed design ported into a GIS dataframe. This integrated data was then applied to a traditionally developed roadway project and used to optimize the location of stormwater management assets. In so doing, the group was able to develop open source data delivery method that allows greater flexibility, efficiency, and optimization of the infrastructure design and planning process, which can be developed collaboratively over geospatial platforms. This data integration offers lifecycle cost benefits, improvements to the safety of the traveling public, and protection of the environment, particularly in groundwater sensitive karst terranes. Disclaimer: The views, opinions, findings and conclusions reflected in this presentation are the responsibility of the authors only and do not represent the official policy or position of the US Department of Transportation/Office of the Assistant Secretar y for Research and Technology, or any state or other entity.

PAGE 49

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 48 Detection of Voids in Karst Terrain with Full Waveform Tomography Khiem T Tran, Clarkson University Michael McVay, University of Florida Trung Dung Nguyen, Clarkson University The paper presents an application of time domain surface based waveform tomography for detection of voids in karst terrain. The measured seismic surface wave fields were inverted using a full waveform inversion (FWI) technique, based on a finitedifference solution of 2 D elastic wave equations and Gauss Newton inversion method. The FWI was applied to real experimental data sets collected from twelve test lines at a karst site in Florida. Two of the test lines were located next to open chimneys to image t heir extents. Ten other test lines were located in an open and flat area without any void indication from the ground surface to detect an unknown void. The inversion results show that the waveform analysis was able to delineate embedded low velocity anoma lies, the void, and highly variable bedrock both laterally and vertically. Locations of the low velocity anomalies were consistent to the known open chimneys observed from the ground surface. The identified void was confirmed by an independent invasive tes t (standard penetration test, SPT). Characterization of Karst Terrain Using Geophysical Methods Based on Sinkhole Analysis: A Case Study of the Anina Karstic Region (Banat Mountains, Romania) Laurentiu Artugyan, West University of Timisoara Adrian C. Ardelean, West University of Timisoara Petru Urdea, West University of Timisoara To understand karst topography, we must determine both the nature and the factors that are defining dissolution processes in soluble rocks, as well as the drainage network resulting from these processes. The goal of this paper is to understand the underground drainage direction configuration and, also, the factors that are involved in surface water drainage of the Anina karstic region. In this study we used two complement ary geophysical methods, spontaneous potential (SP) and ground penetrating radar (GPR), applied in 5 sinkholes with a funnel shaped aspect. Four of these sinkholes are circular and one of them is elongated NW SE direction. Three of the studied sinkholes ar e representing a chain of sinkholes orientated west east. SP data describe the surface drainage, indicating drainage direction and/or moisture accumulation points. The GPR investigation utilizes electromagnetic pulses for the investigation of subsurfac e dielectric properties. GPR offers an image of the underground, showing possible bedding planes, in this case mostly along northsouth orientations. Besides, in two GPR profiles, we could identify an object that could be a cavity, in that point were on SP grid the values indicate small values, pointing out a link between those two geophysical results. Using SP and GPR methods we were able to show that the bottoms of these depressions are retaining more humidity and soil. In addition, the GPR profiles outlined several subsurface objects, at a depth ranging between 20 and 40 meters, which need a more thorough analysis. Our future work is intended to enrich our field data using SP and GPR methods, to compare with our first results. Also, we intend to in tegrate electrical resistivity tomography measurements in our analysis for better subsurface characterization. Thursday October 8th 3:30p m 5:30pm Track A : Karst Management, Regulations and Education __________________________________________ The Cost of Karst Subsidence and Sinkhole Collapse in the United States Compared with Other Natural Hazards David Weary, US Geological Survey Rocks with potential for karst formation are found in all 50 states. Damage due to karst subsidence and sinkhole collaps e is a natural hazard of national scope. Repair of damage to buildings, highways, and other infrastructure represents a significant national cost. Sparse and incomplete data show that the average cost of karst related damages in the United States over the last 15 years is estimated to be at least $300,000,000 per year and the actual total is probably much higher. This estimate is lower than the estimated annual costs for other natural hazards; flooding, hurricanes and cyclonic storms, tornadoes, landslides, earthquakes, or wildfires, all of which average over $1 billion per year. Very few state organizations track karst subsidence and sinkhole damage mitigation costs; none occurs at the Federal level. Many states discuss the karst hazard in their State hazar d mitigation plans, but seldom include detailed reports of subsidence incidents or their mitigation costs. Most State highway departments do not differentiate karst subsidence or sinkhole collapse from other road repair costs. Amassing of these data would raise the

PAGE 50

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 49 estimated annual cost considerably. Information from insurance organizations about sinkhole damage claims and payouts is also not readily available. Currently there is no agency with a mandate for developing such data. If a more realistic estimat e could be made, it would illuminate the national scope of this hazard and make comparison with costs of other natural hazards more realistic. Hazard of Sinkhole Flooding to a Cave Homini n Site and its Control Countermeasures in a Tower Karst Area, South China Fang Guo, Institue of Karst Geology, C AGS Guanghui Jiang, Institue of Karst Geology, C AGS Kwong Fai, Chinese Culture University Andrew Lo, Chinese Culture University Qingjia Tang, Institue of Karst Geology, C AGS Yongli Guo Institue of Karst Geology, C AGS Shaohua Liu, Inst itue of Karst Geology, C AGS Zengpiyan Cave, one of the most important cave hominin sites of the Neolithic in the South of China, was listed on the national register of cultural preservation sites in 2001. Large quantities of precious material in the Zengpiyan site were unearthed since the beginning of the trial excavation in 1973. These materials include hominin skeletal remains, fire pits, human burials, stone implements, tools fashioned from mollusk shells and ani mal or plant fossils. According to the historical record, ancient people lived in caves in the karst plain of Guilin. They moved out of the caves approximately 70 00 years ago. These cave hominin sites provide important material for understanding ancient en vironmental change and human prehistory. However, the exploratory shaft of Zengpiyan had no appropriate treatment after the initial excavation. Groundwater flooded the exploratory shaft due to frequent rises in the water table, resulting in collapse of the exploratory shaft shoring as well as some other serious damages. Even though some rescue and protection measures were taken, for example, slope supporting and backfill treatment, they failed to eliminate the hidden trouble caused by rapid fluctuation of g roundwater levels during the rainy season. Rapid urbanization is also affecting this region. Infrastructure construction of the city changes the hydrogeological conditions of the karst, increasing the area of impervious surface and risks of urban flooding. Moreover, an increase of extreme climate events may lead to frequent flooding of the site by groundwater. Therefore, a focused hydrogeological investigation was carried out to study the status of the site and karst development, and the mechanism of ground water movement at local and regional scales. These surveys include borehole drilling, electrical resistivity surveys, computed tomography (CT) scanning technology, dye tracer tests, groundwater monitoring, and hydrochemistry analysis. The results show that the site is located in the seasonal fluctuation zone of the groundwater. Water level in the karst aquifer is sensitive to rainfall. Continuous rainstorms lead to synchronous rising of the groundwater level in both the cave and the aquifer. In addition, surface runoff and urban sewers cannot discharge smoothly, resulting in surface water backflooding into the cave entrance. Therefore, controlling the recharge of groundwater and the influx of surface runoff, and dredging a groundwater discharge channel are all important in order to reduce the damage of flooding to the archaeological sites. Based on these detailed investigations and research results, countermeasures for flood control and archaeological site protection were put forward. We recommend that the en gineering measures should combine curtain grouting, drain construction, and effective water resources management for the entire basin. Even though the measures are feasible, we cant promise a perfect damage control of the ruins by water due to the complex hydrogeological conditions in the covered karst area. Case Studies of Animal Feedlots on Karst in Olmsted County, Minnesota Martin Larsen, Olmsted County Soil and Water Consv. District A unique area of Olmsted County is located a few miles southeast of Rochester by the small community of Predmore. Surface geology within the Orion Sinkhole Plain is dominated by a large array of sinkholes and limited soil cover over carbonate bedrock of t he Ordovician Stewartville and Prosser Formations. Dye trace studies completed by Eagle and Alexander (2007) have demonstrated that a large portion of the plains groundwater discharges into springs that feed two local trout streams. Land use in the area is mixed. For generations, local farmers have relied on livestock for stable income and profit. To put the 8,000 acre region into perspective, there are approximately 3,600 animal units located at 12 facilities which produce an estimated 74 million pound s of manure per year (United States Department of Agriculture / Natural Resources Conservation Service, 1995) and 10 million gallons of manure contaminated runoff. (Larsen et. al., 2014) 349 known karst features exist of which 316 are sinkholes. (Alexande r, et. al., 1988). Following snowmelt and rain in March of 2013; an incident occurred where an area well was potentially impacted. Investigation revealed manure contaminated runoff was entering groundwater in a newly discovered sinkhole (Larsen, 2013). Loc al citizen concern grew for groundwater quality. Developing relationships with landowners and livestock producers became necessary for protection of water resources and has facilitated research, education and action. A newly formed sinkhole which seasonal ly receives feedlot runoff was studied with ground penetrating radar for repair. Two producers in the region are implementing manure management techniques

PAGE 51

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 50 that are more stringent then regulation. The Wiskow dye trace was completed in spring of 2014. The st udy identified discharge springs that discharge into the Mill Creek trout stream from two vulnerable sinkholes (Johnson et al., 2014). Four springs and four previously unknown sinkholes were identified and mapped. A manure contaminated runoff storage area was constructed in the fall of 2014 by a livestock producer located at a headwater spring of Mill Creek. A filter strip and large manure contaminated runoff system is being designed for construction in 2015. Building great relationships with producers ha s been successful in Olmsted County. Livestock producers are making investments and taking action. Producers are an essential component of the mid western economy and assistance with information, funding and resources will help protect the environment an d keep farms profitable for future generations. Evaporite Geo Hazard in the Sauris Area (Friuli Venezia Giulia Region NE Italy) Chiara Calligaris, Trieste University Stefano Devoto, Trieste University Luca Zini, Trieste University Franco Cucchi, Triest e University Evaporite sinkholes represent a severe threat to many European countries, including Italy. Among the Italian regions, of the area most affected is the northern sector of Friuli Venezia Giulia Region (NE Italy). Here chalks had two main deposi tional periods first in the Late Permian and then during the Late Carnian (Late Triassic). Evaporites outcrop mainly in the Alpine valleys or are partially mantled by Quaternary deposits, as occur along the Tagliamento River Valley. Furthermore, evaporites make up some portions of mountains and Alpine slopes, generating hundreds of karst depressions. This paper presents the preliminary results of the research activities carried out in Sauris Municipality where sinkhole phenomena related to the presence of gypsum are very common. Field investigations were devoted to recognition, mapping and classification of evaporite sinkholes. To recognize sinkhole phenomena, the preliminary steps included the analysis of historical documents collected in archives, the analysis of aerial photos and Airborne Laser Scanning (ALS) surveys. The integration of the above cited activities allowed a preliminary identification of the phenomena, which were later validated by detailed field surveys. All the collected data p opulate a geo database implemented for a project funded by the Geological Survey of Friuli Venezia Giulia Region. The objective of this project is to inventory and classify the sinkholes associated to evaporite rocks. Building C odes to Minimize Cover Coll apses in Sinkhole Prone Areas George Veni, National Cave and Karst Research Institute Connie Campbell Brashear, Bracken Engineering, Inc. Andrew Glasbrenner, Bracken Engineering, Inc. Cover collapse sinkholes are forming with increasing frequency under buildings. Analyses of sinkhole distribution in Beacon Woods, Florida, preliminarily indicate their occurrence is an order of magnitude greater in urban versus undeveloped areas, suggesting the structures themselves are enhancing the collapse process. The most likely causes are induced recharge via at least one of two sources. First, runoff and drainage from roads, structures, and impoundments that is not adequately dispersed will promote sinkho le development. Second, leaking water, sewer, and septic systems beneath or adjacent to a structure will also promote collapse. The process of cover collapse from induced recharge is well understood. However, building codes generally do not require drainag e and structural engineering practices that would reduce induced recharge and thus reduce the risk of collapse. This paper proposes engineering practices that measurably restrict the accidental discharge of municipal water through leaking subgrade drainage systems or the deliberate discharge of stormwater runoff, induced shallow groundwater recharge from retention ponds and septic drainfields, or heavily irrigated land use. We recommend these practices be incorporated into building codes and ordinances to r educe induced sinkhole development in areas prone to cover collapse. Cars and Karst: Investigating the National Corvette Museum Sinkhole Jason S Polk, Western Kentucky University Leslie A North, Western Kentucky University Ric Federico, EnSafe Bria n H am, EnSafe Dan Nedvidek, EnSafe Kegan McClanahan Western Kentucky University Pat Kambes is, Western Kentucky University Michael J Marasa, Hayward Baker On February 12th, 2014, a sinkhole occurred at the National Corvette Museum in Bowling Green, Kentuck y. The collapse happened inside part of the building known as the Skydome and eight Corvettes on display were lost into the void that opened in the concrete floor. In this region of Kentucky, known as the Pennyroyal sinkhole plain, subsidence and cover col lapse sinkholes are commonly found throughout the landscape. This iconic karst region in the United States is also home to

PAGE 52

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 51 Mammoth Cave, the longest cave in the world, and thousands of other caves and karst features. Investigation of the sinkhole collapse began immediately while the Corvettes were extracted from the debris cone inside the void. Techniques used for investigation included water jet drilling, downhole cameras and drone footage, a microgravity surface survey, and mapping of the void and accompanying cave. After exploration of the sinkhole by karst researchers and compilation of the data, the cause of the sinkhole was determined to be a cave roof collapse in a breakout dome. The cave underlying the collapse is about 220 feet (67 m) long and 39 fe et (12 m) wide on average with an average depth of 65 85 feet (2025 m). The structural integrity of the bedrock (thinly interbedded limestone and chert located at a contact between two major limestone units) is lacking in the area. Talus and breakdown are abundant in the cave in which the sinkhole formed. The progression of the roof failure likely occurred over a long span of time, eventually giving way due to a variety of conditions, including speleogenetic and climatic factors. Current remediation is und erway and involves filling the sinkhole with gravel and sand, then installing a micropile supported concrete slab floor under the building. Future changes to the structure will be monitored to detect any activity. __________________________________________ Thursday October 8th 3:30p m 4:30pm Track B: Geophysical Exploration of Karst __________________________________________ Investig ation of a Sinkhole in Ogle Count y, Northwestern Illinois, Using Near Surface Geophysical Techniques Philip J Carpenter, Northern Illinois University Lauren M. Schroeder, Northern Illinois University A sinkhole measuring 40 m in diameter and up to 6.5 m deep occurs within the Nachusa Grasslands, near the town of Franklin Grove, northwestern Illinois. This area, d edicated to prairie conservation and restoration, is owned and operated by The Nature Conservancy. Several meters of unconsolidated sand, gravel, and clay overlie the St. Peter sandstone, beneath which lies karstic Prairie du Chien dolomite. Investigatio ns included EM conductivity profiles, resistivity soundings, 2D resistivity, and ground penetrating radar (GPR), supplemented by conductivity logs, soil cores, and tree core studies. These data indicate the sandstone averages about 5 m deep near the sinkh ole rim and the sinkhole is about 120 years old. Nearby residential wells indicate an average static water level of 11 m below the surface, so the water table currently lies well below the sinkhole floor. GPR sections show abrupt termination of the bedrock reflector near the sinkhole rim, suggesting formation by collapse. Geophysical investigations also identified possible hydraulic conduits associated with the sinkhole. Specifically, GPR profiles, at 50 and 100 MHz, provide the highest resolution images of the subsurface and indicate possible conduits (soil pipes) near the sinkhole rim as diffraction hyperbolas 23 m below the surface. GeoProbeTM conductivity logs showing unusually low conductivity, and sudden probe drops, also suggest the presence of shall ow soil cavities around the sinkhole. However, dye poured into various low spots on the sinkhole floor was never recovered, despi te numerous sampling locations. __________________________________________________ Study on Monitoring and Early Warning of Ka rst Collapse Based on BOTDR Technique Zhende Guan Inst. of Karst Geology, China Univ. of Geosc. X Z Jiang, Inst. of Karst Geology, China Univ. of Geoscience Y B Wu, Inst. of Karst Geology, China Univ. of Geoscience Z Y Pang, Inst. of Karst Geology, China Univ. of Geoscience Brilliouin Optical Time Domain Reflectometer (BOTDR) is a newly developed measurement and monitoring technique, which utilizes Brilliouin spectroscopy and Optical Time Domain Reflectometer (Jiang et al., 2006; Zhang et al., 2009; Xu et al., 2011) to measure strain generated in optical fibers as distributed in the longitudinal direction. This paper introduces the principle and characters of BOTDR technique firstly, and makes an example of karst collapse monitoring at section K14 of highway from Guilin to Yangshuo. And we talk about how to use this technique in underlying karst collapse monitoring in karst highway, discuss environmental factors, like temperature and vehicle dynamic load, how to affect the monitoring results and how t o choose optical fiber type and paving region. At last, we compare the results between using BOTDR and geological radar, and condu ct the safety diagnosis on the experimental road. The application achievements demonstrate that BOTDR is a viable technique fo r the karst collapse monitoring.

PAGE 53

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 52 Pre Event and PostFormation Ground Movement Associated with the Bayou Corne Sinkhole Cathleen E Jones, Jet Prop. Lab., California Inst of Tech. Ronald G Blom, Jet Prop. Lab., California Inst of Tech. We discuss measurements of the precursory and post formation ground displacement in the vicinity of the Bayou Corne, Louisiana, sinkhole made using interferometric synthetic aperture radar (InSAR) and data from the L band UAVSAR instrument. Large precursor y movement was observed at the sinkhole site and shown to be predominantly horizontal in direction, in contrast to sinkhole precursors previously detected with InSAR, all of which indicated vertical deformation. Here we discuss how two opposing imaging dir ections were used to determine the precursory horizontal movement, and use the same technique to look at the progression of post formation ground displacement around the expanding sinkhole during the interval 2012 2014. We find that the Bayou Corne sinkho le has expanded asymmetrically about the initial location, and show that expansion has tracked ground movement observed with InSAR along the margins of the water filled central subsided area. This work shows that InSAR applied to images acquired from multi ple directions can be used to image incipient sinkhole formation over large areas and to track the expected direction of expansion. We discuss how geologists can best use the InSAR technique to quantitatively monitor ground movement associated with sinkho les, particularly in areas where radar rapidly decorrelates, e.g., in Florida or Louisiana. These results demonstrate that InSAR could be used in sinkhole warning systems across a much broader geographical area than previously demonstrated, and for identi fying both precursory and p ost formation surface movement. Thursday October 8th 4:30p m 5:30pm Track B: Modeling of Karst Systems Numerical Simulation of Karst Soil Cave Evolution Long Jia, Institute of Karst Geology Yan Meng, Institute of Kar st Geology Zhen de Gu an, Institute of Karst Geology Li peng Liu, China Inst. of Water Resources and Hydropower This study is focused on numerical simulation of the formation and development of karst soil caves related to cover col lapse sinkholes. The so called karst soil cave refers to the caves formed in the soil layers above bedrock of sinkhole regions. Because the soil caves are formed and developed under groundwater seepage, studying groundwater level changes can help understand soil cave development and collapse. Based on the improved Terzaghi loosening pressure theory and using excess pore water pressure, two kinds of critical groundwater critical groundwater level decline related to soil cave formation groundwater decline related to cave roof collapse. After a soil cave is formed, its evolution can cause uneven displacement and stress redistribution in the overlying soil l ayer. The process of soil cave expansion can be understood by investigating the change in displacement and stress. Numerical simulation of the vertical displacement using FLAC3D shows that the maximum vertical displacement occurs at the arch roof of the so il cave and that the displacement can cause tensile failure of the arch roof. The simulated soil layer displacement is used to determine the soil depth disturbed by the cave by delineating the planes of equal settlement. Analyzing the simulated shear stres s shows that the maximum shear stress occurs at the arch toes and causes shear failure. On the other hand, the zones of low shear stress can be used to evaluate existence of arching effect in the overlying soil layer. By analyzing the plastic zone of the s oil layer, it was found that, in rigid clay, arch roof collapse and tensile failure are the major events that lead eventually to the barrel shaped or bottle shaped forms of collapsed pits. In loose soil, shear failure of the arch toe is the major event that eventually leads to the taper shaped or bowl shaped form of collapsed pits. Generally speaking, stability and size of soil caves can be determined using the three variables of low shear stress area, equal settlement plane, and plastic zone discussed abov e. The numerical simulation of this study is valuable to the monitoring and assessment of sinkhole occurrence.

PAGE 54

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 53 Experimental and Numerical Investigation of Cover Collapse Sinkhole Development and Collapse in Central Florida Xiaohu Tao, Florida State University Ming Ye, Florida State University Dangliang Wang, Florida State University Roger Pacheco Castro Florida State University Xiaoming Wang, Florida State University Jian Zhao, Hohai University The mechanisms of sinkhole formation, development, and collapse are investigated in this study using experimental and numerical methods. Sandbox experiments are conducted to understand how excessive groundwater pumping triggers sinkholes formation. The experimental results indicate that the change of hydrologic conditions is critical to sinkhole development. When seepage force increases due to increase of hydraulic gradient, clay and sand particles start moving downward to form a cavity. The confining unit is of particular importance because the cavity is f irst formed in this layer. Based on the conceptual model developed from the sandbox experiments, the Fast Lagrangian Analysis of Continua (FLAC) code and Particle Flow Code (PFC) are coupled to simulate the sandbox experiments. PFC was used to simulate par ticle movement in the sinkhole area, and FLAC is used for other areas. While the current numerical simulation can simulate the experiment results such as the sizes of the cavity and the sinkhole, the simulation capability is limited by the computing cost o f PFC. More effort of model development is necessary in the future study. Accounting for Anomalous Hydraulic Responses During Constant Rate Pumping Tests in the Prairie Du Chien Jordan Aquifer System Towards a More Accurate Assessment of Leakage Justin L Blum, Minnesota Department of Health The Prairie du Chien Jordan Aquifer system is an important source of drinking water for residents of southeastern Minnesota. Assessment of the hydraulic properties of this aquifer continues to be of interest for wellhead protection and resource evaluation efforts. When performing constant rate pumping tests on wells constructed in the karsted Prairie du Chien Aquifer, anomalous hydraulic responses resulting from cavernous flow are frequently observed. Hydraulic res ponse in the adjacent Jordan Sandstone Aquifer is also commonly distorted because of beddingplane fractures and well development techniques such as blasting and bailing. Resolution of these anomalous responses is important for accurate estimates of leakag e through adjacent resistive layers. Examples are presented with a rational for the analysis process. Thursday October 8th 6:30pm Banquet & Guest Speaker __________________________________________ Hales Bar and the Pitfalls of Constructing Dams on Karst David J. Rogers, Missouri University of Science and Technology Hales Bar Dam was built on the Tennessee River 33 miles downstream of Chattanooga by a private company to generate power in 19051913. The dam site was selected because it was the narro west reach in the downstream end of the Walden Ridge Gorge. The site is underlain by Mississippian Bangor Limestone on the southeast flank of the Sequatchie Anticline. Three different contracts failed to complete the dam because of difficult foundation con ditions. From 19101913 diamond drill core holes were used to explore the site and a series of reinforced concrete caissons 40x45 ft on upstream side and 30x32 ft on the downstream side were installed. Excessive leakage soon appeared near the eastern abutm ent, and gradually increased. Soundings were made in 1914 to ascertain the areas of gross leakage thereafter rags were placed over suction holes on the river bed and concrete pumped over these. Once a leak was stemmed, leakage would resume at other, adjacent locations. The owners tried to stem the leaks by inserting hay bales, old mattresses, chicken wire, and even corsets! In 1919 the owners began drilling grout holes from the inspection gallery within the dam and pumping hot asphalt into the voids. This was followed by the injection of 78,324 cubic feet of hot asphalt grout into the dam foundation, using 6,266 lineal feet of boreholes with average hole depth of 92 ft. By 1922 the problem appeared solved, but leakage gradually resumed between 1922 1929, ri sing to the same level as had been observed previously. In 19301931 a new program of exploration was undertaken, using dyes and oils to identify conduits under the dam. Leakage was found to vary between 100 and 1200 cubic feet per second (cfs). When the dam was acquired by the Tennessee Valley Authority (TVA) in 1939 they employed fluorescein dyes to track the under seepage. Dye tests revealed that the leakage varied between 1720 and 1650 cfs; about 10% of the rivers normal flow. They also noted seepag e boils forming in the gravel bars, which increased each year. The TVA began constructing the most expensive cutoff wall ever

PAGE 55

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 54 built, drilling 750 18 inch diameter holes along the dams centerline and backfilling this with concrete to a maximum depth of 16 3 feet, extending 25 to 103 feet below the river bed. In April 1963 the TVA announced it was abandoning Hales Bar Dam, due to increasing leakage. __________________________________________ Friday, October 9th 8:30 am 11:10pm Engineering and Geotechnical Investigations in Karst Concepts for Geotechnical Investigation in Karst Joseph A Fischer, Geoscience Services Joseph J Fischer, Geoscience Services There seems to be a lack of recognition in the literature that addresses the variety of karst in the United States of America and some of its offshore territories. For example, there are the well known solutioned carbonates of Florida and the Caribbean, b ut there are also the somewhat older, harder carbonates of St. Croix, U.S.V.I. Even Floridas recently deposited karst varies from region to region. There are also the ancient, flat lying carbonates of the interior craton that often have semi horizontal cavities resulting from variations in ground water levels affecting bedding and the contorted rocks of the Appalachians with its apparently chaotic variations in solutioning found across strike and in relation to folds, faults and fracturing. In addition, there are various salt and gypsum deposits in the south and southwest that pose their own problems to man's works. As the geology differs, so does, to some extent, the investigation requirements, investigation techniques and engineering solutions. There is no single set of investigative tools that fit all karst sites. Geophysical investigations are apparently far less suitable for the broken and twisted Appalachian karst than in the flat lying mid continent carbonates or the less contorted karst of Florida. Specific procedures developed for geotechnical investigation in true karst have been documented for many years now. However, it appears that many practitioners are not aware of them or choose not to use them because of the possibility of increased costs; or too often, a lack of geotechnical understanding of the work of others in karstic areas outside their sphere of experience. This paper will attempt to provide a rational geotechnical approach to carbonate rock investigations in the United States while recognizing the inherent variabilities of the targets and the economics of preconstruction investigations; with the understanding that one size does not fit all. Sinkhole Physical Models to Simulate and Investigate Sinkhole Collapses Mohamed Alrowaimi, University of Central Florida Hae Bum Yun University of Central Florida Manoj Chopra, University of Central Florida Florida is one of the most susceptible states for sinkhole collapses due to its karst geology. In Florida, sinkholes are mainly classified as cover subsidence sinkholes that result in a gradual collapse with possible surface signs, and cover collapse sinkholes, which collapse in a sudden and often catastrophic manner. The future development of a reliable sinkhole predi ction system will have the potential to minimize the risk to life, and reduce delays in construction due to the need for post collapse remediation. In this study, different versions of smallscale sinkhole physical models experimentally used to monitor the water levels in a network of wells. This information is then used in a spatial temporal analysis model to study the behavior of the system until the sinkhole collapses. The ultimate goal is to use this process in a reverse manner to monitor an existing ne twork of installed groundwater wells to study the fluctuations in the water levels and use the spatial temporal analysis to predict potential sinkhole collapses. The groundwater levels are monitored using sensors that are hooked up to a high resolution dat a acquisition system. The results of a series of tests conducted using this sinkhole physical model showed that there is a very distinguishable groundwater cone of depression that forms underground before the sinkhole collapses. This cone of depression was studied in its early stages and as it progressed with time. This analysis is used to then investigate the growth of the sinkhole before the surface eventually collapses. The spatialtemporal model showed the development of the groundwater cone of depressi on with time during the development of the cavities within the sediments can be used as a potential signal to identify and isolate the sinkhole location.

PAGE 56

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 55 Monitoring the Threat of Sinkhole Formation Under a Portion of US 18 in Cerro Gordo County, I owa Using TDR Measurements Kevin M O'Connor, GeoTDR, Inc Matthew Trainum, Iowa Department of Transportation Sinkhole formation is a common occurrence in northeast Iowa, and U.S. 18 in Cerro Gordo County was constructed over an area where sinkhole formation had only been locally known. It had not been recorded or identified in the Iowa DNR database at the time. Since 2004, sinkholes have developed along the right of way. Geophysical surveys contributed very little in the identifying the cause. Ho wever a Soil Survey (drilling program) identified numerous voids within carbonate bedrock. The soil borings indicated that shale overlying the carbonate rock has been removed/eroded, and resulted in the development of a karst subsurface through the dissol ution of the carbonate rock. Without removing the structural fill and site soils to expose the rock, it will not be possible to impede the natural processes occurring. An alternative approach was adopted and consisted of: (a) removing the existing pavem ent, (b) installing coaxial cables in trenches excavated within the subgrade, (c) replacing the pavement as double reinforced pavement (including shoulders), and (d) monitoring the cables using Time Domain Reflectometry (TDR). The cables are interrogated several times a day and data is transmitted via cellular modem to Iowa DOT facilities. Among the data transmitted is a log file of deformation activity along each of the cables which is evaluated and an action plan is initiated based on: (a) information in the activity file, and (b) updated plots for each cable. Unexpected behavior has been observed, with activity occurring annually between the months of September and March. Although several explanations have been proposed, there is no definitive correl ation between locations of the activity detected by TDR, sinkhole locations, or geophysical anomaly locations. In spite of this uncertainty, real time remote monitoring for ground movement is continuing. Predicting Compaction Grout Quantities in Sinkhole Remediation Edward D Zisman, Cardno ATC Predicting the required quantity of grout needed to remediate a sinkhole damaged home is a challenging task that involves significant amounts of uncertainty. The difficulty arises from the limited amount of subsu rface information that is available to make subsurface predictions particularly in complex karst environments. In typical sinkhole investigations, our understanding of the subsurface is limited by the three to four data points (borings) that provide a sma ll window into actual subsurface conditions. This information is normally obtained from borings and from information inferred by geophysical surveys. In many cases, the information is not sufficient to make accurate predictions of grout quantities. This paper will discuss the uncertainties in analyzing the many factors that influence grout prediction; it will provide a method of calculating grout quantities and discuss how one may moderate the difficulties in prediction of grout quantities. Examples of case studies are given showing pre grout and post grout information and the lessons learned from these comparisons. A Feasibility Study of the Implementation of a Flowable Fill Material to Prevent Sinkhole Occurrence at the I&W Brine Well Site in Carlsbad New Mexico Chase L Kicker, New Mexico Inst. of Mining and Technology Solution mining at the I&W Brine Well Site in Carlsbad, N ew Mexico has produced a large underground cavity that could result in a catastrophic sinkhole ind ucing collapse. A church, feed store, residences, a highway, and an irrigation canal are potentially in the sinkhole area. The purpose of this research is to demonstrate the feasibility of a solution to stabilize the underground cavity and prevent sinkhole occurrence. Wi th extensive use of relatively inexpensive fillers in a cement based grout, a flowable fill mate rial was developed and tested. The application of this fill material was studied by injecting the flowa ble fill into several types of underground cavities to de termine if the material can be effectivel y implemented as a stabilizing agent. A range of test configurations was created to simulate various cavity scenarios, and the fill material was injected into each cavity to assess how the material responds and eval uate ground stability in each case. This research has shown that maintaining a pressure balance while pumping the flowable fill material into a brine filled cavity an d discharging the brine at the surface is a feasible method to stabilize the cavity. The f lowable fill solidifies in place to prevent sinkhole occurrence. The flowable fill material is recommended for use at the I&W Site. This research has application for stabilizing underground mines to prevent collapse and reduce subsidence. The flowable fill can also be u sed to stabilize foundations in areas of naturally occurring dissolution cavities that cause sinkholes.

PAGE 57

The 1 4th Multidisciplinary Conference on Sinkholes and the Engineering & Environmental Impacts of Karst Page 56 Pre Construction Rock Treatment and Soil Modification Program Using Low Mobility Grout to Mitigate Future Sinkhole Development in a 2,787.1 Square Meter (30,000 SF) Maintenance Facility Steven W Shifflett, US Army Corps of Engineers The US Army required construction of a 2,787.1 square meter (30,000 sq. ft) maintenance facility supported on shallow foundations at the Fort Campbell Military Installation. During the subsurface investigation a seven foot air filled void was encountered i n the bedrock within the building footprint. Electrical Resistivity Imaging (ERI) was conducted in an attempt to determine the lateral extent of the encountered void and to establish the general prevalence of karst features at the site. Due to uncertainty in the subsurface conditions, a rock treatment and soil modification program was developed which consisted of a series of targeted exploratory grout holes advanced in 126 locations in the structural areas of the building footprint. The intent of the progr am was not to prevent the development of a soil dropout, but to improve the foundation support of the structure so that the facility would perform acceptably if a future soil dropout were to occur during the design life of the facility. This was achieved b y targeting each footing with 3 exploratory grout holes. The intent of each grout injection was 1) to identify the top of rock elevation, 2) determine if a karst feature existed, 3) cap the karst bedrock below the footing and treat defects in the rock, and 4) provide localized improvement of soft soils through the use of low mobility grout columns under each footing. Drilling refusal elevations were obtained for every grout hole and were assumed to represent the top of bedrock. Each exploratory hole was clo sely monitored for pressure and volume in 0.61 meter (2foot) stages. Zones where the bedrock had lower elevations or took excessive grout at low pressures were targeted with additional tertiary holes. The tertiary holes were verified with additional SPT sampling. Documented ground improvement was achieved, evident by increased SPT blow counts ranging between 25 to 50+ post treatment. Based on results from this program, lower grouting pressures could have been utilized as part of the refusal criteria to s uccessfully identify and treat karst features. Successful Foundation Preparations in Karst Bedrock of the Masonry Section of Wolf Creek Dam David M Robison, U.S. Army Corps of Engineers Extensive foundation preparations during construction of the Wolf Creek Dam concrete masonry section precluded the need for additional rehabilitation to mitigate seepage through karstic limestone bedrock. Wolf Creek Dam on the Cumberland River in southern Kentucky has become well known for karst related seepage i ssues underneath the embankment section, and yet has had little to no seepage issues associated with the concrete masonry portion of the dam. Post construction efforts to control seepage underneath the embankment began in 1967 and 1968. Emergency grouting commenced and continued through 1970. Between 1975 and 1979 a more permanent solution of a concrete diaphragm cut off wall was constructed through the centerline of the left portion of the embankment section down to competent bedrock. The wall interrup ted the progression of foundation erosion, but post construction monitoring, instrumentation readings, and persistent wet areas downstream showed that seepage paths under or around the wall continued. A second cut off wall upstream of the first was constr ucted between 2007 and 2013, extending nearly the entire length of the embankment and up to 75 feet (22.9 m) deeper than the original wall. Cost of the second wall and other concurrent rehabilitation efforts reached nearly $600 million. Exploratory grout ing beneath the concrete masonry section of the dam in 2012 resulted in low grout volume takes, so no further remediation efforts below the masonry dam were conducted. The original construction photographs and foundation reports for the concrete masonry s ection of Wolf Creek Dam instill confidence that the designers and builders of the monoliths took adequate, if not excessive measures to ensure that all the monoliths were founded on competent bedrock. These measures included extensive borehole investigat ions both prior to and during excavation, efforts to locate, delineate, remove, and clean all karst solution channels, the removal of all loose rock, grouting in the foundation and side vertical faces, large stair step faces on the left abutment, extended excavations to remove soft beds, final manual cleaning of rock surfaces, and the careful documentation of foundation preparations. These measures do not guarantee that seepage issues will not develop under the concrete dam over time, but they do show with reasonable certainty that the monoliths were originally founded on competent bedrock, and that future seepage issues are either unlikely or will be significantly inhibited by the preparation made to the foundation prior to the construction of the concrete monoliths.



PAGE 1

507 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 ACCOUNTING FOR ANOMALOUS HYDRAULIC RESPONSES DURING CONSTANT-RATE PUMPING TESTS IN THE PRAIRIE DU CHIEN-JORDAN AQUIFER SYSTEM TOWARDS A MORE ACCURATE ASSESSMENT OF LEAKAGE Abstract The Prairie du Chien-Jordan Aquifer system is an impor tant source of drinking water for residents of southeast ern Minnesota. Assessment of the hydraulic properties of this aquifer continues to be of interest for wellhead protection and resource evaluation efforts. When per forming constant-rate pumping tests on wells construct ed in the karsted Prairie du Chien Aquifer, anomalous Jordan Sandstone Aquifer is also commonly distorted because of bedding-plane fractures and well develop ment techniques such as blasting and bailing. Resolution of these anomalous responses is important for accurate Examples are presented with a rationale for the analysis process. Introduction The hydraulic properties of fractured and karsted rocks are notoriously variable. (Teutsch & Sauter, 1991) Stan begin to describe various hydraulic characteristics and types or patterns of hydraulic response may be distin guished. This dataset informs the analysis to reduce un certainty in estimated hydraulic properties. Typically, hydraulic conductivity and storativity are de rived from a constant-rate pumping test. In a layered system if observation well data were collected and the information, the hydraulic resistance is used to calculate the rate of pumping-induced leakage to the aquifer. A realistic estimate of the rate of leakage is an important parameter for construction and calibration of regional The Minnesota Department of Health (MDH) has com tests performed on wells constructed in the Prairie du Chien-Jordan Aquifer (PDC-J) system (Figure 1). The tests were conducted for a variety of purposes by drill ing contractors, environmental consultants, and state were performed on high-capacity wells for wellhead protection planning or to satisfy other regulatory permit requirements. Acquisition and analysis of test data con tinues. The extent of the aquifer system in Minnesota is great er than 23,300 sq. kilometers (9000 sq. miles) (Figure 1). This represents a test density of about one test per 310 sq. km. (120 sq. mi). In view of the expense of conducting an aquifer test, the scarcity of appropriately constructed wells from which to obtain water levels and Justin L. Blum Minnesota Department of Health, 625 North Robert Street, St. Paul, Minnesota, 55156, USA, justin.blum@state.mn.us Figure 1. Aquifer Test Locations by Layer.

PAGE 2

508 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE information gained from each test. Prairie du Chien-Jordan Aquifer Regional Hydrogeologic Setting Very nearly all of southeastern Minnesota is underlain by prise the Prairie du Chien-Jordan Aquifer system (Figure Prairie du Chien Aquifer The issue of fracture and solution enhanced (cavernous) (Tipping et al., 2007). Karst landforms, sinkholes, caves, springs, and sinking streams are found where the Prairie du Chien Group is the uppermost bedrock unit. Not all regardless of the type of geologic cover (Tipping et al. 2006). The subaerial exposure of the Prairie du Chien Group has reactivated old features and contributed to the development of new ones. The basal part Prairie du Chien Aquifer, the Oneota Formation, is massive dolos tone with low porosity and few fractures. On a regional basis any head differences that are observed between the Prairie du Chien Group and Jordan Sandstone are attrib uted to the basal Oneota functioning as an aquitard. Coon Valley Member The contact of the Prairie du Chien Group with the old Mossler, 2008). The Coon Valley Member of the Prairie du Chien Group is a dolostone-cemented silty sandstone that formed in topographic lows on the erosional surface on the top of the sandstone. The lateral extent of the Coon Valley Member is limited to the dendritic drainage pattern in which it formed, therefore the occurrence is both limited and quite unpredictable. Jordan Sandstone The Jordan Sandstone is a well-sorted medium-grained sandstone which appears to be fairly uniform over its extent. However, locally it contains lithological varia tions which lower the bulk hydraulic conductivity and bedding-plane fractures which increase the effective conductivity. Description of Aquifer Test Data Earlier compilations of his type of data have been lim ited to the Twin Cities Metropolitan area (Runkel et al. 1999) or selected communities (Runkel & Mossler 2001). The scope of this compilation is the whole extent of the aquifer system in Minnesota. MDH staff col tests conducted by others is also obtained whenever pos sible. Source data from all but eight tests are available. This is critically important because access to the source data has permitted well-documented tests to be reinter preted without having to repeat the test. Aquifer Name Primary Rock Type Nominal Thickness (meter) Decorah-Platteville-Glen wood Aquitard Shale and Do lostone -St. Peter Sandstone (OSTP) Sandstone 30 Basal St. Peter Sandstone (Twin Cities Basin) Siltstone and Sandstone 15 Prairie du Chien Aqui fer (OPDC) Shakopee Fm. Dolostone 85 New Richmond Sandstone Sandstone 10 Oneota Fm. Dolostone 28 Coon Valley Member Sandy Dolos tone Variable Jordan Sandstone (CJDN) Sandstone 30 St. Lawrence Aquitard Dolostone and Shale -Table 1. Hydrogeologic Setting of the Prairie du Chien-Jordan Aquifer System Layers are Shaded. Figure 2. Histogram of Storativity Values from All Tests.

PAGE 3

509 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 The numbers of tests by aquifer layer are shown in Table 2. The geographic distribution of tests is shown in Figure 1. The greatest numbers of tests have been performed in of southeastern Minnesota where the Prairie du Chien Aquifer is the uppermost bedrock. Comparatively few thickness of younger bedrock cover. Descriptive statistics of aquifer properties are shown in Figures 2 and 3. Most tests were conducted in con The box and whisker plots, Figure 3, are standard Tukey dian value and outliers as greater than or less than 1.5 IQR Figure 4 shows a cross-plot of leakage factor vs. data with slight differences in the populations by aquifer layer. A detailed account of the analysis methodology for each test is beyond the scope of this paper as most tests were to outline the general approach to analysis that was de veloped for these data, demonstrate that the opposing assumptions of standard transient and steady-state meth ods are useful to constrain parameter estimates, and how uncertainty of aquifer properties. Well Hydraulics the pumping well as compared to that in the aquifer de Two prominent naturally occurring features of the aqui solution enhanced conduits, cause wells to produce wa such features. In addition, well development practices affect the hydraulic performance of the pumping well. Sandstone wells that have been developed by the blast and bail technique. Dynamite is placed in the well and detonated to break up the rock. Broken rock is then bailed from the well to create a large cavern in the for mation. In some cases hundreds of cubic meters of sand are removed from the well. The actual diameter of this cavern is never known exactly but it is certainly much larger than the diameter of the original borehole from the drillers record. The combination of natural features and an enlarged erted to remove a given volume of water. Thereby, from the perspective of the aquifer properties the well ap pumping well, such as the Coon Valley Member. This of the local difference in bulk hydraulic conductivity. in the analysis certainty in the local aquifer properties is increased. No tests, to-date, show the distinctive hydraulic charac aquifer system. (Gringarten & Witherspoon, 1971) This is the case even in instances where boreholes show near ly simultaneous response to pumping at different wells. Prairie du Chien Aquifer is a good starting point for in terpretation of test data. However, more subtle effects cumulation of test data. permeability fracture intersecting the wellbore will show a characteristic unit-slope on a log-s vs. log-t plot. Similarly, because of increased wellbore storage, a cav ern developed in the formation should cause a unit-slope Aquifer Name Prairie du Chien-Jordan Prairie du Chien Jordan Sub Totals Single well tests 4 1 5 10 Tests with observation wells 26 5 43 74 Total number of tests 30 6 48 84 Table 2. Numbers of tests by layer.

PAGE 4

510 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE on the same plot. (Gringarten, 2008) In the existing da taset, the only early-time deviations from the Theis curve have been of the unit slope, characteristic of well-bore storage. This indicates that the hydraulic effect of both cavernous secondary porosity is to create an enlarged borehole. The dataset of aquifer tests is used to support the de concerned with protecting drinking water quality. As the (Figure 2), the source of water to the system is vertical yses. Analysis Techniques are of two basic types, transient and steady-state. The age). On this basis, all analyses that exclusively use the Theis conceptual model have limited utility for model development in layered systems. tush & Jacob, 1955) assumes that there is no change in storage in the aquifer and that the source of water is from cations to the transient conceptual model to include the effects of leakage were introduced by Hantush (1960), Walton (1960), and Neuman and Witherspoon (1972). Figure 3. Descriptive Statistics for Layers with 30 or More Tests.

PAGE 5

511 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Blended opposing conceptual models have been used for many years, however few practitioners think to use the steady-state analysis techniques unless leakage is clearly indicated during a test. Recharge of some sort must be considered to be present proach is not new. For instance, aquifer tests were used to examine the leaky characteristics of tight formations for natural gas storage (Witherspoon et al., 1967). How ever, the reservoir engineering perspective is not often applied to groundwater problems. Linked Parameters: T, S, and L In the mathematics of the Theis well function, trans missivity (T) and storativity (S) are linked parameters. missivity to the characteristic leakage factor (L). For leaky settings, L, is small on the order of hundreds of large on the order of thousands of meters. The premise of this analysis procedure is that all three parameters are linked. Therefore for a given trans Figure 4. Leakage Factor vs. Storativity by Aquifer.

PAGE 6

512 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE ness of younger bedrock and clay-rich till cover. There fore, the setting should produce a small storativity, in the range of 10 -5 and an observation well, located 479 meters (1570 feet) pumping well. Drawdowns in the pumping well are greater than what would be predicted by the Theis curve match to the ob servation well. The storativity calculated from this data is large for the setting, at 10 -4 Figure 6 shows the data from observation well from this test in more detail with a Theis analysis. It can be seen that the effect of start of pumping/recovery traveled 480 meters in about 30 seconds. This rate of propagation, about 16 meters per second, clearly indicates fracture/ utes) approximates the unit-slope that is characteristic of wellbore storage. The unit-slope departure of early-time data from the Theis curve in nearby observation wells is frequently ob served in this aquifer system. For tests with several ob servation wells the anomalous response is damped and becomes more Theisian in more distant wells. Note missivity, both the storativity and leakage must make sense with respect to the hydrogeologic setting. Quan titatively, storativity and characteristic leakage factor are inversely related. A small storativity should be associ ated with very small volume of leakage, a large leakage factor. Conversely, large storativity should be associated with a large volume of leakage, a small leakage factor. Given this linkage between the parameters T, S, and L, an approach has been developed to deal with the effects tures, and an enlarged borehole as shown by the follow ing examples. Examples of Anomalous Hydraulic Response When observation wells were used, the most effective way to detect and then account for the anomalous pat terns is to plot all data from a test on one graph with the Theis distance-drawdown plot, log-s vs. log-t/r 2. A test was performed on a high-capacity well for an etha nol plant to satisfy state permitting requirements (Cham completed open-hole over both the Prairie du Chien and Jordan portions of the aquifer system. The hydrogeo Figure 5.

PAGE 7

513 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 6. Unit-slope of Early-Time Data. that the half -unit slope trend that indicates a high per meability fracture may only be observed in very early detail. Interestingly, the unit-slope trend is not often observed in the pumping well itself, such as the case with this test even those wells equipped with transducers which rap idly collect data in early-time. It appears that transient mechanical and other well hydraulic issues obscure the unit-slope response in the pumping well. For this test to be able to reliably estimate the characteristic leakage factor. The hydrogeologic setting would cause the leak age factor to be large, probably greater than 6,000 meters (~20,000 feet). A recently completed production test of Rochester 41 cient well. (Blum, 2014) Well 41 is located on the east side of the Rochester basin in an area where the uppermost bedrock is the Prairie du Chien Group with scattered patchy remnants of the St. Peter Sandstone. Well 41 was completed in the Jor dan Sandstone and had been developed by the blast and bail method. Approximately 115 m 3 (150 cubic yards) of sand were removed from the well. Based on the volume of sand removed and the length of open hole, the en hanced borehole radius is at least 1.1m (3.6 feet) The production test was conducted at a constant rate of 4580 liter/minute (1210 gallons per minute) for 24-hours. The driller measured water levels manually at the pumped well. Observation wells were four other nearby Rochester Public Utilities (RPU) water supply wells which were monitored by the utilitys data acquisi ure7). Other known interfering wells were the production wells in the RPU system. The manual data collection at the pumped well precluded any early-time data but those data collected appear to be of high quality. Problems in data collection/reporting were that some transducers in the wells report to nearest 1 to 2 feet rather than the normal 0.01 foot resolution. This is a common problem than the design purpose.

PAGE 8

514 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE For this reason, plots of all water level readings, collect plots use water level readings from the observation wells on a half-hour interval for a more clear view of hydraulic response. four observation wells were affected by cycling of their pumps at some time during the test. In particular, data from well 27 (224212) was not usable as this well was in recovery over the time span of the test and no discern able drawdown caused by Well 41 was found. Wells 21 and 23 cycled after 23-hours into the test and therefore only the pumping-phase data were usable. Limited recovery data were collected from the pumped viceable data were obtained for analysis. The Theis t/r 2 analysis of these data is shown on Figure 8. enced by multi-aquifer well construction. The response Figure 7. Wells Monitored During W41 Produc tion Test. Figure 8. Initial Analysis, Theis t/r2 plot.

PAGE 9

515 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 9. Agarwal Recovery Analysis. Figure Transmissivity m 2 /d Storativity Well Analysis Method 8 445 8.4e-5 Theis t/r 2 9 180 5.3e-1 W41-Agarwal 10 350 8.7e-5 W33-Theis Table 3. Initial Assessment of Aquifer Properties. water. Though, it is common for drawdown at a distant observation well to be smaller than expected, with re spect to the Theis curve. Therefore, the analysis weights the response of the closest Jordan Sandstone well, Well 33, most heavily. A strong indication that anomalous hydraulic responses ability in aquifer properties between wells and analysis techniques (Table 3). The effective radius of the pumped well may be estimat ed by different means depending on the quality of the a range of radial distances were used a semi-log regres sion on the distance-drawdown data using the observa tion wells only provides an initial value of r. Otherwise, well to match the Theis curve is the only recourse. The resulting estimate of effective radius must then be tested with other analysis plots to verify consistency with the hydrogeologic setting. In this case, the distances to the observation wells are not 1000 m from the pumped well. A semi-log distancedrawdown regression is not informative. Trial and error resulted in an optimal value of r, 7.6 m (25 feet) and a value of transmissivity of 287 m 2 /d (Table 4). With the revised wellbore radius visual matches to the typecurves are consistent between analysis techniques (Fig ures 11 and 12). Without the revised wellbore radius of the pumping well, tal match, L (Figure 12). In addition, if a borehole radius of 0.3 m (1 ft) is used for the Agarwal recovery analysis (Agarwal, 1980), this results in a non-credible storativity of 0.5 (Figure 9). Whereas, using a radius of 7.6 m in the calculation, the resulting storativity is large, 8.2e-4, but is physically possible.

PAGE 10

516 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Figure 10. Theis Analysis, W33. The estimate of L from the Hantush-Jacob analysis, Fig ure 13, is not as reliable as that from the de Glee analy sis, Figure 12. This results from the sensitivity of this technique to an error introduced by using wells located at distances greater than 0.2 L. In this case, no wells within 0.2 L were monitored and the Hantush-Jacob analysis is poorly constrained. Conclusion Application of this method depends on the number of observation wells monitored during a test. Given a suf analysis techniques, [log (s) vs. log (t/r 2 ) and log (s) vs. log (r)], produce the most consistent results. Uncertainty by wellbore storage, and solving for consistent values of As described, this procedure accounts for peculiarities of the hydraulic response in wells in this aquifer system to arrive at a consistent transmissivity based on the dif Figure Transmissivity [m 2 /d] Storativity Characteristic Leakage Factor [m] Analysis Method 9 180 8.2E-04 -Agarwal 11 287 8.0E-05 -Theis t/r 2 12 287 -1830 de Glee 13 288 9.9E-05 2380 Hantush-Jacob/Cooper-Jacob Table 4. Aquifer Properties after Accounting for Large Effective Borehole Radius.

PAGE 11

517 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 11. Theis t/r2, Projected to r = 7.6 m. Figure 12. Log-log Distance-Drawdown.

PAGE 12

518 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Figure 13. Semi-log Distance-Drawdown. ferent assumptions of the source of water to the well. If accounted for, the apparent variability of all hydraulic borehole radius, the correlation between the linked pa rameters is weakened. Uncertainty in storativity is low because it is based on a time of propagation of a pressure change in the aquifer that is relatively insensitive to differences in transmis sivity. The least certain parameter is the characteristic leakage factor. This approach to aquifer test analysis properties and enhanced the quality of the overall data set. References Agarwal RG. 1980. A new method to account for producing time effects when drawdown type curves test data. In Proceedings of SPE. 55th Sociey of Petroleum Engineers Annual Technical Conference and Exhibition. Dallas, Texas. Blum JL. 2013. Analysis of the Claremont Renewable Prairie du Chien-Jordan Aquifer. Claremont, Minnesota: Minnesota Dept. of Health. Blum JL. 2014. Analysis of the W41 (796431) Production Test, May 19, 2014 Jordan Sandstone Aquifer. Rochester, Minnesota: Minnesota Dept. of Health. Champion G. 2008. Aquifer Test Report Claremont, Minnesota, Ethanol Production Facility for Claremont Renewable Energy, LLC, Natural Resource Group, LLC (NRG). Method for Evaluating Formation Constants and Geophysical Union, 27, p. 526 534. wateronttrekking door middle van putten. Ph.D. Delft, Netherlands: Delft Technische Hogeschool.

PAGE 13

519 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Teutsch G, Sauter M. 1991. Groundwater modelling in karst terrains: Scale effects, data acquisition Third Conference on Hydrogeology, Ecology, Monitoring, and Management of Groundwater in Karst, ed. J. Quinlan, 17. Nashville, Tennessee. Theis CV. 1935. The Relation Between the Lowering of of Discharge of a Well Using Ground-Water Storage. Trans. American Geophysical Union 16: 519. Tipping RG, et al. 2007. Bedrock geology, topography, and karst feature inventory of Steele, Dodge, Olmsted and Winona Counties, Steele, Dodge, Olmsted and Winona Counties, Minnesota: Minnesota Geological Survey. Available at: http:// purl.umn.edu/123367. Tipping RG, et al. 2006. Evidence for hydraulic heterogeneity and anisotropy in the mostly carbonate Prairie du Chien Group, southeastern Minnesota, USA. Sedimentary Geology, 184 (3-4): 305. Walton WC. 1960. Leaky artesian aquifer conditions in Available at: http://www.isws.uiuc.edu/pubdoc/RI/ ISWSRI-39.pdf. Witherspoon PA, et al. 1967. Interpretation of aquifer gas storage conditions from water pumping tests, New York: American Gas Association. Available at: http://lccn.loc.gov/67063074. Available at: http://repository.tudelft.nl/assets/ uuid:c3e13209-4626-41b9-9038-c223d61e35c4/ deGlee1930.pdf. Gringarten AC. 2008. From Straight Lines to Deconvolution: The Evolution of the State of the Art in Well Test Analysis. SPE Reservoir Evaluation & Engineering, 11 (1). [Accessed April 16, 2013] Available at: http://www.spe.org/ id=EREE&mid=SPE-102079-PA&pdfChronicleId =09014762801630bc. Gringarten AC, Witherspoon PA. 1971. A method of Proceedings of the Symposium on Rock Mechanics, T3-B, p. 1. Aquifers. Journal of Geophysical Research 65: 3713. Hantush MS, Jacob CE. 1955. Steady Three-dimensional Flow to a Well in a Two-layered Aquifer. Trans. American Geophysical Union 36: 286. Jacob CE. 1947. Drawdown Test to Determine the Effective Radius of Artesian Wells. Transactions of the American Society of Civil Engineers 112: 1047170. for Minnesota, Minnesota Geological Survey, University of Minnesota. Available at: http://purl. umn.edu/58940. Neuman SP, Witherspoon PA. 1972. Field determination of the hydraulic properties of leaky multiple aquifer systems. Water Resources Research, 8 (5): 1284. Olcott PG. 1992. Ground Water Atlas of the United States: Segment 9, Iowa, Michigan, Minnesota, Wisconsin. Hydrologic Atlas, p.1 atlas (31 p.) :col. Runkel AC. et al., 2003. RI-61 Hydrogeology of the Minnesota Geological Survey, University of Minnesota. Available at: http://purl.umn. edu/58813. Runkel AC, Kanivetsky R, Tipping RG. 1999. Hydraulic and Hydrostratigraphic Characteristics of the Prairie Du Chien-Jordan Aquifer System in the TwinCities Metropolitan Area, Minnesota., Minnesota Geological Survey, University of Minnesota. Runkel AC, Mossler JH. 2001. Hydrostratigraphic and at nine southeastern Minnesota communities: research in support of wellhead protection, Minnesota Geological Survey, University of Minnesota. Available at: http://conservancy.umn. edu/handle/108833

PAGE 14

520 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

127 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 A COMPARATIVE STUDY BETWEEN THE KARST OF HOA QUANG, CAO BANG PROVINCE, VIETNAM AND TUSCUMBIA ALABAMA, U.S.A. Abstract Some of the most beautiful karst features created by the dissolution of limestone are residual hills with steep or karst. Tower karst to be developed requires a mean an nual temperature of minimum 17 o C to 18 o C and 1,000 to 1, 200 mm/m 2 of annual rainfall (Jakucs, 1977). Two sites matching this criteria were selected: the karst of Hoa Quang District, Cao Bang Province, Vietnam, and Tuscumbia, Colbert County, Alabama, U.S.A. Prelimi nary observations regarding similarities and differences between these two sites are presented in this paper. Introduction The Hoa Quang karst area is located in the northern Viet namese Province of Cao Bang. In 2014, a large num ber of karst springs, caves, sinking streams, and karst hydrogen stable isotope ratios. The pH values are typical for karst waters and ranged from 153.2 to 421.6 S/cm, the total alkalinity as CaCO 3 varies from 125 to 207 mg/L, and the carbon dioxide for the total hardness (as CaCO 3 ) are between 143 and 220 mg/L. The local meteoric water line, based on our measure ments is 2 18 O + 10.45 ( .86) with r 2 =0.998, which is close to the global meteoric water line (GMWL) 2 18 O + 10.35 tercept value differs very slightly from both local and global water lines. Due to the short sampling period, the Gheorghe M. Ponta Geological Survey of Alabama, 420 Hackberry Lane, Tuscaloosa, AL, 35486-6999, U.S.A., gponta@yahoo.com Nguyen Xuan Nam Vietnam Institute of Geoscience and Mineral Resources, 67 Chien Thang Road, Thanh Xuan District, Ha Noi, Viet Ferenc L. Forray Babes-Bolyai University, Department of Geology M. Kogalniceanu, No. 1, 400084, ClujNapoca, Romania, ferenc.forray@ubbcluj.ro Florentin Stoiciu R&D National Institute for Nonferrous and Rare Metals, Biruintei Blvd, 102 Pantelimon, Ilfov, Romania, fstoiciu@yahoo.com Viorel Badalita R&D National Institute for Nonferrous and Rare Metals, Biruintei Blvd, 102 Pantelimon, Ilfov, Romania, viobadilita@yahoo.com Lenuta J. Enache R&D National Institute for Nonferrous and Rare Metals, Biruintei Blvd, 102 Pantelimon, Ilfov, Romania, jenache@imnr.ro Ioan A. Tudor R&D National Institute for Nonferrous and Rare Metals, Biruintei Blvd, 102 Pantelimon, Ilfov, Romania,

PAGE 2

128 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE information provided by the water stable isotopic com position is limited. Carbonate rocks underlie many areas of north Alabama. Karst features can be found around Tuscumbia, in north western Alabama, which is part of the Tennessee-Ala bama-Georgia karst area that is called TAG. TAG has the highest concentration of caves in United States, and home for a few large springs. Tuscumbia Spring is a mu nicipal water supply with a base flow of 1,500 L/s. The field parameters, measured in January 2014, were: pH 6.81, specific conductance 292 uS/cm, and temperature 5.31 o Celsius. In 1989-1990, the Geological Survey of Alabama conducted an extensive investigation in the area, performing dye studies in storm water drainage wells (SDW-1 through SDW-20) to define the re charge area of Tuscumbia Spring. The storm wa ter drainage wells can be a potential source of contamination for the springs. Two rock samples from Vietnam and one from Tuscumbia, Alabama (U.S.A.) were collected and examined using the Xray diffraction (XRD) analysis, microscopic analy orimetry-Thermogravimetry (DSCTG) analysis. The quality of limestones in Vietnam and Tuscumbia (38.7 percent and 39.6 percent versus 31.10 percent calcium concentration) and the amount of precipitation (1,500 to 2,000 mm/m 2 in Vietnam versus 947 mm/m 2 to 1,960 mm/m 2 per year in Tuscumbia) are comparable. frequent tectonic uplifts and a complex geologic pattern result in the tower karst landscape in Vietnam versus a leveled landscape in Tuscumbia, Alabama. Tectonics is the primary driver for the formation of tower karst land scape in Cao Bang Province, Vietnam. Karst of Hoa Quang District, Cao Bang Province, Vietnam Geographic and Climatic Setting Vietnam is in the eastern part of the Indochinese Penin sula (Figure 1, Insert). It covers a total area of 331,210 km 2 of which mountains, mostly covered by forest, rep resent 40 percent. Carbonate rocks are exposed over 60,000 km 2 which represents 18.12 percent of Vietnam (Clements et al., 2006). Cao B ng Province is located in northeastern Vietnam, 270 kilometers north of Hanoi, on the border with China. The province comprises 6,724.6 km 2 and has a popula tion of 632,450. The topography of the region is charac mean sea level (MSL), and karstic plateaus developed be tween 500 and 700 m MSL. The elevation of the town of Cao Bang is 300 meters MSL and it has a temperate cli mate throughout the year. Annual rainfall ranges between 1,500 and 2,000 mm/m 2 and temperature varies from 5 Celsius in December and January to 37 Celsius in July and August. The average temperature in the province is 22 Celsius. Winter temperatures in some areas occasionCelsius. Winter temperatures in some areas occasionCelsius. Winter temperatures in some areas occasion The study area is located in the Hoa Quang Districts. The topography consists of small alluvial plains along the Bac ed by limestone pinnacles and tower karst (Figure 1), sinkholes (dolinas), closed depressions, sinking streams, springs, and large underground streams or submerged passages (Ponta et al., 2013). Regional Geology Northeastern Vietnam is in a tectonic active region lo cated at the boundary between the Indian Plate and Eur asian Plate, which created the Himalayas Mountain Sys tem and the Tibetan Plateau (Strong Wen et al., 2015). The area is underlain by numerous rock types and ages, which have undergone numerous phases of tectonic de Hai, 2009). The site is located a few kilometers north of Ailao Shan-Red River Shear Zone and Song Ma Fault Zone, which divides North Vietnam into two main tec tonic units: The Bac Bo Fold Belt (part of Eurasian Plate/ South China Terrane) and the Indochina Fold Belt (part of Indian Plate/ Indochina Terrane). The Bac Bo Fold Belt is composed by three fold sys tems: Tay Bac, Viet Bac and Dong Bac. Cao Bang Prov ince is located in the Dong Bac Fold System. Cao Bang Province is traversed by two deep fault systems, Lo Gam and Cao Bang Lang Son, which divide the re gion into three parts: the western uplift block (the Viet BAC with the Bong Son anticline), the central part The province is underlain by a variety of rocks ranging in age from Cambrian to Quaternary (Long, 2001). along the NW-SE axis. Between the two deep faults is Triassic age (not shown on Figure 1), including iron and manganese deposits. Northeast of the Cao Bang Lang 1).

PAGE 3

129 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 1. Geological map of Hoa Quang District, Cao Bang Province, Vietnam. Geology after Pham Dinh Long, 2000 (carbonate rocks are shown in white).

PAGE 4

130 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Geology of the Hoa Quang District The Hoa Quang karst area is 15 km long and 14 km wide (210 km 2 ). The central part of the study area is mainly a NW-SE trending faulted anticlinorium plunging north westward. The core of the anticline is formed by Cambri an and Devonian sandstones and siltstones belonging to the Than Sa (C 1-2 ts 2 ) and Luoc Khieu (D 1 lk ) Formations, intercalated with conglomerates, shale, and clay layers of Lower Devonian age (Son Cau Group [D 1 sc ] and Mia Le Formation [D 1 ml 1 two parallel faults one km apart, oriented north-south, parallel with the anticlines axis, which separates the Na Quan (D 1 nq 1-2 ) and Toc Tat (D 3 tt 1-2 ) Formations of Devonian age. In the northeastern part of the study area the Bac Son Formation (C-P bs ) outcrops with siliceous shale, shale-limestone, light gray limestone, and gray limestone (Long, 2001). line, and light gray to dark gray. Bedding is generally 30 to 50 cm thick, oriented NW-SE, dipping 19 o to 25 o SW, E, or NE. Karst Landforms The northeastern part of Vietnam provides a unique cave-forming environment with large rivers draining into the limestone valleys from the mountains. The Hoa Quang karst has developed into a tower karst landscape area underlain by limestones is extensive. Rainfall is abundant and the karsts vertical potential exceeds 200 m. Exokarst landforms are well represented by a variety East of the anticline, the tributaries of the Bac Vong Riv (Figure 2, Cross Section A-A). The Ban Vongs tribu taries collect waters a few hundred meters south of the border with China, and disappear underground through temporary or permanent swallets six times along a 13km path, before resurging in the Hang Ban Nua stream The Hoa Quang area is an old karst where isolated hills (tower karst) remain upon residual plains developed on an 800-m thick limestone unit. As illustrated on cross section A-A (Figure 2) tower karst emerges at bedding planes. These structural features appear on the 1:200,000-scale geologic map. More than likely, on a geologic map at a larger scale (1:10,000 to 1:100,000), more structural features will be documented. The presence of caves, phreatic or vadose, at the contact between towers and alluvial plains (foot-hill caves) are common. In some areas, two levels of cave passages are developed, one as a stream passage at The springs are free draining, recharged by allogenic dritic (a branch-work cave pattern where underground passages are present). The type of tower karst land Figure 2. Cross Section A-A along Bac Vong Rivers tributaries, traversing tower karst landscape in Hoa Quang District, Cao Bang Province, Vietnam; B-B Cross section in Tuscumbia, Alabama, U.S.A.

PAGE 5

131 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 scape developed along the Bac Vongs tributaries and Tra Linh River is residual hills on a planed limestone surface (Ford and Williams, 2007). Sampling Methods Eighteen water samples from cave waters (rimstone pools), rainwater, sinking streams, and springs were col hydrogen isotopes. Selected water-quality parameters were monitored at each sampling point with an YSI 63 instrument. Additionally, a Digital Titrator (Hach Model as alkalinity, total hardness as calcium, and carbon di oxide (CO 2 Ferrous Iron test kit Model IR-18C. UTM coordinates and elevation were recorded with a GPS (Garmin 62S) at each sampling point (Table 1). Groundwater samples from each source were collected and pre-treated as shown below. Samples for cations were collected in 40 mL vials pre-treated with nitric acid. Samples for anions and stable isotopes were col lected in 40 mL vials with no preservatives. Anion analyses were performed at the University of Alabama using a Dionex DX 600 Ion Chromatograph. Geology, University of South Florida, using a PerkinElmer Elan DRC II Quadrupole inductively coupled plasma mass spectrometer (ICP-MS) analytical instru ment. Standards used were formulated stock standards with metals in concentrations from 1,000 mg/L to 10,000 mg/L. The stable isotopes analyses were performed at Stable Ring Down Spectroscopy (CRDS) instrument. To evaluate the limestones from the study area, three rock samples (two from Vietnam, and one from Tus cumbia, Alabama, U.S.A) were collected and examined at R&D National Institute for Nonferrous and Rare Met als Romania using the X-ray diffraction (XRD) analysis, Scanning CalorimetryThermogravimetry (DSCTG) analysis. The DSC-TG thermic analyses were conducted with Setsys Evolution instrument (Setaram). A Zeiss Ax thin section analyses, and the X-ray diffraction XRD) analysis was performed with Bruker D8 Advance dif fractometer. Assessment of Results Water-Quality data pling locations are shown on Figure 1. The elevations of sampling points ranged between 443 m and 641 m MSL. from 0.5 L/s to 800 L/s. The observations were recorded Table 1. Field parameters Hoa Quang District, Cao Bang Province, Vietnam

PAGE 6

132 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE in February 2014, at the end of the dry season, so these Temperatures ranged from 13.5 o to 23.2 o Celsius, which correspond to mean annual air temperature in the area. The pH values ranged from 7.23 to 7.97, typical for ductance are typical for cave waters (between 153.2 and 421.6 uS/cm). Total alkalinity as CaCO 3 varies between 125 and 207 mg/L, the total hardness as CaCO 3 varies between 143 and 220 mg/L, and carbon dioxide (CO 2 ) ranged from 40.8 mg/L to 123.40 mg/L. Laboratory results for anions and cations are provided in Table 2. Calcium concentration ranged from 51.75 to 81.23 mg/L. T sium bicarbonate. Stable Isotopes Isotope fractionation accompanying evaporation from the ocean and condensation during atmospheric trans port of water vapor causes spatial and temporal varia tions in the deuterium and 18 O composition of precipita tion (Dansgaard, 1964). Regional-scale processes such as water vapor transport patterns across landmasses and the average rainout history of the air masses precipitat ing at a given place control the isotopic composition of local precipitation (Ponta et al., 2013). 18 2 H is mainly explained by the general Rayleigh distillation principle. 18 2 H values was recog water line (GMWL). The complex processes that are involved in the isotope fractionation alter the general 18 2 H values. The relationship between 18 O and 2 H in worlds fresh surface/cave waters is described by the global meteoric 2 H 18 O +10. 18 O values of the water samples collected ranged -7.87. The water sample with 18 O value of -2.34 was collected from a rimstone pool in the cave TR33 (Table 1). This sample plots very close to the GMWL, which indicate a close link to a Rayleigh process. Higher precipitation water from warmer months. The local meteoric water line (LMWL) in Vietnam is: 2 18 O + 12.44 ( .25) with r 2 =0.990 (Figure 3) on the basis of 38 measure ments (between 2004 and 2007) from the IAEA rainfall monitoring station in Hanoi (N21 o 2, E105 o 47) (IAEA/WMO, 2015). The local meteoric water line, based on our measure ments (Table 3) is: 2 18 O + 10.45 ( .86) with r 2 =0.998 which is close to the global meteoric wa 2 18 O + 10.35 Table 2. Summary of water-quality data Hoa Quang District, Cao Bang Province, Vietnam

PAGE 7

133 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 3. The relationship between hydrogen and oxygen isotopes in karst waters collected from streams and caves from Vietnam. Additional data used: GNIP (IAEA/WMO, 2015) and GMWL (Craig, 1961) intercept value differs very slightly from both local and global water lines. Due to the short sampling period, the information provided by the water stable isotopic com position is limited. 18 O ( ) (Dansgaard, 1964), calculated for the collected waters from karst areas (Table 3) are typical for continental me teoric waters (deuterium excess values of ~10 ) and have a median value of 11.51 There are some sam ples with lesser or higher deuterium excess values due to local variations in humidity, temperature, and wind speed Karst of Tuscumbia, Alabama Geographic and Climatic Setting The Tuscumbia karst area is located in the northwestern part of Alabama (Colbert County), in the southeastern United States (Figure 4, Insert) along the Tennessee Riv er, at latitude 34 o N. The area has a mild humid climate. The annual precipitation ranges between 947 and 1960 mm/m 2 (1980 through 2014) and the average annual peratures last only a few days. Regional Geology The Tuscumbia karst area is located on the southern teaus physiographic province. The area is underlain by sedimentary formations of Mississippian age limestones, sandstones, and shales). Local geology is typical of the terrain associated with limestone bedrock, overlain by a mantle of clay-rich, unconsolidated material, 5 to 8 meters thick (Fig ure 4). The rocks dip generally toward the southwest at 5 m per kilometer (0.3 o SW). The Fort Payne Chert-Tuscumbia Limestone aquifer system with a thickness of 160 meters is the most im portant water-bearing unit in the Tuscumbia area. It is part of the Mississippian aquifer system that underlies three counties in northwestern Alabama (Colbert, Frank lin, and Lawrence). The Chattanooga Shale (not shown on Figure 4), at the base of the aquifer system, restricts the downward movement of water to lower units, which are not known to be water-bearing (Chandler and Moore, 1991).

PAGE 8

134 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Table 3. Summary of stable isotopes data Hoa Quang District, Cao Bang Province, Vietnam Geology of the Tuscumbia Area The study area is 12 km long and 12 km wide (144 km 2 ). The Fort Payne Chert (not shown on Figure 4) of Lower Mississippian age consists predominantly of medium gray to light gray and white crystalline limestone containing abundant light gray to black Tuscumbia limestone (Mt) of Valmeyeran age is a medium-bedded light bluish-gray hard, dense, fineto medium-grained bioclastic limestone containing abundant light-colored bedded and nodular chert by the Pride Mountain Formation (Mpm) of Cheste rian age and consists of shale, a basal limestone up to The Hartselle sandstone (Mh) is a light gray mas conformably overlies the Pride Mountain Formation site consists of a northwest-trending set, as evidenced by the alignment of closed depressions and sinkholes in the area (Figure 4), and a southwest-trending set iden and Moore in 1991. Karst Landforms The karst plain developed at 152 m MSL in the study area looks like a river bedrock terrace. It is part of a mature karst with low topographic relief except for the southern edge of Pickwick Lake, where limestone bluffs occur (Chandler and Moore, 1991). Exokarst landforms are well represented by a variety of small to ure 2, Cross Section B-B). The sinkholes were formed by the gradual removal, by solution, of the limestone at the base of the Pride Moun tain Formation (Chandler and Moore, 1991). Sinkholes in this area are very effective in collecting surface runoff waters. The drainage is primarily subterranean (John ston, 1933). Spring Creek, with its tributaries, is the only surface stream that crosses the sinkhole plain (Johnston, 1933). Due to urban development, some of the karst features represented on Figure 4 (based on a 1975 7.5-minute

PAGE 9

135 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 4. Geological map of Tuscumbia, Alabama U.S.A. Geology after Szabo 1975. Groundwa ter Flow Directions after Chandler and Moore 1993 (carbonate rocks shown in white).

PAGE 10

136 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE date the construction of dwellings or industrial facilities. Tuscumbia Spring is a municipal water supply source with a base flow of 1,500 L/s. According to the Geological Survey of Alabamas Groundwater Assess ment Program (Real Time Monitoring), between May 2,133 L/s with a minimum of 1,423 L/s and a maximum of 6,211 L/s (Geological Survey of Alabama webpage). ture 15.310 o Celsius. The recharge area of Tuscumbia Spring is 218 km 2 Dye Studies In 1989-1990, the Geological Survey of Alabama conducted an extensive investigation in the area, performing dye trace Seventeen storm water drainage wells were selected for indicate that hydrologic interconnections exist between 17 wells and springs of the Tuscumbia area (Table 4). Strong recoveries of tracers at Tuscumbia Spring were recorded from SDW-11 and SDW-16 (Figure 4, Table 4). The travel-time data (3 hours) and high groundwa water drainage well SDW-11. The water is of the calci um-sodium-bicarbonate-sulfate-chloride type, slightly alkaline and hard, but low in mineral and chemical con stituents (Chandler and Moore, 1993). Rock samples Laboratory results are provided in Table 5. The two rock samples collected in Vietnam (latitude 22 o N., Figure 1) had calcium concentrations ranging from 38.7 percent and 39.6 percent, while the magnesium concentration collected in Vietnam are micritic limestones with veins in Tuscumbia Alabama, U.S.A., (latitude 34 o N., Figure 4) had a calcium concentration of 31.1 percent, silica 8.9 percent, and 0.28 percent aluminum. The thin section reveals that the rock is a siliceous micritic limestone, coarsening to bioclastic wackestone and calcite spar. The assessment revealed chemical, mineralogical and petro graphic differences between the samples, allowing clas stone (Hoa Quang) and siliceous limestone (Tuscumbia). Summary Hoa Quang, Vietnam is at latitude 22 o N., in the peri Alabama, is located at latitude 34 o N., in a subtropi In the area studied in Vietnam, bedding is generally 30 to 50 cm thick, oriented northwest-southeast, and dipping 19 o to 25 o cumbia limestone in Alabama (50 to 100 cm thick) dips gently southwestward at approximately 4.8 to 5.7 m per kilometer (0.3 o The two rock samples collected in Vietnam (Figure 1) have calcium concentrations ranging from 38.7 percent siliceous limestone to 39.6 percent, where as the sample collected in Tuscumbia Alabama, U.S.A., (Figure 4) has a calcium concentration of 31.1 percent (with 8.9 percent silica, and 0.28 per cent aluminum). In both cases, the carbonate rocks In Vietnam, the central part of the study area is mainly a northwest-southeast trending faulted anti clinorium plunging toward the northwest. The west several faults and thrust faults parallel with the an ticlines axis. In Tuscumbia, Alabama, on the 7.5-minute topo publication, relatively dense sets of lineaments ori ented mainly northwest-southeast and northeastsouthwest are interpreted from LANDSAT imagery, high altitude black and white, color photography, and 7.5-minute topographic maps. The amount of precipitation in Tuscumbia, Ala bama, ranged between 947 mm/m 2 to 1,960 mm/m 2 per year versus 1,500 to 2,000 mm/m 2 per year in Vietnam. In the water samples collected in Vietnam, the amount of dissolved carbon dioxide (CO 2 ) ranged from 40.8 mg/L to 123.40 mg/L (Table 1) and cal cium concentration ranged between 51.75 to 81.23 mg/L (Table 2). streams recharged by epigenetic springs and mete The maximum discharge of a spring in Vietnam is The waters in Tuscumbia, Alabama, are collected from 280 km 2 versus 20 to 30 km 2 in Vietnam.

PAGE 11

137 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Table 4. Summary of dye study data Tuscumbia Alabama from Chandler and Moore 1991 Table 5. Summary of rock sample data

PAGE 12

138 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Conclusions The Hoa Quang karst has developed into a tower karst basins. The area underlain by limestones is extensive. Rainfall is abundant and the karsts vertical potential ex ceeds 200 m. Exokarst landforms are well represented As shown on cross section A-A, the tower karst in Viet nam emerges at the intersection of the land surface with features appear on the 1:200,000-scale (medium scale) geologic map. More than likely, when a geologic map at a larger scale (1:10,000 to 1:100,000) is prepared, more structural features will be documented. The base map for the Tuscumbia area is a 1:24,000-scale topographic/geologic map. No fractures or faults are quality of limestones in Vietnam and Tuscumbia (38.7 percent and 39.6 percent versus 31.10 percent calcium concentration) and the amount of precipitation (1,500 to 2,000 mm/m 2 in Vietnam versus 947 mm/m 2 to 1,960 mm/m 2 per year in Tuscumbia) are comparable. frequent tectonic uplifts and a complex geologic pattern result in the tower karst landscape in Vietnam versus a leveled landscape in Tuscumbia, Alabama (Figure 2). Tectonics is the primary driver for the formation of tow er karst landscape in Cao Bang Province, Vietnam. Acknowledgments This material is based upon work supported by the Na tional Speleological Society Research Grant to Gheorghe Ponta. Sponsors for the hydrogeological investigation are the Institute of Geosciences and Mineral Resources Vietnam, Geokarst Adventure, the University of South poca, Romania. References Budel J. 1977. Klima-Geomorphologie. Borntraeger, Berlin. Chandler RV, Moore JD. 1991. Hydrogeologic and water quality evaluation of storm-water drainage wells, Muscle Shoals area, Alabama. Geological Survey of Alabama, Circular 158. Chandler RV, Moore JD. 1993. Fluorescent dye trace evaluation of storm-water drainage wells in the Muscle Shoals area, Alabama. Geological Survey of Alabama, Circular 173. 2006. Limestone karsts of Southeast Asia: imperiled arks of biodiversity. Bioscience 56 (9): 733-742. Craig H. 1961. Isotopic variations in meteoric waters. Science 133: 1702-1703. Dansgaard W, 1964. Stable isotopes in precipitation. Tellus. XVI (4): 436-468. Ford DC, Williams PW. 2007. Karst hydrogeology and geomorphology. London: John Wiley & Sons Press. Geological Survey of Alabama. Groundwater Assessment Program. Real Time Monitoring Program. Tuscumbia Spring: http://www.ogb.state. al.us/waterweb/realtime/tuscumbiasp.aspx IAEA/WMO. 2015. Global network of isotopes in precipitation. The GNIP Database. Accessible at: http://www.iaea.org/water. Jakucs DL. 1977. Morphogenetics of karst regions. New York A Halsted Press Book John Wiley & Sons Press. rocks of northern Alabama, Geological Survey of Alabama. Special Report No. 1. Long PD, editor. 2001. Geology and mineral resources map of Viet Nam, Chinh Xi: Long Tan sheet, 1: 200,000. Hanoi (VN): Department of Geology and Minerals of Vietnam. Ponta G, Onac BP, Nguyen XN. 2013. Karst hydrogeological observations in Cao Bang province (Vietnam): The Tra Linh-Thang Hen Lake Speleological Society Praha. 16th International Congress of Speleology Proceedings 3: 116-120. Isotope patterns in modern global precipitation, Geophysical Monograph 78. In Climate Change in Continental Isotope Records. American Geophysical Union: 1-36. Strong Wen, Yu-Lien Yeh, Chi-Cha Tang, Lai Hop Phong, Dinh Van Toan, Wen-Yen Chang, ChauHuei Chen. 2015. The tectonic structure of the Earth Sciences 107: 26-34 Tuscumbia Quadrangle, Alabama, Geological Survey of Alabama Quadrangle Series Map 6. Tran Thanh Hai 2009. Deformational features of Northeastern Vietnam in: Tran Van Tri, Vu Khuc (Eds) Geology and natural resources of Vietnam, Natural Sciences and technology Publishing House, Hanoi, p. 408-415.



PAGE 1

299 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 A METHOD OF MAPPING SINKHOLE SUSCEPTIBILITY USING A GEOGRAPHIC INFORMATION SYSTEM: A CASE STUDY FOR INTERSTATES IN THE KARST COUNTIES OF VIRGINIA Abstract has been chemically dissolved by acidic groundwater, producing subsurface voids that may lead to sinkholes at the surface if the overlaying soils can no longer support their own weight and collapse. The western counties of Virginia have a high concentration of karst areas due to widespread occurrence of carbonate rock exposures, and their geomorphic development within the Appalachian mountains. As a result, the Commonwealth of Virginia Department of Transportation (VDOT) develop a meth od to determine the roadways and regions most suscep tible to experiencing sinkholes, in an effort to reduce the possibility of reported sinkhole damage to property. While many noninvasive methods exist to detect subsur face voids, such as electric resistivity imaging, micro gravity, ground penetrating radar, and seismic surveys, these methods are time consuming and costly. This study proposes the use of a geographic information system (GIS) to create a susceptibility map of regions in the karst counties of Virginia, and in particular along interstate highways, that are most susceptible to future sinkhole development. Five factors that have previously been shown to play a role in the acceleration of sinkhole formation in Virginia include: bedrock type, proxim ity to fault lines, drainage class, slope of incised river banks, and minimum soil depth to bedrock. The analy sis compares 1:24,000 scale maps of existing sinkholes developed by Virginia Department of Mines Minerals and Energy (DMME) with a series of maps represent to determine which weighted combination is most ap ceptibility map. The layers representing each factor are created using publicly available tabular and spatial data taken from the United States Department of Agriculture (USDA) Soil Survey Geographic (SSURGO) Database, the United States Geological Survey (USGS) National Map, the USGS Mineral Resources Online Data, and the National Weather Service. The methodology used to tabase is highlighted within this paper. Data from the SSURGO database is used to create the bedrock type, drainage class, and minimum soil depth to bedrock lay sinkhole susceptibility in other karst regions, by simply ogy of a particular region. Introduction Karst terrain forms as acidic groundwater interacts with soluble bedrock, during which subsurface draining causes unique solutional patterns to carve into the rocks, forming cavities. The resulting voids introduce the po tential to trigger land subsidence in the event that the bard, 2001). The western counties of the state of Virginia contain abundant karst areas, because of the widespread occur rence of carbonate rock exposures, and their geomor phic development within the Appalachian mountains, ultimately locating the karst areas in long valleys con taining extensive folds and fractures of limestone and dolomite bedrock (Belo, 2003). This folded and faulted by differential weathering of rock units, and provides a natural setting for karst terrain and sinkhole formation as carbonate strata are exposed at or near the surface Sinkholes pose engineering complications and the risk of damaging property and endangering lives if devel oped in a highly populated or well-traveled area. This paper focuses on the natural factors of sinkhole forma tion, and their combination within a geographic infor mation system (GIS) in order to create maps of sink hole susceptibility. While impossible to fully eliminate Alexandra Todd Department of Civil Engineering, University of Virginia, 1333 Webster Street, New Orleans, Louisiana, 70118, USA, alt5ry@virginia.edu Lindsay Ivey-Burden Department of Civil Engineering, University of Virginia, 351 McCormick Road, P.O. Box 400742, Charlottesville, Virginia, 22904, USA, lindsay.ivey@virginia.edu

PAGE 2

300 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE ated through effective implementation of investigative techniques where areas of greater sinkhole susceptibility Due to the public availability of spatial and tabular da tasets provided by agencies such as the United States Geological Survey (USGS), the use of GIS techniques 1999). This investigation proposes a new method of us ing a GIS and data from the Soil Survey Geographic Da tabase (SSURGO) to create layers to predict where those sinkholes might form in an effort to avoid such dangers, vide an inexpensive and quick method of better locating proposed roadway passages to aid in avoiding impact to karst areas (Moore et al., 2008) and determining which roadways may require immediate safety evaluations, ditionally, factors input into the methodology developed ogy of other karst regions with similar data availability. Background There have been few studies aiming to accurately esti mate sinkhole risk due to the lack of detailed datasets, Karst maps were the sole method of assessing subsidence tion does not necessarily [imply] that the whole locality is at risk for land subsidence (Virginia Department of Emergency Management, 2003). Without a well-estab lished set of guidelines that predict probabilities, a true risk determination cannot be formed. In a study on sinkhole distribution in Virginia, Hubbard (2001) determined sinkhole locations by stereoscopic checking any questionable sinkholes. Hubbard deter mined that detected sinkholes mainly occur in regions where carbonate rocks are present, where structural folds to deeply incised rivers and tributaries (Hubbard, 2001). Additionally, it was noted that not all sinkholes can be detected by aerial stereophoto pairs, since aircraft tilt while making other low slope regions appear as sink holes when in fact they were none. Hubbard (2001) es timated that it would take 250 years to map every single sinkhole in Virginias Valley and Ridge province using recent acquisitions of Light and Detection and Ranging (LIDAR) data in Virginia allow for the creation of highly accurate (sub-meter) elevation models that can be used for rapid, precise detection of sinkholes (Doctor and Young, 2013). The growing body of sinkhole datasets has driven scien could be implemented into a GIS to create maps of sink hole susceptibility across broad areas. Hyland (2005) with existing fault lines. in karst regions as much as in non-karst regions because it percolates into the subsurface caves and conduits. Hence a region with rapid surface drainage, or a greater hydraulic gradient and lack of surface water, might im ply subsurface drainage pathways that could potentially lead to the possibility of sinkhole formation in that par ticular area. Smaller surface streams often do not exist or endure as voids accommodate most of the water into the Proximity to these deeply incised rivers remaining above the surface is most likely indicative of a sinkhole suscep tible region, because of the steepened hydraulic gradient areas (Muckel, 2004). Green et al. (2002) decided that sinkhole risk studies should focus on shallower regions of bedrock, conclud ing that the timescale for which sinkholes may develop can be hours to months for shallow depth to bedrock, where it may be decades to centuries with a thicker depth to bedrock. This study aims to create a sinkhole susceptibility analy sis map by combining 5 factors bedrock type, prox imity to fault lines, drainage class, proximity to incised river banks, and depth of the overlying soil into a single representative map spanning the western counties of Vir detail the methodology used to create the bedrock type, drainage class, and depth of overlying soil layers, which base. Methodology Study Area The region of interest in this study involves twenty-sev en counties in Virginia that contain karst terrain, west of

PAGE 3

301 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 includes approximately 29,853 square kilometers and ranges from -83 o 40 to o 19 latitude and from 39 o 27 to 36 o 35 longitude. Data Acquisition and Preparation data was taken from four sources. Bedrock type, depth to bedrock, drainage classes, and county and map unit boundaries were obtained from the SSURGO database (websoilsurvey.nrcs.usda.gov), digital elevation models and the Virginia state boundary were from the USGS National Map (nationalmap.gov), fault lines were down loaded from the USGS Mineral Resources Online Spa tial Data (mrdata.usgs.gov), and rivers were obtained from the National Weather Service website (weather. gov). The analysis created a compiled ranked map de termining regions of potential sinkhole formation based contained a map of regions assigned a Sinkhole Value ranging from 1-15 (1 is low susceptibility and 15 is high based on the corresponding risk factor. The distinct maps were ultimately combined using weights representing the regions, and then the most appropriate combination was statistically determined using a residual sum of squared ity map. The main contribution of this paper, however, is the methodology created to extract relevant sinkhole layer data from the SSURGO database. This database has been created by the National Cooperative Soil Survey over many years and spatially references surface soil data at scales ranging from 1:12,000 to 1:63,360. The database offers a detailed description of the surveyed database to work with as it contains very little unifor mity among entries, varying soil descriptions, and some duplicate entries. In the creation of the bedrock type, nal representative sinkhole susceptibility map produced sort and gather relevant information from the SSURGO tables for the creation of the aforementioned layers. Bedrock Type Bedrock type is a contributing factor to sinkhole for mation since sinkholes have proven to form in regions of relatively pure carbonate rocks. The bedrock type layer was derived from SSURGO tabular data located in the Component and the Component Parent Material (COPM) tables for each of the 27 counties of interest. Desired attributes from the individual tables were com a Python script, converted to pseudo-code for simplicity: # Loop through each countys SSURGO database each map unit and add to a new table # Add a new column to table called Sinkhole Val ue Figure 1. Virginia counties in study area.

PAGE 4

302 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Value for every map unit depending on its parent material Value = 4 materials a Sinkhole Value = 2 hole Value = 3 = 1 COPM table, where formation kind was described as alluvium colluvium or residuum and formation origin ranged through a number of bedrock origins, such as limestone sandstone or shale The Component table percentage. The script removed any duplicate or null entries as well as any entries where the parent material kind was alluvial, since alluvium soils rarely play a role in karst development. If a sinkhole does in fact exist in a region with alluvial deposits, it is more likely a result of one of the other factors in this study (proximity to faults or depth to bedrock, for example) rather than a result of the bedrock type (Hyland, 2005). was added to the resulting table, determined by the parent material origin, assigning pure limestone origins the highest value of 4, limestones and dolomite com binations a value of 3, values containing only partial carbonate rock a value of 2, and entirely clastic and table from the script was imported into Microsoft Excel, per map unit using the component percentage and the Unweighted Sinkhole Value: Component1 Percentage x Unweighted Sink hole Value1 + Component2 Percentage x Unweighted Sink hole Value2 + ... = Final Sinkhole Value for Corresponding Map Unit into 15 equally incremented categories to remain con imported into a Microsoft Access Database to then be with the spatial map units using the Add/Join tool in ArcMap (Figure 2A). Drainage Class Drainage plays a role in predicting sinkhole risk occur rence because it provides information on how rapidly water will drain through the soil type. Drainage class ( drainagecl Com ponent table, composed of values ranging from exces sively drained to very poorly drained A Python script was written creating a table that combined each map unit with its corresponding drainage class. Only 7 different drainage classes were listed, so the re the odd numbers ranging from 1 to 5, and regions that as the odd numbers ranging from 9 to 15, since this most likely meant water was being absorbed into subsurface karstic drainage systems. Zero was assigned to the ex cess regions with no drainage class assignment. Sinkhole Value 7 had no assignment in this layer. (Figure 2C). Depth of Overlying Soil The timescale during which sinkholes may develop is much shorter in regions of shallow depth to bedrock. Thus a sinkhole factor layer representing depth to bed rock was created using SSURGO data found in the Ma punit Aggregated Attribute (MUAGGATT) table. The bedrock minimum depth entry for each map unit along the 27 counties was recorded into a table using a Python script, based on a common map unit key found in the Mapunit and the MUAGGATT tables. The resulting output table was imported into ArcMap, depth ranges were converted to raster and were then re increment below the surface. Shallow soil cover overly ing the bedrock was assigned the highest Sinkhole Value of 15, and the deepest amount of soil cover was assigned the lowest Sinkhole Value of 1 (Figure 2E). Proximity to Fault Lines Hyland (2005) determined a correlation between sink hole formation and proximity to existing fault lines. To Map, the Multiple Ring Buffer tool was used around the fault lines extracted from USGS Mineral Resources Online Spatial Data. Rings ranging from 0 3000 feet from the faults in increments of 200 feet were created Values of 1 through 15. Values were highest at regions closest to the faults, decreasing outwardly as distance in creased from faults (Figure 2B).

PAGE 5

303 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Slope of Incised River Banks A higher risk for sinkhole formation near incised river banks can be attributed to a higher hydraulic gradient determine the degree of incised rivers, Virginia rivers downloaded from USNWS geospatial data were added to the map layer, with a buffer region constructed using the Multiple Ring Buffer tool in ArcMap to account for river widths. Slopes along riverbanks were identi from the USGS National Map and clipped to only ex slopes of the elevation model rasters were computed us ing the ArcMap Slope tool. The total range of slopes were split into 15 equally incremented categories and consequent categories, assigning low slopes in the re gions surrounding rivers to a Sinkhole Value of 1 and high slopes to a Sinkhole Value of 15 (Figure 2D). Statistical Analysis (Creating Weights) sinkhole occurrence in the karst counties of Virginia, twenty-eight different risk maps were created using the vidual risk layer raster images using a series of different chosen weights, as seen in the equation below: (A x Bedrock Type) + (B x Proximity to Faults) + (C x Minimum Depth to Bedrock) + (D x Drain age Class) + (E x Slope of Incised River Banks) = Weighted Combination Susceptibility Map where A, B, C, D, and E are chosen weights assigned to its corresponding combination, and A+B+C+D+E = ments, values were assigned in increments of one tenth. Bedrock type has been shown in existing karst literature hole formation, therefore combinations were chosen giv Figure 2. Five risk layers used in weighted combination.

PAGE 6

304 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE ing bedrock type the highest weight for all combinations except a control combination (combination 1), where each layer was assigned equal weight. Upon completion of the 28 distinctly weighted risk maps, the map that most closely and statistically corresponded to the imported data of mapped sinkhole locations would be chosen. A Python script that could loop through each combination, automating the following steps, was con structed. Spatial data containing existing mapped sink holes was added into the map. Each weighted risk map based on its Sinkhole Level (1-15) so that it could be clipped into the boundary of the existing sinkholes. Us ing the Dissolve tool, the newly clipped risk map was table of the polygon to calculate the total areas of each distinctive Sinkhole Level. In Excel, the areas of sink holes corresponding to each Sinkhole Level, the total area of existing sinkholes, and the percentages of each individual level compared to the total sinkhole area were computed and recorded. For simplicity, the 15 Sinkhole ues 1-3 as a Low Risk Zone, 4-6 as a Medium-Low Risk Zone, 7-9 as a Medium Risk Zone, 10-12 as MediumHigh Risk Zone, and 13-15 as a High Risk Zone. The ideal percentages of existing sinkhole areas per risk pared with the observed percentages using a Residual Sum of Square (RSS) error test. In an ideal situation, there would be no actual sinkholes found in the Low Risk Zone and the highest percentage of actual mapped sinkholes would be found in the High Risk Zone, with a the sum of the total percentages being 100. Hence this investigation predicted that percentages should be 0, 10, Error between the actual and the ideal models exempli combination were computed and ranked. Thus the com bination with the lowest RSS was the value most closely map. Results Final Map with Interpretation which had the smallest RSS when compared with the predicted model, was created using the following equa tion: Weighted Map = (0.6xBedrockLayer) + (0.1xProx imityToFaultLayer) + (0.2xDrainageClass) + (0xSlopeOfIncsiedRiverBanks) + (0.1xDepthToBe drockLayer) Figure 6 also displays a USGS Karst Terrain map beside resulting from chemically dissolved limestone bedrock, conclusions reached. From a visual comparison, the re sulting values make sense, since the regions with karst terrain on the USGS map align with the higher risk re gions on the sinkhole risk map. Sources of Error mate sinkhole susceptibility map were based on a statisti cal comparison between the constructed predicted at risk regions and existing sinkholes. However, to make clear that his sinkholes were mapped as a guideline and not a data using a scale that was 10 times less accurate than the scale the public had requested. Furthermore, the aerial photography used in Hubbards study cannot accurately detect all sinkholes due to aircraft tilt, which creates the illusion of a sinkhole where it may not exist and does source of error in choosing the appropriate weights for the combination of risk layers. By using the USDA Soil Survey data, a degree of error was inevitable due to the fact that data tables from which within the study. Additionally, SSURGO data contained values corresponding to Virginia counties but not all ma The type of data available further limited the scope of this study. The slope of incised riverbanks risk layer was determined using digital elevation models, and raster im ages were created using remote sensing or based on ex isting topographic maps. While these are helpful for ana in karst analyses arise since the directions of subsurface subsurface aquifers are not represented in them (Taylor et al., 2008). It would be much more useful to know the banks, since large slopes on the surface do not neces sarily denote fast paced water travel through subsurface

PAGE 7

305 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 were created and assigned Sinkhole Values based on hy potheses derived from background knowledge and engi value was assigned incorrectly. Conclusion This analysis used a geographic information system and readily available data from the SSURGO database to create a map that represents regions most at risk of sink hole formation in the karst counties of Virginia. While these results serve as a general guideline for mapping karst regions in Virginia, it is important to understand that a risk map created based on generalities cannot be landforms relate to one another, but the combinations and behaviors of the relationship between local, hydro logical, and climactic conditions are numerous (Ford and Williams, 1989). While this map provides a general un gion of interest. In order to apply this technology to new regions, a thorough background of the geology for that region is necessary and factor layers for sinkhole forma applied to produce similar results. References Karst Terrains in Virginia [masters thesis]. Virginia Polytechnic Institute and State University. 48 p. Dai J, Lei M, Lui W, Tang S, and Lai S. 2008. An Guilin, Guangxi Province, China. In: Yuhr LB, Alexander EC Jr, Beck BF, editors. Sinkholes and the Engineering and Environmental Impacts of Karst. Proceedings of the 11th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 2008 Sept. 22-26. Tallahassee, FL. ASCE. p.156-164. Davies WE, Simpson JH, Ohlmacher GC, Kirk WS, and Newton EG. 1984. Engineering Aspects of Karst. In: National Atlas of the United States of America, U.S. Geological Survey, 1 plate, 1:7,500,000 scale. Doctor DH and Young JA. 2013. An Evaluation of Automated GIS Tools for Delineating Karst Sinkholes and Closed Depressions from 1-Meter LIDAR Derived Digital Elevation Data. In: Proceedings of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 2013 Sept. 22-26. Carlsbad, New Mexico. ASCE p. 449-458. Ford DC and Williams PW. 1989. Karst Geomorphology and Hydrology. Chapman and Hall, London. Figure 3.

PAGE 8

306 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE wm.edu/geology/virginia/provinces/valleyridge/ valley_ridge.html. Virginia Department of Emergency Management. 2003. Risk Assessment (HIRA): Section 3-14 Land Subsidence (Karst). Virginia Tech Center for Geospatial Information Technology. 15 p. Available from: http://www.vaemergency.gov/webfm_ send/853/Section3-14-KarstTopography.pdf Whitman D, Gubbels TL. 1999. Applications of GIS Technology to the Triggering Phenomena of Sinkholes in Central Florida. In: Beck BF, Pettit AJ, and Herring JG, editors. Hydrogeology and Engineering Geology of Sinkholes and Karst. p. 67-73. Green JA, Marken WJ, Alexander EC, and Alexander SC. 2002. Karst Unit Mapping using Geographic Information System Technology, Mower County, Minnesota, USA. Environmental Geology 42 (5): 457-461. Hubbard DA Jr. 2001. Sinkhole Distribution of the Valley and Ridge Province, Virginia. In: Geotechnical and Environmental Applications of Karst Geology and Hydrology: Proceedings of the Eighth Multidisciplinary Conference on Sinkholes, 2001 April 1-4. Louisville, KY. ASCE. p. 33-36. Hubbard DA Jr. 1983. Selected Karst Features of the Northern Valley and Ridge Province, Virginia. In Virginia Division of Mineral Resources, Publication 44. Hyland SE. 2005. Analysis of Sinkhole Susceptibility and Karst Distribution in the Northern Shenandoah Valley, VA: Implications for Low Impact Development (LID) Site Suitability Model [masters thesis]. Virginia Polytechnic Institute and State University. 62 p. Ivey, Burden L. 2013. Karst Topography: Noninvasive Geophysical Detection Methods and Construction Techniques. Virginia Council for Transportation Research and Innovation. Final Contract Report. Moore H, McDowell L. The Development and Use of Karst Maps in the Location of Highways in Eastern Tennessee. In: Yuhr LB, Alexander EC Jr, Beck BF, editors. Sinkholes and the Engineering and Environmental Impacts of Karst. Proceedings of the 11th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 2008 Sept. 22-26. Tallahassee, FL. ASCE. p. 680-693. Muckel GB, editor. 2004. Understanding Soil Risks and United States Department of Agriculture, Natural Resources Conservation Service, National Soil Survey Center. Taylor CJ, Kaiser WP, Nelson HL Jr. 2008. Application of Geographic Information System (GIS) Hydrologic Data Models to Karst Terrain. In: Yuhr LB, Alexander EC Jr, Beck BF, editors. Sinkholes and the Engineering and Environmental Impacts of Karst. Proceedings of the 11th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 2008 Sept. 22-26. Tallahassee, FL. ASCE. p.146-155. Valley and Ridge Province, Geology of Virginia [Internet] 2014. [Place of publication unknown]: William and



PAGE 1

167 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 A PROPOSED HYPOGENIC ORIGIN OF IRON ORE DEPOSITS IN SOUTHEAST MINNESOTA KARST E. Calvin Alexander, Jr. Earth Sciences Dept., University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, USA, alexa001@umn.edu Betty J. Wheeler Earth Sciences Dept., University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, USA, whee0023@umn.edu northern Minnesota (Morey, 1998). The high grade iron ore was mostly goethite and hematite and occurred as near-surface, relatively small pods which unconformably filled paleokarst depressions in the Devonian Spillville Formation and the Ordovician Stewartville Formation. Following extensive exploration work that was conducted in the 1930s, two companies carried out mining operations in the Fillmore County district from 1942 to 1968. Cumulative production was 8.1 million tons of iron ore (Bleifuss, 1972 p.498). discontinuous bodies of the nominally Cretaceous Ostrander Gravels. The ore bodies were covered with a few meters of unconsolidated Pleistocene glacial drift and loess and Holocene sediments. The mining was dump trucks. The ore was shipped by rail mainly to mills in the St. Louis area. The iron ores are conventionally mapped as the Iron Hill Member of the Windrow Formation (Andrews, 1958). Andrews (1958) stratigraphic study of the Windrow Formation in the Upper Mississippi Valley, mainly southern Minnesota, southwestern Wisconsin and in northern Iowa, reviewed the literature up to 1958. Based on the literature and his own extensive work, Andrews (1958, p. 597) concluded It seems probable that the Iron Hill member was deposited as a result of reaction of iron-charged waters with carbonate bedrock. Rodney Bleifuss PhD thesis (Bleifuss, 1966) and subsequent publication (Bleifuss, 1972) are the most definitive works on the origin of the iron ores of southeastern Minnesota. Bleifuss thesis work was conducted during the active phase of the iron mining. He observed, studied, and documented many of the iron ore Abstract From 1942 through 1968 there was an active iron ore mining industry in western Fillmore, eastern Mower and southern Olmsted Counties of Minnesota. This iron mining district was 250 miles south of, and the ores were a billion years younger than, the ores of the classic iron mining districts in northern Minnesota. The high grade iron ore was mostly goethite and hematite and occurred as near-surface relatively small pods which unconformably filled paleokarst depressions in the Devonian Spillville Formation and the Ordovician Stewartville Formation. The source of the iron has long been cryptic. The available field and textural evidence is consistent with a hypogenic origin of these iron deposits. Before the current Mississippi River drainage system was incised, regional ground water flow systems could have emerged The waters in the deeply buried aquifers underlying this area currently are anoxic and enriched in dissolved ferrous iron and would have been more so before the regional ground water flow system. When that water emerged into the atmosphere the ferrous iron would and abiotic processes producing the ferric oxide ore at the spring orifices. Numerous springs and seeps in Minnesota are currently building iron oxide deposits at their orifices. Introduction The presence of iron ore deposits in southern Minnesota sedimentary rocks and are often covered by Pleistocene glacial deposits. They are distinctly separated in time and

PAGE 2

168 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The Ore Bodies Location Figure 1 shows the locations of the iron ore leases (MDM, 1941-1970) plotted on top of the bedrock geology of the mining district in western Fillmore (Mossler, 1995), eastern Mower (Mossler, 1998) and southern Olmsted (Olson, 1988) Counties in Minnesota. Figure 1 is an updating of Figure VI-43 in Bleifuss (1972, p. 499). Figure 1 is different from Bleifuss Figure VI-43 only in the bedrock geology, which has been significantly updated. All of the ore bodies were located on what is now interpreted as either the Devonian Spillville Formation or the Ordovician Stewartville Formation. The Spillville is a subdivision of Bleifuss (1972) Cedar Valley Formation. The Stewartville Formation is a subdivision of Bleifuss (1972) Galena Formation. Based on the more recent geologic mapping shown in Figure 1, for the rest of this paper we will update the formation names from Bleifuss (1966, 1972) by substituting Spillville for Cedar Valley and Stewartville for Galena. The Stewartville and Spillville Formations have the greatest secondary karst transmissivity of the geologic units shown on this map. All of the geologic units on Figure 1 regionally dip at a few feet per mile to the southwest. The iron ores are conspicuously not present on the Maquoketa and Dubuque Formations which are stratigraphically between the Stewartville and Spillville Formations. Figure 2 is modified from Andrews (1958) Figure 2 with the names of the geologic units updated to current nomenclature. Although Andrews did not use the word an important part of the process. Andrews (1958, p. 614-615 ) reasoned, based on the work of Krumbein and Garrels (1952), that the iron was transported in an acidic solution (pH less than 7) in the ferrous state and that deposition resulted from an increase in pH of the solution. This increase in pH may be logically attributed to the reaction of the acidic solution with carbonate bedrock and resulted in precipitation of ferric the ferric oxide could be precipitated in this manner both at the surface and by downward-percolating waters ( emphasis added ) in fissures of the underlying carbonate bedrock. bodies as they were being mined. Bleifuss (1972, p. 498) 1. The ores were formed by weathering of the 2. The development of the ore bodies required some supplementary process of concentration, involving migration and local concentration of 3. The age of the Windrow Formation is Cretaceous, and the deposits in the Fillmore County district are correlative with similar lithologic units of 4. Fossil evidence that would positively date the 5. The most likely age of the iron-rich residuum and associated iron ores is Cretaceous. Bleifuss (1972, p. 498) argued to the contrary that his observations and data indicated the ores are Tertiary in age, and that they were developed from the oxidation of a primary marine siderite faces of the Cedar Valley Formation. The origin of the southeastern Minnesota iron deposits has long been cryptic and controversial and remains so. The fundamental issue, on which there is no consensus questions. What was the source of the iron? How did that iron accumulate into mineable ore bodies in the Fillmore County district? The thesis of this paper is that available field and textural evidence is consistent with a hypogenic origin of these iron deposits. Before the current Mississippi River drainage system developed, regional ground water flow systems could have emerged through waters in the deeply buried aquifers underlying this area currently are anoxic and enriched in dissolved ferrous iron and would have been more so before the regional ground water flow system. When that water emerged into the atmosphere the ferrous iron would and abiotic processes producing the ferric oxide ores at the spring orifices.

PAGE 3

169 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 1. Iron mine parcels of the SE Minnesota iron mining district superimposed on the bedrock geology. Modified from Figure VI-43 in Bleifuss (1972).

PAGE 4

170 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Description of the Ore Bodies Figure 3 is a part of Plate 5 from Bleifuss (1966) showing a plan view of a cluster of ore bodies in sinkholes on the Stewartville Formation. Figure 4 is Plate 6 from Bleifuss (1966) showing three cross sections through one of the ore bodies. The following descriptions of ore bodies are repeated here because the original exposures no longer exist. The ore bodies overlie either the Spillville or the Stewartville Formations, and range in thickness from 3 to 30 feet. An under clay which ranges in thickness from a few tenths of an inch to more than two feet is developed between the ore and the underlying carbonate rocks. The ore is locally overlain by decomposed Spillville Formation, residual clays, or sediments of the Ostrander Member of the Windrow Formation. Both the Spillville Formation and the Stewartville Formation beneath the ore generally are fresh, although they may have been changed to a sandy dolomite ranging in thickness from a fraction of an inch to several feet. Although the ore bodies developed on the Spillville and Stewartville Formations are chemically and physically similar, they differ in Formation generally have a greater areal extent, are more uniform in thickness, and have less relief than those on the Galena Formation. Deposits containing more than 50,000 tons of ore were common. In contrast, the ore bodies on the Stewartville Formation are isolated and generally contain much smaller tonnages. Generally, the upper surface of the ore is quite smooth, has a few closed depressions, and a relief rarely exceeding 10 feet. On a large scale, it is somewhat convex beneath the overlying unconsolidated materials. The relief on the carbonate bedrock surface beneath the ore on the Spillville Formation is small, In contrast, the relief beneath the ore on the Stewartville Formation is much greater, and most of the mines show prominent bedrock horses, some of which are more than 30 feet high. (Bleifuss, 1972, p. 501.) Figure 2. Composite stratigraphic section of the Windrow formation as exposed in Fillmore County from Andrews (1958). A = Loess, B = Glacial Drift, C = East Bluff Member (Ostrander Gravels), D = Iron Hill Member (15 massive, concretionary limonite, containing relics of weathered Cedar Valley Limestone (Spillville Formation), E = Cedar Valley Limestone (Spillville Formation) Badly weathered buff limestone, with slump structures produced by solution activity.

PAGE 5

171 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 3. Plate 5 (cropped) from Bleifuss (1966). Plan view of iron ore bodies on the Stewartville Formation. Figure 4. Plate 6 from Bleifuss (1966), cross sections of iron ore body shown in Figure 3.

PAGE 6

172 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Description of the Iron Ores The ore is composed predominantly of the mineral goethite and has minor amounts of Two types of ore are readily identifiable in the field hard ore and soft ore. The term hard ore is applied to that material in which the principal ore mineral is dense, hard, crystalline goethite. Its most striking physical characteristic in place is its coarse, broken rubbly appearance. In typical exposures, it is composed of a mass of broken, closely-packed, angular fragments, one half to two inches across, that are intermixed with nodular masses of goethite as much as 10 inches in maximum dimension. A distinct with individual beds being as much as six inches thick. (Bleifuss, 1972, p. 499-500). The soft ore, in contrast, appears rather massive and structureless in the field, and lacks the rubbly or nodular structure characteristic of the hard ore. In hand specimen, it has a soft punky texture and can be carved easily with a knife. The ore has a high porosity and a low bulk specific gravity. The principal ore mineral is goethite that shows a wide range of color from the bright yellow of ocherous goethite through shades of tan, brown and dark brown, to the brilliant crimson of ocherous hematite. the dark brown ore varieties have much more manganese (about 2.0 percent) than the yellow varieties (about 0.5 percent). (Bleifuss, 1972, p. 499-500). Figure 5 and Figure 6 are black and white images of samples of the hard iron ore. Figure 5 is from Stauffer and Theil (1944, Fig. 6) and Figure 6 is from Andrews (1958, Plate 1A). Both images show the layers of iron ore deposited concentrically around fragments of the limestone bedrock. Both samples are consistent with what would be expected when the iron oxides had been deposited from fluids, which flowed around and reacted with the limestone bedrock. Summary of Relevant Literature Observations 1. Early work on the Fillmore District iron ore deposits viewed the ores as straight forward weathering residues from the underlying country rocks. 2. Andrews (1958, p. 597) argues that the iron ores were deposited as a result of reaction of [acidic] iron-charged waters with carbonate bedrock but doesnt suggest a source of the acidic, iron rich waters. 3. Sloan (1964, p.18) considered the iron ores and associated Ostrander Gravels of Fillmore County to be Cretaceous in age. He observed that the Figure 5. Hard ore deposited around and reacting with carbonate inclusions. The sample is about 10 cm across. (from Stauffer and Theil, 1944, Figure 6). Figure 6. Limonite of the Iron Hill member concentrically surrounding fragments of highly altered Cedar Valley Limestone [Spillville Formation], Spring Valley mine of the Hanna Company, Fillmore County, Minnesota (from Andrews 1958, Plate 1A). The sample is about 18 cm across.

PAGE 7

173 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 To our knowledge none of the numerous springs issuing from the Spillville and Stewartville Formations in the iron ore district of western Fillmore County are currently depositing iron oxides. However, about 45 km west, in western Mower County near Austin, Minnesota, the Cedar River has eroded the thick glacial sediments of central Mower County. The first bedrock there is the Spillville Formation. There are springs in those areas which are currently depositing iron oxides (Green and others, 2002). The sandstone karst of north-central Minnesota (Shade, 2002, Shade and others, 2015) has many springs and seeps that are currently depositing significant amounts of iron oxyhydroxides. Figure 7 is a recent photograph of one such spring. This Sandstone. When sampled on June 14, 2001, (Shade, 2002) the water was a low TDS, Ca (16.5 ppm), Fe (11.6 ppm), Mg (6.5 ppm), Na (2.3 ppm)/bicarbonate (alkalinity = 74 as ppm CaCO 3 ) water. The SO 4 (0.52 iron ores typically occur on a karst topography, sinkholes or caves. 4. Bleifuss (1972, p. 498) argues that the ores were developed from the oxidation of a primary marine siderite faces of the Cedar Valley [Spillville] Formation but doesnt explain the textural evidence that the deposition involved flowing water. 5. The ore deposits are developed only on the Stewartville and Spillville Formations and not on the Maquoketa and Dubuque Formations. 6. The iron ores are in and associated with karst sinkholes and solutionally enlarged fractures and caves. 7. Mystery Cave, the largest cave in Minnesota, is developed in the Stewartville and Dubuque Formations, contains evidence of hypogenic speleogenesis (Klimchouk, 2007) and is overlain by one of the iron ore mines. A Hypogenic Source of the Iron Ores In other papers at this conference and in this paper, we are proposing that hypogenic regional groundwater flow systems have operated, and continue to operate, in southeastern Minnesotas bedrock aquifer systems. The current surface and groundwater flow systems drain to the Mississippi River and its tributaries. Older regional groundwater and surface water drainage patterns, before the current Mississippi River drainage developed, were from east to west and potentially may have been much longer. Deep wells in southeastern Minnesota often produce waters that are very anoxic, enriched in dissolved ferrous iron, with near neutral pHs. Some of the deep wells produce brackish to saline waters which are anoxic and iron rich. The Stewartville and Spillville Formations in Minnesota have high secondary porosity and permeability and are regional aquifer systems. The Decorah Shale aquitard constrains the bottom of the aquifers. The Pinicon Ridge Formation aquitard constrains the top. The Maquoketa and Dubuque Formations act as aquitards to separate the two regional aquifer systems. The Stewartville and Spillville Formations are the natural discharge points, where they reach the surface, for regional groundwater flow systems. Figure 7. Gushing Orange Spring, (MN58:A00002), south of Sandstone, Pine County, Minnesota.

PAGE 8

174 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE about 3 metric tons of iron per year. The entire Fillmore County iron ore district could easily have been produced in the available time by similar springs. Acknowledgments We gratefully acknowledge the positive and constructive reviews that have significantly improved this paper. The and its predecessor, the Legislative Commission on Minnesota Resources, has supported research on southeast Minnesota karst hydrogeology since the 1980s. John Barrys GIS skills produced Figure 1. References Andrews GW. 1958. Windrow Formation of upper Mississippi Valley region a sedimentary and stratigraphic study. The Journal of Geology 66 (6): 597-624. Bleifuss RL. 1966. The Origin of the Iron Ores of Southeastern Minnesota [PhD thesis]. Minneapolis (MN): University of Minnesota. 125 p. Bleifuss RL. 1972. The iron ores of southeastern Minnesota. In: Sims PK, Morey GB editors. Geology of Minnesota: a centennial volume. St. Paul (MN): Minnesota Geological Survey. p. 498505. Green JA, Alexander EC Jr, Markin WJ, Alexander SC. 2002. Plate 10Karst Hydrogeomorphic Units. Geologic Atlas of Mower County, Minnesota, County Atlas Series C-11. St. Paul (MN): Minnesota Department of Natural Resources. Klimchouk AB. 2007. Hypogene speleogenesis: hydrogeological and morphogenetic perspective. Special Paper No. 1 National Cave & Karst Research Institute, Carlsbad (NM). 106 p. Krumbein WC, Garrels RM. 1952, Origin and classification of chemical sediments in terms of pH and oxidation-reduction potentials. The Journal of Geology 60 (1): 1-33. MDM. 1941-1970. Mining directory of Minnesota series. Minneapolis (MN): University of Minnesota Mines Experiment Station. 30 volumes. Morey GB. 1998. Mineral development in Minnesota: past history, present trends, and future possibilities. Report of Investigations 52. St. Paul (MN): Minnesota Geological Survey. 21 p. Morey GB, Balaban N, Swanson L, editors. 1981. Bibliography of Minnesota geology: 1951-1980. Bulletin 46. St. Paul (MN): Minnesota Geological Survey. 143 p. ppm) and Cl (0.43 ppm) were very low. The pH of the water was 6.3. The leaf-covered mound in front of the lady is brown iron oxide that is several feet thick. This area was glacially scoured at the end of the Wisconsinan. The entire accumulation of material therefore must be less than about 10,000 years old. When the anoxic, ferrous iron-enriched groundwaters discharge to the surface, both abiotic and biological insoluble ferric iron and precipitate iron oxyhydroxides. The oxidation of ferrous to ferric iron releases hydrogen ions which rapidly lower the pH and acidify the waters near the surface and at the surface. The acidic waters aggressively react with and dissolve the carbonate bedrock at and near the surface. This enlarges the near surface fractures and creates the karst depressions that fill with the iron ore bodies. The karst depressions are enlarging as they fill with iron ore so the ores collapse on a local scale and produce the ore breccias seen in the mines. If and when a particular hypogenic flow path becomes clogged with iron oxides, the flow will find other nearby paths to the surface and create new iron ore accumulations. Depending on the local geometry, these accumulations of iron oxides can build mounds and/or coalesce to form larger structures. The insoluble iron oxides will then tend to armor the carbonates they cover topography reversals can occur. The age of the Fillmore County iron deposits is very poorly constrained. The ores are pre-Pleistocene and are underlain by Devonian and Ordovician carbonates. Sloan (1964) concluded that the ores were Cretaceous. aged. In either case, there are 10s of millions, if not a 100 million years available for their formation. Nor is there any necessity that the iron ores all formed at the same. Iron depositing springs would likely migrate across the landscape, as surface erosion and regional groundwater flow systems evolve. Estimating the flow of the Gushing Orange Spring shown in Figure 9 at 500 liters/minute and using Shades (2002) chemistry, we calculate that this spring discharges

PAGE 9

175 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Mossler JH. 1995. Plate 2Bedrock Geology. Geologic Atlas of Fillmore County, Minnesota, County Atlas Series C-8, Part A. St. Paul (MN): Minnesota Geologic Survey. Mossler JH. 1998. Plate 2Bedrock Geology. Geologic Atlas of Mower County, Minnesota, County Atlas Series C-11, Part A. St. Paul (MN): Minnesota Geologic Survey. Olsen B. 1988. Plate 2-Bedrock Geology. Geologic Atlas of Olmsted County, Minnesota, County Atlas Series C-3. St. Paul (MN): Minnesota Geologic Survey. Shade BL. 2002. The Genesis and Hydrogeology of a Sandstone Karst in Pine County, Minnesota [masters thesis]. Minneapolis (MN): University of Minnesota, 171 p. Shade BL, Alexander EC Jr., Alexander SC. 2015. The Sandstone Karst of Pine County, Minnesota. These Proceedings. 10 p. Sloan RE. 1964. The Cretaceous System in Minnesota. Report of Investigations 5. Minneapolis (MN): Minnesota Geologic Survey, University of Minnesota. 64 p. Stauffer CR, Thiel GA. 1944. The iron ores of southeastern Minnesota. Economic Geology 39 (5): 327-339. Winchell NH, Upham W. 1884. Geological and natural history survey final report 1872-1882. Volume 1. Minneapolis (MN): Minnesota Geological and Natural History Survey. 697 p.

PAGE 10

176 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

255 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 A SEMI-AUTOMATED TOOL FOR REDUCING THE CREATION OF FALSE CLOSED DEPRESSIONS FROM A FILLED LIDAR-DERIVED DIGITAL ELEVATION MODEL Abstract where surface streams cross under bridges and through culverts. Removal of these false depressions from an el transportation routes (e.g., roads) and streams, and then lowers the elevation surface across the roads to stream surface is corrected to match the approximate location of the National Hydrologic Dataset stream lines, the proce lation thresholds in order to generate stream lines with less contributing area within the watershed. Through liminary results reveal that this new technique provides dams may persist. Problematic areas include extensive road ditches, particularly along divided highways, and Limitations do exist and the results partially depend on the quality of data being input. Of 166 manually identi this tool. After three iterations, 1,735 culverts were elevation dataset, which retains the karst topography for further analysis, and a culvert catalog. Introduction tomated creation of closed depression catalogs are hampered in urban locations. Automated depression de lineation and cataloging methods have been discussed spill points) tends to be preferred due to the ease and speed by which a catalog can be generated. However, behind roads, railways, and other man-made surface features resulting in falsely detected closed depressions. These problems are especially acute when using highresolution topographic datasets such as Light Detection and Ranging (lidar), but are not unique to these DEMs. There are a number of methods to deal with these false depressions, but care has to be taken when working within karst terrain as these depressions are likely to be real and of interest to further research (Zandbergen, Current means to remove topographic barriers use either manual or digital methods. Manual methods require the verts) representing underpasses beneath roads, drive ways, railways, or other obstructions to actual stream expected, these methods are very time consuming. Fol lowing the digital creation of a culvert inventory, this dataset is then used as input into a variety of techniques to cut or burn the culverts into the elevation data. This John Wall Department of Marine, Earth & Atmospheric Sciences, 2800 Faucette Drive, Room 1125 Jordan Hall Campus Box 8208, NC State University, Raleigh, NC, 27695-8208, jwall@ncsu.edu Daniel H. Doctor U. S. Geological Survey, 12201 Sunrise Valley Drive, Reston, VA, 20192, dhdoctor@usgs.gov Silvia Terziotti U. S. Geological Survey, 3916 Sunset Ridge Road, Raleigh, NC, 27607, seterzio@usgs.gov

PAGE 2

256 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE burning method normally uses the entire stream net work to hydrologically condition the surface (e.g., Mi or can create steep canyons in the hydrologically cor rected elevation data potentially causing errors in fur ther analysis, particularly when streams represented as vector lines do not accurately match the elevation sur face. Current methods are too aggressive for karst ter rain resulting in true depressions being scrubbed from true depressions. Additionally, other tools such as those developed by Poppenga et al. (2010) require thresholds (i.e. area and depth) be set which could result in smaller, The goal of this study was to develop an ArcGIS toolbox to identify potential culvert locations which could then be enforced into the elevation data, thereby removing false closed depressions within karst terrain. This is done stream lines which are then buffered and used to lower the elevation across the man-made topography. This der to retain the karst depressions. A semi-automated ap proach is taken with user-provided data and thresholds. Study area and previous work The study area is the Boyce 7.5-minute quadrangle pre dominantly covering Clarke County with smaller por tions of Warren and Fredrick Counties in Virginia (Figure 1). The region spans roughly 150 km 2 Located within an extensive karst region of the Great Valley physiographic province of the Appalachian mountain range, the Boyce quadrangle covers part of the Shenandoah River drain age basin. Details on the geology of the quadrangle can be found study area has resulted in a mature dissected karst sur face of moderate to low relief, with 90 m total elevation range and a mean elevation of 180 m above sea level. Sinkholes and other karstic depressions generally oc cur as a result of cover-collapse or suffosion processes Figure 1. (A) Digital Elevation Model of the Boyce 7.5-minute quadrangle with a draped hillshade. thresholds (400,000 m 2 and 10,000 m 2 ) used.

PAGE 3

257 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 meters. Doctor and Young (2013) presented an evaluation of an method. This method uses the difference between a compared to manually-delineated closed depressions. They concluded that the primary hindrance to a fullyautomated process for identifying closed depressions face where streams pass beneath transportation routes. For that study, stream underpasses, usually in the form work, and were manually added as an input layer used to through elevation obstructions such as those created by transportation routes. Although this approach was effec tive, it was tedious and was not successful in identify ing all possible culverts within the quadrangle thereby impacting the closed depression catalog count and mor phometrics. Thus, a new automated method to identify stream underpasses was deemed necessary. Methods The method presented here for reducing the creation of false closed depressions was developed using ArcGIS tools. For this work, ArcGIS 10.2.2 was used, but the models should work with any 10.X version of ArcGIS. Additionally, a license for the Spatial Analyst Tools is necessary. Three vector datasets were acquired covering streams, railways and road networks along with a high-resolution lidar raster data set. Stream data were acquired from the United States Geological Surveys (USGS) National Hydrologic Dataset (NHD). Railways and roads were acquired from the USGS National Map. All vector data coarser resolution than the lidar data. This resulted in the 2). The resulting corrected vector data was used as input into the tools. The lidar dataset was acquired between 1 March and 9 March 2011. Acquisition took place dur ing leaf-off conditions and after snow cover melted. The vertical RMSE was 9.0 cm while the point spacing was 1.0 m. The Hydrocutter toolbox contains two tools: Hydro and Cutter. Together, these tools can be run iteratively in a semi-automated way employing user thresholds. Hydro mined using the Fill, Flow Direction, and Flow Accumu lation tools provided by ESRI within the Spatial Analyst Hydrology toolbox. The result of the Hydro tool is a vector dataset which can be used as an input into Cutter. Cutter topographically enforces culverts where streams pass beneath topographic highs, generally along trans closed depressions. all set to the DEM provided by the user within the tool environments automatically so that they do not need to be set by the end-user. 3 and described in detail here. Transportation vector da tasets are intersected with an initial NHD stream vec tor dataset. This results in two point datasets (one for railways and one for roads, respectively) which are then merged into a single point dataset representing all inter sections with streams (i.e. culverts). These culverts are the widest railway or road. For this study, a 25 m buffer diameter around each point was used. The circular buf fer polygons are then used to clip the stream vector line to provide stream segments where the streams cross the transportation routes. A Zonal Fill tool is then used to ments (ESRI, 2012). This results in a raster layer which Figure 2. Data quality is important to consider when running Hydrocutter toolbox. Note the is sues with the railway presented here. The initial railway data (dashed red line) is offset from the edited railway (dark brown solid line).

PAGE 4

258 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE is used to enforce this minimum elevation value into the DEM thereby hydrologically conditioning it. This conditioned DEM is used as input into the Hydro tool. The only additional information that must be pro vided when running the Hydro tool is a threshold value accumulation value of 400,000 m 2 was found to reason ably approximate the lengths of the NHD stream vector lines and was therefore used for the initial iteration of this tool. The Hydrocutter toolbox was run on the lidar data cov ering the Boyce quadrangle. The initial iteration of the Cutter tool used the NHD streams and edited roads and railway data as input. The resulting DEM was then used threshold of 400,000 m 2 was used to approximate the A second hydrologic conditioning iteration was carried 2 streams and transportation routes, and then the Hydro 10,000 m 2 After the DEM had been reconditioned to account for intersections between the stream lines of 10,000 m 2 2 and validation. Given the different approaches, the closed depressions described here and those of Doctor and Young (2013) are compare these datasets. Results Using the Hydrocutter toolbox approximately 14 times as many culverts and stream intersections with transpor cident with culvert lines from Doctor and Young (2013) The 2,000 m 2 have any obvious geomorphic expression of surface run off, such as channels, swales, gullies, etc. Thus, stream m 2 a result, the 10,000 m 2 channels as well as creating an inventory of potential culverts (Figure 4). are within 25 m of the Hydrocutter culverts, 80 of these found 260 intersections between the NHD stream vector tions were indicative of possible stream culverts. These culverts were burned into the original DEM, and used to generate a new stream vector line dataset that was more representative of the actual lidar-derived elevation mod el. The next iteration used lidar generated stream lines 2 thus mimicking Figure 3. Hydrocutter toolbox.

PAGE 5

259 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 4. (A) Red points indicate the initial intersections found between NHD stream data with roads and railway data. These are compared to (B) green and yellow points which are the inter sections found between the 400,000 m 2 and 10,000 m 2 stream lines respectively and transporta extend further upstream and therefore intersect a greater number of roads. (C) A detail of these differences is illustrated between initial intersections and subsequent iterations of the Hydrocutter tool resulting in differences of stream dataset quality highlighted particularly in the left half of the panel.

PAGE 6

260 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Conclusions The Hydrocutter toolbox provides two products use that cause false closed depressions resulting from the impoundment of water behind man-made features such as roads. The second is an inventory of potential culverts which is not only useful to karst studies, but the broader geologic, hydrologic, and environmental management communities. Nevertheless, there are several known issues and pieces of cautionary advice that come with using this tool. First, needs to be large enough to allow impoundments to be breached. This can be problematic when road widths are variable and span a greater width than the buffer distance 2 was routes when compared to the topographic expression of value, meaning that a user should determine a reasonable by examining the correspondence between the elevation model and the stream lines generated. Third, errors due to poor vector data quality compared to the lidar data can be propagated through the analysis. Therefore, high accuracy of the vector data used as the initial inputs to the process is important. If good vector datasets are un available, manual editing to 1:24,000 scale vector data to match the lidar elevation model might be necessary, as was done here. Fourth, some re-routing of stream lines can occur between iterations of the Hydrocutter tools. of intersections (Figure 4). The process outlined here generally improves the overall lematic in areas where the initial road or NHD stream data is poor. Thus, as with many tools, the quality of the data being input into the tool is inherently representative of the data quality coming out of the tool. Although further quantitative comparison is necessary, the Hydrocutter toolbox is better at breaching man-made impoundments while preserving the natural closed de pression landscape within karst terrain (Figure 5). This is vital to creating closed depression catalogs generat ed from lidar datasets rather than statewide inventories which adequately represent the closed depression popu stiehl, 2014). improve the manner in which the location of a stream pathway is delineated across an impoundment. Using the current Cutter tool, the pathway follows the pre-existing stream line that crosses an obstruction. If the stream vector is not in the correct location (i.e. where a culvert or underpass is), then errors may result which could be propagated through the analysis. Clipped streams seg ments may not fully connect the actual stream channels in the lidar surface. Figure 6 illustrates a possible so lution which would employ a least-cost path approach likelihood of connecting the lowest point on either side of the obstruction (Poppenga et al., 2010). Future work will focus on implementing a hybrid meth od between the Cutter tool and least cost path technique. This will allow for more accurate connections between false closed depressions. Figure 5. sions. Note that some of the manually identi closed depression using the results of Hydro cutter suggesting a coalescence of closed manual interpretation.

PAGE 7

261 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Acknowledgments and Disclaimer We wish to thank John Young (USGS), Alan H. Rea (USGS), and Brian Bruckno (Virginia Dept. of Trans portation) for helpful reviews. Although these data have been processed successfully on a computer system at the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the poses, nor shall the act of distribution constitute any such warranty. The USGS or the U.S. Government shall not be held liable for improper or incorrect use of the data described and/or contained herein. Any use of trade, and does not imply endorsement by the U.S. Govern ment. References Doctor DH, Young JA. 2013. An evaluation of auto mated GIS tools for delineating karst sinkholes and closed depressions from 1-meter lidar-derived digital elevation data. In: Land L, Doctor DH, Ste phenson JB, editors. Sinkholes and the Engineering and Environmental Impacts of Karst: Proceedings of the Thirteenth Multidisciplinary Conference, May 6-10, Carlsbad, New Mexico: NCKRI Sym posium 2. Carlsbad (NM). National Cave and Karst Research Institute, p. 449-458 Edmundson RS, Nunan WE. 1973. Geology of the Ber ryville, Stephenson and Boyce Quadrangles, Vir ginia. Virginia Division of Mineral Resources Re port of Investigations 34, Charlottesville, Virginia. 112 pp., 3 plates, 1:24,000 scale. ESRI. 2012. ArcGIS Help 10.1 Zonal Fill (Spatial Analyst). Available online at http://resources.ar cgis.com/en/help/main/10.1/index.html#/Zonal_ torial. Center for Research in Water Resources University of Texas at Austin. Available online at Removal.html (accessed March 16, 2015). Lindsay JB, Creed IF. 2006. Distinguishing actual and artefact depressions in digital elevation data. Com puters & Geosciences 32: 1192. Maidment DR. 2002. Arc Hydro: GIS for water resourc es. Esri Press. Mitasova H, Mitas L, Brown WM, & Johnston DM. 1999. Terrain modeling and soil erosion simula tions for Fort Hood and Fort Polk test areas. Geo graphic Modeling and Systems Laboratory, Uni versity of Illinois at Urbana-Champaign. Poppenga SK, Worstell BB, Stoker JM, Greenlee SK. 2010. Using selective drainage methods to extract Figure 6. Future development for this toolbox is illustrated by (A) The original elevation surface lated with the minimum elevation value within the buffer zone to create hydrologic connectivity. The red areas indicate false depressions that remain, (C) The least cost path through the buffer

PAGE 8

262 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE digital elevation data. U.S. Geological Survey Sci Tarboton DG. 2012. Terrain analysis using digital el evation models (TauDEM). Logan, UT. Available online at http://hydrology.uwrl.usu.edu/taudem/ taudem5.0/index.html (accessed March 16, 2015). Wall J, Bohnenstiehl DR. 2014. Power-law relationship of sinkholes and depressions within karst geology 63rd Annual Meeting Southeastern Section of the Geological Society of America, Blacksburg, Vir ginia, 10-11 April. Zandbergen PA. 2010. Accuracy considerations in the analysis of depressions in medium-resolution Li dar DEMs. GIScience & Remote Sensing 47 (2): 187.



PAGE 1

471 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 BUILDING CODES TO MINIMIZE COVER COLLAPSES IN SINKHOLE-PRONE AREAS Abstract Cover-collapse sinkholes are forming with increasing frequency under buildings. Analyses of sinkhole distri bution in Beacon Woods, Florida, preliminarily indicate their occurrence is an order of magnitude greater in ur ban versus undeveloped areas, suggesting the structures themselves are enhancing the collapse process. The most likely causes are induced recharge via at least one of two sources. First, runoff and drainage from roads, structures, and impoundments that is not adequately dispersed will promote sinkhole development. Second, leaking water, sewer, and septic systems beneath or process of cover-collapse from induced recharge is well understood. However, building codes generally do not require drainage and structural engineering practices that would reduce induced recharge and thus reduce the risk of collapse. This paper proposes engineer ing practices that measurably restrict the accidental discharge of municipal water through leaking subgrade drainage systems or the deliberate discharge of storm water runoff, induced shallow groundwater recharge irrigated land use. We recommend these practices be incorporated into building codes and ordinances to reduce induced sinkhole development in areas prone to cover-collapse. Introduction Tragedy struck. Mr. Jeff Bush was sleeping in his home the night of 27 February 2014 in Seffner, Florida, when a sinkhole opened under his bedroom and swallowed him. His body was never recovered. Sinkhole collapses are often viewed with fascination and a certain excitement about their potential danger. should not be underestimated. But what was the cause of that collapse? Media interviews mostly pointed to groundwater pumping, which could have lowered the water table, piping soil and sediment downward, resulting in the collapse. This is certainly plausible, as are natural processes unrelated to human activities. However, the opening of the sinkhole directly beneath the Bush home may indicate another origin. Cover collapse sinkholes from induced recharge are perhaps the most widespread type of collapse. Col lapse from groundwater withdrawal is more widely induced recharge collapses occur frequently in Florida and far beyond. The location of the Seffner collapse, under the Bush home, suggests a recharge-induced origin potentially from leaking water or sewer pipes, Unfortunately, evidence of the mode of this and other collapses is often lost in the collapse, non-forensic ex tion. Many books and papers describe geotechnical measures to remediate sinkholes and prevent them once subsid the 13 volumes of Sinkhole Conference proceedings, 1984-2013). This paper promotes a proactive approach that we believe will prevent some collapses from ever occurring. We begin by describing the general causes of recharge-induced collapses. Next we use a case study to demonstrate how and why they occur with greater frequency under and around buildings, making them potentially deadlier than other types of sinkholes. While the case study area is in Florida, this paper is not George Veni National Cave and Karst Research Institute, 400-1 Cascades Avenue, Carlsbad, New Mexico 88220-9215, USA, gveni@nckri.org Connie Campbell Brashear Bracken Engineering, Inc., 2701 West Busch Boulevard, Suite 200, Tampa, Florida 33618, USA, conniecb@brackenengineering.com Drew Glasbrenner Bracken Engineering, Inc., 2701 West Busch Boulevard, Suite 200, Tampa, Florida 33618, USA, dglasbrenner@brackenengineering.com

PAGE 2

472 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE focused on Florida so details of some geologic process es are kept general to be relevant to most regions where cover collapses occur. We end by proposing building Origin of Recharge-Induced Cover Collapses Cover collapse sinkholes form where thick mantles of soil, regolith, and/or sediment overlay cavernous bed rock. With time, these unconsolidated sediments move down into conduits in the bedrock. If the sediments move slowly and have little structural strength, the land surface above gradually subsides. Where the sediments are more structurally competent and may also move more rapidly, a cavity develops within until it becomes unstable and abruptly collapses. The sediment movement results from gravity and the that have been well described and understood for many if the water table of a karst aquifer extends into overly ing sediments, declines in groundwater levels below the bedrock will wash some sediment into bedrock conduits. Other sediment will slump and fall into the cavities due to its increased water-saturated weight. Re peated rises and falls in the water table will carry more away and through the aquifer. Eventually, subsidence or collapse may be seen at the surface. This paper focuses on the second mechanism of sedi ment movement: induced recharge. In this case, the karst water table generally remains within the bedrock, below the unconsolidated sediments. The sediments move into the bedrock conduits when water sinks consistently into the ground at a particular location, saturating the sediment and carrying it downward. As or intermittently, underlying sediment will be lost into the aquifer and a subsidence or collapse sinkhole will eventually form (Figure 1). Many induced sinkholes form along highways, where swales). Most sinkholes in non-karst areas are induced, often by leaking water and sewer pipes that can supply water to saturate sediments and carry sediment away to create cavities, the most famous and fatal of which were up to 30 m in diameter by 60 m deep, formed in Guatemala City in 2007 and 2010 and killed six people (Hermosilla, 2012). induced by homes and other buildings. Recharge from trates soil along the buildings foundation, putting the building at risk from subsidence or collapse. Unseen induced recharge also occurs beneath buildings from leaking water and sewer pipes. While leaking water pipes might be indicated by water usage on monthly are not detected without direct testing (Figure 1B). With the Seffner and other collapses occurring directly beneath buildings, the possibility that these buildings and their infrastructure caused some of the collapses must be considered, especially as global increases in population in sinkhole-prone areas may be putting more people at risk. Case Study: Beacon Woods, Bayonet Point, Pasco County, Florida To test the hypothesis that roads and other urban in frastructure may result in a greater frequency of cover Figure 1. A: Initial conditions susceptible to cover collapse by induced recharge. B: Ex amples of induced recharge sources creating a regolith cavity that may collapse.

PAGE 3

473 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 collapses than undeveloped land, we selected Beacon Woods as our study area. Located approximately 50 km northwest of Tampa, Florida, the district of Beacon Woods is at the north end of the community of Bayonet Point, about 6 km east of the Gulf of Mexico. Beacon Woods study area covers roughly 8 km 2 of residential and light commercial developments planned in the 1970s and mostly built out by 1980. With the exception of a golf course, the land use is relatively homogenous 19 to the west, Hudson Avenue to the north, Fivay Road and Little Road to the east, and State Road 52 to the south. This study area was selected because of the age of the residential developments, the detailed prop erty/permit records readily available online, the karst topography, and the mapped cave and sinkhole system running its length. The blue lines overlain on Figures 2 sections of this cave system. Historical aerial photography of the vicinity was down loaded and reviewed, the oldest dating from 1941 when very few buildings had been constructed in the area (Figure 2) (Aerial Photography: Florida Collection [In Photo Look-Up System [Internet]). According to the pre-development topography, the ground surface ranged in elevation from about 6 m (20 feet) along Fivay Road (all elevations relative to the National Geodetic Vertical Datum). In the western half of the study area, the shallow geo logic unit is the Suwannee Limestone (Arthur, 1993). While a few limestone outcrops occur in this part of the study area, most observed limestone is in the form of saprolitic boulders, many excavated and placed as decorations during development. The cover soils are in occasional isolated pockets. Moving east, an over lay of sand dunes (Arthur, 1993) provides topographic relief, guiding surface drainage. The multitude of closed underlying limestone. ments in surveying and mapping have improved in accuracy, there is an undeniable gradual increase in the areal extent of surface waters evident across the three Survey since 1954 (Figure 3). Most of the wetlands Figure 2. system (blue) and study area (yellow) overlay on pre-development ground surface (United States Department of Agriculture Natural Resource Conservation Service., 2015. [Cave outline adapted with permission from map by Southeast Exploration Team, personal commu nication]). Figure 3 study area (yellow) overlay on pre-develop ment topography (U.S. Geological Survey, 1954).

PAGE 4

474 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE as drainage canals and stormwater retention ponds or greenspace in the most current topographic map (Figure 4). The large Beacon Woods Cave System is present beneath the center of the study area, trending, and its provide entrance into the system for both exploration and a considerable volume of surface water from Bear Creek. The mapped portions of the cave system have an average water depth of 46 m, although some sections are considerably shallower. The lower sections of the cave contain brackish water, with diffuse haloclines at A resurgence of the cave system is approximately 4.3 km north-northwest of Smokehouse Pond (Figure 3), at the former Hudson Springs, while dye tracing in the 1960s linked a further upstream yet impassable sinkhole with a spring in the Gulf of Mexico (Wetterhall, 1965). Compaction grouting of residences affected by collapse or subsidence in the vicinity of the underlying cave system has a high potential, by design, of intersecting lesser possibility of directly impacting a cave passage. While the impact to endangered cave biota cannot be understated, the more pressing economic concern may be the potential to trigger a collapse of a large section of cave and drastically reducing the capacity for drainage of the basin. To correlate the occurrence of cover-collapse sinkholes, the predominant type in the region, with urban infra structure, we compiled potential ground settlement investigation and ground settlement repair permits last 25 years in the Pasco County Florida Public Access to Permit Applications database (Pasco County Florida Public Access to Permit Applications [Internet]). Our analysis found about 750 investigations and 650 repairs documented with the Pasco County Building Depart permits recorded for this 25-year time period. The permits and their locations are plotted as red squares on the aerial photograph in Figure 5. Since 2008, the Pasco County Building Department After extensive data mining we determined that a stag gering $33,801,000 dollars (US) have been spent from March 2008 to March 2015 in compaction or slurry grouting, chemical grouting, and/or underpinning the structures within the 8 km 2 study area. Also, the subsid ence incident database compiled by the Florida Geo logical Society as of October 2014 was sorted for the study area and the limited descriptions were evaluated for applicability to our analysis (Subsidence incidents Figure 4. study area (yellow) overlay on post-develop ment topography (U.S. Geological Survey, 2012). Figure 5. study area (yellow) overlay on 2014 Google Earth aerial imagery; red squares and yellow circles identify areas of subsidence or col lapse.

PAGE 5

475 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 reported to the Florida Geological Survey [Internet]). Included in Figure 5, the yellow circles indicate the locations deemed valid karstic subsidences. In general, these reported incidents are for areas that are in public spaces and/or undeveloped wetlands. The aerial extent of developed land has been estimated at 7 km 2 and undeveloped land at 0.78 km 2 The number of sinkhole activity reports per unit area of developed land, 57/km 2 4/km 2 creases the occurrence of induced collapse and subsid ence, perhaps by more than an order of magnitude. Our especially in that sinkholes in undeveloped areas are less likely to be reported and recorded than sinkholes which impact human infrastructure. Proposed Building Code to Minimize Recharge-Induced Collapses Considering the above results, we believe that build ing codes for sinkhole-prone areas should address the likely dramatic increase of induced collapses around roads and urban infrastructure with measures that will reduce or prevent their occurrence. For the study area, we reviewed the historic aerial photography and found soils imported to raise grades below foundations and roadways in order to install a system of curb and drain age pipe to channel surface runoff to isolated retention ponds within nested residential streets or to long swales include: roof runoff street drainage from curb to culvert automatic lawn irrigation systems (operational without a rain gauge or if broken and leaking) leaking plumbing below or beside buildings obsolete or unrepaired shallow irrigation wells unlined stormwater ponds leaking swimming pools, and ated from a review of the current 2010 Florida Build and Foundations (International Code Council, Inc., it is written for general adaptation to any building or roadway construction code in any karst area prone to cover collapse sinkholes. Under .5, Investigated conditions, many deleterious soil and groundwater conditions (even seismic) are detailed or reserved, but cover-collapse isnt among them. There may be future avenues available with an addition of a hypothetical Section 1803.5.13 Cover-Collapse Risk Zones: Where historical photographic evidence of cov er-collapse sinkholes exist or where geophysi cal surveys and subsurface explorations at the a registered design professional to demonstrate that the intended construction will measurably restrict the accidental discharge of municipal water through leaking subgrade drainage sys tems, the deliberate discharge of stormwater runoff from buildings, roads, parking lots, or other constructed impervious cover, induced shallow groundwater recharge from retention land use such as plant nurseries, golf courses swimming or decorative pools. Some terms in the above proposed code would need statutes. The proposed code inherently encourages to develop stormwater runoff management options that areas. This could be accomplished with lined shallow life habitat between neighborhoods. Also, specifying by requiring municipal sewer lines and transfer sta induced recharge-related subsidence as well as improve water quality. For the homeowner currently living in catchment system at the discharge point of each gutter the risk of soil settlement or erosion on their property. Conclusions Sinkhole collapse and subsidence are often treated as sensational fascinating events of mysterious origin by the media. That perception has often masked the fact that the sinkhole process is well understood and action close to $6 million/year have been spent in that 8 km 2 study area in sinkhole and subsidence remediation and associated repairs. The February 2014 collapse under

PAGE 6

476 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE York, New York: American Society of Civil Engineers. Subsidence incidents reported to the Florida Geological Survey [Internet]. Updated October 24, 2014. [Florida Geological Survey]: Florida Geological SIRs_database.htm United States Department of Agriculture Natural Resource Conservation Service. 2015. Custom dated September 23, 2014. U.S. Geological Survey. 1954. Port Richey Quadrangle, Florida-Pasco County. U.S. Geological Survey. 2012. Port Richey Quadrangle, Florida-Pasco County, 1954 photorevised 1988, 1998 and 2012. Wetterhall WS. 1965. Reconnaissance of springs and sinks in west-central Florida. Florida Geological Survey, Report of Investigations No. 39. the Bush home in Seffner, Florida, proved the cost can be much greater. Further, our research indicates a potentially greater than order of magnitude increased frequency of sinkhole and subsidence development in association with urban infrastructure over undeveloped land, strongly indicat ing a causal relationship likely due to induced recharge. While additional research will better quantify this rela tionship, we believe our results thus far should prompt government agencies charged with public protection through roadway and building codes to adopt codes that such as through the example code we have provided. References Aerial Photography: Florida Collection [Internet]. [University of Florida George A. Smathers Libraries / Florida Department of State, State Library and Archives of Florida] [cited March 8, map. Arthur J. 1993. Geologic map of Pasco County. Florida Geological Survey. Florida Department of Transportation Aerial Photo Look-Up System [Internet]. [Florida Surveying and Mapping] [cited March 8, 2015] AerialPhotoLookUpSystem/ Galloway D, Jones DR, Ingebritsen SE. 1999. Land subsidence in the United States. US Geological Survey Circular 1182. Hermosilla RG. 2012. The Guatemala City sinkhole collapses. Carbonates and Evaporites 27 (2): 103107. International Code Council, Inc. 2010. Florida Building ecodes_support/Free_Resources/2010Florida/2010 Florida_main.html Newton JG. 1987. Development of sinkholes resulting from mans activities in the eastern United States. US Geological Survey Circular 968. Pasco County Florida Public Access to Permit Applications [Internet]. [Pasco County Building Department]: [cited March 8, 2015]. Available Default.aspx Sitar N, editor. 1988. Karst terrains: exploration, foundation, design and performance, and remediation measures. New York, New York: American Society of Civil Engineers. Sowers GF. 1996. Building on sinkholes: design and construction of foundations in karst terrain. New



PAGE 1

477 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CARS AND KARST: INVESTIGATING THE NATIONAL CORVETTE MUSEUM SINKHOLE Abstract On February 12th, 2014, a sinkhole occurred at the Na tional Corvette Museum in Bowling Green, Kentucky. The collapse happened inside part of the building known as the Skydome and eight Corvettes on display were lost region of Kentucky, known as the Pennyroyal sinkhole plain, subsidence and cover collapse sinkholes are com monly found throughout the landscape. This iconic karst region in the United States is also home to Mammoth Cave, the longest cave in the world, and thousands of other caves and karst features. Investigation of the sink hole collapse began immediately while the Corvettes were extracted from the debris cone inside the void. drilling, downhole cameras and drone footage, a micro gravity surface survey, and mapping of the void and ac companying cave. After exploration of the sinkhole by karst researchers and compilation of the data, the cause of the sinkhole was determined to be a cave roof col lapse in a breakout dome. The cave underlying the col lapse is about 220 ft. (67 m) long and 39 ft. (12 m) wide on average with an average depth of 65-85 ft. (20-25 m). The structural integrity of the bedrock (thinly interbed ded limestone and chert located at a contact between two breakdown are abundant in the cave in which the sink hole formed. The progression of the roof failure likely occurred over a long span of time, eventually giving way due to a variety of conditions, including speleogenetic and climatic factors. Current remediation is underway under the building. Future changes to the structure will be monitored to detect any activity. Introduction On February 12th, 2014, a sinkhole occurred at the National Corvette Museum in Bowling Green, Warren County, Kentucky. The collapse happened inside part of the building known as the Skydome, which is a large circular structure connected to the main building where rare Corvettes are displayed (Figure 1). On the day of Jason S. Polk Western Kentucky University, 1906 College Heights Blvd., Bowling Green, KY, 42101, USA, jason.polk@wku.edu Leslie A. North Western Kentucky University, 1906 College Heights Blvd., Bowling Green, KY, 42101, USA, leslie.north@wku.edu Ric Federico EnSafe, Inc., 1148 College Street, Bowling Green, KY, 42101, USA rfederico@ensafe.com Brian Ham EnSafe, Inc., 220 Athens Way, Ste 410, Nashville, TN, 37228, USA, bham@ensafe.com Dan Nedvidek EnSafe, Inc., 1148 College Street, Bowling Green, KY, 42101, USA, dnedvidek@ensafe.com Kegan McClanahan Western Kentucky University, 1906 College Heights Blvd., Bowling Green, KY, 42101, USA, kegan.mcclanahan@wku.edu Pat Kambesis Western Kentucky University, 1906 College Heights Blvd., Bowling Green, KY, 42101, USA, pat.kambesis@wku.edu Michael J Marasa Hayward Baker, Inc., 53 Century Blvd., Suite 200, Nashville, TN, 37214, USA, MJMarasa@HaywardBaker.com

PAGE 2

478 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the collapse, eight Corvettes on display were lost into the In this region of Kentucky, known as the Pennyroyal sinkhole plain, subsidence and cover collapse sinkholes are commonly found throughout the landscape (Ford and Williams 2007, Palmer 2007, North et al. 2014). This is an iconic karst region in the United States and home to Mammoth Cave, the longest cave in the world, and hun dreds of other caves and karst features. In Bowling Green, sinkhole collapses are not uncom mon, with many having been mapped by the Kentucky Geological Survey and others that occur on a regular basis, though usually on a smaller scale and often unre ported (KGS 2015). There are also over 200 documented caves in Warren County, Kentucky (KSS 2014), lending merit to the potential for sinkholes and cave collapses this potential, sinkholes do not usually pose a high threat to the community. Sinkhole Types In many classic karst settings, most sinkhole collapses form from regolith arch failure, where cohesive soils, usually dense clays, create an arch that eventually gives way from below due to undermining caused by spall ing of materials into voids during water movement (i.e. types of sinkhole often form rapidly and can cause im mediate threat to life and property if they occur in de veloped areas. Other common sinkholes in south-central Kentucky occur as subsidence landforms, or closed de pressions, where surface soils slowly spall into cavities below over long periods of time. These features create a gentle depression in the landscape that sometimes has a throat in the center, but usually pose little risk to life or property. A third type of sinkhole, though less common, results from cave roof collapse. Though not as well documented, these types of sinkhole collapses are found throughout the study area, including past ex amples like the Dishman Lane sinkhole (Kambesis et al. 2003) and several collapsed cave entrances, such as Figure 1. Initial sinkhole collapse: The original collapse on the day it occurred at the National Corvette Museum in Bowling Green, Kentucky.

PAGE 3

479 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Lost River Caves main entrance and Crumps Cave in Smiths Grove, Kentucky (Crawford et al. et al. 2013). As structural integrity of the overlying bed rock weakens through geologic time, the cave roof forms a breakout dome, or cantilevered dome, which eventu ally fails in a similar manner as a cohesive soil arch, with a sudden, catastrophic collapse of the roof and overly ing materials into the cave below (Loucks 2007). These types of sinkholes are hard to detect and predict, but ex amples can be found throughout the world of both small and large feature of this type. Study Area The National Corvette Museum is located in the Pen nyroyal Plateau, which encompasses the Pennyroyal sinkhole plain of south-central Kentucky. The Sky dome structure sits atop the contact between the Ste. Genevieve and St. Louis limestones (Figure 2), both of which are Mississippian aged formations that comprise a including hosting many of Mammoth Caves passages (Palmer 1981, Palmer 2007). These formations are often thinly bedded at the contact and dip gently toward the north. The Corydon ball chert layer is present and lo cated at the contact between the Ste. Genevieve and St. Louis limestones. Sinkholes, both shallow and deep, are prominent features throughout the landscape and vary in depth based on the rock unit in which they form (Howard west are large subsidence sinkholes. The southern sink hole holds water throughout the year due to compacted clays or an impermeable chert layer at its bottom, or a combination of both. weathered limestone bedrock at the soil-rock interface, which provides a discontinuous boundary of pinnacleand cutter-type limestone features. Soil cover consists of thin, heterogeneously distributed Baxter and Crider soils. Baxter soils are often found on hillslopes and are elly-silt loam, sometimes overlying the Crider red silt loams and clay loams covering the region (Soil Survey Staff 2015). nally drained to the subsurface through various conduits until reaching the water table below. The regional aqui fer level is often located up to 160 to 200 ft. (48 to 62 m) below the surface. Precipitation averages 51 inches (1,300 mm) per year and the climate is humid-subtropi cal, though the average annual temperature is near 55F (13C). Investigation Methods The sinkhole investigation began immediately after the if there were additional voids or passages beneath the sinkhole, but the data were inconclusive. Micropile drill ing to support the Skydome structure and the spire in its center, which provides structural support commenced and provided an opportunity for additional investigation. Downhole cameras were used in the holes drilled for the micropiles in an attempt to identify any additional cavities or passages extending from the main sinkhole drill logs, which provided basic information on depth to competent bedrock and additional voids surrounding the main collapse. Subsequently, a microgravity survey on a 10-foot grid (3-m) was conducted in the area surround ing the sinkhole and a buffer area outside the Skydome. Compilation of these data indicated the need for further exploration due to the possibility of additional voids ex tending from the main sinkhole opening away from the debris cone. Prior to exploration and mapping, construction com pany Scott, Murphy & Daniel, LLC worked with karst scientists from Western Kentucky University (WKU) to develop a plan for removing the eight Corvettes and ened and fractured concrete slabs cantilevered beyond the edge of the hole and over loose sediments. Construc tion workers were trained on vertical caving techniques and the Corvettes were removed using a combination of techniques to strap them to 36-ton (36,000 kg) crane Figure 2. Local Geology at the Skydome: The Skydome is located near the contact be tween the Ste. Genevieve and St. Louis lime stones on the Pennyroyal sinkhole plain. Data from the KGS Map Information Service.

PAGE 4

480 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE booms and lift them from the sinkhole. All eight were recovered with varying amounts of damage, with some completely destroyed from the large breakdown that fell into the hole on top of them (Figure 3). hole, a team from WKU and EnSafe explored the sink hole and discovered additional cave passages (Figure 4). A grade 5 cave survey was undertaken to map the sink hole and cave passages. This was tied in to the surface engineering survey in order to better determine possible areas of concern and to develop a remediation plan for outside the building as needed. Discussion The oval-shaped opening to the sinkhole measured ap proximately 36 ft. (11 m) at its longest axis, which is investigations, combined with visual clues, revealed the possibility of additional cave passage extending from the sinkhole to the north and south. Exploration and survey of the sinkhole concluded that it was a collapsed por tion of a cave roof that failed and formed the sinkhole opening. The cave measures 220 ft. in length (67 m) and averages 40 ft. (12 m) in width. The deepest location is dome. The passages extend beyond the Skydome struc ture and trend toward existing sinkholes to the north and south, likely already collapsed portions of the same relic survey, which indicated the potential for voids in these areas. Inside the cave, the debris cone consists mainly of large breakdown and weathered limestone, with some overly ing soils. The limestone comprising the cave walls was thinly bedded. The sinkhole is located at the contact be tween the St. Louis and Ste. Genevieve, where the Co rydon ball chert layer (present in the ceiling, Figures 5 and 6) created conditions for instability over the span of the breakdown dome that failed. As the layers of rock collapsed over time, the breakdown dome continued to likely was already fairly thin prior to its construction, as this would have been a slow geologic process. Addi tional wedging from calcite mineral formation between Figure 4. Entrance to southern cave pas sage: Brian Ham of EnSafe, Inc. stands at the entrance of the southern cave passage. The cantilevered breakout dome can be seen as the sinkhole walls migrate upward. Figure 3. Remains of a Corvette: After extrac tion from the sinkhole, some Corvettes were repaired, while others were damaged beyond and private funds, and the others are on dis play at the Museum. Figure 5. Chert layer: Remnants of a layer of Corydon ball chert, which appears to be cleaved off from where the surrounding lime stone has broken away, revealing the impuri

PAGE 5

481 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 the limestone beds likely created further weakness in the structure of the rock. There are many stylolites found within the limestone beds as well, along the planes of which additional points of weakness exist. These lateral along their path, where delamination and cracking can occur (Heap et al. 2013), as was observed in the cave. Over time, a breakout dome formed and migrated up ward, thinning the bedrock support and creating a situa tion wherein structural failure and the sinkhole collapse was imminent. Hayward-Baker, Inc. (Nashville, Tennessee) was con sulted and a plan was formulated to repair the sinkhole and Skydome using micropiles to support a concrete slab and a one-foot (0.3 m) thick concrete slab was poured, then a double layer of metal sheet pilings were laid hori of the sheet pilings is to block the manufactured sand is to provide support to micropile drilling equipment and temporary support for structural concrete slab con struction. Holes were cut in the sheet pilings to drill the micropiles down to the bedrock. The micropiles are 7 inches (17 cm) in diameter (Figure 7) and will be placed at an average depth of 141 ft. (43 m), with a range of 130 provide support. The 46 micropiles are installed on a 20 are 23 existing micropiles in place under the footprint of the building and the spire in the center. Once the construction is complete, the Skydome struc micropiles and thus protected from any further collapse or subsidence should either occur in the future. A 4-foot toring and also to integrate into an exhibit. The Corvette Museum is working on an interactive exhibit to provide visitors an educational experience about the sinkholes development and repair, as well as information about karst landscapes. Conclusion The sinkhole at the National Corvette Museum on Feb ruary 12th, 2014 caused damage to the eight rare Cor vettes that fell in the hole. Fortunately, the cars were recovered and study of the sinkhole was possible during this process. Investigation of the sinkhole using multiple methods revealed it was part of a larger cave system and limited to the vicinity of the Skydome. Compilation of the various data, cave map, and observations provided the explanation of the sinkhole resulting from the failure of the caves roof, which was made up of thinly-bedded and impure limestone. While not as common, these types Figure 7. Micropile under the Skydome: A mi cropile breaching the northern cave passage that was installed to support the Skydome structure before the remediation of the sink hole was initiated. Figure 6. Thinly bedded limestone: The walls of the cave passage in the northern section of cave leading up to the sinkhole entrance. Thinly bedded St. Louis limestone layers are clearly visible.



PAGE 1

455 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CASE STUDIES OF ANIMAL FEEDLOTS ON KARST IN OLMSTED COUNTY, MINNESOTA Martin Larsen Olmsted County Soil and Water Conservation District, 1485 Industrial Drive NW Rm 102, Rochester, MN 55901, USA, Martin.Larsen@mn.nacdnet.net and mapped. A manure contaminated runoff storage area was constructed in the fall of 2014 by a livestock producer located at a headwater spring of Mill Creek. A filter strip and large manure contaminated runoff system is being designed for construction in 2015. Building great relationships with producers has been successful in Olmsted County. Livestock producers are making investments and taking action. Producers are an essential component of the mid-western economy and assistance with information, funding and resources will help protect the environment and keep farms profitable for future generations. Introduction A large array of sinkholes known as the Orion sinkhole plain is found south of Interstate 90 and north of the Abstract A unique area of Olmsted County is located a few miles southeast of Rochester by the small community of Predmore (Figure 1). Surface geology within the Orion Sinkhole Plain is dominated by a large array of sinkholes and limited soil cover over carbonate bedrock of the Ordovician Stewartville and Prosser Formations. Dye trace studies completed by Eagle and Alexander (2007) have demonstrated that a large portion of the plains groundwater discharges into springs that feed two local trout streams. Land-use in the area is mixed. For generations, local farmers have relied on livestock for stable income and profit. To put the 8,000 acre region into perspective, there are approximately 3,600 animal units located at 12 facilities which produce an estimated 74 million pounds of manure per year (United States Department of Agriculture / Natural Resources Conservation Service, 1995) and 10 million gallons of manure contaminated runoff. (Larsen et al., 2014) 349 known karst features exist of which 316 are sinkholes (Alexander et al., 1988). incident occurred where an area well was potentially impacted. Investigation revealed manure contaminated runoff was entering groundwater in a newly discovered groundwater quality. Developing relationships with landowners and livestock producers became necessary for protection of water resources and has facilitated research, education and action. A newly formed sinkhole that seasonally receives feedlot runoff was studied with ground penetrating radar for repair. Two producers in the region are implementing manure management techniques that are more stringent then regulation. The Wiskow dye trace was completed in spring of 2014. The study identified discharge springs that discharge into the Mill Creek trout stream from two vulnerable sinkholes (Johnson et al., 2014). Four springs and four previously unknown sinkholes were identified Figure 1. A state location map of the Orion Sinkhole Plain (Larsen, 2015).

PAGE 2

456 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Resources Conservation Service, 1995) and according to site evaluations, 39 million liters (10 million gallons) of manure contaminated runoff (Larsen, et al., 2014). 349 known karst features exist of which 316 are sinkholes (Alexander, et al., 1988). Although karst geology is widespread across Olmsted County, unique challenges exist for the protection of water resources by livestock producers in this area. There are engineering complications for the design of manure storage structures that may hold millions of gallons of manure and manure-laden runoff. Proper investigation and siting must take place so that the risk of soil collapse Operational challenges exist for livestock producers proximity to sinkholes. The entire area is underlain by a karst system with widespread dendritic, underground drainage that discharges to springs and streams. Proper manure management techniques must be developed and runoff entering the groundwater system. Kinney Creek and on the east by Mill Creek (Johnson et al., 2014). Surface geology of the Orion Sinkhole Plain is dominated by a large number of sinkholes and limited soil cover over carbonate bedrock of the Ordovician Stewartville and Prosser Formations (Figures 2 and 3). Dye trace studies completed by Eagle and Alexander (2007) and Green (2004) have demonstrated that a significant portion of the groundwater from the sinkhole plain discharges into springs that feed two local trout streams (Figure 5). Land-use in the area is mixed. For generations, local farmers have relied on livestock for stable income and profit. The highly variable soil types are not well suited for farmers to depend exclusively on a cash grain cropping system. Within the 3,200 hectare (8,000 acre) region, there are approximately 3,600 animal units. Twelve facilities produce an estimated 34 million kilograms (74 million pounds) of manure per year (United States Department of Agriculture / Natural Figure 2. Map of the Orion Sinkhole Plain with Case Studies and Feedlots Identified (Larsen, 2015).

PAGE 3

457 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 that water in the fall of 1978 (Landherr, 1979, written communication). More recently, an incident occurred on March 11, 2013, where a private well on the northern edge of the Orion sinkhole plain was potentially impacted with manurecontaminated runoff. A series of climatic events created a perfect storm for runoff. Rainfall occurred early in the winter of 2012 2013 when there was limited snow cover in the region. As the winter progressed, little additional snow fell and to illustrate beneficial relationships between livestock producers, engineers, scientists and regulators that have had positive results and are essential. There is not a complete understanding of how current rules and engineering practices are performing in such dynamic karst environments. In order to enhance these standards and practices, parties must interact in a collaborative manner so that information, knowledge and feedback is shared. Case Study 1 Concern for Groundwater and Dye Tracing to Determine Groundwater Flow Concern for groundwater contamination resulting from livestock production overlying karst systems trepidations commonly arise when new animal operation proposals are being considered. For example, in the late 1970s, two large swine facilities were proposed in the county. A contentious discussion between farmers, regulators, and other property owners focused on karst systems and groundwater quality. At one of the proposed sites, an Eyota area farmer constructed an un-permitted concrete block manure storage pit (which was excavated into the Galena Limestone). Following construction of the storage pit, a nearby private well in the Galena Limestone allegedly experienced dark-colored water Figure 3. Illustration of the stratigraphy in the Orion Sinkhole Plain (Alexander, et al., 1988). Figure 4. A pail of foaming water collected from the contaminated well. Foam can commonly occur in water contaminated with human or animal waste (Schmidt, 2013).

PAGE 4

458 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE temperatures dropped dramatically. The rain and snowmelt that occurred on March 8 10 was not able to infiltrate the ground that was sealed by ice and thick frost. Flowing water carried large amounts of manure contaminated runoff from farmland throughout southeast Minnesota. During the week prior to March 11, 2013, a farmer applied cattle bedding pack on a field southern Eyota Township in compliance with all applicable federal, state and local laws. The field does not contain any obvious sinkholes. For the most part, soil is thin, the field has less than 5 feet of soil cover overlying the Prosser Limestone that is visible in outcrops within the field. Soil texture is mixed with sands and sandy loam inclusions. According to a neighboring landowner, their well water became discolored and odorous following the manure applications and subsequent runoff event (Figure 4). The well, which is 43 meters deep, was constructed in the early 1950s has 6 meters of steel casing. It was drilled in the highly fractured and porous Ordovician Prosser Formation. It does not penetrate the lower Decorah Shale, which serves as a regional aquitard (Figure 3). County and Minnesota Pollution Control Agency response lasted many weeks. Clearly significant amounts of runoff left the application field and moved thousands of meters into drainage ways and neighboring properties. The movement was tracked, documented and mapped with GIS tools and software. Vulnerable wells in the surrounding area were monitored and tested. It was not until a later snowmelt event on April 3, 2013, when manure contaminated runoff was discovered entering groundwater (Larsen, 2013). Runoff entered the ground water through a subtle, newly discovered sinkhole. The swallow hole was only 5 amounts of runoff. The Wiskow dye traces were completed in spring of 2014 as a result of the manure contaminated runoff incidents of March 2013. Many participating agencies were involved. The traces were completed by Jeff Green and Scot Johnson of the Minnesota Department of Natural Resources (DNR) Ecological Water Resources Division, Martin Larsen of the Olmsted County Soil and Water Conservation District and E. Calvin Alexander, Jr. of the University of Minnesota Earth Sciences Department. Two sinkholes were identified in the Prosser formation to be used in the trace. One, the small newly identified located at 557550 E / 4865569 N 3.9 m. On May 6, called fluorescein) was introduced into the sinkhole. A skid loader excavated a basin in the area of the very small swallow hole. 1,100 liters (300 gallons) of fresh next (Figure 6) followed by 2,300 liters (700 gallons) of water. Initially, drainage was very slow and the water into the ground. (Johnson et al., 2014). Another sinkhole, approximately 2 kilometers south of the Wiskow property was located on the Applen property to introduce dye (MN55:D00282) found at 557516 E / 4865569 N 3.3 m. The sinkhole is approximately 21 meters in diameter, 9 meters deep with a 1 meter (200 gallons) of water followed by 435 grams 17.7 wt additional water (Johnson, et al., 2014). Twelve springs resurging from the Prosser and Cummingsville Formations in the area were identified Figure 5. A map of the Orion Sinkhole Plain area with the locations of previous dye traces and the dye vectors that were assigned during the 2014 Wiskow Dye Traces (Larsen, 2014).

PAGE 5

459 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 for sampling. Two vulnerable domestic wells were monitored. Rhodamine WT was traced from the Applen sinkhole (MN55:D00282) to the newly mapped Gasner Farmyard Spring (MN55:A00566). The dye took less than one day to reach the spring at a minimum velocity of 922 meters (3,025 feet) per day. Dye was detected in the spring for the remainder of the sampling period. Uranine C was traced from the Wiskow sinkhole (MN55:D00983) to the newly mapped Pagel Spring (MN55:00567). The dye took two days to reach the spring at a minimum approximate velocity of 278 meters (913 feet) per day. Rhodamine WT or Uranine C was not detected at any other sample location. Surface water entering the karst system through the sinkholes studied in the Wiskow Dye trace does not appear to leave the Prosser Formations. Dye was not detected in the lower Cummingsville Formation springs. Case Study 2 Studying a Sinkhole for Pollution Abatement In the spring of 2013, a new sinkhole formed in proximity to an existing feedlot in the Orion Sinkhole Plain. Initially in the spring of 2014 the sinkhole was reactivated at the surface. The farmer then contacted the Olmsted County Soil and Water Conservation District for assistance in (NRCS) personnel and an Olmsted County Soil and Water Conservation District employee met on site to perform a detailed survey of the sinkhole. Soil borings were taken to determine the depth to bedrock around the perimeter of the sinkhole. Multiple cross sections were taken with ground penetrating radar (Figures 7 and 8). NRCS recommended the sinkhole be fully excavated to bedrock and locate the conduit system. The conduits would be covered with select stone drain material followed by a non-woven geotextile liner. The largest portion of the sinkhole filled with compacted engineered soils. Case Study 3 Blue Ridge Confined Animal Feeding Operation Schoenfelder Farms owns a large confined animal feeding operation (CAFO) named Blue Ridge. Blue Ridge is centrally located in the sinkhole plain. The feedlot was initially permitted by the Minnesota Pollution Control Agency (MPCA) in 1991. The Schoenfelders have been researching solutions to modify the site so that it meets current pollution control laws. According to federal Figure 6. Technician introducing Uranine C into the shallow excavated Wiskow sinkhole (Larsen, 2014). Figure 7. Technicians surveying the sinkhole with a ground penetrating radar (Larsen, 2014).

PAGE 6

460 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE status and, therefore must eliminate or collect all water leaving the feedlot. Many proposals have been presented by Schoenfelder Farms and their private engineers to the MPCA, but location challenges have delayed completion of environmental corrective action for three years. In attempts to find a suitable site for a large manure storage structure, many locations have been evaluated at Blue Ridge. Three primary challenges exist for locating the proposed storage area. Minnesota feedlot rules require the following: Due to the proximity of sinkholes within the plain, a liquid manure storage area (LMSA) cannot be constructed on most of the Schoenfelder property. According to Minnesota feedlot rules LMSAs must not be located within 91 meters (300 feet) of any sinkhole (Minn R. 7020.2005 Subp. 1.). than 950,000 liters (250,000 gallons) when four or more sinkholes exist within 305 meters (1,000 feet). The required volume for the facility is significantly more than 950,000 liters, and many sites would not allow for the 22 million liter (6 million gallon) structure that is needed at the Blue Ridge Site (Minn R. 7020.2100 Subp. 2. A). Minnesota feedlot rules require a minimum vertical separation to bedrock of 3 meters (10 feet) when the site has a capacity of 1,000 or more animal units (1 slaughter steer is equal to 1 animal unit). Most of the sinkhole plain has soil cover of less than 1.5 meters (Minn R. 7020.2100 Subp. 2. B. (3)). Since the feedlot is located at the northern edge of the plain, the only potentially suitable location is 1,000 feet north in an open field. One small corner of the Schoenfelder property appears to allow for separation to sinkholes and vertical separation to bedrock. In the fall of 2014, test pits were dug with an excavator to determine separation to bedrock (Figure 9). As of March 2015 research is continuing to find a solution that meets Minnesota Feedlot Rules and is a viable option for the Schoenfelders. Case Study 4 A Constructed LMSA to Protect Surface and Groundwater A manure contaminated runoff storage area was constructed in the fall of 2014 by a livestock producer with 275 animal units located at the headwater spring of Mill Creek (Figure 10). The structure was built to protect the stream from manure contaminated runoff. The LMSA also stores manure so that the producer can application considerations are described further in the next section.) Prior to construction manure contaminated runoff discharged directly into the creek. Figure 8. GPR Image of the sinkhole. The lighter areas are dense material in the bottom of the sinkhole, possibly indicating is has been filled before with debris (England, 2014). Figure 9. A 3 meter (10 foot) test pit dug at Blue Ridge to determine the depth to bedrock. Shortly after excavation that terminated at the bedrock surface, the hole filled with approximately 2 meters of water (Larsen, 2014).

PAGE 7

461 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 a problem arise with the concrete liner. Perimeter drain tile is also installed around the footing of the LMSA. The purpose of the drain tile is two-fold. It controls groundwater around the LMSA and collects any potential leachate. The drain tile water is also inspected at the same location as the interior form-adrain (Fryer, 2015, written communication). Considerations of Manure Application on Karst Manure application on the land surface is not necessarily a bad thing if it is done carefully and all laws, regulations and best management practices are followed. Crop nutrient prices have escalated exponentially in the past five years and manure contains costly nutrients that crops require. Detailed nutrient management plans assist producers with application method, timing and rates applied. Following the severe runoff incidents in March application standards that are being presented to livestock producers in Olmsted County. In most cases the standards exceed Minnesota feedlot rule requirements. The standards are intended to limit the risk of manure contaminated runoff reaching ground water in karst areas. In regards to manure applications on snow covered or to be most vulnerable (Figure 11). Collaborating to identify and communicate these risks has been productive. In general the consensus with producers is that 91 meter (300 feet) may not adequate for protection, and that the entire watersheds of sinkholes should not receive manure applications. The LMSA was re-designed when it was discovered that soil was very thin. The redesign significantly increased the cost of the structure for the producer. Permitting and contractor services also took longer than expected. The LMSA was constructed so that contaminated water from the existing feedlot flows directly into the structure, and diversions on each end keep contaminated water from bypassing the system. The producer can also push manure into the LMSA from the existing feedlot. A concrete ramp is used so the landowner can thoroughly clean out any solids. Where the capacity of the feedlot is less than 300 animal units, Minnesota feedlot rules require LMSAs to be constructed with a minimum of 1.5 meters (five feet) vertical separation to bedrock. Because there 7020 feedlot rules allowed construction. In order to alleviate environmental concerns associated with storing manure within 18 centimeters to bedrock, the LMSA was constructed with two liners. The first is 15 centimeters of reinforced concrete. The concrete was designed to withstand pressures from larger cleaning equipment. Waterthem water tight. The concrete is underlain by a 60 mil LPDE liner that is attached 15 centimeters below the top of the 1.8 meter sidewalls of the structure. Between the liners, Form-a-drain tm allows gravity drainage of any potential leachate from the concrete to a surface tile outlet. The outlet serves as an inspection point that is regularly monitored should Figure 10. Completed LMSA that intercepts manure contaminated runoff leaving the feedlot edge before it reaches Mill Creek on the downstream (Larsen, 2014).

PAGE 8

462 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Do not winter apply manure when weather forecasts include rain and snow melt. Regarding manure applications, producers should use standards above 7020 rule requirements, including more restrictive setbacks during winter applications completed after December 31. Avoid application to aged snow pack that is crystalline in nature and contains high moisture content. Modify tillage practices to reduce the impact to exposed bedrock outcrops. Contour tillage should be used at all times. Regarding newly acquired application acres: properties before spreading manure at new sites, explain manure management practices and ask about well types and locations so that pro-active steps can be taken to reduce groundwater contamination risk. Keep manure application rates low at first when spreading at a location that is new to the operation or has had little or no recent manure history. When new land is acquired, talk to previous operators, if available, about abandoned wells, shallow bedrock, sinkholes or places where water disappears into the ground. Provide educational workshops to all livestock producers within active karst and close bedrock soils to communicate that vigilance and management is essential for protection of water resources. Continued Challenges The challenges continue for livestock producers protecting water resources in karst systems. Preventing all contamination of ground water is expensive and very difficult to achieve. However, it does not alleviate the responsibility that landowners have to greatly reduce the potential for contaminants entering aquifers. Acute, catastrophic events may occur when karst features are unknown, when they develop quickly, or when they are underestimated by engineers and regulators. Tools such as electrical resistivity imaging will continue to be more efficient in detecting hidden karst features. However, implementing the new geophysics tools may not be required for construction of a manure storage area. More recommendations were developed: In karst areas (where carbonate bedrock is present), producers should identify and understand: Areas where bedrock limits tillage or drain tile installation or drainage ditch depth. Presence of sinkholes, springs, closed depressions in or near fields that may fill and drain rapidly. Sinkhole application setbacks are required by feedlot rules (Minn. R. 7020.2225, Subp 6. A-C.) Areas where runoff water or tile discharge disappears into the ground. Known or potential locations of abandoned wells or cisterns. Feedlot rules contain well setbacks including wells that are abandoned. (Minn. R. 7020.2225, Subp 6. A-C.) Locations of wells used for any private, water supply wells Historical water quality problems in the area, such as shortand long-term water quality/quantity problems related to the ground -water system. Figure 11. Karst lands and agriculture fields that drain to springs from open sinkholes (Larsen, 2014).

PAGE 9

463 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 flow or enlarged solution. When engineered soils are placed on top of the bedrock, the material between the stored manure and bedrock is homogeneous with known compaction and density. There is, however a risk that the interaction between surface and groundwater flow is impacted, creating preferential conditions for sinkhole development. Conclusion The four case studies outlined provide excellent educational, informational and examples for future engineering and operational proposals. Case Study 1 is a detailed example how risk of water pollution exists, even when all applicable manure application guidelines are followed. The dye tracing study that followed the runoff incident indicates groundwater flow and resurgence. Case Study 2 shows that dynamic karst terrain is difficult to predict, and that karst features may appear quickly. Survey and studies are valuable for remediation if the new or existing karst features present an elevated risk of groundwater contamination. many bedrock formations are typically spaced 9 30 are usually sediment filled as seen in Figure 12. They may also be enlarged through solution and become an integrated conduit system capable of carrying large sinkhole formation. Proposed LMSAs may be 90 meters in dimension and Minnesota feedlot rules require: A minimum of two soil borings within the boundaries of the proposed manure storage area for the first one-half acre of surface area. A minimum of one additional location is required for each additional one acre of surface area for the manure storage area. (Minn. R. 7020.2100, Subp 4. A. (2)) The soil borings are only one small picture of the karst system below and rarely provide enough information to indicate the presence of karst features. Figure 13 provides an illustration of an LMSA constructed over carbonate bedrock to meet current Minnesota feedlot rules. Clearly there is a risk of missing potential dangers beneath soil covers greater than 10 feet. Bedrock removal, in order to create required separation has occurred in many instances in Olmsted County (Figure 14). The outcomes of removing bedrock in order to achieve separation are not clear. Potential advantages may be that a clean bedrock surface can be inspected for areas of preferential Figure 12. Fracture traces in the Ordovician Prosser Formation. Rochester Township, Olmsted County (Peter et al., 1976). Figure 13. Illustration of an LMSA constructed over carbonate bedrock with an undiscovered void (Larsen, 2015).

PAGE 10

464 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Larsen MR. 2014. Technicians surveying the sinkhole. Unpublished photograph. Larsen MR. 2014. Technician introducing Uranine C. Unpublished photograph. Larsen MR. 2015. Illustration of an LMSA constructed over carbonate bedrock. Unpublished illustration. Larsen MR. 2014. Limestone bedrock in process of removal. Unpublished photograph. Larsen MR. 2014. Completed LMSA. Unpublished photograph. Larsen MR. 2014. A 3 meter test pit. Unpublished photograph. Larsen MR. 2015. A state location map of the Orion Sinkhole Plain. Unpublished digital GIS map assembled by querying data from Minnesota Department of Transportation. Available from http://www.dot.state.mn.us/maps/gdma/gis-data. html Larsen MR. 2015. Map of the Orion Sinkhole Plain. Alexander EC Jr., Maki GL. 1988. Sinkholes and sinkhole probability. Plate 7 in Geologic Atlas of Olmsted County. Olsen BM. 1988. Bedrock Geology. Plate 2 in Geologic Atlas of Olmsted County. Larsen MR. 2015. Map of the Orion Sinkhole Plain with locations of previous dye traces. Unpublished Jr., Maki GL. 1988. Sinkholes and sinkhole probability. Plate 7 in Geologic Atlas of Olmsted County. Olsen BM.1988. Bedrock Geology. Plate 2 in Geologic Atlas of Olmsted County. Green JA. 2004. Burnap Farm Dye Trace. Eagle SD, Alexander EC Jr. 2007. Morehart Farm Dye Trace. 2 July Johnson SB, Green JA, Larsen MR, Kasahara BJ, Alexander, EC Jr. 2014. Wiskow Dye Traces Olmsted County, Minnesota. Larsen MR. 2015. Karst lands and agricultural fields. Alexander EC Jr., Maki GL. 1988. Sinkholes and sinkhole probability. Plate 7 in Geologic Atlas of Olmsted County. Green JA. 2004. Burnap Farm Dye Trace. Eagle SD, Alexander, EC Jr. 2007. Morehart Farm Dye Trace. 2 July. Larsen MR, Williams R. 2014. MinnFARM Models of feedlots in Orion Township. Unpublished. Peter W, Woxland W. 1976. Fracture traces. Unpublished photograph. United States Department of Agriculture / Natural Resources Conservation Service. 1995. RCA Issue Brief #7 [Internet]. Available from http:// www.nrcs.usda.gov/wps/portal/nrcs/detail//?cid= nrcs143_014211 Schmidt S. 2013. A pail of foaming water. Unpublished photograph. Case studies 3 and 4 are examples of real investment into environmental protection from multi-generation family farms. The setting and engineering of practices may be extensive, costly and take time. Ongoing manure application and engineering challenges need to be continually discussed and researched between livestock producers, engineers, scientists and regulators so that pollution reduction practices can be designed, implemented and funded. Building excellent relationships with producers has been successful in Olmsted County. Livestock producers are making investments and taking action. Producers are an essential component of the mid-western economy and assistance with information, funding and resources will help protect the environment and keep farms profitable for future generations. References Alexander EC Jr., Lively RS. 1995. Karst aquifers, caves, and sinkholes. Lively RS, Balaban NH, editors, Text supplement to the Geologic Atlas Fillmore County. County Atlas Series C-8, Part C. Minnesota Geological Survey. p. 10-33. Alexander EC Jr., Maki GL. 1988. Sinkholes and sinkhole probability. Plate 7 in Geologic Atlas of Olmsted County, Minnesota. County Atlas Series, Atlas C-3. Minnesota Department of Natural Resources, St. Paul. Eagle SD, Alexander EC, Jr. 2007. Morehart Farm Dye Trace. 2 July 2007. England M. 2014. GPR Image of the Sinkhole [Ground Penetrating Radar Derived Image]. Unpublished. Green JA. 2004. Burnap Farm Dye Trace. Unpublished. Johnson SB, Green JA, Larsen MR, Kasahara BJ, Alexander EC, Jr. 2014. Wiskow Dye Traces, Olmsted County, Minnesota. Larsen MR. 2013. Eyota Township manure runoff incident. Unpublished. Figure 14. Limestone bedrock in process of removal from the floor of an LMSA (Larsen, 2014).



PAGE 1

387 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CHARACTERIZATION OF KARST TERRAIN USING GEOPHYSICAL METHODS BASED ON SINKHOLE ANALYSIS: A CASE STUDY OF THE ANINA KARSTIC REGION (BANAT MOUNTAINS, ROMANIA) Laurentiu Artugyan Department of Geography, West University of Timisoara, Blvd. V. Parvan 4, Timisoara 300223, Timis, Romania, Adrian C. Ardelean Department of Geography, West University of Timisoara, Blvd. V. Parvan 4, Timisoara 300223, Timis, Romania, adrian.ardelean@e-uvt.ro Petru Urdea Department of Geography, West University of Timisoara, Blvd. V. Parvan 4, Timisoara 300223, Timis, Romania, petru.urdea@e-uvt.ro Our future work is intended to enrich our field data using SP and GPR methods, to compare with our first results. Also, we intend to integrate electrical resistivity tomography measurements in our analysis for better Introduction surface. The largest and most compact area of carbonate Synclinorium, situated in the SW of the country, in the unit called Banat Mountains. Karst terrain results from rock masses dissolution, having as a consequence an effective underground flow (Waltham et al., 2005). To understand karst topography, we must determine both the nature and the factors that are defining dissolution processes in karst soluble rocks as well as the drainage network resulted from these processes. (Ford, Williams, 2011). of dissolution that geological substrate has undergone locally (Shofner et al., 2001). The fractures and their orientation in a karstic area give important knowledge regarding the drainage network, due to the fact that the karst system depends highly upon them (Chalikakis et al., 2011). The study case of this paper is located in one of Banat Mountains subunits, Aninei Mountains. This approach is a comparative study using spontaneous potential (SP) and ground penetrating radar (GPR) as geophysical Abstract To understand karst topography, we must determine both the nature and the factors that are defining dissolution processes in soluble rocks, as well as the drainage network resulting from these processes. The goal of this paper is to understand the underground drainage direction configuration and, also, the factors that are involved in surface water drainage of the Anina karstic region. In this study we used two complementary geophysical methods, spontaneous potential (SP) and ground penetrating radar (GPR), applied in 5 sinkholes with a funnel shaped aspect. Four of these sinkholes are circular and one of them is elongated NW-SE direction. Three of the studied sinkholes are representing a chain of sinkholes orientated west-east. SP data describe the surface drainage, indicating drainage direction and/or moisture accumulation points. for the investigation of subsurface dielectric properties. GPR offers an image of the underground, showing possible bedding planes, in this case mostly along northsouth orientations. Besides, in two GPR profiles, we point were on SP grid the values indicate small values, pointing out a link between those two geophysical results. Using SP and GPR methods we were able to show that the bottoms of these depressions are retaining more humidity and soil. In addition, the GPR profiles between 20 and 40 meters, which need a more thorough analysis.

PAGE 2

388 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE (2012). Other goals for GPR applications in karstic Anchuela et al. (2010), for sinkholes detection near al. (2011) combined GPR with different techniques for using GPR and ERT. Al-fares et al. (2002) developed a study for a karst aquifer structure involving also GPR measurements. A study that involves both methods used in this paper was done by Carpenter et al. (2013) near Cancun, Mexico. In Romania, geophysical methods are not often used for karst investigations, even if there are many interesting karstic regions. There are two papers using resistivity methods, Mafteiu (1991) and Mitrofan et al. (2008), vertical electrical sounding (VES) using Schlumberger and pole-dipole arrays. plane of the fracturing effect that predetermines the Plateau he identified the border between fissured limestone and compact limestone. Mitrofan et al. (2008) managed to delineate with succes a concealed flow path in Hercules spring (Cerna Valley). Anina karstic region to understand the underground drainage direction and the factors that are involved in surface water drainage. This study is based on field data collection during five field campaigns, from May 2013 to November 2014. Study Area The Anina karstic region is situated in the South-West of Romania, within the Banat Mountains, as shown shaded in yellow in Figure 1. Geologically, the study area is located in the central part of most compact, homogeneous structure covered by carbonate a suspended karstic plateau without surface water drainage, located in the northern part of the Anina karstic region (Figure 2). methods. SP is a passive and an electrical geophysical method, which quantifies naturally occurring electrical fields at the Earths surface. The self-potential surveying is based upon measuring the spontaneous or natural potentials developed in the earth by electrochemical actions between minerals and subsurface fluids or by electrokinetic processes involving the flow of ionic fluids (Sharma, 2002). Also SP in the subsurface is caused by a number of processes that are not well understood at this time (Reynolds, 1997). In recent years the SP method has found increasing use in geothermal, environmental, and engineering applications to help locate and delineate sources associated with the movement of thermal fluids and groundwater. The spontaneous potential method has been used for GPR is a non-destructive geophysical tool that can produce a continuous profile in cross section or record features underground without drilling, boring, or digging. GPR profiles are normally used to assess the location and depth the continuity of the natural subsoil conditions (Apel and radargram) is very similar to a seismic reflection profile. Acquisition of data by means of GPR is based on the propagation, reflection and distribution of high-frequency to the underground. Using GPRs in karst areas partially covered with alluvial deposits is not very common, mainly due to alluvial deposits, clay content, which is hindering the penetration depth of GPR systems (Anchuela et al., 2008). Therefore, the results obtained will depend upon the type of soil and its degree of saturation, compaction, mineralogy, and also on the frequency of antennas used (Anchuela et al., 2009). If the study area contains clayey soils, it is recommended the GPR method should not be used in the sinkhole investigation (Zisman et al, 2013). Studies to detect cavities using the GPR method were done by Chamberlain et al. (2000), Kadioglu and Ulugergerli

PAGE 3

389 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 There are many surface karstic landforms that may be seen: sinkholes, sinkholes vallyes, blind valleys, dry valleys, karrens, and karren fields (Figure 3). Field Methods knowledge of both surface and those forms of underground features, and application of the geophysical methods are an option to study the subsurface in connection with the surface landforms. One of these methods, which is also used in the analysis of the groundwater, especially in karst areas, is spontaneous potential (SP). The second method that completes our geophysical approach is GPR. electrodes, a fixed electrode and a mobile one. The measurements were made with a digital multimeter, Voltcraft VC 850. We measured SP at 11 sites, repeating measurements 2 or 3 times, in different seasons and atmospheric conditions for comparison purposes. Our approaches for SP measurements are represented by profiles with N-S and E-W orientations and grids. Each Figure 1. Location of the study area in Romania and in the Banat Mountains. The karst area is shaded in yellow. Figure 2. and Anina karstic region within the structural Synclinorium. Figure 3. a. karren field; b. sinkhole valley; c. dry valley; d.,e. karrens.

PAGE 4

390 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE 2013 (Figure 5b), where we can notice that the middle of the sinkhole is accumulating humidity (red) and the drainage direction is mostly north-south, as we will obtain also in GPR measurements for bedding planes orientation. In May 2013 (Figure 5a) SP measurements indicate that on the boundaries of the sinkhole there is a direction of water infiltration, due to negative anomalies. At the bottom of the sinkhole, larger values suggest water retention, a function of the flat terrain and also based on the deep soil cover. In May 2013 the values where more different, alternating negative values (especially on the border of the sinkhole, where karrens are present) with positive values (in the middle of the sinkhole). The SP values obtained in November 2013 are more homogeneous, due to weather conditions. The campaign was done after many months of uniform precipitations. In Figure 5b is more obvious the bottom of the sinkhole, where are the largest values, indicating the stagnancy of water (Artugyan and Urdea, 2014). electrode was placed inside a hole, 10 cm deep in the soil and after 1 minute we noted the value indicated on the voltmeter (in mV) and then we moved the mobile electrode. The station spacing for the mobile electrodes was 5 m. For the GPR method, we used a MALA RAMAC Because we developed our study in a karstic area, our goal was to identify voids on the radargrams below the locations of the SP anomalies. Because in the study area other geophysical studies are missing, we chose as a first approach to use an antenna that give a deeper penetration in the subsoil, trying to have better image plateau. Based on previous study on limestone and due to the fact that the RTA antennas are of compact type, no Common Midpoint (CMP) or Wide Angle Reflection and Refraction (WARR), thus an overall wave velocity of 0.12 m/ns was used for the depth conversion of the penetration of 46 meters and 1 profile also with the 25 section of this paper, we present the results obtained. the most representative in these measurements. The measurements presented in this paper were made near Results Site 1 SP results for the sinkhole in Site 1 were obtained in May 2013 and November 2013. Measurements obtained after the first campaign express mostly negative values, with some positive values. In the autumn 2013 all SP values were positive, excepting one borehole where we obtained a negative value. The field measurements were interpolated based on the using ArcGIS 10 software developed by ESRI ( http:// www.esri.com/software/arcgis ), obtaining the raster presented in Figure 5. Water flow in this sinkhole has similar direction as the tectonic fault orientation, N-S or NW-SE. This observation is highlighted in the contoured images, but is clearer in the data obtained in November Figure 4. Location of GPR measurement sites.

PAGE 5

391 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Site 2 Site 2 features a large sinkhole, with the east-west diameter of 70 meters and the north-south diameter by 60 meters. The SP measurements were made in two campaigns, October 2013 and November 2014. In October 2013 the SP values present negative values, with some positive anomalies, showing that at that period, after several weeks without precipitation, the higher resolution, but in our investigation on these sites didnt help us too much, being more difficult to interpret the radargrams and losing in depth. After we antenna because we needed a deeper penetration trying to observe in the underground certain cavities, bedrock bedding planes, fractures or maybe the groundwater level. profile longer than sinkholes diameters, to observe the difference in the radar signals. The first radargram (Figure 7) shows very well the bottom of the sinkhole and also the slopes, and at the end of it, we notice the buried karrens or small voids. Besides, we notice a continuous signal that we consider as a bedding plane of the area. Figure 6. Site 1 and the location of GPR profiles. Figure 5a. Spontaneous Potential measurements in Site 1 in May 2013. Figure 5b. Spontaneous Potential measurements in Site in November 2013.

PAGE 6

392 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the highest values, indicating moisture accumulation and water retention due to thick soil. On N-S direction, we may observe a negative anomaly at 24 m from the start of the profile in both campaigns, which could indicate a void in the underground which quickly drains water from the surface. Even if on N-S direction the profiles are not very smooth, we can notice that except those negative anomalies at 24 m, the bottom of the sinkholes indicate the largest values, showing the tendency to retain moisture for a long time. The aspect of this sinkhole (Figure 10) gives as the interpretation of the GPR results: in the middle there is a large accumulation of materials (organic, soil) and on the slopes the karrens are also observed on the GPR profiles. surface is not well moistured, being favorable for rapid flow into the underground. In November 2014 the values are positive, indicating that the drainage is more stable, the soil is more saturate with water. The middle of the sinkhole presents the highest values on E-W direction. On the E-W orientation, in the middle of the sinkhole the SP presents values indicating accumulation, as we expect to obtain in the middle of these karstic depressions where the bottom is filled with thinner soil and the humidity presents higher values (Figure 8). The same situation is observed also in November 2014, but with higher precipitation in the area and the SP values are negative, indicating that the surface drainage is more unstable. Again SP values indicate that in the middle this geophysical method shows accumulating moisture. Also, we can notice that in both profiles there are two negative anomalies at 9 and 12 meters from the starting point of the profile and at 63 meters. These anomalies are showing that at that point the drainage is more rapid, being possibly linked to certain voids underground. On the north-south orientation (Figure 9) the profile is more sinuous, presenting many negative anomalies, but we notice that in both campaigns at the point located at 24 meters there is a negative anomaly, possibly indicating a void in the underground where water is more rapidly drained. The north-south profile is not very expressive for this sinkhole, the bottom of it being not very obvious as for the east-west profile. This fact could indicate that on the north-south orientation the fractures in the bedrock are more developed, determining a certain behavior in water drainage. If we take into account the fault main orientation in the area, NNW-SSE, maybe we could find an explanation for the aspect of those N-S profiles. We can observe that in both profiles in the middle of the sinkhole is well highlighted on E-W orientation, with Figure 7. GPR profile (25 MHz antenna) in Site 1 on east-west direction. Figure 8. Spontaneous Potential measurements at Site 2 (east-west orientation). Figure 9. Spontaneous Potential measurements at Site 2 (north-south orientation).

PAGE 7

393 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 the sinkhole, with highest radar signals in the middle, but the second GPR profile, for N-S orientation (Figure 12), has many anomalies between the surface and 25 meters depth. These anomalies are also observed on the SP profiles. Site 3 Site 3 is a chain of three sinkholes with west-east direction. Sinkhole 1 is the smallest one and less deep, the second one is the largest one, and Sinkhole 3 is the deepest one of this chain of sinkholes. Sinkhole 3 consists of a circular sinkhole, a funnel shaped one, with a swallow hole in the middle. SP values were obtained at the beginning of October 2014, after several days of precipitation. The results show that in the middle, in the smaller sinkhole the water is staying for a long time, indicating moisture accumulation (larger values of SP data). Also, near the edges of the sinkhole, where karrens are present, the drainage is more rapidly, due to these rocks and thin soil. SP values indicate that the surroundings the swallow hole inside the large one presents the tendency of rapid flow into the swallow hole direction (smaller values of SP measurements) (Figure 13). For GPR profiles, this site means a chain of three sinkholes (Figure 14) where we intend to observe on the radargrams the boundary of these karstic depressions Because is known that clay may perturb the GPR signals, we intend to employ in this site study electrical resistivity tomography, to describe accurately the underground and validate the GPR results. Also, we observe that as on the first reflection radargram we could identify the countinuos GPR signal, considering that it should be also bedding planes, this area being a GPR profiles for Site 2 are designed to better explain the SP results and to give an image of the underground of this sinkhole. We can notice that GPR profiles are similar to both SP profiles, meaning that the W-E profile, shown in Figure 11, describes a smooth hyperbola for Figure 10. GPR measurements at Site 2. Figure 11. GPR profile at Site 2 on west-east direction. Figure 12. GPR profile at Site 2 on north-south direction.

PAGE 8

394 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE radargrams present a signal between 30 and 40 meters depth for Sinkhole 3 (Figure 16), a deeper sinkhole, and between 20 and 30 meters for Sinkhole 2 (less deep than being present on both radargrams that are parallel to the terrain at a distance of several meters. The velocity was established based on previous study that applied GPR Combining the SP measurements, with the GPR results, we can point out that it could be certain void in the middle of this site. We observe that the signal present in the middle of the last sinkhole in Figure 14 was obtained in the N-S profile (Figure 17). For the second sinkhole we notice that there are signs that we include in the buried rocks, but we rise the question if is not also clay padding, due to the fact that under that 10 meters there is no GPR signal (Figure 18). Discussion The results of SP measurements indicate in most of the cases that there is a direction in the water circulation (based on the negative values of SP measurements), but we also obtained positive values during the dry season, most of them being measured during August and September, after large dry periods. We observe that on the karstic plateaus, starting from May the soil was very dry and hard, with very small absolute values of SP, but also with positive values in the middle of the dolines, suggesting moisture accumulation areas. GPR radargrams indicate bedding planes at depths between 20 and 30 meters, all these profiles being along north-south orientation. On one of the radargram we and if there could be certain cavities because the last 2 sinkholes (from West to East) present swallow holes. The first radargram of this site (Figure 15), comprising all the three sinkholes is indicating very well the boundary into the underground of those depressions, the first sinkhole and the second one being more closely, and the boundary between the second one and the third one, that are separated by a dirt road. Also, we can notice that karrens present mostly on slopes of the first and the third sinkhole are observed on the GPR signals as buried rocks and also the radar signals indicate that in the middle of the third sinkhole, the largest and the deepest of these three depressions, could be a cavity or a void, based on previous GPR results in karstic areas (for example 2012). The funnel aspect of this sinkhole is favourable for a vertical cavity development. Again, we observed that countinous GPR signal that could be considered as bedding planes, but for the radargram presented in Figure 15, these are smaller than and not as obvious as in previous sinkholes. We also obtained two radargrams on N-S orientation for sinkhole number 2 and 3 of this chain of sinkholes. Both Figure 13. SP and GPR measurements at Site 3. Figure 14. GPR measurements at Site 3 (Sinkhole 1 and Sinkhole 2 in this picture).

PAGE 9

395 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 In this study we used two complementary geophysical methods, spontaneous potential (SP) and ground penetrating radar (GPR), applied in 5 sinkholes with a funnel shaped aspect. Four of these sinkholes are circular and one of them is elongated NW-SE direction. Three of the studied sinkholes are representing a chain of sinkholes orientated west-east. SP describes the surface drainage water indicating the tendency in the drainage direction or accumulation points. On the other hand, GPR describes the subsurface using the response of the signal sent by the radar antenna. There are limitations in both methods, but they have been successfully applied in several sites for karst topography Conclusions and Future Work GPR offers an image of the underground, showing possible bedding planes, mostly along north-south orientation. The north-south direction of the identified bedding planes are according to the main faults orientation of the studied area, NNE-SSW. Due to this aspect, we consider that the bedding planes are mostly observed on north-south profiles. Besides, in two GPR cavity, below and anomaly on SP grid. observe a possible void or a cavity at 20 meters depth in the west to east profile. At the same depth we notice also on the north-south profile that the GPR signal point out an anomaly in the underground. There are two profiles that are pointing out some discontinuities and possible cavities. One of these profiles was measured over a chain of three sinkholes and this profile at between 6 meters and 25 meters depth, shows some anomalies that indicate differential radar signal The homogeneous aspect of radargrams indicates that Figure 15. GPR profile at Site 3 for a chain of 3 sinkholes, from west to east. Figure 16. GPR measurements at Sinkhole 3 of Site 3. Figure 17. Sinkhole 3 of Site 3: GPR north-south profile.

PAGE 10

396 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the perspective of the doline triangle using GPR Examples from Central Ebro Basin (Spain). Engineering Geology 108 (3-4): 225-236. http:// Juan A. 2010. A geophysical survey routine for the detection of doline areas in the surroundings 114 (3-4): 382-396. enggeo.2010.05.015 D, Gil-Garbi DH. 2013. Actual extension of sinkholes: Considerations about geophysical, geomorphological, and field inspection techniques (NE Spain). Geomorphology http://dx.doi. frequency ground penetrating radar (GPR) in mapping strata of dolomite and limestone rocks for ripping technique. International Journal of Surface Mining, Reclamation and Environment 19 (4): 260-275. http://dx.doi. org/10.1080/13895260500275418 Artugyan L, Urdea P. 2014. Groundwater drainage monitoring and karst terrain analysis using Spontaneous Potential (SP) in Anina Mining Area (Banat Mountains, Romania). Preliminary study. Proceedings of Karst without Boundaries Dubrovnik, p. 157-164. Bautista, Rosa M. 2013. Ground-Penetrating Radar, Resistivity and spontaneous potential investigations of a contaminated aquifer near Cancn, Mexico, Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, p. 231-237. Chalikakis K, Plagnes V, Guerin R, Valois R, Bosch FP. 2011. Contribution of geophysical methods to karst-system exploration: an overview. Hydrogeology Journal 19: 1169-1180. http:// dx.doi.org/10.1007/s10040-011-0746-x Using SP and GPR methods we were able suggest that the bottoms of these depressions are retaining more humidity and soil. In addition, the GPR profiles outlined 20 and 40 meters, which need a more thorough analysis. Our future work is intended to enrich our field data using SP and GPR methods, to compare with our first results. Also, we intend to integrate electrical resistivity tomography (ERT) measurements in our analysis for a The ERT measurements that we intend to use in the future should provide a complete image of the subsurface, and with interpreted air-filled voids on the radargrams Acknowledgments We would like to thank those students and friends who helped us in the data field acquisition campaigns, who have been a real support in obtaining these results. This work has been supported from the strategic grant Postdoctoral programs of excellence for highly qualified human resources training for research in the field of Life sciences, Environment and Earth Science cofinanced by the European Social Fund within the Sectorial Operational Program Human Resources Development 2007. References Analysis of the karst aquifer structure of the Lamalou area (Herault, France) with ground penetrating radar, Journal of Applied Geophysics 51: 97-106. Juan A. 2008. Mapping subsurface karst features with GPR: results and limitations. Environmental Geology 58 (2): 391-399. http://dx.doi. org/10.1007/s00254-008-1603-7 Figure 18. Sinkhole 2 of Site 3: GPR north-south profile.

PAGE 11

397 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Mafteiu M. 1991. Contributions to the investigation Romania) by means of resistivity measurements. Theoretical Applied Karstology 4: 65-76. investigations by means of resistivity methods in karst areas in Romania. Environmental Geology 55: 405-413. Nouioua I, Rouabhia AEK, Fehdi CH, Boukelloul ML, Gadri L, Chabou D, Mouici R. 2013. The application of GPR and electrical resistivity tomography as useful tools in detection of sinkholes in the Cheria Basin (northeast of Algeria). Environmental Earth Science 68: 1661-1672. http:// dx.doi.org/10.1007/s12665-012-1859-9 Oradea. 444 pp. Robert T, Dassargues A, Brouyre S, Kaufmann, O, Hallet V, Nguyen F. 2011. Assessing the contribution of electrical resistivity tomography (ERT) and self-potential (SP) methods for water well drilling program in fractured/karstified limestones. Journal of Applied Geophysics 75 (1): 42-53. Reynolds JM. 1997. An Introduction to Applied and Environmental Geophysics 1st edition. Wiley. in earth dams using the self-potential method. Engineering Geology 82: 145-153. Sharma PV. 2002. Environmental and engineering geophysics. Cambridge University Press Shofner GA, Mills HH, Duke JE. 2001. A simple map index of karstification and its relationship to sinkhole and cave distribution in Tennessee. Journal of Cave and Karst Studies 63 (2): 67-75. Stevanovic Z, Dragisic V. 1998. An example of identifying karst groundwater flow. Environmental Geology 35 (4): 241-244. Waltham T, Bell F, Culshaw M. 2005. Sinkholes and and construction. Springer, Berlin, 382 p. Zisman ED, PE, PG, M ASCE, Clarey D. 2013. If its weight of hammer conditions, it must be a sinkhole?, Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, p. 45-52. Chamberlain AT, Sellers W, Proctor C, Coard R. 2000. Cave detection in limestone using ground penetrating radar. Journal of Archaeological Science 27: 957-964. http://dx.doi.org/10.1006/ Imaging subsurface cavities using geoelectric tomography and ground-penetrating radar. Journal of Cave and Karst Studies 67 (3): 174-181. Ford D, Williams P. 2011. Geomorphology Underground: The Study of Karst and Karst Processes. In: Gregory KJ, Goudie AS, editors. The SAGE Handbook of Geomorphology, SAGE Publications Ltd., London, 469-486. risk of subsidence of a sinkhole collapse using ground penetrating radar and electrical resistivity tomography. Engineering Geology 149-150: 1-12. Guichet X, Jouniaux L, Catel N. 2006. Modification of streaming potential by precipitation of calcite in a sandwater system: laboratory measurements in the pH range from 4 to 12. Geophysical Journal International 166: 445-460. http://dx.doi. J, Guerrero J. 2011. Integrating geomorphological mapping, trenching, InSAR and GPR for the A review and application in the mantled evaporite karst of the Ebro Valley (NE Spain). Geomorphology 134: 144-156. Jardani A, Revil A, Santos F, Fauchard C, Dupont JP. 2007. Detection of preferential inflitration self-potential and EM-34 conductivity data. Geophysical Prospecting 55: 749-760. Straface S, Johnson T. 2009. Reconstruction of the water table from self-potential data: A Bayesian Approach. Ground Water 47 (2): 213-227. http:// Jouniaux L, Maineult A, Naudet V, Pessel M, Sailhac P. 2009. Review of self-potential methods in hydrogeophysics. Comptes Rendus Geo-science 341 (10-11): 928-936. crte. 2009.08.008 Kadioglu S, Ulugergerli EU. 2012. Imaging karstic cavities in transparent 3D volume of the GPR data set in Akkopru dam, Mugla, Turkey. Nondestructive Testing and Evaluation 27 (3): 263-271. Lange LA. 1999. Geophysical studies at Kartchner Karst Studies 61 (2): 68-72.

PAGE 12

398 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

549 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CONCEPTS FOR GEOTECHNICAL INVESTIGATION IN KARST Joseph A. Fischer Geoscience Services, 1741 Route 31, Clinton, NJ 08809 Joseph J. Fischer Geoscience Services, 1741 Route 31, Clinton, NJ 08809 Karst Variability Carbonate bedrock is found throughout the world. Spectacular examples of true, pinnacled karst are found in China and the namesake plateau in Slovenia and Italy. However, all carbonates are not the same, although this distinction is often overlooked by investigators. They range in strength and character from the soft offshore corals of the Caribbean Islands and Florida to marbles. More confusion is added by having the older, hard, but flat-lying carbonates of the central US, the hard, stressed, folded and fractured limestone, dolomites and marbles of the Appalachian Mountain chain and the soluble gypsum (evaporite) of the southwestern US (which can Florida). As can be deduced from the extensive portion of the US underlain by karst (Figure 1), development growth has likely forced administrators, politicians, the public, as of building atop karst. The result has sometimes been environmentally aware regulation and better technical understanding to address the problems posed by this variable and generally disguised environment. We can no longer conscientiously drill three or four test borings one boring per mile to address the engineering concerns along a roadway or transmission line in karst and assume that we have all the information necessary for evaluation and design of structures. Geotechnical analyses and recommendations are not the same for all conventional (non-karst) sites, but they are even more varied and complicated for karst sites. In A) The potential variations in physical properties across and below a site, B) the applicability and appropriateness of the available suite of site investigative tools, and C) the availability or lack of potential planning and/or engineering solutions to cover the uncertainties that will likely exist at the karst site in question. Abstract There seems to be a lack of recognition in the literature that addresses the variety of karst in the United States of America and some of its offshore territories. For example, there are the well-known solutioned carbonates of Florida and the Caribbean, but there are also the somewhat older, harder carbonates of St. Croix, U.S.V.I. Even Floridas recently deposited karst varies from region to region. There are also the ancient, flat-lying carbonates of the interior craton that often have semiwater levels affecting bedding and the contorted rocks of the Appalachians with its apparently chaotic variations in solutioning found across-strike and in relation to folds, faults, and fracturing. In addition, there are various salt and gypsum deposits in the south and southwest that pose their own problems to mans works. As the geology differs, so does, to some extent, the investigation requirements, investigation techniques and engineering solutions. There is no single set of investigative tools that fit all karst sites. Geophysical investigations are apparently far less suitable for the broken and twisted Appalachian karst than in the flatlying mid-continent carbonates or the less contorted karst of Florida. Specific procedures developed for geotechnical investigation in true karst have been documented for many years now. However, it appears that many practitioners are not aware of them or choose not to use them because of geotechnical understanding of the work of others in karstic areas outside their sphere of experience. This paper will attempt to provide a rational geotechnical approach to carbonate rock investigations in the United of the targets and the economics of pre-construction not fit all.

PAGE 2

550 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE karst or providing a means of development of the site conscientiously, not necessarily economically. Geology For simplicity, this paper will attempt to crudely divide this presentation into three groups of karst. Old, mid-continent, generally flat-lying carbonates, Old, folded and faulted Appalachian carbonates, and Recent, coralline limestone. For those interested in a more precise division, we suggest the United States Geological Survey (USGS) the Framework of the New National Karst Map (Weary et al., 2008). Generically, the differences are age and degree of tectonism. The USGS further differentiates US karst types by thickness of the overburden and precipitation. The overburden thickness of concern to the geotechnical community are generally less than The intention of this paper is to point out the difficulties of performing geotechnical investigation in karst as a result of the differences in bedrock ages, degree of deformation and perhaps most important, the degree of tectonism experienced by the bedrock in different tools used or the manners in which geotechnical investigations are performed should be the same for all types of karst. It should be noted that our experience has been in limestone (CaCO 3 ), dolomite (CaMg(CO 3 ) 2 ) and/or marble, we will allow others to comment on gypsum (CaSO 4 2 O) and other evaporites. Another aspect that must be considered is the existence of local or State ordinances regulating either the construction or impact allowed at karst sites. These regulations, where they exist, can have different intentions. For example, many municipal limestone ordinances in Pennsylvania are primarily directed toward inhibiting New Jerseys generally toward limiting construction on Figure 1. National Karst Map showing portions of US underlain by karst (from Tobin and Weary, 2004).

PAGE 3

551 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 the same orogenic forces, but did experience some of the stress fields caused by several openings and closing of the proto-Atlantic Ocean. Geologic Concerns The occurrence of sinkholes (dolines) swallowing buildings, automobiles, farm equipment and people settlement of structures (including dams) and sinkhole occurrences in roadways, in backyards, below swimming pools, farms, manure storage facilities, railroad structures, fuels storage areas and bridge abutments. and compromising stormwater detention/retention/ infiltration systems, thus allowing contaminants to reach ground water supplies. large construction can be exceptions. The recent New National Karst Map (Figure 1) also includes the areas underlain by evaporite karst, which are not covered in this paper. All of the aforementioned karst types of concern were originally deposited in warm, relatively shallow seas. Deposition or coral growths continue today in the warm waters of the Atlantic Ocean and Caribbean Sea. The older carbonates (Cambro-Ordovician-aged) have been Mountains, a series of orogenies resulting in faulting, folding and fracturing (and some metamorphism) along what was the Atlantic coast approximately 300+ million years ago. The mid-continent carbonates are generally the same age Figure 2. Representation of the simplistic divisions of United States karst (Section A from Schmertmann and Henry, 1992).

PAGE 4

552 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Geotechnical Engineering Sinkholes are an obvious concern in areas underlain by carbonate bedrock. Pictures of huge holes in the ground swallowing cars and houses make for big news. However, much of the older rocks can be quite hard when protected from weathering. For example Mammoth Cave in Kentucky and the Natural Bridge in Virginia. Considering the expected lifetime of many construction appropriate approach. The unconfined compressive strength of Cambro-Ordovician limestones and dolomites can be on the order of 10,000 to 15,000 pounds per square inch. As a result, the roof over a cavity can likely be well-documented. However, the weak, recent corals of Florida and the Caribbean are not as friendly. Flying over the flat lands of Florida it is possible to see large, circular, water-filled sinkholes dotting many areas. Aerial photographs of such a Florida landscape can be most diagnostic, especially if linear patterns and/or frequency of occurrence can be determined. These images can also be troubling to an owner, developer or geotechnical investigator. However, even the relatively weak, recent corals of Florida and the Caribbean can support significant loads where not compromised by solutioning. Although the Schmertmann and Henry representation of Figure 2 (Section A) may somewhat exaggerate concerns at a Florida site, the concerns remain. There are a number of engineering solutions to founding structures, roads, utilities, detention/retention basins and tunneling in karst. The basic problem is evaluating the subsurface conditions satisfactorily and to define the solution in a reasonably economic manner. So, it becomes somewhat of a balancing act. There is a need to find suitable materials to carry the proposed loads during the economic life of the structure. The problem is more complicated than it would be at most non-karst sites. One of the problems of founding on Appalachian karst over a more conventional site is compounded by the variability, both laterally and vertically, of the seams and fractures, and the general subsurface conditions. The material properties of these contorted rocks can vary significantly over short distances. The irregularity of the Appalachian, as well as some recent, softer Florida-type bedrock surfaces presents additional concerns in its effect upon the ground surface, further complicating geotechnical evaluations (Figure 2). Long and sometimes sinuous conduits are common in the central USs flat karst (e.g., Mammoth Cave). Both lateral and vertical variations in the overburden materials (thickness and properties) are more common in the Appalachians. After a visit to almost any commercial water conditions over time have had on the bedrock. These differences in rock type, age of deposition and the range of effect from tectonism have to be considered in geotechnical investigation and for potential remedial solutions in these various environments. Geotechnical Evaluation The first step in developing a program of investigation is to expand upon the knowledge of the geology in the area or site of concern. State and Federal agencies generally have a wealth of information concerning subsurface conditions. These data can include: 1. Bedrock types and their expected depth below grade. 2. Drillers logs and well yields from specific locations or geologic formations. These data can provide clues to the degree of fracturing and or solutioning found within various formations. However, we must be aware of drillers classifications such as gray rock. 3. The existence and density of caves, sinkholes and disappearing streams. 4. The existence of known or suspected faulting, antiforms and synforms (i.e., where the bedrock where increased solutioning is likely to be experienced. 5. Bedrock strengths and quality of overburden. 6. Textural classifications that can provide a clue to material solubility. Be aware that similar rocks and soil types do not necessarily have the same formation names from State to State.

PAGE 5

553 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 filtration in overburden soils and increasing the speed that contaminant can reach a receptor. Preliminary Site Evaluation In any limestone investigation, the best bang for the buck is usually the results of the initial stages of a site study. The first step is a review of any available data from federal and state sources including environmental reports and studies performed for nearby sites. Aerial photos of the site are, whether from aircraft or satellites, highly valuable and often available from archives. A series of aerial photos taken over time can show changes in vegetation, landforms, farming practices, etc. For example, why is a tree standing alone in the middle of a cultivated field? Karst features can develop over time and then later masked by farming practices. Aerial photos taken in the early spring (before tree cover) and during wet years (e.g., Figure 3) can show changes in moisture that can be quite telling. Persistent linear and circular features are particularly suspicious if noted in photos taken over time or on LiDAR. Even drought-induced crop lines have been used as a tool in delineating potential sinkhole locations (Panno et al, 2013). These features should be further investigated by a site reconnaissance to help in identifying any noted features for use in a subsurface model. In the past, sinkholes had many uses, garbage control aids. Areas that have been mined or quarried can susceptibility to solutioning. In Appalachian karst the ground elevation variations can be more severe than observed from a windshield reconnaissance. Standing on a high point overlooking a site or flying over a large site at low altitude can be very informative, particularly within more flat-lying areas. We have observed sites as pockmarked with sinkholes as a World War I battlefield. geologist/geotechnical engineer can develop a great deal of insight from a well-rounded site evaluation performed early in the site selection or development process. This initial phase, even though quite economical, can be used to determine the intensity of subsequent site investigations and/or whether a site or route should be abandoned. Florida and mid-continent karst concerns can be caused by the lateral and vertical movement of ground water, with the effects being greater over shorter periods in the softer Florida and Caribbean karst. Can the investigator sample enough locations on the site, either by direct or indirect means, to provide an appropriate support solution for such a variable subsurface? Imagine the difficulties of investigating a site such as the one shown on Figure 3 for the development of a satisfactory model of the subsurface. The many variables related to different karst areas of the US make it virtually impossible to answer all site related questions definitively before the start of construction. Hence, any construction-related planning should consider contingencies for increased costs for both inspection and the possibility of additional or supplemental remediation. In addition, the construction process itself can create unstable or weakened conditions. Often, there is poor control of surface and ground water during construction of a facility. Excavation at a site can remove a protective layer of low permeability soils over solutioned rock, and pond water in these compromised areas. Ground water can travel along the top of the bedrock surface until it finds an entrance into a cavity, eroding soils from directly atop the bedrock and increasing the area of concern. Changing the hydraulic head and/or flow rate at a construction site, either by cut or fill, might alter otherwise stable conditions. The effects of changing the hydraulic conditions at karst sites are exemplified by the failure or remediation of many dams built atop karst. Also, the potential for ground water contamination is much higher in karstic environment by decreasing Figure 3. Aerial photograph of western New Jersey karst site.

PAGE 6

554 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE variations. Differences in results from the drilling program may be caused by geologic differences affecting the electrical properties of the subsurface materials, modeling parameters, and orientation of the electrode array. Apparently, even the geophysicist could not correlate the survey results into a coherent model in Appalachian karst. This geophysical investigation did subsurface conditions in preparation for any kind of development. Thus, it appears that geophysical procedures alone will not yield the answers to many karst concerns and can, at best, be interpreted with the aid of test borings and probes drilled by experienced, cooperative personnel under the technical direction of experienced field personnel including geophysicists. Test pits can be performed in a conventional manner and can be very informative if portions of the rock surface can be exposed. Potential signs of solutioning can be deduced from near surface-effects. Is the weathered bedrock relatively uniform and straightforward or is there evidence of leaching or groundwater movement? Are the remains of an old, filled sinkhole obvious in the pit walls or bottom (Figure 4)? Is relict bedding distinguishable in the pit wall? Test borings should be drilled using rotary-wash techniques without the use of drilling mud whenever possible so that drilling water loss depths and quantities can be monitored. In clean, sandy soils this may not be possible, but drilling with augers and periodically introducing water (say between samples) or the use of a light mud where necessary Site/Route Investigation If the site or selected route seems economically and technically viable, then it is likely that some additional geotechnical study will be required. Obviously, the investigation. It is often advisable to phase future investigations. The information most desired at any karst site is the distribution and dimension of soil voids and bedrock cavities, and whether those cavities or voids are filled with water or soil. Also necessary is some knowledge of the bedrock surface variations. The expected variations in the bedrock surface will differ largely from relatively flat mid-continent rocks to Floridas variable and generally soft carbonates and the even more erratic, hard Appalachian rocks. Another consideration is what effect the planned construction would have on local ground water supplies, which can be influenced even from considerable distances in a water-filled cave system. Dye studies performed by knowledgeable and aware professionals seems to be the only way of assessing possible ground water concerns and travel paths that new construction atop karst can effect. Exploring caves (spelunking or diving) by the more adventurous investigators can be very useful, although dangerous. A host of geophysical procedures have been espoused as an effective investigative tool. These techniques include seismic reflection, refraction and tomography, electrical conductivity and resistivity, self-potential, ground penetrating radar, gravity, and Spectral Analysis of Surface Wave (SASW) methods. Apparently, the best results in the use of geophysics at karst sites have been a combination of geophysical procedures coupled with test borings (Benson, et al, 1998). The efficacy of investigating with a single geophysical tool, using air-track probes and test borings to calibrate the results in Appalachian karst, is unfortunately exemplified by the following statements by one geophysicist in his report of a resistivity survey: Generally, resistivity data is very good with good repeatability and trends that correlate well from by: The results of the survey show several different subsurface conditions. Detected by the survey are possible sinkholes and possible depth to bedrock Figure 4. Evidence of past and ongoing sinkhole formation in house foundation excavation.

PAGE 7

555 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 enough quality subsurface information to be able to generate accurate excavation and backfilling costs. A pinnacled bedrock surface makes this very difficult. For example, the Maryland State Highway Administration (MDSHA) responded to a large sinkhole immediately 110 feet long by 30 to 35 feet wide and 35 feet deep at the throat. In an effort to keep the highway embankment stable, the sinkhole was quickly and partially backfilled with some 2,700 cubic yards (cy) of quarry waste. Unfortunately, while drilling to place grout, the rock surface was revealed to be quite variable, necessitating an additional 2,045 cy of grout to fill the subsurface cavity. The maximum bedrock depth encountered was 100 feet (Martin, 2004). Even with the local geologic information available from highway construction and local quarry operators, they could not anticipate the extent of the weakened or missing subsurface materials. B) Transfer Construction Loads to Sound Rock Transferring loads from weakened subsurface areas to those capable of supporting loads is another founding alternative in karst. Bridging openings in the bedrock have been used. A common foundation solution can be the use of driven piles or caissons, particularly with the present day ability to drill through the pile or caisson shaft in order to evaluate the quality of the founding materials below the pile tip and to possibly introduce grout if conditions warrant. Pin piles have often been used satisfactorily because the pre-drilling used for their installation allows an increase in knowledge of the subsurface conditions at the pile location. However, grouting the pin piles to bond them to the sides of the hole can require large amounts of grout and there can remain unsupported lengths through Mathematical models have been or can be developed to assess the load-carrying ability of cave or sinkhole roofs in order to provide a requisite number and type of deep foundations to be used. However, it is probably more economical to perform a one-time investigation/foundation solution such as drilling and grouting, or installing a pile foundation to resistance depths, then drilling through the pile should allow for drilling fluid losses to be monitored while keeping the boring open. Drilling water lost at the top of the rock usually indicates a down-gradient channel or the gradual erosion of soil to pinnacles. Conventional soil sampling techniques are generally adequate. Although providing water to the drilling site can become a logistical problem, it can be Encountering a karstic bedrock surface can be quite weathered, broken, a bedrock pinnacle or an erratic boulder, or saprolite below sound rock? Coring most carbonate rock is best done using doubleor tripletube, split core barrels. At least one spare core barrel should be on hand as the variable conditions that can be encountered are often hungry for drilling equipment The information that can be observed from cores derived from a split core barrel is far more representative of the actual bedrock conditions and well worth the increased expense of its use. Fracture frequency and orientation is more easily observed and fracture and cavity filling is often captured in the barrel, along with highly weathered hammering the core from a non-split barrel. Again, experienced drillers and competent inspectors are essential. Foundation Design Considerations Most foundation solutions are available for use once grade, B) transfer construction loads to sound rock or bypassing the area of concern, C) densifying overburden materials and D) grouting of cavities with non-shrinking materials. Whatever concept is chosen, the execution should be flexible and hopefully cost-effective. A) Excavate to Sound Rock and Backfill The simplest is excavation to sound materials and returning the area to grade with compacted fill (sometimes after dental grouting of bedrock openings) or even lean concrete if the excavation is shallow enough. The unfortunate part of such a program is the need for

PAGE 8

556 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Conclusions There are many hundreds of square miles of the United States underlain by karstic soils and bedrock. Unfortunately, all karst is not the same, though this has that have swallowed houses, cars and people, but even if an expert has been contacted, the geotechnical concerns that may have existed are treated in passing. Geotechnical practitioners must be more communicative when dealing with property/facility owners, planners and engineers. Clients and designers should be made aware of the possible dangers lurking below as well the impact the karstic subsurface conditions could possibly have on their plans. It is difficult to understand why more municipalities do not have appropriate limestone ordinances as some have been tested in court and proven legal. The same can be said few appear to understand the difficulties that can result from the existence of carbonate bedrock below a site. Obviously, experienced consulting is necessary to ensure sound construction in karst and in the development of limestone ordinances. These ordinances should be directed toward the varying conditions of an individual karst site, as well as the differences in karst from region to region as discussed herein. As much data is available to the geotechnical engineer prior to planning a subsurface investigation, this should Determining an appropriate investigation program for a karst site is dependant upon this knowledge, which can References Benson RC, Kaufmann RD, Yuhr L, Martin D. 1998. Assessment, prediction and remediation of karst conditions on I-70, Frederick, MD. In Proceedings of the 49th Highway Geology Symposium, 1998 Federal Highway Administration. 1995. Dynamic Compaction, Geotechnical Engineering Circular No. 1. FHWA-SA-95-037. Martin DA. 2004. Case History of the South Street Sinkhole, Frederick, Maryland. In Proceedings of the 55th Highway Geology Symposium, 2004 Panno SV, Luman DE, Kelly WR, Alschuler MB. 2013. The use of drought-induced crop lines as a tool shaft. The quality of the subsurface data needed to develop such a model for arch support and to install an extensive deep foundation would likely be more economical if the data were obtained simultaneously with foundation installation than in two separate phases with modeling time in between. There is no doubt that liaison between the designers and the on-site foundation installers are necessary in such cases. C) Densifying Overburden Materials The densification of the overburden materials is compaction (DDC) and compaction (i.e., low-mobility) grouting. DDC is simply the dropping a very large weight from a fixed height to impart a design energy to densify near-surface soils, collapse soil voids and clog bedrock surface openings with overburden soils. If the rock is shallow, cave roofs may also be collapsed. The writers have found that dynamic compaction (dynamic destruction in karst?) operations, coupled with test borings and or probes, have been successfully used in both defining the need for site remediation the preferred remediation. Dynamic compaction can be performed as a lower cost alternative to increased investigation while possibly solving many of the problems that would be encountered during construction. Low-mobility grouting is often used to densify weak overburden soils in preparation for construction loads and has the benefit of also filling, at least partially, encountered voids and cavities. The drawback to such grouting is if voids and cavities get extensive, grout costs can escalate substantially. It is also a difficult method to use in sites underlain by moist clayey soils because of the slow rate of pore pressure decay. D) Grouting Cavities High-mobility grouting, with site mixed grout generally provide the most flexible and cost-effective means of remediating karstic subsurface conditions. The use of commercial foaming agents can reduce material costs yet yield strengths of 4,000 pounds per square foot or more, depending upon the constituents of the mix. That is usually more than enough strength to reinforce bedrock cavities and grouting program does require a grout crew and supervising beast below and are ready to respond.

PAGE 9

557 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Doctor, DH, Stephenson, JB, editors, Proceedings of the 13th Multidisciplinary Conference on Sinkholes & the Engineering and Environmental Mexico. Carlsbad (NM): National Cave and Karst Research Institute. p. 53-59. Schmertmann JH, Henry JF. 1992. A design theory for compaction grouting. Grouting, Soils Improvement and Geosynthetics, ASCE Geotechnical Publication 30 (1): 215-228. Tobin BD, Weary DJ. 2004. Digital Engineering Aspects of Karst Map: A GIS Version of Davies WE, Simpson JH, Ohlmacher GC, Kirk WS, and Newton EG, 1984, Engineering Aspects of Karst: U.S. Geological Survey, National Atlas of the United States of America, Scale 1:7,500,000. Weary DJ, Doctor DH, Epstein JB, Orndorff RC. the framework of the new national karst map. US Geological Survey Scientific Investigations Report 2008-5023.

PAGE 10

558 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

31 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CONDUIT FLOW IN THE CAMBRIAN LONE ROCK FORMATION, SOUTHEAST MINNESOTA, U.S.A. John D. Barry Minnesota Department of Natural Resources, Division of Ecological & Water Resources, 500 Lafayette Road, St. Paul, MN 55155-4040, U.S.A., john.barry@state.mn.us Jeffrey A. Green Minnesota Department of Natural Resources, Division of Ecological & Water Resources, 3555 9th St. NW Suite 350, Rochester, MN, U.S.A., 55901, jeff.green@state.mn.us Julia R. Steenberg Minnesota Geological Survey, 2609 Territorial Road, St. Paul, MN 55114, U.S.A., and01006@umn.edu dominated units generally have bedding-parallel and vertically oriented apertures less than a few centimeters. The process by which the bedding-parallel secondary mechanically developed. However, interstitial carbonate cement within these units leads to the possibility of dissolution being a minor factor in the formations groundwater flow characteristics. These dye traces were conducted at three different sites across a twenty-three kilometer distance and are evidence that the siliciclastic Lone Rock Formation has a conduit-flow component similar to that found in carbonate karst aquifers. Introduction Southeastern Minnesota is underlain by a sequence of Cambrian to Devonian sedimentary bedrock layers that were deposited in a broad structural depression known as the Hollandale Embayment. In general, the older rocks are dominated by siliciclastic materials and the younger rocks dominated by carbonates (Mossler, 2008). A hydrogeologic framework that describes four prominent karst systems for southeastern Minnesota (Runkel et al., 2013) is based largely on the work of Alexander and Lively (1995), Alexander et al. (1996), and Green et al. (1997, 2002). These include the Devonian Cedar Valley, the Upper Ordovician Galena-Spillville, the Upper Ordovician Platteville Formation, and the Lower Ordovician Prairie du Chien Group. The karst systems defined in the framework meet the traditional criteria associated with karst, an integrated mass-transfer system in soluble rocks with a permeability structure dominated by conduits the circulation of fluid (Klimchouk and Ford, 2000). Abstract The karst lands of southeast Minnesota contain more than one hundred trout streams that receive perennial are karst aquifers. Field investigations into the flow characteristics of these formations have been conducted using fluorescent dyes to map groundwater springsheds water resource protection. Recent field work has focused on the Cambrian Lone Rock Formation, a siliciclastic unit consisting of very finegrained sandstone and siltstone with minor beds of shale and dolostone. The formation is mapped within tributary valleys of the Mississippi River throughout southeastern Minnesota and southwestern Wisconsin. Overlying the Lone Rock is the Cambrian St. Lawrence Formation. into stream sinks where the upper St. Lawrence is the bedrock unit closest to the land surface. At three of these sinking stream locations, dye was recovered emanating from springs located in the basal St. Lawrence or from Dye-breakthrough velocities calculated using passive charcoal detectors ranged between 21-214 meters/day at one location and 88-153 meters/day at another. At a third site, automatic water samplers were placed at a spring that had been previously demonstrated to be connected to a St. Lawrence stream sink through dye tracing. In that trace, an eight-hour sampling frequency determined the dye-breakthrough velocity was 314 meters/day. Based on outcrop and borehole observations in Minnesota, secondary pore networks in siliciclastic-

PAGE 2

32 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE 2014). The topography is dominated by a broad plateau of resistant dolostone of the Ordovician Prairie du Chien Group (OPDC). The OPDC is one of the four karst systems identified in southeastern Minnesota and contains many solution-enhanced fractures and cavities such as sinkholes, large conduits, and caves. A region-wide high permeability Tipping et al., 2006). Catastrophic failures of sewage treatment ponds have occurred where this high permeability The resistant dolostone layer that dominates the plateau is dissected by numerous narrow valleys, especially in the eastern region of the counties near the Mississippi River. In general, the stratigraphy underlying the OPDC is dominated by more easily weathered sandstone, siltstone, and shale layers that are prevalent on slopes and valley floors. It is in these non-carbonate layers that the focus of this report occurs. outcrops of the lower Jordan Sandstone through upper Tunnel City Group in the Twin Cities metropolitan area shows that vertical to subvertical fractures tend to preferentially terminate at specific bedding contacts (Runkel et al., 2014). Changes in lithology at these interfaces are also likely responsible for large head In the last seven years, investigators in southeastern rapid flow in two of the siliciclastic dominated formations of the Cambrian system: the St. Lawrence and Lone Rock Formation (Green et al., 2008, 2012). The St. Lawrence consists of well-cemented, thin-to-medium beds of siltstone, dolomitic siltstone, very fine-grained sandstone and shale. The Lone Rock, a formation within the Tunnel City Group, is mostly composed of a fine-grained sandstone and siltstone with interbedded shale and dolostone (Mossler, 2008). Groundwater flow velocities of 35 meters/day (115 feet/day) recorded in many dye traces through these units are consistent with conduit flow. However, the lack of large conduit networks in outcrop and borehole observations makes classifying of flow through tracing and breakthrough curves for these units is more consistent with the definition of pseudokarst, landscapes with morphologies resembling karst, and/or may have a predominance of subsurface drainage through conduit type voids, but lack the element of long-term evolution by solution and physical erosion (Kempe and Halliday, 1997). Recent work has parallel apertures in siliciclastic bedrock of southeastern Minnesota are connected through an anastomosing network of apertures, clustered along discrete (<2m) stratigraphic intervals and found at depths exceeding 200 meters (Runkel et al., 2015). The apertures are more commonly associated with distorted bedding interpreted to reflect dewatering features that occurred shortly after burial when the rock was only partially lithified. Therefore, these voids are unlikely to be primarily the result of dissolution as karst is traditionally defined (Stewart et al., 2012 and Runkel et al., 2015). This paper focuses on the geologic and hydrogeologic setting and results of three recent traces in these siliciclastic units conducted in Houston and Winona counties (Figure 1). Geologic and Hydrogeologic Setting of Dye Traces In Houston and Winona counties, bedrock units from the Upper Cambrian through the Upper Ordovician are generally within 15 meters of the land surface and are capped by unconsolidated sediments such as loess, Figure 1. Regional geologic setting of Paleozoic rocks in southeast Minnesota and locations of the dye traces presented in this report. Upper Devonian units are depicted with blue hues in the southwest corner of map. The Ordovician units are depicted with orange hues. Lower Cambrian units are depicted with gray and light brown hues in eastern portion of map. Geologic map from Runkel et al., 2013.

PAGE 3

33 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 was timed and mapped based on when and where it reemerged at a spring. Dye was recovered through water grab samples, passive dye receptors (packets of coconut charcoal informally referred to as bugs) and, for the Bridge Creek trace, an automatic water sampling device. differences noted at a recently installed multilevel groundwater monitoring multiport system that intersects (Runkel et al., 2014). In deep bedrock settings, the Cambrian St. Lawrence Formation (CSTL) has a low bulk vertical conductivity a marked anisotropy (Runkel et al., 2003). The low bulk vertical hydraulic conductivity for the CSTL in deep settings has influenced the general understanding of the hydrostratigraphic properties of the unit and it is generally CSTL as an aquitard is included in the Minnesota Well Rules handbook where it specifically states that a stratum at least 10 feet [3.05 m] in vertical thickness of the St. Lawrence is a confining layer (Minn. Dept. of Health, 2011). In southeastern Minnesota where bedrock is shallow and is cut by deeply incised valleys, the CSTL can have highly variable bulk vertical conductivity and high bulk valleys, the ability of the CSTL to behave as an aquitard is tenuous. A geologic column for Houston County (Figure 2) shows properties for each of the units (Steenberg, 2014a). into either aquifer or aquitard based on their relative permeability. Layers assigned as aquifers are permeable and easily transmit water through porous media, fractures or conduits. Layers assigned aquitard have lower permeability that vertically retards flow, effectively hydraulically separating aquifer layers. However, layers designated as aquitards may contain high permeability bedding plane fractures conductive enough to yield large quantities of water. Methods The dye traces and geochemical data presented in this report were conducted to further delineate springsheds St. Lawrence Formation, and describe surface water to groundwater interactions in the counties. Traces focused on locations where surface water was known to be sinking and where local landowner permission was granted. Fluorescent dye was poured into a sinking stream or sinkhole. From there its flow through a conduit system Figure 2. Geologic and hydrogeologic attributes of Paleozoic rocks in southeastern Minnesota. Modified from Steenberg, 2014a.

PAGE 4

34 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE into a discrete sinking stream point on Bridge Creek (MN28:B00006) located in the CSTL. Stream discharge at the time was estimated to be 0.008-0.014 cubic meters per second (0.3-0.5 cubic feet per second). Eosin dye was detected at levels high enough for positive quantification at a number of sampling sites. Dye was detected at Rostvold Spring which emanates from the Lone Rock (CTLR) 20 28 days later. Assuming a straight-line distance from the stream sink (MN28:B00006) to Rostvold Spring, this translates to a minimum peak groundwater velocity ranging from roughly 146 to 205 meters/day (480 ft/day to 670 ft/ day). Dye was also detected in charcoal receptors at the Bridge Creek Outlet, the Frauenkron Crossing, the Frauenkron well, and the Bolster spring pond, all within the CTLR. Dye detection for the Frauenkron well was determined using passive detectors placed in their toilet tank reservoir. The well owners had previously stopped using the well for potable water after having multiple gastrointestinal issues. The well has since Analyses of water samples and charcoal detectors were performed at the University of Minnesota Department of Earth Sciences Hydrochemistry Laboratory using a direct water samples from springs and streams for two of the traces using a Dionex ICS-2000 Ion Chromatography System. These samples were collected under base flow Stratigraphic interpretations of spring positions are based on a combination of field outcrop examination and correlations to water well records from the County Well Index (Bauer and Chandler, 2014). Dye Tracing Results Bridge Creek Bridge Creek in Yucatan Township, Minnesota was the site of dye traces in 2012 and 2013 (Figure 3). It receives discharge from a number of springs that emanate from the basal Jordan Sandstone. The 2012 trace consisted of the introduction of 1.104 kilograms (kg.) of Eosin dye Figure 3. Bridge Creek site sampling locations and dye flow vectors for 2012-2013. Dye flow vectors for the 2012 trace are shown in black. Dye flow vectors for the 2013 trace are shown in yellow. See Figure 2 for definition of map symbols. Note: trace vectors not drawn to dye receptor locations that integrate upstream waters. Geologic map from Runkel et al., 2013.

PAGE 5

35 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Girl Scout Camp Creek The Girl Scout Camp Creek trace was run synchronously with the 2013 Bridge Creek trace. Although not fieldchecked, the creek is assumed to begin at springs emanating from the Jordan Sandstone. The trace was initiated on September 12, 2013, with the introduction of 1.057 kg of Rhodamine WT dye into a sinking pool in the CSTL (MN28:B00004) located in a tributary valley of Girl Scout Camp Creek. Stream discharge at the time was estimated to be 0.007 cubic meters per second (0.25 cubic feet per second). Rhodamine WT dye was detected at levels high enough for positive identification at multiple sampling sites including the Whispering Hills Spring (CTLR) within 15 days and at the Peterson Spring (CTLR) within 14 days (Figure 5). An early breakthrough time for the St. Lawrence nose spring was not recorded due to mammals destroying passive detectors early in the trace. In southeastern Minnesota, numerous CSTL and CTLR springs resurge at promontory points at the toe referred to as nose springs in this paper. Dye was recovered at the passive detectors at the CSTL nose spring between September 27 and October 17. Assuming a straight line distance from stream sink (MN28:B00004) to Whispering Hills Spring, this translates to a minimum peak groundwater velocity of roughly 88 meters/day (290 ft/day) and to Peterson Spring, a minimum peak groundwater velocity of roughly 153 meters/day (503 ft/day). been abandoned and replaced. Borehole geophysics conducted during well abandonment found evidence of fracture flow in the middle to upper Tunnel City from the caliper and fluid resistivity logs. The year following the 2012 Bridge Creek trace, 24.1 centimeters (9.5 inches) of rain fell in the Bridge Creek watershed between June 21 and 25, 2013. Overland stream flow during this event heavily altered the geomorphology of Bridge Creek, causing large shifts in the streams thalweg. Following the precipitation event, surface water was no longer making it down the valley and reconnaissance was conducted to determine where the stream was losing water. Surface water was found to be sinking into a location roughly 610 meters (2000 feet) upstream from MN28:B00006 in the CSTL. A stream sink had previously been speculated to be in this general vicinity based on black and white aerial imagery from 1991. An additional trace was conducted on September 12, 2013 because of this dramatic shift in the sinking stream location. The second dye trace introduced 1.073 kg of Uranine dye into a sinking pool and discrete sinking stream point on Bridge Creek (MN28:B00005) (Figure 3). Stream discharge at the time was estimated to be 0.003 cubic meters per second (0.1 cubic feet per second). Passive receptors and three automatic samplers programmed to sample at eight hour intervals were deployed during the first seventy days of the 2013 dye-breakthrough curves (Figure 4). Uranine dye was detected at levels high enough for positive identification at both the passive and active sampling sites. Uranine dye was detected in the automatic samplers at Rostvold Spring (CTLR) 15 days later. Assuming a straight line distance from the stream sink (MN28:B00005) to Rostvold Spring, this translates to a minimum peak groundwater velocity of roughly 314 meters/day (1,031 ft/day). This velocity is consistent with previous traces in the St. Lawrence (Green et al., 2012). Dye was also detected in charcoal detectors at Frauenkron Crossing, Jerry Lee Spring, and Bolster Pond Spring Outlet. Eosin dye used in the 2012 trace was additionally detected a year later during the 2013 trace at Rostvold Spring, Bridge Creek Outlet, Jerry Lee Spring, and Frauenkron Crossing. The long tail on dye recovery observed in this trace is consistent with dye recovery in previous St. Lawrence traces (Green et al., 2008, 2012). Figure 4. Breakthrough curve for the 2013 Bridge Creek dye trace. Dye was introduced into stream sink (28:B00005) located roughly 4,716 meters (15,469 feet) in a straight line distance from Rostvold Spring. Dye arrived at the spring 15 days after input.

PAGE 6

36 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The 2012 trace consisted of the introduction of 1.166 kg of Uranine dye into a discrete sinking stream point on upper Campbell Valley Creek (MN85:B0020) located in the CSTL. Stream discharge at the time was estimated to be 0.00006.0003 cubic meters per second (0.002 0.011 cubic feet per second). Uranine dye was detected at levels high enough for positive identification at multiple sampling sites including the Barnhardt 1 and 2 springs (CSTL) and Pagels Big Spring and Power Spring (CTLR). Assuming a straight line distance from the stream sink (MN85:B0020) to the Barnhardt springs, the minimum peak groundwater velocity is roughly 21 meters/day (69 ft/day). Assuming a straight line distance from the stream sink (MN85:B0020) to Pagels springs, the minimum peak groundwater velocity ranges between 55 and 214 meters/day (180 ft/day-702 ft/day). Despite being located within the lowermost Lone Rock Formation, these velocities are still consistent with previous traces in the St. Lawrence and uppermost Lone Rock Formation. The locations of the Strand nose spring and bank spring were not known during the active monitoring stage of this trace and they were not visited unit March 4, 2014. At that point passive dye receptors were deployed and water chemistry was collected. Passive detectors showed no evidence of dye at the Strand springs. Campbell Valley Creek Dye traces were conducted on Campbell Valley Creek in Pleasant Hill Township, Minnesota in 2012 and 2013. Campbell Valley Creek starts at a spring that emanates from the basal Jordan Sandstone and sinks within 50 meters (164 feet) into the upper St. Lawrence Formation (Figure 6). Farther down the valley, water resurges at two small perennial St. Lawrence Formation springs. The Campbell Valley Creek also sinks into the Lone Rock Formation within 10 meters (32 feet) downstream and resurges at the basal Lone Rock Formation farther down the valley. Figure 5. 2013 Girl Scout Camp Creek site sampling locations and dye flow vectors (yellow). Note: trace vectors not drawn to dye receptor locations that integrate upstream waters. Geologic map from Runkel et al., 2013. See Figure 2 for definition of map symbols.

PAGE 7

37 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Nitrate concentrations greater than 1 part per million (ppm) are greater than background conditions and possibly indicate that an aquifer has been impacted by activities on the land surface (Minn. Dept. of Health, 1998 and Wilson, 2012). Nitrate concentrations greater than 3 ppm indicates that an aquifer has been impacted by activities on the land surface (Minn. Dept. of Health, 1998). Chloride concentrations of greater than 5 ppm can also be used to indicate that an aquifer has been impacted by activities on the land surface. Multiple investigators have used Cl/Br ratios to identify chloride sources to In general, samples with chloride-to-bromide ratios above 300 are waters that have been elevated by human activity. Anion chemistry collected during the Bridge Creek and Girl Scout Creek dye traces show elevated levels of nitrate, chloride, and chloride-to-bromide ratios in the groundwater of upper stratigraphic units and surface waters of these watersheds (Table 1). These The 2013 trace consisted of the introduction of 1.843 kg of Rhodamine WT dye into the creek upstream from the stratigraphically lower sinking stream reach located in the Lone Rock Formation (MN85:X0037). Stream discharge at the time was estimated to be 0.008.014 cubic meters per second (0.3.5 cubic feet per second). The dye input points and receptor locations for both the 2012 and 2013 traces are shown in Figure 6. Dyes were detected at levels high enough for positive quantification at Pagels Big Spring and Pagels Power Spring. Minimum peak groundwater velocity to the Pagel Springs is roughly 33 meters/ day (108 ft/day), assuming a straight line distance from the sinking stream reach (MN85:X0037) to the springs. Geochemical Results and Discussion Elevated levels of nitrate and chloride can be used as geochemical indicators of recent human influence on groundwater. They can be attributed to the application application. We used the following classification scheme to assign the term elevated to these anion species. Figure 6. 2012-2013 Campbell Valley Creek site sampling locations and dye flow vectors. Dye flow vectors for the 2012 trace are shown in yellow. Dye flow vectors for the 2013 trace are shown in black. See Figure 2 for definition of map symbols. Note: trace vectors not drawn to dye receptor locations that integrate upstream waters. Geologic map from Runkel et al., 2013.

PAGE 8

38 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE moves rapidly downward through enhanced fractures in the Prairie du Chien until it encounters the basal Jordan Sandstone (Figure 7). From there, lateral flow across the basal Jordan resurges as springs in incised valleys that form the headwaters of many creeks. The nitrate concentrations of these headwater streams are elevated above background conditions. In incised valley settings, rapid vertical flow in a stair step like pattern appears to continue through formations underlying the Jordan, with sinking streams identified in both the St. Lawrence and Lone Rock Formations (Figure 7). Groundwater emerging from springs progressively deeper in the geologic section show mixing of nitrate poor water from distant sources up-gradient of the incised valleys. In general, nitrate concentrations of these springs show moderately high mixed signatures or low signatures that are consistent with un-impacted groundwater. Tritium values of groundwater collected in nearby Wabasha and Fillmore counties show that vintage water, elevated species are likely due to land use activities on the agricultural landscape in these and surrounding watersheds and are similar in concentration to levels found in a regional nitrate investigation of southeastern Minnesota (Runkel et al., 2013). No water samples were collected for anion analysis for the Campbell Valley trace. Analytical results of water collected for these investigations verify the presence of anthropogenic signatures from land use in groundwater in this geologic setting. In southeastern Minnesota, wells open to the Prairie du Chien Group are more than twice as likely to yield water with a nitrate concentration above 2 ppm (30.3 percent) than are wells open only to the Jordan Sandstone (12.3 percent) (Runkel et al., 2013). The marked difference in nitrate concentrations of the Prairie du Chien and the Jordan is attributed to the ability of the lower Oneota to retard nitrate-enriched water downward into the Jordan. However, near the edges of the Prairie du Chien plateau, groundwater Table 1. Anion chemistry of grab samples collected in the Bridge Creek and Girl Scout Creek Watersheds. Note: ug/g = mg/L = ppm

PAGE 9

39 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 curves of the Galena and Prairie du Chien formations. Breakthrough curves in these karst units are generally asymmetric with rapid peaks and tails that last hours to days for the Galena and days to weeks for the Prairie du Chien (Alexander, E.C., Jr., oral commun., 2015). The groundwater velocities calculated for the Bridge Creek, Girl Scout Camp Creek, and Campbell Valley Creek traces are consistent with previous traces in the St. Lawrence and Lone Rock where peak rates range from 35 to 600 meters/day (Green et al., 2012). The connection between surface water and groundwater found in these dye traces has human health and contaminant transport implications that are related to aquitard integrity and aquifer susceptibility in this unique setting. In incised valley settings in southeastern Minnesota, both the St. Lawrence and Lone Rock may be compromised by connectivity to surface water and the ability of water to travel at rapid rates. Conclusions Dye trace results, outcrop observations, water chemistry, and data gleaned from borehole geophysics demonstrate in excess of 50 years old, generally is present underlying the St. Lawrence Formation (Petersen, 2005, Zhang and the St. Lawrence impedes downward flow and acts as an aquitard. Dye traces completed from sinking streams in the St. Lawrence and Lone Rock all show similar breakthrough curves. The curves exhibit a rapid breakthrough, rapid rise to a peak, followed by very long (months to years) tails. In February 2015, passive dye receptors were placed at springs where positive dye detection occurred for each of the traces detailed above. The receptors were in place for a two week period and recovered Uranine dye from the 2013 Bridge Creek trace at the Jerry Lee Spring and recovered dye Rhodamine WT dye from the 2013 Girl Scout Camp Creek trace at the Peterson Spring. These dyes were recovered roughly 510 days following their introduction. No dye was recovered from the Campbell Valley trace at Pagels Power Spring. Breakthrough curves from traces in the St. Lawrence and Lone Rock formations are fundamentally different from Figure 7. Cross section of the Girl Scout Camp Creek trace with surface water and spring nitrate levels. Green shading used to depict regional nitrate levels from Runkel et al., 2013.

PAGE 10

40 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Land L, Doctor DH, Stephenson JB, editors. NCKRI Symposium 2 Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, published online by NCKRI, Carlsbad, NM, p. 285-292. Alexander EC, Jr., Green JA, Alexander SC, Spong RC. 1996. Spring sheds, Plate 9. In: Lively RS, Balaban NH, editors. Geological Atlas of Fillmore County, Minnesota: Part B, County Atlas Series published by the Minnesota Department of Natural Resources. Alexander EC, Jr., Lively RS. 1995. Karstaquifers, caves and sinkholes. In: Lively RS, Balaban NH, editors. Text supplement to the Geologic Atlas, Fillmore County,Minnesota: Minnesota Geological Survey County Atlas Series C-8, Part C, p. 10-18. Anderson JR, Runkel AC, Tipping RG, Barr K, Alexander EC, Jr. 2011. Hydrostratigraphy of a fractured urban aquitard: Geological Society of America Field Guides 24, pp 457-475. Bauer EJ, Chandler VW. 2014. Data-base map: plate 1, Geologic Atlas of Houston County, Minnesota, Minnesota Geological Survey County Atlas C-33, 4 pls. scale 1:100,000. Davis SN, Whittemore DO, Fabryka-Martin J. 1998. Uses of chloride/bromide ratios in studies of potable water: Ground Water 36 (2): 338-350. Green JA, Alexander EC, Jr., Marken WG, Alexander SC. 2002. Karst hydrogeomorphic units: pl. 10 of Falteisek J, editor, Geologic atlas of Mower County, Minnesota, Minnesota Department of Natural Resources, Division of Waters County Atlas C-11, pt. B, scale 1:100,000. Green JA, Mossler JH, Alexander SC, Alexander EC, Jr. 1997. Karst hydrogeology of Le Roy Township: Mower County, Minnesota, Minnesota Geological Survey Open File Report 97-2. 2 Plates, Scale 1:24,000. Green JA, Runkel AC, Alexander EC, Jr. 2012. Karst conduit flow in the Cambrian St Lawrence Formation, southeast Minnesota, USA. Carbonates and Evaporites 27 (2): 167-172. Green J, Luhmann A, Peters A, Runkel A, Alexander EC, Jr, Alexander S. 2008. Dye tracing within the St Lawrence confining unit in southeastern Minnesota. In: Yuhr L, Alexander EC, Jr, Beck B, editors. Sinkholes and the Engineering and Environmental Impacts of Karst, American Society of Civil Engineers, Proceedings GSP 183, p 477-484. Kempe S, Halliday WR. 1997. Report on the discussion on pseudokarast, in Proceedings of the 12th International Congress of Speleology, v. 6, that the St. Lawrence and the Lone Rock have attributes of both an aquifer and an aquitard dependent upon depth of burial and landscape setting. Dye traces have shown that conduit networks throughout the St. Lawrence in in cised valley settings are connected to conduit networks in the underlying Lone Rock Formation and that these cal hydraulic conductivity within the relatively low per meability rock matrix. The conduit networks investigated here allow for rapid subsurface drainage but do not exhibit the behavior of solutionally enlarged networks common in karst (rapid rising and losing limbs on breakthrough curves). Instead, breakthrough curves for traces in the St. Lawrence and Lone Rock formations had very long losing limbs that last from months to years. This demonstrates the lack of large scale karst conduit networks and supports recent work suggesting that these features are mechanical in origin and formed very early on in the rocks history. H counties will be furthered through the Department of Natural Resources portion of the County Atlas Program. Roughly 100 samples from wells and 20 from springs in anions, trace metals, tritium and stable isotopes. These data will allow us to better resolve the interaction of recent and older regional groundwater, especially in the incised valleys of these counties. Acknowledgments The work presented in this report could not have occurred without the permission of landowners who graciously allowed access to their property. This effort was conducted as part of the Innovative Springshed Mapping for Trout Stream Management-Phase II as funded by the Minnesota Environment and Natural Resources Trust Commission on Minnesota Resources (LCCMR). Calvin Alexander, Jr. of the University of Minnesota Earth Sciences Department performed sample analysis and interpretation. Special thanks are given to Holly Johnson for her graphic editing assistance. References Alexander EC, Jr., Runkel AC, Tipping RG, Green JA. 2013. Deep time origins of sinkhole collapse failures in sewage lagoons in SE Minnesota In:

PAGE 11

41 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 p., 1 map in pocket. Steenberg JR. 2014a. Bedrock Geology, plate 2, Geologic Atlas of Houston County, Minnesota, Minnesota Geological Survey County Atlas C-33, 4 pls., scale 1:100,000. Steenberg JR. 2014b. Bedrock Geology, plate 2, Geologic Atlas of Winona County, Minnesota, Minnesota Geological Survey County Atlas C-34, 4 pls., scale 1:100,000. Stewart ZW, Cowan CA, Runkel AC. 2012. Sediment deformation in the Jordan Sandstone and influences on modern hydrogeology: Geological Society of America Abstracts with Programs 44 (7): 556. Wilon JT. 2012. Water-quality assessment of the Cambrian-Ordovician aquifer system in the northern Midwest, United States: U.S. Geological Survey Scientific Investigations Report 2011, 154 p. Worthington SRH. 2003. A comprehensive strategy for understanding flow in carbonate aquifers. / Speleogenesis and Evolution of Karst Aquifers 1 (1), www.speleogenesis.net. 8 pages, re-published from: Palmer, AN, Palmer, MV, Sasowsky, ID. (eds.), Karst Modeling: Special Publication 5, The Karst Waters Institute, Charles Town, West Virginia (USA), 30-37. Zhang H, Kanivetsky R. 1996. Bedrock hydrogeology, plate 6, Geologic Atlas of Fillmore County, Minnesota, Minnesota Department of Natural Resources County Atlas C-8, 4 pls. scale 1:100,000. Klimchouk AB, Ford DC. 2000. Types of karst and evolution of hydrogeologic setting. In: Klimchouk AB, Ford DC, Palmer AN, Dreybrodt W, editors. Speleogenesis evolution of karst aquifers: Huntsville, Ala., National Speleological Society, p. 45-53. Lusardi BA, Adams RS, Hobbs HC. 2014. Surficial Geology, plate 3, Geologic Atlas of Houston County, Minnesota, Minnesota Geological Survey County Atlas C-33, 4 pls. scale 1:100,000. Minnesota Department of Health. 1998. Guidance for mapping nitrate in Minnesota groundwater, 20 p. Minnesota Department of Health. 2011. Rules HandbookA Guide to the Rules Relating to Wells and Borings, Saint Paul, MN, accessed January 21, 2015 at http://www.health.state.mn.us/divs/eh/ wells/ruleshandbook/ruleshandbook.pdf for Minnesota: Minnesota Geological Survey Report of Investigations 65, 76 p., 1 pl. Panno SV, Hackley KC, Hwang HH, Greenberg SE, Krapac IG, Landsberger S, OKelly DJ. 2006. Sources in Ground Water. Ground Water 44 (2): 176-187. Petersen TA. 2005. Hydrogeologic cross sections, plate 9, Geologic Atlas of Wabasha County, Minnesota, Minnesota Department of Natural Resources County Atlas C-14, 3 pls. scale 1:100,000. Runkel AC, Cowan CA, Stewart ZW, Jacobson WZ. 2015. Origin of Hydraulically Significant Bed Parallel Macropore Networks in Siliciclastic Bedrock: Geological Society of America Abstracts with Programs 47 (6): 36. Runkel AC, Tipping RR, Green JA, Jones PM, Meyer OFR14-04, Hydrogeologic Properties of the St. Lawrence Aquitard, Southeastern Minnesota. Minnesota Geological Survey. 2013. Physical hydrogeology of the groundwatersurface water system of southeastern Minnesota and geologic controls on nitrate transport and stream baseflow concentrations: Minnesota Geological Survey report delivered to the Minnesota Pollution control agency, Contract number B50858 (PRJ07522). Runkel AC, Tipping RG, Alexander EC, Jr., Green JA, Mossler JH, Alexander SC. 2003. Hydrogeology Minnesota. MGS Report of Investigations 61, 105

PAGE 12

42 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

211 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 CREATION OF A MAP OF PALEOZOIC BEDROCK SPRINGSHEDS IN SOUTHEAST MINNESOTA Jeffrey A. Green Minnesota Department of Natural Resources, Division of Ecological & Water Resources, 3555 9th St. NW Suite 350, Rochester, MN, U.S.A., 55904, jeff.green@state.mn.us E. Calvin Alexander, Jr. Morse-Alumni Professor Emeritus, Earth Sciences Department, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, alexa001@umn.edu hundreds-of-meters to kilometers-per-day range in all of the bedrock aquifers tested. The width and duration of tails of breakthrough curves in these conduit flow systems vary with the bedrock aquifers. The Galena Group has Full Widths at Half Maximums (FWHMs) of a few hours and tails that are down to background in a few days. The Prairie du Chien Group also has FWHMs of hours but has tails that continue for weeks. The St. Lawrence and Lone Rock Formations have FWHMs of months to years. Introduction Springs are natural discharge points for groundwater systems. They provide baseflow for streams and are critical sources of cold, relatively constant-temperature water for trout streams. In southeast Minnesota (Figure 1) springs are commonly found emerging from the deeply incise the water-bearing bedrock layers. In this region, many springs flow from carbonate (limestone, dolostone) and carbonate-cemented siliciclastic layers. These formations dissolve in slightly acidic groundwater and have developed integrated systems of conduits that allow groundwater to flow quickly through the aquifers Abstract Springs are groundwater discharge points that serve as vital coldwater sources for streams in southeast Minnesota. The springs generally emanate from Use of systematic dye tracing began in the 1970s and continues through the present as a standard method for investigating karst hydrology and to map springsheds. The work was accelerated in 2007 because of increased funding from the State of Minnesotas Environment and Natural Resources Trust Fund. A compilation springshed map of dye traces conducted over the last several decades has been assembled for the region. In southeast Minnesota, the springs are the outlets of conduit flow systems in both carbonate and siliciclastic bedrock aquifers. Conduit flow dominates groundwater transport in carbonate aquifers and is an important component of groundwater flow in siliciclastic aquifers. Conduit flow in aquifers occurs independently of the presence or absence of surface karst features. The springsheds of these springs have three interacting components: Groundwater Springsheds (analogous to classic karst autogenic recharge areas), Surface Water Springsheds (analogous to classic karst allogenic surface runoff areas), and Regional Groundwater Springsheds. Surface Water Springsheds can be up to several orders-ofmagnitude larger than the Groundwater Springsheds to which they contribute water. Surface Water Springsheds can feed surface flow into one or several stream sinks. Those multiple stream sinks may be in one or more Groundwater Springsheds. The leading edges of dye tracing breakthrough curves typically show groundwater flow velocities in the Figure 1. Location map of the state of Minnesota. The study area is indicated by black shading in the inset.

PAGE 2

212 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE karst of Fillmore County, but there were caveats. Within the county, there are allogenic areas where perennial surface streams flow and then sink to become sinking streams, but they are either: 1) perched on patches of relatively impermeable glacial deposits on top of the 3) in areas where the first water table is at or near the surface. There are autogenic areas where surface water events or spring snow melts, and allogenic flows that sink in sinking streams or sinkholes. But in these areas most of the recharge was found to be through the thin soils rather than into sinkholes or sinking streams (e.g., Hallberg et al., 1984). Large areas exist over shallow carbonate bedrock with no perennial surface water flow, but where there are few if any visible surface areas prompted innumerable, incredibly unproductive debates and negotiations over how many sinkholes are necessary in a given area to define it as a karst. east and north of Fillmore County in stratigraphically lower carbonates, sandstones and siliciclastic rocks, the allogenic and autogenic models became increasingly aquifers extend westward far beyond the extent of the allogenic and autogenic basins. Regional groundwater from these areas was shown to contain increasingly larger components of flow from the stratigraphically lower springs. In Green and Alexander (2014) and Green et al. (2014) our conceptual model is described to include three sources of spring flow, i.e., three (at least conceptually) mappable components to each springshed. These three components are shown diagrammatically in Figure 2. Groundwater Springsheds (GwS): Areas where precipitation infiltrates from the land surface to the first conduit-flow dominated aquifer and flows to the spring. GwSs are mapped mainly based on groundwater flow connections demonstrated by tracing techniques. These flow velocities are much faster than typically expected based on porous media models. Our GwS can include classic autogenic basins in addition to areas that dont fit various definitions of autogenic basins. and emerge in the springs. Many other springs emerge from siliciclastic units which contain no significant carbonates. Springs emanating from siliciclastic bedrock aquifers share some of the key hydrologic properties while not exhibiting all of the characteristics of traditional carbonate karst. In order to conserve and protect springs and the surface water bodies they supply, it is necessary to understand the geologic setting and from where the water is derived. The University of Minnesota (U of M), the Minnesota Department of Natural Resources (DNR), and a group of experienced local cavers have been mapping springsheds in southeast Minnesota since the 1970s. The working hypothesis of these groups is that dye traces provide fundamentally valuable scientific information about how groundwater moves in the subsurface. This can ultimately provide a better understanding of these systems, which is needed to formulate public policies and guide professional decisions. Funding from the State of Minnesota Environment and Natural Resources Trust Fund (ENRTF) has allowed these researchers to by releasing fluorescent organic dyes into sinkholes or sinking streams to determine the general flow paths to springs. This information was assembled to produce a Southeast Minnesota (Green and Alexander, 2014). This map displays the springsheds which were mapped and verified as of June 2014 at a 1:150,000 scale. It also incorporates work done as part of the Fillmore County Geologic Atlas (Alexander et al., 1996). Green et al. (2014) presented the methodology used to define springsheds shown on the map. This paper discusses some of the results of that effort. A springshed can be defined as those areas within groundand surface-water basins that contribute to the discharge of a spring (Copeland, 2003). Dye tracing is the preferred method for documenting karst springsheds. Kingston (1943) reported the earliest known dye traces in Minnesota karst, which were conducted in 1941 in response to a typhoid fever outbreak in 1940 in Harmony, Fillmore County, Minnesota. When coauthor E. Calvin Alexander, Jr. began tracer studies in Fillmore County in the late 1970s, the theoretical model was that of the classic autogenic and allogenic recharge components to karst water systems. This model was reasonable in the well-developed Galena

PAGE 3

213 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 The boundaries of these three types of springsheds do not necessarily correspond to those of overlying surface watersheds. In addition, the groundwater springshed boundaries are dynamic, changing as groundwater levels rise and fall, in response to wetter and drier cycles of precipitation. Hydrostratigraphic Background southeast Minnesota is Cambrian to Devonian in age (Mossler, 2008) and was deposited as a series of shallow, fluctuating seas advanced and retreated across what is now southeast Minnesota (Figure 3). The Cambrianage rocks are primarily sandstone, siltstone and shale while the Ordovician and Devonian rocks are primarily carbonate (limestone and dolostone) and shale. Bedrock layers that serve as the source of water for the many springs found in southeast Minnesota are displayed on the stratigraphic column (Figure 3), The Decorah Shale (dark blue-gray in Figure 4 and 5) is a regional aquitard and is the boundary between the Galena limestone karst system and the Prairie du Chien karst system (Runkel et al., 2014). Many springs discharge Surface Water Springsheds (SWS): Areas where water flows across the land surface in perennial streams (streams that typically flow all year, perhaps except under unusually dry conditions) before sinking into the subsurface at the edges of GwSs. These areas include, but are not limited to, classic allogenic basins and are mappable using conventional surface topography maps and digital elevation models coupled with field work to identify sinking streams. Regional Groundwater Springsheds (RGS): Areas contributing to a springs flow that are beyond or underneath the GwSs and the SWSs. Given Minnesota, the RGSs can extend many kilometers to flow to a spring. Conceptually they can be mapped from sufficiently detailed regional potentiometric maps, but such maps rarely exist in this region. The volumetric contribution of flow to a spring from from geochemical and isotopic water quality data, even if the geographic extent of the RGS is unknown (e.g., Runkel et al., 2014). Figure 2. Springshed block diagram. A surface water springshed is the watershed of a perennial or intermittent stream flowing across the land surface that descends into stream sinks or sinkholes. Those sinking points mark the beginning of a groundwater springshed carrying groundwater flow to a downgradient spring. The black lines represent the general direction of groundwater flow through enlarged openings in the subsurface, such as vertical and horizontal joints and fractures. The blue lines depict flow from sinkholes and stream sinks. The gray-shaded linear features along two of the blue lines represent larger conduits carrying groundwater flow. Regional groundwater springshed flow (shown as regional flow in the diagram) is a third component of the flow of a spring. Water enters the groundwater system by way of continued infiltration from the surface and can come from areas under or far beyond the surface water springsheds. Regional sources of water to a spring have a significantly longer underground residence time and can provide a large fraction of spring discharge.

PAGE 4

214 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE from the lower Galena because the water cannot move through the relatively impervious, underlying Decorah shale. The St. Lawrence Formation (green in Figure 6) is an important regional bedrock unit from which many springs discharge and where disappearing streams (stream sinks) contribute water to springs (Green et al., These rock units are covered by varying thicknesses of regolith (mixtures of tills, outwash deposits and loess). Areas with less than 15 meters of regolith have been the focus of springshed mapping efforts because sinkholes and disappearing streams are primarily found in those areas. Surface karst features are not common where the regolith is thicker than 15 meters (brown areas in Figures 4, 5, and 6). The areas where the thickness of regolith is greater than 15 meters are typically in three settings: isolated areas in the blufflands regions, alluvium deposits west. The hydrostratigraphy of southeastern Minnesota is further complicated by the extensive development of plane parallel fractures, and subtle structural features (Runkel et al., 2014). In addition, Alexander et al. (2013) subaerial unconformities in this section have played in the development of the extensive, pervasive secondary porosity and permeability in these rocks. Methods The GwSs were defined primarily from sinking stream or sinkhole to spring connections established by tracing techniques. Tracing technology and techniques have evolved significantly from the late 1970s to the present. Tracing agents have included cations, anions, conductivity, environmental stable and radioactive heat, flow pulses and sediments. The bulk of the tracing work, however, has used fluorescent dyes as the tracers. Several water tracing fluorescent dyes have been used, but our standard dyes were Rhodamine WT, sulforhodamine B, eosin and fluorescein/Uranine. Initial fluorescent dye traces in the 1970s were conducted using visual dye detection. In the 1980s and early 1990s, Turner Designs Model 10-005 Field Filter Figure 3. Stratigraphic column. This stratigraphic column shows the bedrock of southeast Minnesota, highlighting lithostratigraphic attributes (crystalline basement rock shown in hachured grey). The carbonate bedrock has been partially dissolved to form karst aquifers. The coarse clastic bedrock stores and transmits large volumes of water in the primary pore space in the sandstone matrix. The Decorah formation has a low primary permeability that does not allow water to infiltrate readily. Therefore, water is forced to flow laterally to numerous springs at or near its upper contact. The St. Lawrence formation has enhanced secondary porosity and permeability in the form of bedding plane fractures and conduits, which is a critical factor for spring occurrence. Adapted from Mossler (2008).

PAGE 5

215 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Scanning Spectrofluorophotometer became our standard dye analysis instrument and that continues to be used at the present (Alexander, 2005). The dye traces have used a mixture of sample collection methods, such as integrating charcoal detectors (bugs) and direct water sampling was the availability of auto-sampling devices for timed direct water sampling, starting in the 1980s. Both direct water sample and bug trace techniques are still used. Multiple dyes have been used since the 1990s to trace from different sinkholes or stream sinks at the same time. The dyes were introduced into the groundwater systems through sinking streams, snow melt running into sinkholes, and dry sinkholes. In the latter case, the dyes were flushed with water from a tanker truck (typically 1800 liters). Successful dye traces were conducted in the Lithograph City Formation, Spillville Formation, Galena Group, Prairie du Chien Group, St. Lawrence Formation, and the Lone Rock Formation (Figure 3). The SWSs were mapped topographically where they contribute surface runoff to a mapped stream sink. The upstream boundaries of surface water basins were identified using watershed maps derived from digital elevation models (DEM) created with Light Detection and Ranging (LiDAR) data or topographic maps (Minnesota Department of Natural Resources, 2014). The catchment area upstream of the stream sink was identified using an ArcGIS tool that selects the surface watersheds upstream of a given point. Results The Green et al. (2014) map displays GwSs and SWSs identified across southeast Minnesota. The GwS boundaries are outlined in red and record conditions at the time that the dye traces were conducted. SWSs are outlined in yellow. Common boundaries of neighboring springsheds represent surfacewater or groundwater divides. Sinkholes or stream sinks that were used as dye-trace input points are dye was later detected in the springs that were being monitored. The dye-trace vectors (black arrows) are the diagrammatic depiction of the groundwater flow routes. Galena Group Springsheds Figure 4 is reproduced from Inset 2, an enlarged portion of the Green and Alexander (2014) map showing GwSs and SWSs in the well-developed Galena Group karst of west central Fillmore County. More dye traces have been conducted in the Galena than in all of the other bedrock units combined. Mohring and Alexander (1986) additional traces were included in the Alexander et al.(1996) map which was a precursor of the Green and Alexander (2014) map. Luhmann et al. (2012) report multi-tracer experiments in the area shown in Figure 4. Many, but not all, of the flow vectors in Figure 4 enter the subsurface at sinking streams or sinkholes in the Dubuque and Stewartville Formations and their spring resurgences are in the lower Cummingsville Formation above the Decorah Shale. The Stewartville and Prosser formations of the Galena Group exhibit the densest sinkhole development in Minnesota. Groundwater flow through these units is and fractures. Groundwater flow paths from sinkholes and stream sinks to springs in the Galena Group often cross surface watershed boundaries. The low permeability of the rock matrix generally provides negligible flow. Multiple dye-tracing investigations have demonstrated breakthrough travel velocities of 1.6 to 4.8 km per day (Green et al., 2005). Tracer breakthrough curves through the Galena Group are only slightly asymmetric with small, exponentially decreasing tails. The breakthrough curves typically have Full Widths at Half Maximums (FWHM=the time duration of the breakthrough curve measured between the points on the concentration axis which are at half of often back to background levels in a few days. Dye traces from sinkholes and stream sinks on the boundaries of multiple groundwater springsheds have demonstrated that those boundaries are dynamic: groundwater flow directions can change in response to ranging from several hundred hectares to many square kilometers (Green et al., 2005), have thin glacial cover over bedrock, and commonly have sinkholes. Stream sinks and sinkholes serve as direct recharge points for surface water to enter the limestone aquifer and are good

PAGE 6

216 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Figure 4. Map of springsheds in the Galena Group limestone karst. (Inset 2 from Green and Alexander, 2014). The inferred groundwater flow route vectors are diagrammatic depictions of the connections discovered by dye tracing from a dye introduction point (sinkhole or stream sink) to a spring. The dye traces are used to delineate the boundaries of Groundwater Springsheds. While most of the sinkholes and stream sinks transmit water to a single spring, there are several that connect to multiple springs in different directions. These traces mark the boundaries between Groundwater Springsheds under the hydrologic conditions at the time of the dye trace.

PAGE 7

217 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 indicators of conduit flow in the subsurface. Recharge water also infiltrates through the sediment cover and into the carbonate bedrock. The landscapes between the sinkholes or stream sinks and the springs to which they are connected may lack surface-karst features but subsurface. Prairie du Chien Group Springsheds Figure 5 is an enlarged portion of the Green and Alexander (2014) map showing the Prairie du Chien Group GwS and SWSs of the West Branch of Duschee Creek Valley in east central Fillmore County, Minnesota. The Prairie du Chien Group is stratigraphically lower than the Galena Group, and is the first bedrock in a broad band stretching east and north of the Galena Group subcrop. It extends from the Iowa border northwestward and underlies much of the Twin Cities Metropolitan Area at a relatively shallow depth. The stratigraphically intervening Platteville, Glenwood and St. Peter Formations tend to be present as the first bedrock in only narrow bands which are not extensive at the surface. Typically they are primarily seen along incised stream valleys where a portion of the Galena Group remains as a cap rock. much of southeast Minnesota. The visible surface karst features are sparse and are typically further apart than Chien subcrop have little if any perennial surface water flow and groundwater contamination issues are widely present. Alexander et al. (2013) describe the karst conditions that led to the catastrophic sinkhole collapse of three Municipal Waste Water Treatment Lagoons on the Prairie du Chien. The Duschee Creek Valley area in Figure 5 is of interest because the two large springs at the north end of the valley are the water sources of the Lanesboro State Fish Hatchery, which produces brown and rainbow trout for stocking programs in Minnesota. Thus, these springs provide the water source for an enormously important statewide economic resource. Dye traces from a large sinkhole 3.4 km south southeast of the Hatchery Spring demonstrated rapid breakthroughs (flow velocity of several km/day) at the Hatchery Springs with FWHMs of hours to a day. But the tails of the breakthrough curves extended for weeks (Wheeler, 1983). This pattern is one of four flow patterns we have observed in traces in the Prairie du Chien Group and is the most common pattern in this hydrologic system. In 1991, co-author Jeff Green located a critical stream sink in the West Branch of Duschee Creek Valley, 3.7 km south of the Hatchery Spring. Dye traces in 1991 and of storm water contaminants to the Hatchery Springs. An engineered sealing of that stream sink and rerouting of the stream away from the stream sink eliminated the storm water contamination at the Hatchery Spring (Kalmes and Mohring, 1995). A second pattern was observed in a 1984 well-to-well trace in the Shakopee Formation of the Prairie du Chien Group (Alexander and Milske, 1986). A breakthrough flow velocity of approximately 2 km/year was observed with a broad, smooth peak persisting about two months, followed a second two-month smaller peak. A third pattern was observed in dye traces conducted at the Oronoco Landfill in Olmsted County (Alexander et al., 1991). We observed an initial rapid breakthrough spike at monitoring wells, followed by an exponential drop off with a half-life of a few days. The initial spike was followed by a series of precipitation-driven, decreasing spikes over the next couple of years. The fourth pattern is no pattern. About 15 percent of the dye traces attempted in the Prairie du Chien were never detected at any monitoring point (Green and Alexander, 2011). St. Lawrence and Lone Rock Formation Springsheds Figure 6 is reproduced from Inset 4, an enlarged portion of the Green and Alexander (2014) map showing the Fillmore and western Houston Counties, Minnesota. Groundwater flow through the St. Lawrence is through conduits that include modified bedding-parallel fractures, nonsystematic vertical fractures, and through the bedrock matrix (Runkel et al., 2003). Dye tracing has shown that, while not exhibiting all of the characteristics of traditional carbonate karst, the St. Lawrence Formation has a karst-like conduit flow component. Multiple dye-

PAGE 8

218 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Figure 5. Map of Springsheds in the Prairie du Chien Group carbonate karst. An enlarged section of Green and Alexander (2014). This figure shows the GwS and SWSs of the two large springs that are the water sources of the Lanesboro State Fish Hatchery. The two southern dye input points (a sinkhole and a stream sink) have been shown by dye tracing to be connected to both springs. The northernmost dye point is a well that had dye introduced below the surface in the annulus prior to grouting. That dye was only detected at the west spring.

PAGE 9

219 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 of magnitude faster than velocities expected for aquifers dominated by porous media flow, and many orders of magnitude faster than expected flow velocities of rocks mapped as aquitards. The breakthrough curves of the St. Lawrence traces rise rapidly to a peak and then fall with FWHMs of months. The dye has been consistently detected at springs for one to three years after it was tracing investigations in the St. Lawrence Formation have demonstrated breakthrough travel velocities of 90 to 600 meters per day (Green et al., 2008, 2012,). These are about an order of magnitude lower than the observed breakthrough travel velocities in the Galena limestone, the best-developed carbonate karst in Minnesota (Green et al., 2005). These velocities, however, are a few orders Figure 6. Map of Springsheds in the St. Lawrence and Lone Rock pseudokarst, (Inset 4 from Green and Alexander, 2014). The vertical dashed line is the Fillmore County/Houston County boundary. The streams in the Surface Water Springsheds are perennial that sink into stream sinks in the St. Lawrence Formation. Dye traces from these stream sinks have been detected at St. Lawrence and Lone Rock Formation springs. The Groundwater Springsheds boundaries are based on the dye traces and on field observations that indicate the hillsides along the valleys contribute water to the flow system.

PAGE 10

220 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE GwS boundaries are extended into the upland interfluves between the stream sinks and springs. The lack of large conduit networks in outcrop and borehole observations makes classifying the St. Lawrence and Lone Rock as karst problematic. Flow curves for these units is more consistent with their being classified as pseudokarst. Pseudokarst has been described as landscapes with morphologies resembling karst, and/or may have a predominance of subsurface drainage through conduit type voids, but lack the element of long-term evolution by solution and physical erosion (Kempe and Halliday, 1997). Recent work in the siliciclastic bedrock of southeast bedding-parallel apertures are connected through an anastomosing network of apertures, clustered along discrete (<2m) stratigraphic intervals and are found at depths exceeding 200 meters (Runkel et al., 2015). The apertures are commonly associated with distorted bedding which is interpreted to reflect dewatering features that occurred shortly after burial when the rock was only partially lithified. These voids, then, are unlikely to primarily be the result of solution as karst has been traditionally defined Summary Springs are the discharge points for groundwater systems. Dye tracing to delineate springsheds has been an ongoing activity since the 1970s in southeast Minnesota. This tracing work has been compiled into a new map Southeast Minnesota (Green and Alexander, 2014). This map displays both delineated groundwater and surface water springsheds in carbonate karst bedrock formations, and also in pseudokarst bedrock formations. The map is available on the Minnesota Department of Natural Resources website. These dye traces collectively demonstrate that in southeast Minnesota: 1. Springs are the outlets of conduit flow systems. 2. Conduit flow dominates groundwater transport in these carbonate and siliciclastic bed rock aquifers independent of the presence or absence of surface karst features. released into a stream sink (Green et al., 2012). The St. Lawrence has few mapped sinkholes (one to three per county) and no known caves. Its primary features are stream sinks and springs. Surface-water springsheds that are thousands of hectares may drain into one set of stream sinks in a valley. Streams commonly sink into the upper St. Lawrence in valleys but the terminal sinking point often moves up and down the valley depending on stream stage. In these settings, the streams are a series of pools and riffles with the pools functioning as the stream sinks. In locations where the streams flow along bedrock exposures, the stream sinks are typically discrete points. There are also streams that lose flow but do not totally disappear as they cross the upper St. Lawrence. Dye tracing has demonstrated that most of the groundwater springsheds align with surface topography but there are examples where dye traces have crossed surface watershed divides. Multiple springs may be connected to single sinking points of streams in valleys. The Lone Rock Formation, a fine siliciclastic unit, is the bedrock unit below the St. Lawrence. The Lone Rock is mostly composed of a fine-grained sandstone and siltstone with interbedded shale and dolostone (Mossler, 2008). Dye traces from St. Lawrence stream sinks at three sites in southeast Minnesota have been detected at Lone Rock springs (Barry et al., 2015). These traces have demonstrated that the Lone Rock also has a karstlike conduit flow component. Observed breakthrough travel velocities were between 21-214 meters/day using passive charcoal detectors and 314 meters/day using over one year later at springs monitored from one of the tracing sites (Barry et al., 2015). In Figure 6, the GwSs includes the area between the stream sink where dye was introduced into St. Lawrence, and the St. Lawrence or Lone Rock springs where it was detected. Our field observations and spring flow measurements of St. Lawrence and Lone Rock springs have shown that these springs respond rapidly (less than 24 hours) to large precipitation events, although they have no visible changes et al., 2011). We have interpreted these observations to mean that water is infiltrating into the hillsides above bedrock units which increases spring discharge while not impacting spring turbidity. To account for these phenomena and the strong regional groundwater flow component, the

PAGE 11

221 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Series C-8, Minnesota Department of Natural Resources, St. Paul, MN. http://www.dnr.state. mn.us/waters/programs/gw_section/mapping/ platesum/ fi llcga.html Alexander EC Jr, Huberty BJ, Anderson KJ. 1991. Olmsted County Dye Trace Investigation Final Report. Report to Olmsted County prepared by Donohue and Associates, February 1991, 7 loose leaf binders. Alexander EC Jr., Milske JA. 1986. Dye tracing studies of the Fountain Minnesota Sewage System. Proceedings of the Environmental Problems in Karst Terranes and Their Solutions Conference, National Water Well Associations, Dublin, Ohio, p. 249-262. Alexander EC Jr, Runkel AC, Tipping RG, Green JA. 2013. Deep time origins of sinkhole collapse failures in sewage lagoons in SE Minnesota. In: Land L, Doctor DH. Stephenson JB, (editors) NCKRI Symposium 2 Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, published on-line by NCKRI, Carlsbad, NM, p. 285-292. ISBN 9780-9795422-7-5 Alexander SC. 2005. Spectral deconvolution and quantification of natural organic material and fluorescent tracer dyes. In: Beck B, (editor), Sinkholes and the Engineering and Environmental Impacts of Karst: Proceedings of the Tenth Multidisciplinary Conference, San Antonio, 24-28 September 2005, ASCE Geotechnical Special Publication 144, Amer. Soc. Civil Engineers, Reston, VA, p. 441-448. http://dx.doi. org/10.1061/40796(177)47 Barry JD, Green JA, Steenberg JR. 2015. Conduit flow in the Cambrian Lone Rock Formation, Southeast Minnesota, U.S.A. These proceedings, 12 p. Copeland, RE. 2003. Florida spring classification system and spring glossary. Florida Geological Survey: Special Publication 52, 17 p. Green JA, Alexander EC Jr. 2011. Dye tracing observations from the Prairie du Chien Group in Minnesota: 2011 Geological Society of America meeting, Abstracts with Programs, Abstract 60-11 43 (5): 167. Karst Springsheds in Southeast Minnesota. Minn. Dept. Natural Res., St. Paul, MN. Pdf available from: http:// fi les.dnr.state.mn.us/waters/ groundwater_section/mapping/springshed/ springshed_map.pdf Green JA, Alexander SC, Alexander EC Jr. 2005. Springshed mapping in support of watershed management. In: Beck B, (editor). Sinkholes and the Engineering and Environmental Impacts of Karst: Proceedings of the Tenth Multidisciplinary 3. The springsheds of these springs have three components: Groundwater springsheds (analogous to classic karst autogenic recharge areas), surface water springsheds (analogous to classic karst allogenic surface runoff areas) and regional groundwater springsheds. 4. Surface water springsheds can be up to several orders-of-magnitude larger than the groundwater springsheds to which they contribute water. A surface water springshed can feed surface flow into one or several stream sinks. Those multiple stream sinks may be in one or more groundwater springsheds. 5. Breakthrough curves typically show groundwater flow velocities in the hundred-of-meters to kilometers per day range in all of the bedrock aquifers tested. 6. The widths and tails of breakthrough curves in these conduit flow systems vary with the bedrock aquifers. The breakthrough curves of Galena Group have FWHMs of a few hours and tails that are down to background in a few days. Breakthrough curves of the Prairie du Chien Group in the Duschee Creek Springshed also have FWHMs of hours but have tails that continue for weeks. St. Lawrence and Lone Rock Formations breakthrough curves have FWHMs of months to years. Acknowledgements The authors gratefully acknowledge very constructive critical reviews of this paper by Chris Smart and two anonymous reviewers. We also acknowledge and thank hundreds of colleagues, co-workers, students, friends and land owners who have participated in the dye traces over the decades. Minnesota Environment and Natural Resources Trust Fund on Minnesota Resources (LCCMR) and its predecessors. Special thanks are given to Holly Johnson of the DNR for her graphic editing assistance, Ruth MacDonald of the DNR, and Betty Wheeler of the University of Minnesota for their thorough editing of the manuscript. References Alexander EC Jr, Green JA, Alexander SC, Spong RC. 1996. Springsheds, Plate 9, Geologic Atlas of Fillmore County, Minnesota: Part B, County Atlas

PAGE 12

222 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE 6584.2010.00737.x Minnesota Department of Natural Resources, 2014, National Hydrography Dataset (NHD) HighResolution Minnesota: last accessed August, 2014 at https://gisdata.mn.gov/dataset/waternational-hydrography-data Mohring E, Alexander E. 1986 Quantitative tracing of karst groundwater flow: Southeastern Minnesota, North Central, U.S.A. Proceedings of the 5th International Symposium on Underground Water Tracing, Athens, Instit. of Geol. and Mineral Expl., Athens, Greece, p. 215-227. for Minnesota: Minnesota Geological Survey Report of Investigations 65: 76 p., 1 pl. http://purl. umn.edu/58940 Runkel AC, Cowan CA, Stewart ZW, Jacobson WZ. 2015, A Possible Origin of Hydraulically Significant Bed Parallel Macropore Networks in Siliciclastic Bedrock: Geological Society of America Program 47 (5): 55. http:// www.geosociety.org/Sections/nc/2015mtg/ documents/2015_NC_program.pdf 2014. Geologic controls on groundwater and surface water flow in southeastern Minnesota and its impact on nitrate concentrations in streams: Minnesota Geological Survey Open File Report 14-02: 154p. http://conservancy.umn.edu/ handle/11299/162612 Runkel AC, Tipping RG, Alexander EC Jr, Green JA, Mossler JH, Alexander SC. 2003. Hydrogeology Minnesota: Minnesota Geological Survey, Report of Investigations 61: 105 p., 1 map in pocket. http://purl.umn.edu/58813 Stewart ZW, Cowan CA, Runkel AC. 2012. Sediment deformation in the Jordan Sandstone and influences on modern hydrogeology: Geological Society of America Abstracts with Programs. 44 (7): 556. https://gsa.confex.com/gsa/2012AM/ fi nalprogram/abstract_209806.htm Wheeler BJ. 1983. Groundwater Tracing in the Duschee Creek Karst Basin in Southeast Minnesota. Plan B MSc Paper, Geology & Geophysics Dept., Univ. of Minn., Minneapolis. 74 p. Conference, San Antonio, 24-28 September 2005, ASCE Geotechnical Special Publication 144, Amer. Soc. Civil Engineers, Reston, VA, 403-409. Green JA, Barry JD, Alexander, EC Jr. 2014. Bedrock Springs of Southeastern Minnesota. on Minnesota Resources. Minn. Dept. Natural Resources, St. Paul, MN, 48 p. Available from: http:// fi les.dnr.state.mn.us/waters/groundwater_ section/mapping/springshed/springshed_ assessment_protocols.pdf Green JA, Luhmann AJ, Peters AJ, Runkel AC, Alexander EC Jr, Alexander SC. 2008. Dye tracing within the St. Lawrence confining unit in southeastern Minnesota In: Yuhr LB, Alexander EC Jr, Beck BF (editors). Sinkholes and the engineering and environmental impacts of karst: American Society of Civil Engineers, Proceedings ASCE GSP 183: 477-484. http://dx.doi. org/10.1061/41003(327)45 Green JA, Runkel AC, Alexander EC Jr. 2012. Karst conduit flow in the Cambrian St. Lawrence confining unit, southeast Minnesota, USA: Carbonates and Evaporites 27 (2): 167-172. http:// dx.doi.org/10.1007/s13146-012-0102-9 Hallberg GR, Libra RD, Bettis EA III, Hoyer BE. 1983. Hydrogeologic and water quality investigations in the Big Spring Basin, Clayton County, Iowa. Iowa Geological Survey, Open-file Report 84-4, 321 p. Kalmes A, Mohring E. 1995. Sinkhole treatment to improve water quality and control erosion in southeastern Minnesota. In: Beck BF, Pearson, the fifth multidisciplinary conference on sinkholes and the engineering and environmental impact of karst, Gatlinburg, TN, 2-5 April 1995: 265-272. Kempe S, Halliday WR. 1997. Report on the discussion on pseudokarst, in Proceedings of the 12th International Congress of Speleology, v. 6, Basel, Kingston SP. 1943. Contamination of water supplies in limestone formation. J. Amer Water Works Assoc 35: 1450-1456. Luhmann AJ, Covington MD, Alexander SC, Chai SY, Comparing conservative and nonconservative tracers in karst and using them to estimate flow path geometry. Journal of Hydrology 448-449: 201-211. Luhmann AJ, Covington MD, Peters AJ, Alexander SC, Anger CT, Green JA, Runkel AC, Alexander EC, Jr, 2011, Classification of thermal patterns at karst springs and cave streams: Groundwater 49 (3): 324-335.


Description
National Cave and Karst Research Institute Symposium 5
Sinkholes and the Engineering and Environmental
Impacts of Karst
PROCEEDINGS OF THE FOURTEENTH MULTIDISCIPLINARY
CONFERENCE
October 5 through 9, 2015
Rochester, Minnesota
EDITORS:
Daniel H. Doctor,
United States Geological Survey Reston, Virginia,
USA
Lewis Land,
National Cave and Karst Research Institute Carlsbad, New
Mexico, USA
J. Brad Stephenson,
CB&I Federal Services Knoxville, Tennessee,
USA
Program
The 14th Sinkhole Conference Program with Abstracts
--
Field Trip Guides
Ordovician Karst of Southeast Minnesota Field Trip
Guidebook/
Calvin Alexander, Jeff Green, Tony Runkel and John
Barry
Living with Karst in Rochester, MN/
Jeffrey S. Broberg
--
PROCEEDINGS
Introduction and Table of Contents
--
INVITED SPEAKER:
Hales Bar and the Pitfalls of Constructing Dams on Karst
(abstract)/
J. David Rogers
--
UPPER MISSISSIPPI VALLEY KARST AQUIFERS:
Karst Hydrogeologic Investigation of Trout Brook/
Joel T. Groten and E. Calvin Alexander, Jr
.
Human Impacts on Water Quality in Coldwater Spring,
Minneapolis, Minnesota/
Sophie M. Kasahara, Scott C. Alexander and E. Calvin
Alexander, Jr
Hydrologic and Geochemical Dynamics of Vadose Zone
Recharge in a Mantled Karst Aquifer: Results of
Monitoring Drip Waters in Mystery Cave, Minnesota/
Daniel H. Doctor, E. Calvin Alexander, Jr., Roy
Jameson and Scott C. Alexander
Conduit Flow in the Cambrian Lone Rock Formation,
Southeast Minnesota, USA/
John D. Barry, Jeffrey A. Green and Julia R.
Steenberg
Using Nitrate, Chloride, Sodium, and Sulfate to Calculate
Groundwater Age/
Kimm Crawford and Terry Lee
Karst Spring Cutoffs, Cave Tiers, and Sinking Stream
Basins Correlated to Fluvial Base Level Decline in
South-Central Indiana/
Garre A. Conner
Driftless Area Karst of Northwestern Illinois and its
Effects on Groundwater Quality/
Samuel V. Panno and Walton R. Kelly
Seeps and Springs at a Platteville "Observatory" on the
River Bluffs (abstract)/
BJ Bonin, Greg Brick and Julia Steenberg
--
KARST HYDROLOGY:
Hydrogeological Dynamic Variability in the Lomme Karst
System (Belgium) as Evidenced by Tracer Tests Results
(KARAG project)/
Amael Poulain, Gaetan Rochez and Vincent
Hallet
Recharge Area of Selected Large Springs in the
Ozarks/
James W. Duley, Cecil Boswell and Jerry
Prewett
Hydrological and Hydrogeological Characteristics of the
Platform Karst (Zemo Imereti Plateau, Georgia)/
Zaza Lezhava, Nana Bolashvili, Kukuri Tsikarishvili,
Lasha Asanidze and Nino Chikhradze
Tracer Studies Conducted Nearly Two Decades Apart
Elucidate Groundwater Movement Through a Karst Aquifer in
the Frederick Valley of Maryland/
Keith A White, Thomas J. Aley, Michael K. Cobb, Ethan
O. Weikel and Shiloh L. Beeman
Dye Tracing Through the Vadose Zone Above Wind Cave,
Custer County, South Dakota/
James Nepstad
A Comparative Study Between the Karst of Hoa Quang, Cao
Bang Province, Vietnam and Tuscumbia, Alabama, USA/
Gheorghe M. Ponta, Nguyen Xuan Nam, Ferenc L Forray,
Florentin Stoiciu, Viorel Badalita, Lenuta J. Enache and
Ioan A. Tudor
Hydrochemical Characteristics and Formation Mechanism of
Groundwater in the Liulin Karst System, Northwestern
China/
Min Yang, Feng'e Zhang, Sheng Zhang, Miying Yin and
Guoqing Wu
--
KARST GEOLOGY:
Karst Paleo-Collapses and Their Impacts on Mining and the
Environment in Northern China/
Gongyu Li and Wanfang Zhou
The Sandstone Karst of Pine County, Minnesota/
Beverley Lynn Shade, E. Calvin Alexander, Jr. and
Scott C. Alexander
A Proposed Hypogenic Origin of Iron Ore Deposits in
Southeast Minnesota Karst/
E. Calvin Alexander, Jr. and Betty J.
Wheeler
Down the Rabbit Hole: Identifying Physical Causes of
Sinkhole Formation in the UK/
Tamsin Green
Relay Ramp Structures and Their In uence on Groundwater
Flow in the Edwards and Trinity Aquifers, Hays and Travis
Counties, Central Texas/
Brian B. Hunt, Brian Smith, Alan Andrews, Douglas
Wierman, Alex S. Broun, and Marcus O. Gary
Goliath's Cave, Minnesota: Epigenic Modification and
Extension of Preexisting Hypogenic Conduits/
E. Calvin Alexander, Jr., Scott C. Alexander, Kelton
D.L. Barr, Andrew J. Luhmann, and Cale T. Anger
--
GIS Databases and Mapping of Karst
Regions
Creation of a Map of Paleozoic Bedrock Springsheds in
Southeast Minnesota/
Jeffrey A. Green and E. Calvin Alexander,
Jr.
Media, Sinkholes and the UK National Karst Database/
Vanessa J. Banks, Helen J. Reeves, Emma K. Ward, Emma
R. Raycraft, Hannah V. Gow, David J. R. Morgan and Donald
G. Cameron
Shallow Depressions in the Florida Coastal Plain: Karst
and Pseudokarst/
Sam B. Upchurch, Thomas M. Scott, Michael C. Al eri
and Thomas L. Dobecki
Sinkhole Vulnerability Mapping: Results from a Pilot
Study in North Central/
Florida Clint Kromhout and Alan E. Baker
A Semi-Automated Tool for Reducing the Creation of False
Closed Depressions from a Filled LiDAR-Derived Digital
Elevation Model/
John Wall, Daniel H. Doctor and Silvia
Terziotti
History and Future of the Minnesota Karst Feature
Database/
Robert G. Tipping, Mathew Rantala, E. Calvin
Alexander, Jr., Yongli Gao and Jeffrey A. Green
Legacy Data in the Minnesota Spring Inventory/
Gregory Brick.
Development of Cavity Probability Map For Abu Dhabi
Municipality Using GIS and Decision Tree Modeling/
Yongli Gao, Raghav Ramanathan, Bulent Hatigoplu, M.
Melih Demirkan, Mazen Elias Adib, Juan J. Gutierrez,
Hesham El Ganainy and Daniel Barton Jr.
Evaluation of Cavity Distribution Using Point-Pattern
Analysis/
Raghav Ramanathan, Yongli Gao, M. Melih Demirkan,
Bulent Hatipoglu, Mazen Elias Adib, Michael Rosenmeier,
Juan J. Gutierrez, and Hesham El Ganainy
A Method of Mapping Sinkhole Susceptibility Using a
Geographic Information System: A Case Study for
Interstates in the Karst Counties of Virginia/
Alexandra L. Todd and Lindsay Ivey-Burden
Finding Springs in the File Cabinet/
Mason Johnson and Ashley Ignatius
New Methodologies and Approaches for Mapping Forested
Karst Landscapes, Vancouver Island, British Columbia,
Canada (abstract)/
Tim Stokes, Paul Griffiths and Carol Ramsey
--
Contamination of Karst Aquifers
Evaluation of Veterinary Pharmaceuticals and Iodine for
Use as a Groundwater Tracer in Hydrologic Investigation
of Contamination Related to Dairy Cattle Operations/
Larry 'Boot' Pierce and Honglin Shi
Karst Influence in the Creation of a PFC Megaplume/
Virginia Yingling
Tracking of Karst Contamination Using Alternative
Monitoring Strategies: Hidden River Cave Kentucky/
Caren Raedts and Christopher Smart
Spatiotemporal Response of CVOC Contamination and
Remedial Actions in Eogenetic Karst Aquifers/
Ingrid Y. Padilla, Vilda L. Rivera and Celys
Irizarry
Determination of the Relationship of Nitrate to Discharge
and Flow Systems in North Florida Springs/
Sam B. Upchurch
--
Geophysical Exploration of Karst
The Million Dollar Question: Which Geophysical Methods
Locate Caves Best Over the Edwards Aquifer? A Potpourri
of Case Studies from San Antonio and Austin, Texas,
USA/
Mustafa Saribudak
--
Rollalong Resistivity Surveys Reveal Karstic
Paleotopography Developed on Near-Surface Gypsum
Bedrock/
Lewis Land and Lasha Asanidze
Integration and Delivery of Interferometric Synthetic
Aperture Radar (InSAR) Data Into Stormwater Planning
Within Karst Terranes/
Brian Bruckno, Andrea Vaccari, Edward Hoppe, Scott T.
Acton and Elizabeth Campbell
Detection of Voids in Karst Terrain with Full Waveform
Tomography/
Khiem T. Tran, Michael McVay and Trung Dung
Nguyen
Characterization of Karst Terrain Using Geophysical
Methods Based on Sinkhole Analysis: A Case Study of the
Anina Karstic Region (Banat Mountains, Romania)/
Laurentiu Artugyan, Adrian C. Ardelean and Petru
Urdea
Investigation of a Sinkhole in Ogle County, Northwestern
Illinois, Using Near-Surface Geophysical Techniques/
Philip J. Carpenter and Lauren M. Schroeder
Study on Monitoring and Early Warning of Karst Collapse
Based on BOTDR Technique/
Zhende Guan, X. Z. Jiang, Y. B. Wu, Z. Y.
Pang
Pre-Event and Post-Formation Ground Movement Associated
with the Bayou Corne Sinkhole/
Cathleen E. Jones and Ronald G. Blom
The Application of Passive Seismic Techniques to the
Detection of Buried Hollows/
Michael G. Raines, Vanessa J. Banks, Jonathan E.
Chambers, Philip E. Collins, Peter F. Jones, Dave J.
Morgan, James B. Riding and Katherine Royse
Using Electrical Resistivity Imaging to Characterize
Karst Hazards in Southeastern Minnesota Agricultural
Settings (abstract)/
Toby Dogwiler and Blake Lea
Karst Management, Regulation, and
Education
The Cost of Karst Subsidence and Sinkhole Collapse in
the United States Compared with Other Natural
Hazards/
David J. Weary
Hazard of Sinkhole Flooding to a Cave Hominin Site and
its Control Countermeasures in a Tower Karst Area,
South China/
Fang Guo, Guanghui Jiang, Kwong Fai Andrew Lo,
Qingjia Tang, Yongli Guo and Shaohua Liu
Case Studies of Animal Feedlots on Karst in Olmsted
County, Minnesota/
Martin Larsen
Evaporite Geo-Hazard in the Sauris Area (Friuli Venezia
Giulia Region NE Italy)/
Chiara Calligaris, Stefano Devoto, Luca Zini and
Franco Cucchi
Building Codes to Minimize Cover-Collapses in
Sinkhole-Prone Areas/
George Veni, Connie Campbell Brashear and Andrew
Glasbrenner
Cars and Karst: Investigating the National Corvette
Museum Sinkhole/
Jason S. Polk, Leslie A. North, Ric Federico, Brian
Ham, Dan Nedvidek, Kegan McClanahan, Pat Kambesis,
Michael J. Marasa and Hayward Baker
LEPT, A Simplified Approach for Assessing Karst
Vulnerability in Regions by Sparse Data; A Case in
Kermanshah Province, Iran/
Kamal Taheri, Milad Taheri and Fathollah
Mohsenipour
--
Modeling of Karst Systems
Accounting for Anomalous Hydraulic Respones During
Constant-Rate Pumping Tests in the Prairie Du
Chien-Jordan Aquifer System Towards a More Accurate
Assessment of Leakage/
Justin Blum
Numerical Simulation of Karst Soil Cave Evolution/
Long Jia, Yan Meng, Zhen-de Guan and Li-peng
Liu
Experimental and Numerical Investigation of Sinkhole
Development and Collapse in Central Florida/
Xiaohu Tao, Ming Ye, Xiaoming Wang, Dangliang Wang,
Roger Pacheco Castro, JIan Zhao
Evaluation of First Order Error Induced by
Conservative-Tracer Temperature Approximation for
Mixing in Karstic Flow/
Philippe Machetel and David A. Yuen
Study on he critical velocity of groundwater to
form subsidence sinkholes in karst area/
Fuwei Jiang, Mingtang Lei, Jian-ling
Dai
--
Engineering and Geotechnical Investigations
in Karst
Concepts for Geotechnical Investigation in Karst/
Joseph A. Fischer and Joseph J. Fischer
Sinkhole Physical Models to Simulate and Investigate
Sinkhole Collapses/
Mohamed Alrowaimi, Hae-Bum Yun and Manoj
Chopra
Monitoring the Threat of Sinkhole Formation Under a
Portion of US 18 in Cerro Gordo County, Iowa Using TDR
Measurements/
Kevin M. O'Connor and Matthew Trainum
Predicting Compaction Grout Quantities in Sinkhole
Remediation/ Edward D. Zisman
Pre-Construction Rock Treatment and Soil Modification
Program Using Low Mobility Grout to Mitigate Future
Sinkhole Development in a 2,787.1 Square Meter (30,000
SF) Maintenance Facility/
Steven W. Shiffett
Successful Foundation Preparations in Karst Bedrock of
the Masonry Section of Wolf Creek Dam/
David M. Robison.
Hydrocompaction Considerations in Sinkhole
Investigations/
Edward D. Zisman and Stephen West
--




PAGE 1

381 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DETECTION OF VOIDS IN KARST TERRAIN WITH FULL WAVEFORM TOMOGRAPHY Khiem T. Tran Assistant Professor, Clarkson University, Department of Civil and Environmental Engineering, 8 Clarkson Ave., P.O. Box 5710, Potsdam, NY 13699-5710, USA, ktran@clarkson.edu Michael McVay Professor, University of Florida, Department of Civil and Coastal Engineering, 365 Weil Hall, P.O. Box 116580, Gainesville, FL, 32611, USA, mcm@ce.ufl.edu Trung Dung Nguyen Ph.D. student, Clarkson University, Department of Civil and Environmental Engineering, 8 Clarkson Ave., P.O. Box 5710, Potsdam, NY 13699-5710, USA, nguyent@clarkson.edu As reviewed by Plessix (2008) and Vireux and Operto (2009), the full waveform inversion (FWI) approach offers the potential to produce higher resolution images of the subsurface by extracting information contained in the complete waveforms rather than approaches using only the dispersive properties of Rayleigh waves or firstarrival signals. Nasseri-Moghaddama et al. (2007), for example, have clearly shown that the recorded responses at the surface can carry valuable information regarding the below the surface. However, FWI is computationally intensive, requiring a full solution of the governing wave equation. Many algorithms for waveform inversion have been developed and applied to synthetic and real seismic data in large-scale (kilometer-scale) domains experiments surface waves can clearly separate from body waves and be removed in the inversion process. However, at shorter length scales (meter-scale), it is difficult to separate body waves from surface waves, and only a few studies of waveform inversion involving both body and surface waves have been performed for near-surface investigations on synthetic data (Gelis et laboratory data (Bretaudeau et al., 2013). A 2-D full waveform inversion (2-D FWI) technique (Tran and McVay, 2012) was reported which inverted both body and surface waves in the case of real experimental data. The technique includes forward modeling to generate synthetic wavefields and employs the GaussNewton inversion method to update model parameters Abstract This paper presents an application of time-domain surfacebased waveform tomography for detection of voids in karst terrain. The measured seismic surface wave fields were inverted using a full waveform inversion (FWI) technique, based on a finite-difference solution of 2-D elastic wave equations and the Gauss-Newton inversion method. The FWI was applied to real experimental data sets collected from twelve test lines at a karst site in Florida. Two of the test lines were located next to open karst chimneys to image their extent. Ten other test lines were located in an open and flat area without any void indication from the ground surface to detect an unknown void. The inversion results show that the waveform analysis was able to delineate embedded low-velocity anomalies, a void, and highly variable bedrock both laterally and vertically. Locations of the low-velocity anomalies were consistent to the known open chimneys observed from the ground surface. The unknown identified void was confirmed by an independent standard penetration test (SPT). Introduction Embedded void detection in a site usually begins with non-destructive testing (NDT), as NDT data can provide general subsurface conditions over a large volume of materials. At suspicious locations (anomalies), more involved invasive methods such as the Cone Penetration Test (CPT) or Standard Penetration Test (SPT) are then conducted to obtain more detailed information. Various approaches have been developed and employed ground-penetrating radar and traditional seismic wave methods. These methods have both advantages and limitations in identifying and quantifying voids.

PAGE 2

382 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE portion of the site. The chimneys were formed due to sinkhole activities. Line 1 was at the grid line G6G42, next to open chimneys 1 and 2. Line 2 was at the grid line 28AK, perpendicular to line 1, and next to open chimneys 2 and 3. Due to safety concerns, the two test lines were conducted about 1 m away from the chimneys. Line 1 (G6G42) was conducted using a linear array for a total receiver spread of 34.5 m (station 0.75 m to 35.25 m). The seismic energy was created by striking a 150-mm square metal plate with a 90 N sledgehammer. Twenty-five shots at 1.5 m spacing were recorded, for a total shot spread of 36.0 m (station 0.0 m to 36.0 m). Line 2 (28AK) was conducted using the same 24 geophones at a spacing of 1.2 m for a total receiver spread of 27.6 m (station 0.6 m to 28.2 m). Twenty-five shots at 1.2-m spacing were recorded for a total shot spread of 28.8 m (station 0.0 m to 28.8 m). Unlike line 1, the geophone spacing of 1.2 m (instead of 1.5 m) was For line 1 data analysis, a proper initial model is required to avoid the inversion being trapped in local minima. For simplicity, an estimate of the initial model was established via a spectral analysis of the measured data. A linear increasing S-wave velocity from 200 m/s at the surface to 400 m/s to a depth of 18m (half of test line length) over a length of 36 m was considered. The initial P-wave velocity for the domain was calculated from the S-wave velocities assuming that the initial Poissons ratio throughout the domain was 0.25. The mass density throughout the model was kept constant at 1,800 kg/m 3 for all inversions. Three inversion runs were performed for frequency ranges with the lowest frequency range. The medium of 18 m 36 m was divided into about 1200 cells of 0.75 m 0.75 m. During the inversion, S-wave and P-wave velocities of cells were updated independently, and each run was stopped after 20 iterations when the observed waveform data and the estimated waveform data were similar. The final results are shown in Figure 2A for analysis of shown in Figure 2A for comparison. The final inverted S-wave profile (Figure 2A, top) shows two low-velocity and vertical variations in limestone boundaries (Vs > 800 m/s) at the bottom of profile. Evidently these anomaly until the residual between predicted and measured surface velocities are negligible. For the forward modeling, the classic velocity-stress staggered-grid finite difference solution of 2-D elastic wave equations in the time domain (Virieux, 1986) are used in combination with perfectly matched layer boundary conditions (Kamatitsch and Martin, 2007). For model updating, the between the estimated responses obtained by forward simulation and the observed seismic data. Virtual sources and a reciprocity principle are used to calculate partial derivative wavefields (via the gradient matrix) to reduce computational time. Observed and estimated wavefields are convolved with appropriate reference traces to remove the influence of source signatures. The inversion technique is independent of sources, or source signatures are not required to be measured during field testing. Any source signatures (e.g., sine, triangle, or Ricker wavelets) having the same central frequency of the measured data can be used for inversion. See Tran and McVay, 2012, for details. embedded voids/anomalies in karst terrain. The inversion was carried out independently for Pand S-wave velocities in each cell with the mass density of the medium assumed constant. Application The 2-D FWI scheme was applied to a real test site to highly variable subsurface profiles and embedded anomalies. The test site was a Florida Department of Transportation (FDOT) retention pond located in Newberry, Florida. From invasive tests, the site consisted of medium dense, fine sand and silt 2 to 5 m thick, limestone varied from 2 m to 10 m in depth (Tran and Hiltunen, 2011). The site was divided into 26 parallel northsouth survey lines equally spaced 3.0 m apart. The lines were labeled A through Z from west to east across the site, and each line was about 200 m long, with station 0 m located at the southern end of each line. Twelve test lines were conducted at southern and northern portions of the site. Southern Portion of the Site Two test lines were conducted next to open chimneys in the unconsolidated sediments (Figure 1) at the southern

PAGE 3

383 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 3 are also shown in Figure 2B for comparison. The final inverted S-wave profile (Figure 2B, top) shows a lowhigh lateral and vertical variations in limestone boundaries (S >800 m/s) at the bottom of profile. A valley of lowvelocity area was found at distance 8 m near chimney 3. The inverted P-wave profile (Figure 2B, bottom) was consistent with the estimated S-wave profile. For further verification of the inverted profiles, S-wave velocity profiles from two different perpendicular lines that intersected are shown in Figure 3. The intersection was at distance 22 m of line 1 and distance 18 m of line 2. The similarity of two independent S-wave profiles suggested consistency and credibility of the FWI. locations were the same as those of the chimneys (i.e., 12 m and 21 m). Note that the exact depths of chimneys were not measured due to safety concerns. The inverted P-wave profile (Figure 2A, bottom) was consistent with the estimated S-wave profile. Chimney both S-wave and P-wave images. Chimney 2 of about 1-m diameter was not shown, due to 3-D effects. To needed to be closer to the chimney, and data at higher Data analysis for line 2 was similar to that of line 1. The medium of 14.4 m 28.4 m was divided into about 1200 cells of 0.6 m 0.6 m. Three inversion runs were performed for frequency ranges with central frequencies Figure 1. Southern portion. Clockwise from upper left: (A) Test location diagram; (B) Chimney 1 photo; (C) Chimney 2 photo; (D) Chimney 3 photo.

PAGE 4

384 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Four inversion runs were conducted for frequency ranges with central frequencies of 6, 10, 15, and 20 medium was divided into cells of 0.75 m x 0.75 m. During inversion, S-wave and P-wave velocities of cells were updated independently, and each run was stopped are shown in Figure 4. The final inverted S-wave profile (top) shows a void embedded at about 6 to 9 m depth (S-wave velocity less than 50 m/s), along with high lateral and vertical variations in weathered limestone (S-wave velocity more than 600 m/s) boundaries at the bottom of profile. The inverted P-wave profile (bottom) is consistent with the S-wave profile. To verify the seismic test results, a Standard Penetration Test (SPT) was conducted at the predicted void location (distance 18 m) by the sponsor (FDOT) three weeks after the seismic test, and the SPT N values are shown in Figure 4B. It is interesting that a void exists at this location that was embedded at about 4 by air) or very low (void filled by raveled soil). Although the 2-D FWI showed the useful capability to locate the void, the predicted depth (6 to 9 m) is deeper than the Northern Portion of the Site As the northern portion of the site was an open and flat area with no indication of a void on the ground surface, ten test lines were conducted along lines K through T. Each line was conducted using a linear array of 24 total receiver spread of 34.5 m (station 0.75 m to 35.25 m). Twenty five shots at 1.5 m spacing were recorded, for a total shot spread of 36.0 m (station 0.0 m to 36.0 m). Acquired seismic data from all ten test lines were embedded void are presented here. Figure 2. Southern portion: S-wave and P-wave velocities (m/s) of (A, top) Line 1 and (B, bottom) Line 2. Figure 3. Comparison of inverted S-wave velocity at the intersection of two lines (distance 22 m of line 1 and distance 18 m of line 2).

PAGE 5

385 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 inversion successfully identified complex subsurface and highly variable limestone surfaces at the bottom of profiles. The inverted results are consistently identified known open karst chimneys in the unconsolidated sediments observed from the ground surface. The independent inverted S-wave velocity profiles at the intersection of two perpendicular test lines are similar, suggesting consistency and credibility of the full waveform inversion technique. The identified void was confirmed by an independent standard penetration test (SPT). For the cases presented, full waveform inversion is computationally practical, as the results obtained were all achieved in about three hours of computer time on a standard laptop computer. References Brenders AJ, Pratt RG. 2007. Full waveform tomography for lithospheric imaging: Results from a blind test in a realistic crustal model. Geophysical Journal International 168 (1): 133-151. Bretaudeau F, Brossier R, Leparoux D, Abraham O, Virieux J. 2013. 2D elastic full-waveform imaging of the near-surface: application to synthetic and physical modelling data sets. Near Surface Geophysics 11: 307-316. Cheong S, Pyun S, Shin C. 2006. Two efficient steepest-descent algorithms for source signaturefree waveform inversion. Journal of Seismic Exploration 14: 335-348. Choi Y, Alkhalifah T. 2011. Source-independent timedomain waveform inversion using convolved wavefields: Application to the encoded multisource waveform inversion. Geophysics 76 (5): R125-R134. elastic waveform inversion using Born and Rytov formulations in the frequency domain. Geophysical Journal International 168: 605-633. Kamatitsch D, Martin R. 2007. An unsplit convolutional perfectly matched layer improved at Geophysics 72 (5): SM155-SM167. Nasseri-Moghaddama A, Cascante G, Phillips C, Hutchinson DJ. 2007. Effects of underground cavities on Rayleigh waves-Field and numerical experiments. Soil Dynamics and Earthquake Engineering 27: 300-313. Plessix RE. 2008. Introduction: Towards a full waveform inversion. Geophysical Prospecting 56: 761-763. real depth of the void (4 to 7 m), this is mostly attributed to the discrepancy between the estimated waveform data (plane strain) and the measured data (non-plane strain). Non -plane strain measured data is due to the 3-D void and applied point loads (hammer strikes). Conclusions An application of Gauss-Newton inversion of full seismic elastic waveforms is presented for a highly Figure 4. Northern portion: (A, top) S-wave and P-wave velocities (m/s) and (B, bottom) SPT N value.

PAGE 6

386 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Ravaut C, Operto S, Improta L, Virieux J, Herrero A, DellAversana P. 2004. Multiscale imaging of complex structures from multifold wide-aperture seismic data by frequency-domain full-wavefield tomography: Application to a thrust belt. Geophysical Journal International 159: 1032-1056. Operto S, Virieux J. 2011. Shallow-structure inversion. Geophysics 76 (3): R81-R93. Sheen DH, Tuncay K, Baag CE, Ortoleva PJ. 2006. Time domain GaussNewton seismic waveform inversion in elastic media. Geophysical Journal International 167: 1373-1384. Shipp RM, Singh SC. 2002. Two-dimensional full wavefield inversion of wide-aperture marine seismic streamer data. Geophysical Journal International 151: 325-344. Tran KT, Hiltunen DR. 2011. Inversion of FirstArrival Time Using Simulated Annealing. Journal of Environmental and Engineering Geophysics 16: 25-35. Gauss-Newton Inversion of 2-D Full Seismic Waveform in Time Domain. Soil Dynamics and Earthquake Engineering 43: 16-24. Virieux J. 1986. PSV wave propagation in heterogeneous media: velocitystress finitedifference method. Geophysics 51 (4): 889-901. Virieux J, Operto S. 2009. An overview of fullwaveform inversion in exploration geophysics. Geophysics 74 (6): WCC1-WCC26.



PAGE 1

347 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DETERMINATION OF THE RELATIONSHIP OF NITRATE TO DISCHARGE AND FLOW SYSTEMS IN NORTH FLORIDA SPRINGS Abstract The Suwannee River Water Management District has collected quarterly discharge and water quality data from 30 1 st and 2 nd magnitude springs in the Suwannee River Basin since 1998. These data were collected quarterly well into the late 2000s and constitute a valuable Trend and correlation analyses were used to compare the relationships of NO 3 + NO 2 (nitrate in this paper), trations increase as discharge from the spring increases. (2 springs with poor data) have relationships where high discharge was related to lower nitrate concentrations. Twenty percent of the springs had positive correlations Most important in terms of understanding the plumbing relations. Introduction The Suwannee River Water Management District is lo cated in north Florida (Figure 1). There are more known than any other comparable area of North America. These source to north-central Florida. In recent years, many of these springs have experienced increases of nitrate concentrations in discharge from the These increases in nitrate concentrations are caus ing eutrophication in some springs, spring runs, and receiving waters (rivers and streams) (Florida As a result of concern for increasing nitrate concentra tions in spring discharge, Floridas Department of Envi ronmental Protection and Floridas water management districts initiated comprehensive water-quality sampling programs that continue today. The Suwannee River Wa ter Management District began intensive spring sam pling in 1998. This paper reports on an analysis of these data from 1998 through 2007. Spring Discharge Considerations The springs all occur in Oligocene or Eocene limestones that constitute the upper Floridan aquifer. Dolostone is with the water is limestone. The springs vary in setting from vents or groups of vents situated at the heads of spring runs up to several kilo meters in length to vents located on the margins of their receiving waters, primarily the Suwannee and Santa Fe rivers. Sam B. Upchurch Figure 1. Springs of the Suwannee River Water Management District by magnitude.

PAGE 2

348 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Virtually all of these springs included in this investiga tion have been affected by increasing nitrate concen ground nitrate concentrations. Upchurch (1992) reported that the mean nitrate as NO 3 in rainfall was 0.97 mg/L (standard deviation = 1.01, n = 1373 samples), and in the upper Floridan aquifer groundwater in the Suwan nee River Water Management District, he reported the median nitrate (NO 3 as N) to be < 0.05 mg/L. Nitrate concentrations in many of the springs discussed in this paper exceed the background concentrations observed in regional Floridan aquifer water. short-term variability is typically less than the long-term variability. In this paper, the relative behavior of nitrate, Temporal trends, while present, do not affect the correlations discussed herein. For example, the relative Springs (Figure 3) indicates a positive correlation that is = 0.05 regardless of seasonal and long-term temporal trends in discharge or other causes of seasonality. Hypotheses and Methods It has long been understood that water derived by dif and pH as a result of longer residence times and rock/ water contact than rapidly recharged water in conduitness, and other rock-water interaction indicators in dif discharge (Basset and Ruhe, 1973) as a result of dilution under high discharge conditions. In the eogenetic karst of Florida, this pattern is often the springs in the Suwannee River drainage basin are known to be connected to swallets and in-stream siphons have unknown connections with surface-water sources. located on the shores of the rivers (Figure 1) and, de pending on head/stage relationships, the springs act as estavelles. In order to identify springs that are dominated by diffuse bank-storage events, correlation and trend analyses were surrogate for calcium, hardness, and other rock/water interaction indicators, with spring discharge (m 3 /s). As a result of concerns about the sources of nitrate in the springs, correlation and trend analyses were also used to compare nitrate (NO 3 + NO 2 ) with discharge and spe Methods 2 ) and alpha level ( Figure 3. Relationship of nitrate concentrations Levy County, Florida. Figure 2. conductance at Manatee Springs, Levy County, Florida.

PAGE 3

349 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 As noted above, unless stated otherwise, the term nitrate refers to nitrate plus nitrite (NO 3 + NO 2 ) concentrations. Hypothesis It is proposed that the relationships between discharge, each spring. Figure 4 presents the hypotheses used to ra the spring or it may represent discharge of riverine water stored in the spring conduit system and, to some degree, intergranular porosity in the eogenetic limestone. The A positive correlation would suggest that water contained minimal dilution. This relationship may be complicat Floridan aquifer water for irrigation. of nitrate and spring discharge. If nitrate concentrations increase as discharge increases, the nitrate is contained in a surface-water source. A negative correlation sug gests that the nitrate is stored in the soil or rock column Model C (Figure 4) shows the proposed relationships be and nitrate (an analyte that originates at or near the land surface by human activity). A positive correlation be tween the two analytes suggests that the source of the ni trate is similar to the source of the total dissolved solids the nitrate is stored in the soil or rock column or applied upon irrigation. A negative correlation suggests that the nitrate is derived from a surface-water source and is en tering the aquifer through a swallet or siphon or by backMany of the springs in the Suwannee River basin are estavelles that take water when river levels are high. This complicates interpretation of the conduit-dominat ed models. If the spring discharges relatively elevated nitrate concentrations without high river levels, one can assume that the nitrate is derived from within the spring discharge occurs in the waning phases or shortly after an episode of high river stage during which bank stor age and estavelle action in the spring occurred, then the surface-water source may not be within the springshed. Results trend analyses. Figure 5 illustrates the relationship of Figure 4. Models suggesting the relationships of charge and with each other.

PAGE 4

350 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE charge For the most part, there is an expected negative correla tance. This pattern is consistent with the relationship springs in the middle Suwannee River (Figure 5) that showed positive trends can be explained by the proxim ity of the springs to several large farms that irrigate with the spring water. Trends of Nitrate Concentrations with Dis charge Figure 6 depicts the trends of nitrate versus discharge in springs in the Suwannee River basin. With two excep tions (Table 1), the springs show either a positive trend gests that the sources of nitrate are related to surface wa ter entering the limestone aquifer through swallets, si phons, or sinkholes that do not have associated streams, or drainage of riverine water from estavelles. Model B (Figure 4) shows expected patterns of nitrate concentrations as a function of spring discharge. If the nitrate is stored in soil or rock, the Bassett and Ruhe pattern should appear. That is, nitrate should be slowly charge through conduits would be expected to dilute the nitrate. Conversely, rapid recharge of elevated nitrate surface water through a conduit or release of elevated ni trate riverine water stored in an estavelle should produce a trend with a positive slope. Most of the springs had positive slope relationships be tween nitrate concentrations and discharge (Figure 6). The two exceptions with negative slopes are located in an area of heavy irrigation and, apparently, high fertil Figure 7 illustrates the positive trend relationship in a dominated by riverine recharge. Much of the water dis charging from this spring enters the aquifer via a siphon et al., 2007). In this system, discharge from the spring increases as stage in the river rises, so the nitrate concen Table 1. Summary of correlations of discharge, Correlation Number of Springs Positive Correla tion No Sig Correla tion Negative Correla tion Discharge v. nitrate 22 20 2 Discharge v. Sp. Conduc tance 9 17 20 Sp. Conduc tance v. Nitrate 21 25 6 Figure 5. sus discharge in springs of the Suwannee River system. Figure 6. Trends of nitrate (NO 3 + NO 2 ) versus discharge in springs of the Suwannee River basin.

PAGE 5

351 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 pared to Nitrate Concentrations conductance through time, there is an interesting pat in the springs. According to the hypotheses (Figure 4C), events. Almost half of the springs with statistically sig Figure 9 shows an example of the pattern of nitrate and relation between the analytes, and Figure 10 illustrates the behavior of these analytes through time. This spring, Ginnie Spring (Figure 1), is located in a resort with campgrounds and is down gradient from several dair ies and other sources of nitrate. In this spring, elevated nitrate concentrations that are more-or-less synchronous conductivity river water appears to have diluted the con centrations of groundwater discharging from the spring. Figure 7. Trends of nitrate (NO 3 + NO 2 ) versus discharge in Columbia Spring, a known resur gence of the water derived from the Santa Fe River. The spring is located in Columbia County. Figure 8. Trends of nitrate (NO 3 + NO 2 ) versus nee River basin. Figure 9. Trends of nitrate (NO 3 + NO 2 ) versus from Ginnie Springs, a second magnitude spring on the south side of the Santa Fe River in Gilchrist County. Figure 10. conductance at Ginnie Springs. Low con centrations of the analytes results during high discharge events.

PAGE 6

352 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE In contrast, those springs with negative correlations be known swallets or siphons. In these springs, elevated within the springshed. A few springs that exhibit negative correlations between swallets or siphons. One example is Poe Springs, a sec ond magnitude spring located in a county park on the Santa Fe River in Alachua County. This spring (Figures 11 and 12) is located down gradient from farms with row crops and pasture land. (Figure 1) in 1998 and 2004. From this relationship with elevated nitrate water (Figure 12) represents discharge In other words, the data suggest that either Poe Springs is an estavelle or discharged water stored in the conduit system by a siphon located up gradient from the spring. Summary and Conclusions Investigation of the relationships between discharge, wannee River basin suggests that the interactions of the of nitrate release from springs. These data, which are de rived from a doubly porous aquifer, clearly suggest that a land use within the springshed may be erroneous. Comparison of trends of time-series data allow for sort ing out the myriad processes that affect spring discharge and water quality. In addition to measuring the discharge and water-quality parameters within a spring or spring run, it is advised that river discharge and stage relation ships be considered and that inventories of swallets and siphons be included as part of a spring nitrate study. References Bassett JL, and RV Ruhe. 1973. Fluvial geomorphology in karst terrain. In: Morrisawa M, editor. Fluvial geomorphology. Binghamton, NY: State University of New York, p. 75-89. Butt PL, Morris TL, and Skiles WC. 2007. Swallet/ resurgence relationships on the lower Santa Fe River, Florida. High Springs (FL), Karst Environmental Services. Florida Springs Task Force. 2000. Floridas springs: Strategies for protection & restoration. Tallahassee, Florida Department of Environmental Protection. Harrington D, Maddox G, and Hicks R. 2010. Florida springs initiative monitoring network report Florida Department of Environmental Protection, Bureau of Watershed Restoration, Ground Water Protection Section. Figure 12. conductance at Poe Springs. Variations in the Figure 11. Trends of nitrate (NO 3 + NO 2 ) versus from Poe Springs, a second magnitude spring on the south side of the Santa Fe River in Ala chua County.

PAGE 7

353 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Hornsby D, Ceryak R. 1998. Springs of the Suwannee River Basin in Florida. Suwannee River Water Management District, WR99-02. of sources of nitrate in spring waters, Suwannee River Basin, Florida. US Geological Survey Open File Report 98-0069. Stevenson J, Pinowska A, Wang Y-K. 2004. Ecological condition of algae and nutrients in Florida springs. Florida Department of Environmental Protection, Tallahassee. Upchurch, SB. 1992. Quality of waters in Floridas aquifers. In: Maddox GL, Lloyd JM, Scott TM, Upchurch SB, and Copeland R, editors. Florida Ground Water Quality Monitoring Program -Vol ume 2, Background Hydrogeochemistry, Florida Geological Survey, Special Publication No. 34, Ch. IV, pp. 12-52, 64-84,90-347. Upchurch SB, Chen J, Cain CR. 2007. Trends of nitrate concentrations in waters of the Suwannee River Water Management District, 2007. Live Oak (FL), Suwannee River Water Management District. White W.B., 1988. Geomorphology and hydrology of karst terrains. New York (NY): Oxford University Press.

PAGE 8

354 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

277 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DEVELOPMENT OF CAVITY PROBABILITY MAP FOR ABU DHABI MUNICIPALITY USING GIS AND DECISION TREE MODELING Abstract Cavity collapse and settlement due to the presence of and other engineering problems in certain areas within the Abu Dhabi City Municipality (ADM). A cavity prob ability map helps to identify regions that are more sus ceptible to the formation of cavities by identifying and tion. Information relating to cavities was cataloged and reviewed based on available data from the Geotechni cal Information Management System (GIMS), which is a consolidated geotechnical database developed by the ADM. Geological and geotechnical subsurface condi tions are obtained from previous site investigation cam paigns performed in the ADM region. All geotechnical, geological, and cavity related datasets are stored in a GIS geodatabase system. Based on detailed literature Gachsaran Formation, cavity density, cavity thickness and distance to nearest neighbor. A decision-tree model based on cavity distribution was developed for cavity velopment are lithostratigraphic position or bedrock geology and depth to the soluble Gachsaran Formation. Implementation of the decision-tree model in ArcGIS resulted in a cavity probability map. This cavity prob ability map is mainly based on existing borehole data. Areas not fully mapped by boreholes must be re-evalu ated for cavity risk when new borehole data is available. Low Probability, Low to Moderate Probability, Moder ate to High Probability, High Probability, and Very High Probability areas were delineated in the probability map. Introduction The Abu Dhabi Municipality (ADM) area has under in the last two decades (UPC, 2007). Almost the entire way over pre-existing, coastal barrier and supratidal sab kha sediments (Price et. al, 2012). During the process of infrastructure development and extension of Abu Dhabi Yongli Gao Center for Water Research, Department of Geological Sciences, The University of San Antonio, One UTSA Circle San Antonio, TX, 78249, USA, yongli.gao@utsa.edu Raghav Ramanathan RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, raghav.ramanathan@rizzoassoc.com Bulent Hatipoglu RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, bulent.hatipoglu@rizzoassoc.com M. Melih Demirkan RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, melih.demirkan@rizzoassoc.com Mazen Elias Adib Town Planning Sector, Abu Dhabi City Municipality, P. O. Box 263, Abu Dhabi, UAE, Mazen.Adib@adm.abudhabi.ae Juan J. Gutierrez RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, juan.gutierrez@rizzoassoc.com Hesham El Ganainy RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, Hesham.elganainy@rizzoassoc.com Daniel Barton Jr. RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, Daniel.Barton@rizzoassoc.com

PAGE 2

278 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE face problems including cavities and collapse features have been encountered (Tose and Taleb, 2000). Cavity parts of the Municipality (Mouchel, 2012). The Gachsa ran Formation, which is composed of interlayered mud stone and gypsum, underlies all of the ADM area and is known to be vulnerable to cavity formation in the area. The mudstone and gypsum beds within the upper part numerous sinkholes have been reported, particularly in In recent years, Geographic Information System (GIS) are used for manipulation and management of spatial data. There have been studies that apply GIS as a tool to identify or highlight regions that are more prone to relative probabilities of cavity occurrences in the ADM using GIS tools. Geological and Geographic Background The study area in ADM is approximately 11,000 km 2 It includes the mainland urban area of Abu Dhabi in addi tion to the coastal islands. The coastal area is relatively above sea level to the east and southeast across an arcu ate escarpment trending from Mafraq in the south to Al Shahama in the north (Price et al., 2012). The near surface geology of coastal Abu Dhabi Islands vial deposits overlying variably cemented Pleistocene sands (Macklin et. al, 2012). Most solution cavities occur further inland in regions such as Shakbout City, Zayed City, and regions surrounding the Abu Dhabi In ternational Airport as shown in Figure 1. Inland geol ogy of the ADM consists of Aeolian sand, active sabkha sequences, dune-bedded sandstone, marine developed carbonate mudstone and sandstone, and evaporite de posits (Tose and Taleb, 2000). The ADM is underlain by the Gachsaran Formation that is part of the Neogene system (Alsharhan and Narin, 1997). The Gachsaran Formation is a thick evaporitic basinal succession that was deposited in a shallow marine/brackish setting with input from a nearby land source indicated by plant mat ter. It is well known from offshore oil wells, but is only poorly exposed onshore in the Abu Dhabi Area where it is recorded in numerous temporary excavations and boreholes that have penetrated up to 100 m of interbed ded mudstone and gypsum (Farrant et al., 2012a). Small exposures occur around Mafraq, Shakhbout City, Shaha ma, Al Bahya, and along the foot of the Dam Formation escarpment around the Al Dhafra Air Base at Al Maqa trah (Farrant et al., 2012a, b). Evaluation of the lithological sections indicated that ground excavations had periodically intercepted open id circulation was commonly reported on drilling logs. Borehole data indicated that most of these cavities occur close to the top of rock, often at the interface between the derlying mudstone and gypsum. The data also showed that the cavities are most prevalent where the Gachsaran is closest to the surface. This formation of cavities is believed to be formed by groundwater movement along the interface of the mudstone and gypsum layers forming cavities that are more vulnerable to collapse in the vicin ity of the top of bedrock. The source for location and information of cavities for the study is mainly from a borehole database main tained by the Municipality of Abu Dhabi. The database consists of 21,257 geotechnical borings (Geotechnica, 2014). This borehole database is called Geotechnical Information Management System (GIMS). The GIMS for Abu Dhabi City supports a consolidated geotechni cal database in accordance with internationally accepted standards. A preliminary geodatabase was developed to Figure 1. Cavity distribution in Abu Dhabi Municipal ity is concentrated in regions such as Zayed City, Shak bout City, regions around the Abu Dhabi Airport and Al Falah.

PAGE 3

279 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 manage spatial data acquired during the data collection and extracted from the GIMS database. GIS Geodatabase In the last decade, GIS and database management sys spatial data relating to geologic, geotechnical and karstic 2005). Spatial data manipulation in GIS environments is a key function of any GIS application (Demers, 1997). There are numerous advantages to manage spatial infor mation and GIS data layers in a geodatabase environ ment as it allows for coordinated relationships between feature classes, which enable the creation of domains thereby reducing errors during data entry (Orrnsby and Burke, 2004). A geodatabase enables storage in a single datasets (FLNRO, 2013). The geodatabase supports a model of topologically integrated feature classes, simi lar to the coverage model. It also extends the coverage model with support for complex networks, relationships (MacDonald et al., 2001) Information Management System (GHIMS) was de karstic features, such as cavities, surface subsidence, and presence of soluble bedrock formations, in addition to data storage and management were performed in Arc Catalog and all data manipulations were performed in ArcMap. The GHIMS geodatabase is a tool developed to ture classes, raster catalogs, and raster classes together for a variety of geologic, hydrogeologic, and risk as sessment maps. In addition to basic geology (lithology, cavity location, etc.), the geodatabase includes damaged buildings and roads survey data, susceptibility of cavity in the region, solution cavities under the surface (Tose of the GHIMS geodatabase. Discussing all datsets stored within the GHIMS geoda tabase is not within the scope of this paper. Only layers that store information relevant to the solution cavities, which serve as input data is discussed. The KARST_ CAVITY (KC) feature dataset consists of six feature classes as shown in Figure 2. There are four .point fea ture classes and two polygon feature classes. Cavity_col lapsibility (KC_CVT_CLLPSB) is a point feature class that shows the distribution of stable or unstable cavities The stability of cavities depends on a series of stability ference model using a software called FLAC 3D. This analysis is outside the scope of this paper. Halite_Bhs (KC_HALITE_BH) is a point feature class that provides Name Type BOREHOLES BH 2 CUT_FILL CF 2 8 8 6 KC 6 2 3 2 54 HR 6 IR 2 Table 1. Major components of the GHIMS geodatabase are listed here. The GHIMS geodatabase is suitable for storing and managing a variety of geologic, hydrogeologic, and risk assessment related information.* indicates data layers that are relevant to this paper.

PAGE 4

280 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the locations of boreholes that have halite or rock-salt listed in the geology description of the boring logs. The halite or salt layer is an evaporite crust that is susceptible to dissolution. Old_Risk_Map (KC_OLD_RSK_MP) is a polygon feature class that represents the existing cavity risk map developed based on previous studies (Tose and Taleb, 2000). This feature class has been used only as a reference and is not used in the development of the cavity probability map. Salt_Layer (KC_SLT_LR) is a polygon feature class that provides the possible spatial extent of cavity risk map and from querying geologic descriptions provided in the GIMS borehole database. Void_Depths (KC_VD_DPTH) is a point feature class, which stores information relating to the presence of cavities or voids, and the depth to these cavities or voids based on data from borehole log descriptions from the GIMS database. Water_Loss (KC_WTR_LOSS) is a point feature class, which stores information relating to the event of water borehole log descriptions from the GIMS database. This layer could indicate probable locations of subsurface for presence of cavities, this layer is also used for refer ence only. Parameters Contributing to Cavity Forma tion Karstic cavities are geologic features that result from water erosion in soluble rocks over time due to seasonal associated seepage forces. The developed void system results in randomly shaped cavities that vary widely in caves encountered during construction of infrastructure string drop as documented in boring logs. The formation and collapse of the karstic cavities may be triggered by changes in the groundwater regime, changes in surface drainage, and construction work or urban development. In the Abu Dhabi area, irrigation inland and construction related dewatering within the urban area is likely to be one of the key triggers for sinkhole development via en GIMS borehole database using SQL. The Shakbout City cavities occurred include the southeastern Zayed City, the Abu Dhabi International Airport and the Al Falah ar eas. A small number of cavities were sparsely distributed in other areas. However, some boreholes indicate the presence or multiple cavities at different depths. In such cases the cavity closest to the surface is used for the cav ity risk assessment. Eliminating multiple cavities in the same boreholes, the total dropped to 729 cavities nearest to the surface. Bedrock solubility, depth to Gachsaran analysis were used as contributing factors in the forma tion of cavities. Depth to Gachsaran Formation The Gachsaran Formation, which is composed of inter layered mudstone and gypsum, underlies all of the ADM and is known to be vulnerable to cavity formation in the area. The mudstone and gypsum beds within the upper part of the Gachsaran Formation are prone to dissolu ation of the lithologic sections indicates that ground ex cavations have periodically intercepted open voids in the Figure 2. KARST_CAVITY (KC) feature dataset stores all information relating to location of sub sur face cavities, locations of collapsible cavities, and information on the presence of evaporate layers susceptible to dissolution

PAGE 5

281 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 is commonly reported on drilling logs. Borehole data indicate that most of these cavities occur close to the top of bedrock, often at the interface between the over ing mudstone and gypsum. The data also shows that the cavities are most prevalent where the Gachsaran is clos est to the surface. This formation of cavities is believed to be formed by groundwater movement along the inter face of the mudstone and gypsum layers forming cavities that are more vulnerable to collapse in the vicinity of the top of bedrock. In other areas such as Abu Dhabi Island and Al Falah, cavities have been encountered within the stratigraphically higher sand and sandstone layers, as well as at the interface with the Gachsaran. Figure 3 shows a histogram of the distribution of cavi ties in relation to the depth to Gachsaran formation at the cavity location. It is evident that the closer to the surface of the Gachsaran Formation the more likely the forma tion of cavities. Figure 4 shows the extent and depth to the Gachsaran Formation. Cavity Density Cavity density provides the number of cavities present per square kilometer. The cavity density is calculated using the Point Density tool under the Spatial Analyst toolbar in ArcMap application. The Point Density tool calculates the density of point features around each out around each raster cell center, and the number of points that fall within the neighborhood are added together and divided by the area of the neighborhood (Silverman, 1986). Figure 5 shows the cavity density output calcu lated in ArcMap. Cavity Size lateral directions, but since the source of cavity data is discreet points, cavities are assumed as two dimensional 0.1 m in thickness to 3 m in thickness. Cavities with thickness greater than 3 m were also observed although these were few in number compared to the total dataset. The largest cavity encountered is around 17.5 m thick. Figure 3. Histogram showing the distribution of cavities in relation to the depth to Gachsaran formation. Figure 4. The areal extent and vertical depth of the Gachsaran Formation below ground sur face level. Figure 5. Cavity density raster output created using location of cavities as input source in Ar cMap environment using the point density tool.

PAGE 6

282 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE smaller than 1 m were found to be generally stable, as supported by the numerical analyses results. Figure 6 shows the histogram for cavity distribution with respect Point Pattern Analysis Pattern analysis is the study of the spatial arrangement of point features in two-dimensional space. A pattern analysis usually demonstrates if a distribution pattern is random, dispersed, or clustered (Gao et al., 2005). In ad dition, a distribution pattern containing clusters of high were conducted for cavities in different lithological ma terials and geographical clusters. Figure 8 demonstrates a histogram of the distance to the nearest cavity for all nearest cavity is linearly increasing within the Gachsa ran Formation as shown in Figure 9. The overall Distance to Nearest Neighbor (DNN) distri bution of all cavities does not follow Poisson, Normal, Figure 6. Cavity size variation represented in histogram format. Majority of the cavities fall between 0.1 m to 1 m in thickness. Figure 7. The spatial distribution of cavities based on cavity size. Figure 8. Histogram and cavity distribution with respect to distance to nearest cavity. The dis tribution of cavities with distance to nearest cavity greater than 160 m follows a normal dis tribution. Figure 9. through the 9th nearest cavity in the Gachsa ran formation.

PAGE 7

283 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 or Log-Normal distributions. However, the distribution of the DNN for all cavities more or less follows normal distribution once DNN is greater than 160m. Decision Tree Model and Implementation One of the important advantages of geoinformatics techniques is that it can be used to extrapolate the oc currence of local events over a wider territory using sta tistical methods and predict the possibility of occurrence of these local events over an expanded territory. Geoin formatics technology or GIS applications can be used to develop multi-parametric models that can make predic tions based on a set of training examples. Several stud ies have developed multi-parameter models based on multi-scenario considerations to make predictions on the 2004, Kaufmann, 2008) The purpose of multi-parametric model in assessing cav spatial data mining aids in discovering spatial patterns among various contributing parameters (Shekhar and Chawla, 2002). A study in Great Britain uses a detailed and sub dividing regions into high, moderate, and low per, 2007). A similar scoring system was developed in Missouri by assigning scores to sub classes of multiple parameters such as depth to water table, bedrock char acteristics, proximity of nearest sinkhole, and distance to nearest structure from existing sinkholes (Kaufmann, 2008). A more rigorous multi-parameter model is the frequency ratio model. Parameter maps that are used in the collapse susceptibility analyses are divided into four groups such as: geological and hydrological, topographical, land use, and vegetation cover. Each of these parameters is further calculated as a function of the frequency of cavities oc Another common multi-parametric modelling approach and Yolkin, 2004). In the probabilistic approach, sink hole or cavity collapse risk is expressed in terms of the period (for example, during the service life of a building) on the studied territory, which may cause impermissible deformation of structures, or in terms of the probability P that there will be no such sinkholes (reliability), i.e., P = 1 Ps. Decision tree models are one of the most widely used ston, 1992). A decision tree model uses a top-down ap Hu et al., 2001). Each node indicates a test condition fol lowed by the next node all the way to the last node (Tan et al., 2005). In this study the decision tree model is im plemented to develop a cavity probability map given the study are and extent. The decision tree method is more suitable for integrated and regional scale assessments of complicated phenomena such as occurrence of cavities (Hu et al., 2001). Based on the contributing parameters listed in this study a decision tree model was developed as shown in Figure 10. The primary controls on cavity development were lithostratigraphic position or bedrock geology and depth to the soluble Gachsaran Formation. very important role in cavity distribution and formation Figure 11 represents the various spatial data manipula tions performed in ArcMap to create the input layers insoluble bedrock units based on their susceptibility to dissolution. The depth to Gachsaran Formation raster layer was queried from the GHIMS geodatabase and re depth to Gachsaran Formation less than 30 m and pix els representing values of depth to Gachsaran Formation greater than or equal to 30 m. Similarly cavity density raster layer and cavity thickness Figure 11. To create the input layer for distance to near est cavity the mean and standard deviation of DNN were Using the Buffer tool in ArcMap environment raster lay ers indicating boundaries within 210 m, 400 m and 600 m were created and were combined using the Union tool in ArcMap. Using Model Builder tool in ArcMap, the decision tree model was implemented using the input layers shown in Figure 11. A pictorial representation of the model built to calculate the cavity probability map is shown in Figure 12. Results Implementation of the decision tree in ArcGIS resulted in a cavity probability map. Figure 13 shows the cavity probability map developed for the ADM area. The cavity

PAGE 8

284 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Figure 11. ing the implementation of the decision tree represents the process to cerate the input lay probability map divides the study area into regions of low probability, low to moderate probability, moderate to high probability, high probability, and very high prob ability. The descriptions of these probability areas are as follows. LOW PROBABILITY Areas underlain by the soluble Gachsaran Formation and the depth to the Gachsaran Formation is equal to or greater than 30m are shown on the map as having low probability for cavity development. LOW TO MODERATE PROBABILITY Areas underlain by the soluble Gachsaran Formation and the depth to the Gachsaran Formation is less than 30m are shown on the map as having low to moderate probability for cavity development. The cavity density is less than one cavity per square kilometer. The expected future cavity development is generally low in these ar eas, but is moderate where small cavity clusters have developed. MODERATE TO HIGH PROBABILITY Areas in which cavities are a routine part of the sub surface and the minimum cavity density is 1 cavity per square kilometer. Higher probability cavity clusters are Figure 10. Decision tree model created to assign cavity risk probability for the ADM region. The decision tree includes characteristics of bedrock geology, depth to the Gachsaran Formation, cavity density, cavity size, and distances to the nearest cavities in the ADM area.

PAGE 9

285 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Figure 12. cavity probability map. Figure 13. The cavity probability map developed in ArcMap environment based on decision tree modeling technique.

PAGE 10

286 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE usually contained with the moderate to high probability. The minimum distance to the nearest cavity is 400 m for smaller cavities (less than 3m in thickness) and 600m for larger cavities (greater than and equal to 3m). HIGH PROBABILITY Areas in which cavities are a common part of the sub surface and the minimum cavity density is 1 cavity per square kilometer. The minimum distance to the near est cavity is 210 m for smaller cavities (less than 3m in thickness) and 400m for larger cavities (greater than and equal to 3m). New cavities are expected to form in these areas. VERY HIGH PROBABILITY Areas in which cavities are dominant features of the sub surface and the minimum cavity density is 1 cavity per square kilometer. The minimum distance to the nearest cavity is 210 m and at least a large cavity (greater than and equal to 3m) occurs within these areas. Four of these clusters containing extremely large cavities (greater than and equal to 10m) would be very susceptible for future cavity development. Discussion and Conclusions The cavity probability map, when compared with earlier, tion of cavities (Tose and Taleb, 2000). However, other use and topography, as well as anthropogenic changes to landscape and groundwater were not considered in the study due to the lack of data availability. This cav ity probability map is mainly based on existing borehole data. Areas not fully mapped by boreholes need to be re-evaluated for cavity risk once new borehole data are available. Also, in this study cavities are assumed as dis continuous 2D features, while in reality cavities tend to develop and propagate in vertical and lateral directions. References Abu Dhabi Urban Planning Vision 2030. 2004. Abu 2007]. p. 22. Available from: http://www.upc. gov.ae/template/upc/pdf/abu-dhabi-vision-2030revised-en.pdf Alsharhan AS, Nairn AEM. 1997. Sedimentary basins and petroleum geology of the Middle East. 1st ed. Amsterdam (AE): Elsevier Science B. V. p. 441. c 1997 Kluwer Academic Publishers. p. 21-30. Cooper AH. 2007. The GIS approach to evaporate-karst 53: 981-992. DOI 10.1007/s00254-007-0724-8 Demers MN. 1997. Fundamentals of geographic information systems. Wiley, New York, p. 486 Guangxi Province, China. In: Yuhr LB, Alexander CE Jr., Beck FB, eds. In: Yuhr LB, Alexander CE Jr., Beck FB, editors. Sinkholes and the Engineering Special Publication No. 183, p. 156-164. Farrant, AR, Ellison, RA, Merritt, JW, Merritt, JE, Newell, AJ, Lee, JR, Price, SJ, Thomas, RJ, And Leslie, A. 2012. Geology of the Abu Dhabi 1:100 000 map sheet, United Arab Emirates. (Keyworth, Nottingham: British Geological Survey). ISBN 978 085272 725 6 Farrant, AR, Ellison, RA, Merritt, JW, Merritt, JE, Newell, AJ, Lee, JR, Price, SJ, Thomas, RJ, And Leslie, A. 2012. Geology of Al Wathba 1:100 000 map sheet, United Arab Emirates. (Keyworth, Nottingham: British Geological Survey). ISBN 978 085272722 5 File geodatabase standards for data creation, publication, and distribution. 2013. British Columbia: Ministry of Forests, Lands and Natural Resource Operations gov.bc.ca/local/dbc/docs/geo/services/standardsassessment in Minnesota using a decision tree model. Environmental Geology, 54(5): 945-956. assessment in Minnesota using a decision tree model. Environ Geol 54:945 Gao Y, Alexander EC Jr. 2003. A mathematical model for a sinkhole probability map in Fillmore County, Minnesota. In: Beck BF (eds) Sinkholes and the engineering and environmental impacts of karsts. Proceedings of the ninth multidisciplinary conference. Huntsville, Alabama, September 6, ASCE Geotechnical Special Publication, no. 122, p. 439. General Geotechnical Investigation Report (Draft), Spektra Jeotek. 2012. Presidential Affairs Building, Khalifa City-B., February 16 2012.

PAGE 11

287 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Geotechnical Risk Map, Presidential Affairs 44 Plots Khalifa City-B, Spektra Jeotek. 2011. Abu Dhabi, UAE. Hu RL, Yeung MR, Lee CF, Wang SJ, Xiang JX. 2001. Regional risk assessment of karst collapse in Tangshang, China. Environmental Geology 40: 1377-1389. Jiang X, Lei M, Li Y, Dai J. 2005. National-scale risk Beck BF (ed) Sinkholes and the engineering and environmental impacts of karst. Proceedings of the tenth multidisciplinary conference San Antonio, Texas, September 24, ASCE geotechnical special publication, no. 144, p. 649. Kaufmann JE. 2008. A statistical approach to karst Alexander CE Jr., Beck FB, editors. Sinkholes and the Engineering and Environmental Impacts of American Society of Civil Engineers. p. 257-268. Koutpov VM, Mironov OK, Tolmachev VV. 2007. areas using GIS technology. Environmental Geology 54: 957-962 Lei M, Jiang X, Li Y. 2001. New advances of karst collapse research in China. In: Beck BF, Herring JG (eds) Geotechnical and environmental applications of karst geology and hydrology. Proceedings of the eighth multidisciplinary conference on sinkholes and the engineering and environmental impacts of karsts. Louisville, KY, 1, April, A. A. Balkema, Lisse, p. 145. Macklin S, Ellison R, Jason M, Farrant A, Leon L. 2011. Dubai, UAE. Bulletin of Engineering Geology and the Environment 7: 1-19. MacDonald A, Brown T, Andrade J, Hoel E, Bailey J. 2001. Building a Geodatabase: GIS by ESRI. ESRI User Manual 2001. Mitchell TM. 1997. Machine learning. McGraw-Hill, New York, p. 414. Mouchel Report. 2012. Abu Dhabi Capital District Emirati Neighborhood Package 2A and 3. Abu Dhabi. Ormsby T, Napoleon, E, Burke, R, Groesl C, Bowden, L. 2008. Getting to know ArcGIS Desktop: Basics of ArcView, ArcEditor and ArcInfo, 2nd Edition, p. 592. Price SJ, Farrant AR, Leslie A, Terrington RL, Merrit J, Entwisle D, Thorpe S, Horabin C, Gow H, Self bedrock geological model of the Abu Dhabi urban area, United Arab Emirates. British Geological Survey Commercial Report, CR/11/138. assessment of karst risk on the local and regional Engineering Geology for Infrastructure Planning in Europe: A European Perspective (Lecture Notes in Earth Sciences), June 14, Lecture Notes in Earth Sciences, no. 104, p. 760. Shekhar S, Chawla S. 2002. Spatial databases: a tour. Prentice Hall, p, 300. Silverman, B.W. 1986. Density Estimation for Statistics and Data Analysis. New York: Chapman and Hall. Tan P-N, Steinbach M, Kumar V. 2005. Introduction to data mining. Addison Wesley, Reading, USA, p. 769. The Geotechnica. 2004. Equipe Group, p. 27. Available from: http://www.equipegroup.com/pdfs/ theGeotechnica_August_2014.pdf Tennant EW. 2007. A sample geodatabase structure for managing archaeological data and resources with ArcGIS. Technical Briefs in Historical Archaeology 2: 12-23 of probabilistic methods for predicting sinkhole danger. Power Technology and Engineering 37: 350-353 Tose JA, Taleb A, 2000. Khalifa city b ground conditions program. In: Mohamed & Al Hosani, editors. Georngineering in Arid Lands. Balkerna, Rotterdam. p. 75-81 Addison-Wesley, Reading, USA, p. 423 karst depression in gypsum: A case study from Sivas basin (Turkey). Engineering Geology 90: 89103.

PAGE 12

288 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

177 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DOWN THE RABBIT HOLE: IDENTIFYING PHYSICAL CAUSES OF SINKHOLE FORMATION IN THE UK Tamsin Brittany Green School of Geography, University of Leeds, University Road, Leeds, West Yorkshire, LS2 9JZ, gy12tg@leeds.ac.uk due to the highly soluble nature of the limestone (Waltham et al., 1997). This study explores how surficial features exacerbate dissolution, and aims to demonstrate that bedrock characteristics are not the most important factors. In addition, the aim is to assess the relative importance of the surficial features. Base map data and geological data were obtained from EDINA DigiMap and the British Geological Survey respectively (BGS License number 2014/143 ED British Geological Survey NERC. All rights reserved. Edina DigiMap Crown Copyright/database rights 2015. An Ordnance Survey/ (Datacentre) supplied service). This data, together with surface feature data from ArcGIS, enabled the and low susceptibility areas. Within this report, predictions are estimated on a spatial scale, following the sensitivity associated with temporal prediction. Therefore, chronological data is not accounted for, and the maps only present visual future susceptibility on a two-dimensional level. slope, curvature, and altitude within the two areas play in exacerbating subsidence, and whether one factor in particular may be more critical. This analysis assesses the relative importance of each variable, to ultimately create a reliable spatial sinkhole susceptibility map. Due to the topical, public, and media interest in sinkhole collapse, this research is of significant importance in todays environment, economy, and society. Sinkholes Beck, 1992). Most relevant papers date back to the Purdue, 1907). These predominantly focused on preexisting cavities, where the limestone cave systems that were once mining sites, initiated subsidence. Recent literature now focuses more upon the range of external factors that exert pressure on these vulnerable locations, due to the increasing availability of modern technological Abstract Heavy precipitation in the UK in February 2014 induced ground subsidence and consequently a rapid increase in the frequency of sinkhole occurrences. These new the causes of the increased occurrence by investigating the relative importance of various surficial factors. Malham and the Mendips are two areas of particular interest, since both are underlain by limestone bedrock and are susceptible to subsidence. This is due to limestone form an extensive network of karstic caves. It was therefore useful to compare two sites of similar geology, both from the Triassic and Jurassic periods, as this controlled the amount of presently exposed limestone from past glacial retreat, for accurate comparison of susceptibility. Susceptibility maps of the two areas were created by integrating GIS application and statistical methods to develop algorithms to address the issue of dissolution. The maps aim to identify the physical surficial conditions, in addition to heavy precipitation that exacerbates subsidence development. Statistical testing of the GIS data indicated that in Malham, slope is the most significant parameter (Kruskal-Wallis, most significant parameter (Kruskal-Wallis, H= 20.44, appeared less statistically significant with fewer values reported from post-hoc Mann-Whitney U tests. This integrated geological mapping and statistical approach areas within the UK. Introduction This study investigates the physical and surficial causes of sinkhole formation in both Malham and the Mendips, in the UK. These are two areas underlain predominantly by limestone bedrock, and are highly prone to dissolution,

PAGE 2

178 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Study Site Two 25km 2 areas consisting of pre-existing doline points were extracted from the BGS GIS database for each Malham, and North-west Mendip Hills (Figures 5 and 6). susceptibility based on the slope, aspect and curvature of the surrounding areas, determined by the values as grouped in Table 1. These sites were selected based on the predominant presence of limestone bedrock defining these two areas, in addition to known sinkhole activity. Methods of Study This research relied purely on secondary data obtained physical formation, and spatial distribution of sinkholes. model and BGS doline data for each area was imported into ArcGIS. The doline points provided extensive information based on count, type, shape and distribution of the points. Distance between the points was calculated using point cluster analysis, and each point was then corresponded to its bedrock class that it was underlain by. The identity tool further enabled the partnering of each point with its related topographical slope, curvature and altitude values, and enabled the integration of statistical testing and GIS. Curvature can be defined as the degree to which a surface is curved, and can be strongly linked with trends of faults and topographical fractures (Stecchi et al., 2007). The curvature is the second derivative of the elevation surface, which was run on a 3x3 cell scale determined This paper aims to narrow down the importance of particular surficial factors. Based on a range of the ability to destroy lives and local communities. The on the current ambiguity of such existing surficial causes will provide insight into avoiding a potentially unsafe environment. Though many variables involving subsidence formation have been previously investigated in research, no definitive answers have been concluded following the arbitrary nature of et al., 2002) and the fairly novel area of sinkhole research. This paper determines the importance of each factor, rather than concluding the generic causation of multiple factors. Furthermore, many studies focus on higher susceptibility to dissolution, though carbonate This study focuses on limestone carbonate karst areas the two areas have similar geology, with only surficial differences for susceptibility mapping. Though it can apparent spatial dichotomies even within the UK, it was necessary to focus on a local scale, in order to identify detailed causes, rather than wider, regional causes. This study identifies gaps in literature by creating a susceptibility map comparing two areas on a local scale, using purely surficial and physical factors in order to obtain as much detail and understanding as possible. they have been predominantly single-site based and al., 2009). This paper aims to further this research by creating a comparative map of two areas based upon multiple surficial factors. This study is meant to be useful in mapping the safety 2008), and therefore aims to implement a preventative measure, and create local awareness of specific conditions that may aggravate subsidence (Farrant and Cooper, 2008). Table 1. Susceptibility key for mapping areas. Level Description High Flat slope <3 On limestone High altitude Considerate Gently sloping <6 Mudstone Intermediate altitude Moderate Slopes <9 Siltstone and interbedded rocks Intermediate altitude Low Steep slope >9 to 5 Non-porous bedrock e.g. sandstone Low altitude

PAGE 3

179 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 of a correlation present based on altitudinal values. In Malham, the highest frequency of 64 sinkholes occurs at 510m above sea level, though this is not the highest elevation. The second highest peak consists of 52 sinkholes at 380m above sea level. In the Mendips, the frequency also varies, with a peak count of 39 sinkholes at 290m above sea level. Malham bedrock clustering did however return with a statistical significance (Kruskal-Wallis, H=23.82, p<0.001) (Figure 3). Bedrock clustering in the Mendips returned with no statistical difference across different bedrock classes (Kruskal-Wallis, H=8.39, p=0.078) (Figure 4). Slope, altitude and curvature each presented a difference with Kruskal-Wallis testing and a further post-hoc Mann-Whitney U for individual bedrock class pairing. The results were more varied for Malham, to remove any minor topographical hollows or peaks factor was reclassified into low-high susceptibility groups based on natural breaks, which was used to highlight the values where points naturally clustered. The susceptibility map was consequently created through using the raster calculator tool, multiplying each surficial layer together. The output was reclassified into four levels of susceptibility. Results of Study variable returned as not normal, so non-parametric tests were performed throughout. All Mann-Whitney U from Figures 1 and 2 that there is a higher frequency of sinkholes on flatter slopes than steep slopes, and on linear curvatures than concave or convex. There is less Figure 1. Frequency bar charts in Malham (left) presenting the number of sinkholes (total: 400) on each driver of slope, curvature and altitude respectively. Figure 2. Frequency bar charts in the Mendips (right) presenting the number of sinkholes (total: 161) on each driver of slope, curvature and altitude respectively.

PAGE 4

180 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Mendips (Figure 4) due to the wider range of sinkholes that exist outside the predicted common thresholds. The map above (Figure 5) was created using the raster calculator by combining slope, curvature and altitude. The high-susceptibility values were defined as <5.7 for slope, >413m above sea level for altitude based on reclassified natural data breaks with sinkhole frequency. Each reclassified layer was input into the raster calculator where they were combined to provide an output layer. This was reclassified again into the four classes. presented a difference. In the Mendips, conglomerate appeared to show the strongest difference between each other class within all tests for the surficial factors. Chert also showed significance when tested for altitudinal difference (Figure 4). Sandstone presents the widest range of curvature values, limestone presents the smallest mean range, with 121 on extreme curvature values, accounting for the large number of sinkholes apparent on limestone bedrock. More outliers are evident in Malham (Figure 3) than the Figure 3. Malham doline point clustering on each bedrock class presented by boxplots on each surficial factor: slope, curvature and altitude respectively. Figure 4. The Mendips doline point clustering on each bedrock class presented by boxplots on each surficial factor: Slope, curvature and altitude respectively.

PAGE 5

181 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Discussion Susceptibility Maps The models above (Figures 5 and 6) present high to low susceptibility of sinkhole formation in Malham and the Mendips. It is encouraging that the high frequency of preThe high-susceptibility values for slope were defined as <3.7 for slope, -0.37 to 0.16 1/100 altitude. The same method as stated above was used here also. Figure 5. Spatial susceptibility model for Malham based on high-susceptibility values of the surficial factors, and pre-existing doline points.

PAGE 6

182 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE demonstrating that such local scale features clearly exacerbate sinkhole existence. However it is also clear that some sinkhole points occur on the considerate susceptibility areas. This is because the threshold of the slope, curvature and altitude values do not define the exact points that sinkholes can only occur on, but instead provide an indication of the most vulnerable existing sinkholes is mapped on the highly susceptible red areas, whilst the orange areas have few sinkhole densities, and green and yellow areas have few to none. Based on pre-existing locations of current sinkholes and local topographical features of slope, altitude and curvature, the different areas demonstrate potential wider locations of future sinkhole development. Thus, Figure 6. Spatial susceptibility model for the Mendips based on the values of the surficial factors, and pre-existing doline points.

PAGE 7

183 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 can be directly correlated with slope where instability occurs in carbonate karst. Although it is evident gentler slopes are more susceptible than steeper slopes, perhaps the characteristics relating to carbonate karst slopes exacerbate this, in addition to the physical slope angle. Though my results demonstrated gentle slopes to be more prone to sinkholes, Stecchi et al. (2009) reported building destruction initiated by ground subsidence on these gentle slopes. It is interesting to define this study, it is clear that slope boundaries are highly this study is built on. For example, sinkhole frequency is high on slopes ranging from 0-6 in Malham, but only 0-2 in the Mendips, thus the gentle boundary could not be equally applied. Furthermore, Glade (2005) presents slopes >2 as the most active due to high erosion and weathering processes, though this can only be sitespecific to his study. A critical and potentially useful theory investigated by Sass (2007) looks into the surface depth to bedrock measurements on slopes. This is interesting as although bedrock depth is not considered in this report, it could explain the presence of certain bedrock types in would reveal deeper underlying bedrock layers, than a flat slope comprised of the original first layer bedrock, that is a target for further deposited material. This could therefore explain in the Mendips for example, why bedrock such as sandstone, a predominantly low permeability rock (Ward and Morrow, 1987), is more more resistance clayey material, (Franklin and Chandra, 1972) is more prominent on flat slopes, if sandstone was originally at deeper depths than other bedrock. Stecchi et al. (2009) also present slope to only be connected with ground movements, and not topography, Rahimi and Alexander, 2013) note that the visual surface depressions and hollows indicate subsidence and sinkhole development. Curvature It was anticipated that sinkhole frequency would be higher on concave curvatures than linear or convex curvatures, due to the heavy pooling of precipitation areas, though considerate to low areas still need to be considered regardless. The threshold values also largely vary on a spatial scale, even within two similar geological areas within the UK, and so these varying boundaries need to be site-specific. Results statistically proved local factors of flat slopes, higher altitudes and linear curvatures to be more susceptible to sinkhole formation, with the exacerbation of precipitation. Though it was expected that concave curvatures would be highly susceptible, my statistical results found linear curvatures to be more so. However, difficulty does arise with curvature analysis through the necessary filtering. These results were also consistent with findings from Farrant and Cooper (2008), Simms and Ruffell, (1989) local factors exacerbate subsidence can however provide an approximate and valuable insight into potentially vulnerable locations, with the suitable underground conditions. Furthermore, in standardising approximations, the highsusceptibility percentage of the total slope and altitudinal values can be quantified. The areas combined present the total range. Slope A difference in sinkhole frequency was clearly evident, with flat slopes containing more sinkholes than steep slopes. This finding was expected. Farrant and Cooper (2008) suggested that dissolution pipes and irregular rockhead form on flat slopes, ideal for karst formation, whereas steep slopes promote erosion. This finding is additionally supported in the context of landslides in Glade (2005) and Cooper (2008), where steep slopes induce ground instability and consequential rockfall, thus also causing erosion indirectly. It is also interesting to note that Farrant and Cooper (2008) claim slope to be an irrelevant factor when considering evaporite subsidence on gypsum and salt bedrock, as the karstic rock is rarely exposed to the surface. In contrast, Santo et al. (2007) highlights the critical relationship between carbonate karst and local slope stability. This is an issue where widespread presence of carbonate karst and a high availability of dissolution to the slopes through hydrogeology induce subsidence. This is therefore a relevant study, as sinkhole formation

PAGE 8

184 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE It is interesting to note the disparities among theories involving the promotion and prohibition of sinkhole development. UWSP (n.d.) explains how higher altitudes promote stronger weathering processes, and therefore expose the karstic bedrock further. In addition to this, it is also noted by Santo et al. (2007) that ice predominantly be applicable to the Mendips where average altitude is approximately 700m above sea level, far higher however than Malham. In addition, the Mendips are also comprised of anticlines and periclines, both, which promote accelerated erosion, thus further exposing bedrock below (BGSb, n.d.). The conflicting theory adheres to the concept that exposed karst limits development (Simms and Ruffell, 1989). Research presented a strong variance in the frequency of sinkholes at a range of altitudes, with no distinct pattern, and so both theories are seen to be noteworthy. Whilst it is more strongly believed that erosive processes limit sinkhole development, rather than enrich it, exposed limestone karst is generally rare (Beck, 1986). However, Beck (1986) does note that exposed limestone initiates rapid recharge for sinkholes, so perhaps it is this indirect correlation that can cause a certain type of would therefore explain the variety of results returned from ArcMap data extraction. There is a clear local dependence on altitude for sinkhole formation, likely resulting from multiple changes in soluble rock such as land surface cutting and local precipitation. However, the underlying cause for altitude acting as a driver of spatial distribution, is due to sinkholes developing in soluble rock and salt far below sea level, up to high elevations in rock permafrost. Bedrock Clustering Whilst some studies have previously identified sinkhole distributions through nearest neighbor analysis (Gao et clustering of points on each bedrock class. It was anticipated that limestone, being the most prevalent bedrock to sinkhole 2002), would comprise the highest number of sinkholes, with the most clustering. It was expected that limestone would comprise the flattest slopes, highest altitudes and most linear curvatures. However, limestone did not show as much of a significant difference as expected at either site (Figures 3 and 4), in contrast to that of conglomerate in the Mendips (Figure 4), which appeared to be the most statistically different (BGSc, n.d.), and consequent weight exerted on the ground surface. However, this is not the case in this study, as linear curvatures presented the highest frequency. This apparent contradiction can be explained by the erosive nature of rainfall (Simms and Ruffell, 1989), exposing karst and limiting sinkhole development, rather than promoting it. Tharp (2002) similarly presents findings of low curvature being most prone to sinkhole formation due to hydraulic fracturing. In contrast to this, Stecchi et al. (2009) found high curvature values to correlate with fractures and fault lines. This therefore highlights the difficulty in obtaining accurate curvature analysis, based on the high level of filtering generally needed to take into account the wider landscape, rather than minor topographical changes. Though Stecchi et al. (2009) claimed that no filtering was needed, due to the smoothness of the raw data, Sullivan et al. (2007) and Bergbauer and Pollard (2003) state the necessity of filtering data, in order to avoid any problems relating to the dependence on the sample grid, and consequent focus on minor topographical disparities, as opposed to wider changes. A further notable link with curvature, are fault lines 2002). This is due to the ability of curvature to predict the distribution of deformation (Bergbauer and Pollard, 2003), and its close relationship to geology. Vendeville (1991) points out the criticality of curvature analysis, underlies the studys choice to compare two sites of similar geology from the Jurassic/Triassic geological time periods. Malhams range of concave curvature values, in contrast to the Mendips could therefore present a general sinking in to be indicative of sinking. Altitude Although some high altitudinal values correlated with high sinkhole frequency, there was a wide range of variance in the data. Whilst a pattern is less notable in the Malham (Figure 1), this is most likely due to the wider range of data in the Mendips, and overall higher number of sinkholes present in Malham (Figure 1), so the spread is wider. In the Mendips, the distinct drop in frequency at its highest lowest altitude, which highlights issues of changeability across spatial scales.

PAGE 9

185 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 References the development of sinkholes and Karst in Florida, U.S.A. Environmental Geology and Water Sciences 8 (2): 5-18. Bergbauer S, Pollard DD. 2003. How to calculate normal curvatures of sampled geological surfaces. Journal of Structural Geology 25: 277-289. British Geological Survey (BGSa) [Internet] n.d. [Place of publication unknown]: Increased incidence of 2014 26 May]. Available from: http://www.bgs. ac.uk/caves/sinkholes/feb2014.html British Geological Survey (BGSb) [Internet]. n.d.. [Place of publication unknown]: Geological September]. Available from: http://www.bgs. ac.uk/mendips/geology/geological_structure.htm British Geological Survey (BGSc). [Internet] n.d. [Place [cited 2014 20 September]. Available from: http:// www.bgs.ac.uk/caves/sinkholes/home.htm British Geological Survey (BGSd). [Internet]. n.d. [Place [cited 2015 20 February]. Available from: http:// www.bgs.ac.uk/mendips/rocks/triass_rocks.htm Cooper AH. 2008. The GIS approach to evaporite-karst 53: 981-992. Doctor DH, Young JA. 2013. An evaluation of automated GIS tools for delineating karst sinkholes and closed depressions from 1-meter LiDAR-derived digital elevation data. In: (Land L, Doctor DL, Stephenson JB, editors) NCKRI Symposium 2 Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, published on-line by NCKRI, Carlsbad, NM, p. 449-458. Elrod MN. 1898. The geological relations of some St. Louis group caves and sinkholes. Proceedings of the Indiana Academy of Science p. 258-267. management. Quarterly Journal of Engineering Geology and Hydrogeology 41: 339-356. Florea LJ, Paylor RL, Simpson L, Gulley J. 2002. Karst GIS advances in Kentucky. Journal of Cave and Karst Studies 64 (1): 58-62. Franklin JA, Chandra, R. 1972. The Slake-Durability Test. The International Journal of Rock Mechanics and Mining Sciences 9: 325-341. to all other bedrocks. However conglomerate consists of limestone fragments (BGSd, n.d.), and so can be said to hold similar characteristics to pure limestone, so the theory is not wholly disproved. In Malham, slope appears to present a stronger statistical altitude also portrays strong importance. Alternatively, curvature values across both sites appear to have the least importance in determining sinkhole formation, with the least values presenting a statistical significance. This could be explained by the ease of error created, and difficulty in analysing curvature, as noted in Bergbauer and Pollard (2003) and Sullivan et al. (2007). In the Mendips, the most important driver of subsidence appears to be altitude, presenting the strongest statistical difference across bedrock. Though there is much conflict over whether high altitude influences sinkhole development n.d.), perhaps it is more directly the difference in altitude across the different bedrock classes that define development. Overall, it can be said that slope and altitude have a mutual importance as drivers influencing subsidence, though each portrays different significance within each site. This is due to the change in driver importance across varying spatial gradients, highlighting the need to adapt to the change in sitespecific drivers. Though even within the UK Malham and the Mendips present a difference in driver importance, confident conclusions can be drawn, based on the similarities. Conclusions Three conclusions can be drawn from this investigation: 1. Slope and curvature appear to be the most given the difficulty in curvature analysis however, slope is the most reliable and critical factor. 2. the detail of temporal data. However, this research has presented the dichotomies present even across a small spatial gradient within the UK, which still creates some ambiguity in the surficial drivers. 3. Susceptibility mapping has proved useful for the future, though studies must be aware of the spatial inconsistencies present. This therefore encompassing larger areas and temporal data.

PAGE 10

186 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE 2007. Karst processes and slope instability: some investigations in the carbonate Apennine of Campania (southern Italy). In: Parise M, Gunn in Karst Areas: Recognition, Analysis and Mitigation. Geological Society, London, Special Publications 279: 59-72. Sass O. 2007. Bedrock detection and talus thickness assessment in the European Alps using geophysical methods. Journal of Applied Geophysics 62: 254-269. Simms MJ, Ruffell AH. 1989. Synchroneity of climate change and extinctions in the Late Triassic. Geology 17: 265-268. Stecchi, F, Antonellini M, Gabianelli G. 2007. Curvature analysis used to map subsidenceGeophysical Research Abstracts 9. Stecchi F, Antonellini M, Gabianelli G. 2009. Curvature analysis as a tool for subsidence-related risk Geomorphology 107: 316-325. Sullivan EC, Marfurt KJ, Blumentritt C, Ammerman collapse features in the Fort Worth Basin (USA). Geological Society, London. 277: 187-203. Tharp TM. 2002. Poroelastic analysis of cover-collapse drawdown. Environmental Geology 42 (5): 447-456. University of Wisconsin-Stevens Point (UWSP) [Internet]. n.d. [Place of publication unknown]: Available from http://www4.uwsp.edu/geo/ Upchurch SB, Littlefield JR Jr. 1988. Evaluation of Data for Sinkhole-Development Risk Model. Environmental Geology and Water 12 (2): 135-140. Vendeville B. 1991. Mechanisms generating normal fault curvature: a review illustrated by physical models. In: Roberts AM, Yielding G, Freeman B, editors. The Geometry of Normal Faults. Geological Society Special Publication 56: 241-249. Vineyard JD, Williams, JH. 1967. A foundation problem in cavernous dolomite terrain, Pulaski County, Missouri. Proc. 18th Ann. Highway Geol. Symposium, Purdue Univ. Eng Bull, Eng Ext Ser. 127 51 (4): 49-59. Waltham T, Simms MJ, Farrant AR, Goldie HS. 1997. Karst and Caves of Great Britain. 1st ed. Chapman & Hall. karst terrains. Quarterly Journal of Engineering Geology and Hydrogeology 41: 291-300. P, Cendrero A. 2009. Evaluating and comparing methods of sinkhole susceptibility mapping in the Ebro Valley evaporite karst (NE Spain). Geomorphology 111 (3-4): 160-172. Gao Y, Alexander EC Jr., Barnes RJ. 2005. Karst database implementation in Minnesota: analysis of sinkhole distribution. Environmental Geology 47: 1083-1098. with geomorphology. Geomorphology 66: 189-213. Identification, prediction and mitigation of Environmental Geology 53 (5): 1007-1022. Johnson, KS. 1997. Evaporite karst in the United States. Carbonates and Evaporites 12 (1): 2-14. collapse sinkholes in the Tournaisis area, southern Belgium. Engineering Geology 65: 117-124. Murray Jr GH. 1968. Quantitative Fracture StudyThe American Association of Petroleum Geologists Bulletin 52 (1): 57-65. editor. Sustainability of the Karst Environment, International Interdisciplinary Scientific Conference, Plitvice Lakes, Croatia, 23-29 September 2009. UNESCO, p. 155-162. perspectives on the environmental impacts and 235-237. Purdue AH. 1907. On the origin of limestone sinkholes. Science 26 (656): 120-122. Rahimi M, Alexander EC Jr. Locating sinkholes in LIDAR coverage of a glacio-fluvial karst, Winona County, MN. In: (Land L, Doctor DL, Stephenson JB, editors) NCKRI Symposium 2 Proceedings of the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, Carlsbad, New Mexico, published on-line by NCKRI, Carlsbad, NM, p. 469-480. Geological risk assessment of the area surrounding Altamira Cave: A proposed Natural Risk Index and Safety Factor for protection of prehistoric caves. Engineering Geology 94 (3-4): 180-200.

PAGE 11

187 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Ward JS, Morrow NR. 1987. Capillary Pressures and Gas Relative Permeabilities of Low-Permeability Sandstone. Society of Petroleum Engineers 2 (3). Wilson WL, Beck BF. 1992. Hydrogeological Factors Affecting New Sinkhole Development in the Orlando Area, Florida. Groundwater 30 (6): 918930.

PAGE 12

188 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

63 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DRIFTLESS AREA KARST OF NORTHWESTERN ILLINOIS AND ITS EFFECTS ON GROUNDWATER QUALITY Samuel V. Panno Illinois State Geological Survey, 615 E. Peabody Drive, Champaign, IL, 61820, USA, s-panno@illinois.edu Walton R. Kelly Abstract The bedrock aquifer of the Driftless Area of northwest ern Illinois is Ordovician Galena Dolomite. Previous work by the authors and others showed that the geology tion of karst. Bedrock in the study area has been shown the county are dominated by solution-enlarged crevices from 0.4 inches to 3 feet (1.0 cm to 0.9 m) wide within most road cuts and quarries examined. Other karst fea tures include cover-collapse sinkholes ranging from 3 to 25 feet (0.9 to 7.6 m) in diameter overlying Galena Do lomite, karst springs and crevice caves. A preliminary evaluation of the groundwater quality Jo Daviess County in the Driftless Area of northwest ern Illinois was conducted to assess the susceptibility of the Galena Dolomite aquifer to surface-borne contami nants. This was done by evaluating available groundwa ter quality data from published sources and the Illinois State Water Survey Groundwater Quality Database (i.e., wells and springs), and also by sampling 11 private wells try, dissolved organic carbon, stable isotopes and triti um. We found that groundwater in the study area is of a Ca-Mg-HCO 3 type as would be expected for an aquifer dominated by dolomite. In parts of the county, the upper part of the carbonate-hosted aquifer contains elevated concentrations of chloride, nitrate and potassium. Likely contamination sources are anthropogenic and include human waste. The presence of these contaminants sug gests movement of surface-borne contaminants into the aquifer and into wells screened at depths ranging from 65 to 150 feet (20 to 46 m). taminants (greatest concentrations nearest the surface) within the fractured and creviced aquifer to a depth of about 300 feet (91 m). Nitrate-N (NO 3 -N) concentrations in karst springs are typically between 10 and 13 mg/L, but can exceed 30 mg/L. Because the predominant land use in the study area is row-crop agriculture, it is likely that much of the NO 3 the 11 water well samples, NO 3 -N concentrations ranged from < 0.04 (detection limit) to 5.4 mg/L and concentra 3 N concentrations in groundwater containing less than 3 TU were below detection (0.04 mg/L), and above 3 TU, NO3-N and tritium were positively correlated. This relationship suggests a nonpoint source of N and denitri ) concentrations in karst springs were between 15 and 25 mg/L above background concentrations (1 to 15 mg/L). Water wells samples had lower Cl concentrations with 7 of 11 wells below background (ca. 15 mg/L), although the concen tration in the shallowest well was 110 mg/L and was probably derived from road salt. Overall, the groundwa ter quality of the Galena Dolomite aquifer in Jo Daviess County is what would be expected in an open, dolomitedominated karst aquifer. Introduction Groundwater in karst regions of the Midwestern US is typically Ca-HCO 3 to Ca-Mg-HCO 3 type water that can have relatively high levels of surface-borne contami nants, especially at shallow depths. Typical contami nants in wells and springs include sodium (Na), potas sium (K), chloride (Cl ), nitrate-nitrogen (NO 3 N) and 2007). The Driftless Area of northwestern Illinois was mapped in the mid-1990s as karst by Weibel and Panno (1997) and Panno et al. (1997). Domestic wells in Jo Daviess County get their water primarily from the Galena Do lomite at depths less than 250 feet (76 m) (Frankie and and Panno 2015), its effects on groundwater quality in the area are uncertain, although Panno and Luman (2008) explored this to a limited extent. They found that Cl and NO 3 -N concentrations in private and pub lic wells (most are cased more than 100 feet below the surface) in the county were as high as 55 and 31 mg/L, respectively. These concentrations are well above the

PAGE 2

64 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE upper background threshold concentrations (1 to 15 mg/L for Cl and 0 to 2.5 mg/L for NO 3 -N) for freshwa ter aquifers in the northern two-thirds of Illinois (Panno et al. 2006a, b). Potential sources include both urban and rural contaminants such as road salt (Cl ), nitrogen fertil 3 Cl and NO 3 NO 3 -N and Cl was observed. In order to determine the susceptibility of the shal low karst aquifer to groundwater contamination in the study area, we gathered water quality data from historic sources (Illinois State Water Survey (ISWS) Groundwa ter Quality Database), a recent study of six karst springs in northeastern Jo Daviess County (Maas 2010), and from sampling 11 domestic shallow water wells across cations and anions, and conduct a preliminary evalua tion of the existing water quality in the Driftless Area. sense of the residence time of groundwater in the study area and compare tritium to selected ions that may be introduced by surface-borne contamination. Geology and Hydrogeology Jo Daviess County lies within the Driftless Area of glacial drift that covers the bedrock of most of the up per Midwestern U.S. (Figure 1). However, glacial till is exposed along U.S. Route 20 between Galena and East communication, 2014), as well as a small area of gla cial till in an isolated portion of the far eastern part of the county. Bedrock consists of Middle-Ordovician age (443 490 Ma) carbonate rocks of the Galena-Platteville Group, thin remnants of the Ordovician-age Maquoketa Shale, and Silurian dolomite (412 443 Ma) whose re sistant rock caps the mounds and highlands of the county (Figure 2). Tectonic compression and extension occurred in this area during and following the formation of the Wiscon sin Arch that began in Cambrian time (490 543 Ma) and continued to be active in late Silurian or Devonian time (354 417 Ma) (Nelson 1995). As a result of com pression and extension, bedrock along the Wisconsin et al. (1959) stated that All the rock formation in the [Upper Mississippi Valley mining] district [most of Jo Daviess County] contain well-developed vertical and in much as 300 feet (91 m) vertically. Joints are especially well developed in the Galena Dolomite. Trowbridge and Shaw (1916) stated that crevices within the Galena Dolomite are frequently encountered in wells, and drilling tools are sometimes lost in them. In the mid-1990s, the Illinois State Geological Survey most of Jo Daviess County as having a very high aquifer sensitivity because fractured dolomite bedrock aquifers lie beneath the glacial drift or loess. Areas where dolomite bedrock is exposed are most sensitive. The main karst water-bearing formation in the county is the Galena-Platteville Group. Silurian dolomite is also Ekberg (2008) subdivided the secondary porosity of the Galena-Plattville Group into matrix, fracture and conduit porosity. These subdivisions are supported by spring hy Figure 1. Karst areas of Illinois, USA. The study area is the Driftless Area in the northwestern Panno 1997).

PAGE 3

65 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 drographs and drawdown curves from aquifer tests that support a triple porosity aquifer. The fracture porosity and northwest-trending vertical fractures (consistent with Heyl et al. 1959), and bedding planes (Ekberg 2008). Panno and Luman (2008) examined sinkholes and the abundant secondary porosity (crevices) exposed along road cuts and in quarries in eastern Jo Daviess County and concluded that the Galena Dolomite constitutes a karst aquifer. While cover-collapse sinkholes are present over the Galena Dolomite, they are typically small due Figure 2. 2000). Figure 3. Crevice in Galena Dolomite along road cut just east of Jo Daviess County. Pho tograph by S.V. Panno. Figure 4. Sinkholes are relatively small in Jo Daviess County due to the relatively shallow soils. Within croplands, these features are eas ily buried.

PAGE 4

66 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE to the thin soils of the area and are easily buried (albeit temporarily) in croplands (Figure 4). However, because of their presence, the county falls into a medium to high category of aquifer vulnerability as outlined by Lindsey et al. (2010). METHODS Groundwater Chemistry Groundwater chemistry data from private wells and karst springs in eastern Jo Daviess County were avail able through the ISWS Groundwater Quality Database described in Panno and Luman (2008), and from Maas (2010), respectively. In addition, 11 domestic shallow wells were sampled across the county (one in Septem ammonia, dissolved organic carbon (DOC), D/H and 18 O/ 16 O isotopic ratios, and bacterial indicators. Ground water samples from the 10 wells sampled in June 2014 parameters (temperature, pH, dissolved oxygen (DO), sured using a multisonde (Hydrolab, Loveland, CO). through 0.45-m membranes, placed in polyethylene or glass bottles and stored at 4 o C prior to analysis. All oratory, and kept refrigerated until analysis. Analyses for inorganic chemicals were conducted in accordance with standard methods (APHA, AWWA, WEF 1999) at the ISWS using standard methods. Sterile techniques were used for samples collected for bacterial indicators (total coliform and Eschericia coli). ers before transporting to the analytical laboratory the same day. Bacterial indicators were determined using the Colilert method (IDEXX 2013) at the City of Dubuque Laboratory, Dubuque, Iowa. tlund and Dorsey 1977) and liquid scintillation counting as described in Hackley et al. (2007). Results are report ed in tritium units. Results and Discussion Other Hydrologic Features Springs are a common feature throughout Jo Daviess County and the locations of some have been mapped by Reed (2008) and Maas (2010). The only available data on the chemical composition of springs in the county are from Maas (2010) for six springs in northeastern Jo Da viess County within the Warren Quadrangle. The springs gation and are consistent with discharge of groundwater along bedrock crevices where the overburden thins near stream valleys. Groundwater, under hydrostatic pressure, would be able to breach land surface in low-lying areas with relatively thin overburden (usually near streams). Bedrock springs typically appear to be large circular to elliptical depressions with small sections breached, providing openings through which the spring water dis charges to a nearby stream. Chemical Composition of Groundwater Because spring water in karst regions is an amalgam of groundwater from various sources, the chemical compo ground concentration ranges of constituents and trends for contaminants (Table 1). The overall chemical com position of groundwater and relationships between and among selected cations and anions were used to identify the source(s) of contaminants. Background concentra tion ranges of selected ions are useful for comparing nat ural compositions of surface water and groundwater with waters that have been affected by anthropogenic and/or natural contamination. When making such comparisons, one must be aware that, for example, concentrations of Na, Cl and NO 3 -N that are somewhat elevated above background do not constitute water that is harmful to however, indicate that surface-borne contaminants from land-use activities have entered groundwater and will ultimately discharge to surface waters. Further, it has been shown that elevated concentrations of Na and Cl can be deleterious to vegetation (e.g., Panno et al. 1999), aquatic organisms (e.g., Kelly et al. 2012), can impart a salty taste to drinking water at Cl concentrations ex ceeding 250 mg/L, and elevated Na concentrations in drinking water may be a problem for people with high blood pressure (USEPA 2014). Nitrate-N concentrations greater than 10 mg/L in drinking water have been shown to cause methemoglobinemia (blue-baby syndrome) and may be linked to stomach cancer (ORiordan and Ben tham 1993). For the purposes of this investigation, we considered concentrations exceeding the upper end of background as anthropogenic tracers that may be used to investigate aquifer recharge areas, recharge rates and groundwater movement through the underlying karst aquifer. Groundwater in Jo Daviess County is a calcium-mag nesium-bicarbonate (Ca-Mg-HCO 3 -) type groundwater with elevated concentrations of Cl and NO 3 -N in some

PAGE 5

67 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 areas (Panno and Luman 2008). The background con centration range for Cl in shallow groundwater in north ern and central Illinois is between 1 and 15 mg/L (Panno et al. 2006a). The range for background concentration of NO 3 -N in Illinois is between 0 and 2.5 mg/L (Panno et positions of groundwater from wells and springs, the distribution of these ions and relatively high DO con centrations (4.7 to 8.7 mg/L) in the underlying aquifer system. Potential sources of Cl and NO 3 -N include road sources in open aquifer systems such as this are espe cially vulnerable to surface-borne contaminants. There is little or no attenuation of contaminants discharged wells, springs, and streams down-gradient of contamina tion sources can show effects within a few days or even in karst aquifers may result in contaminants becoming concentrated in conduits (Field 1993). Maas (2010) sampled six springs in the eastern part of Jo springs discharges from open, oxygenated systems typi cally containing elevated levels of Cl and NO 3 -N. Ni trate-N concentrations were commonly greater than 10 mg/L and as high as 30 mg/L. Spring water from all six springs was undersaturated with respect to calcite and dolomite which suggests that the springs are either domi nantly shallow groundwater or are a mixture of deep and shallow groundwater. The lack of saturation with respect Galena Dolomite is an ongoing process in this area. The elevated concentrations of NO 3 -N found in all but one of the springs are similar to those of tile drain waters of Illinois. Tile drains have been described by Schilling and Helmers (2008b) as analogous to karst drainage basins with regard to nutrient losses in an agricultural watershed. The elevated nutrient concentrations in the springs suggest that the springs studied by Maas (2010) may be affected by recharge water containing relatively high concentrations of surface-borne contaminants. The the springs may be estimated by the changes in tempera spring water temperatures. All but one spring were af fected by seasonal changes. fers, groundwater pathways may be discrete conduits/ crevices and/or bedding planes fed by numerous, smaller crevices within carbonate bedrock. Because springs are discharge points for groundwater, they may be fed by a shallow component containing surface-borne contami nants from a variety of land uses, whereas other inputs may originate from deeper, usually less contaminated, sources. The percentages of each component source can springs and with time/season. nants are presented in Table 1. These concentrations are from Panno et al. (2006a, b) and groundwater samples from the karst regions of southwestern Illinois, and also inferred from the water quality data from the spring Table 1. The range of concentrations and estimated background threshold values for parameters and ions determined from springwater samples collected in eastern Jo Daviess County by Maas (2010). Threshold values are the upper bounds of background concentra tions and were estimated based on previous work by Panno et al. (2006a, b)*, unpublished data from the sinkhole plain of southwestern Illinois (Panno, ISGS), and spring data from Hicks Spring** (Maas 2010) located in eastern Jo Daviess County. ND = Not determined.

PAGE 6

68 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE which appears to be the least affected by recharge events and surface-borne contaminants (Maas 2010) (Table 1). There are distinct differences among the six springs based on concentrations of indicators of surface-borne contamination which include K, Na, Cl and NO 3 -N. Maas (2010) came to slightly different conclusions us that relied on ion concentrations in soils and carbonate rock to determine whether ion concentrations of the springs were above or below background threshold val ues. He concluded that Na and sulfate (SO 4 2), as well as Cl and NO 3 N were anomalous and were derived from surface sources. Because the geology of Jo Da viess County is not solely carbonate rock, other strata (e.g., shale) could have contributed Na and SO 4 2to the groundwater. Potential contaminants in the area include diammonium phosphate, potash, and, on a more local basis, hog and dairy manure. The McPhillips Spring is in the vicinity of a dairy where manure is applied to the cations, June 2013). Because households in this area are also a possible contaminant (Panno et al. 2007). Preliminary work by Panno and Luman (2008) on avail able water quality data from public water wells from Jo Daviess and Stephenson Counties showed that Cl con centrations ranged from less than 1 to 55 mg/L and NO 3 N concentrations from less than 0.1 to 31 mg/L. centrations to absence of NO 3 -N at depths greater than 400 feet. Background concentrations may have a low er upper threshold than the estimated 15 mg/L Cl and the 2.5 mg/L NO 3 -N (Table 1). Additional estimates for background concentrations of Cl and NO 3 -N (based on municipal well data) are slightly lower, about 1 to 10 mg/L and 0 to 2 mg/L, respectively. The deeper ground water in the area from wells that intersect the St. Peter Sandstone have Cl concentrations between < 1 and 3 mg/L in Jo Daviess County and about 1 mg/L in the Galena area. Chloride concentrations in aquifers at that depth are anomalously low with the lowest concentra tion (0.6 mg/L) being only slightly greater than the concentration of Cl in rainwater (0.1 to 0.3 mg/L) and slightly less than in soil water (0.7 to 1.7 mg/l) (Panno evapotranspiration) and rock water interactions typically increase the Cl concentration of groundwater up to 15 mg/L. Recharge of glacial melt water would be capable point where they would impart little Cl to the ground water via rock-water interactions. This effect is apparent in part of the Mahomet Aquifer in east central Illinois (Hackley et al. 2010). The pH values of spring water sampled by Maas (2010) were always below 7.0 which are consistent with the spring water being undersaturated with respect to calcite and dolomite. Water in equilibrium with calcite or dolo mite would have a pH value of about 7.5 as noted earlier. These lower pH values indicate that groundwater from the springs is not in equilibrium with the carbonate rock supporting the interpretation of an open system. trical conductance of the water that can be easily taken does yield an indication of the presence of contaminants to groundwater of surface water. The addition of contam inants (e.g., road salt) can increase the SC proportionally. The approximate background values in southwestern Il linois and northwestern Illinois groundwater are 650 and 700 S/cm, respectively based on available data from Maas (2010) and Panno (ISGS, unpublished data). Potassium and Cl concentrations in spring water sug gest that at least two of the springs may have been con waste. Similar effects were seen in tile drain-fed stream water samples from the same area during the same pe riod (Maas 2010). Groundwater samples from the other springs all fall below the K background threshold of 0.6 mg/L (Table 1). Nitrate-N concentrations for most of the springs were greater than those of shallow groundwater of Illinois and far exceeded the background value of 2.5 mg/L (Table 1). Nitrate-N concentrations ranged between 2.9 and 30 with all but one sample exceeding 10 mg/L. These con centrations exceeded those found in tile drains of central Illinois measured by the authors (e.g., ranging from 0.51 to 23.1 mg/L, median = 11.3 mg/L (Panno et al. 2005)) and in tile-drain fed surface streams in Jo Daviess Coun ty (i.e., ranging from 4.19 to 14.6 mg/L, median = 9.31 mg/L) (Maas 2010). The elevated levels are most like in one of the springs were very low, equivalent to what Panno et al. (2006b) determined to be indicative of back ground conditions (< 0.4 mg/L). Whereas it is possible that the low NO 3 -N concentration in this spring is the tion of Cl and K suggest that it is more likely that the spring represents background concentrations for shallow

PAGE 7

69 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 groundwater in the area. Hicks Spring appears to be the least affected by anthropogenic sources and recharge events among the six springs sampled by Maas (2010), All springs are in close proximity to land supporting row-crop agriculture and one is in close proximity to a livestock operation. However, the actual recharge areas of the springs are not known. Consequently, it is not pos sible at this time to use land use as an indicator of the sources of surface-borne contaminants in any more than a general way. Chloride and NO 3 -N concentrations in Hicks Spring are well below their background thresh olds. Because background concentrations for Cl and NO 3 cause the ions covary, the data suggest that these springs are being fed by an aquifer with a steady input of these ions. The data also suggest that the ions are entering the ent. A plot of Na vs Cl for all spring data shows no dis cernible covariance, indicating that road salt (NaCl) is not the likely source of either ion in the springs. How ever, there are not enough chemical data to determine the source(s) of these ions at this time. Overall, the degree of contamination of the spring water with regard to NO 3 -N is greater than typically found in shallow groundwater of Illinois. This, plus the fact that concentrations of K and Cl exceed background concen trations in all but one spring indicates that N-based fertil groundwater systems. The high DO concentrations in the spring samples (Maas 2010) suggest rapid movement of water into the subsurface which would limit attenuation as organic matter is reduced, resulting in anoxic condi NO 3 -N concentrations. The thin soils and the presence of macropores and sinkholes in this area appear to promote rapid recharge to bedrock aquifers, a common feature of karst regions. Shallow Wells within the Galena Dolomite of Illinois have been sampled on a systematic basis. The wells ranged from 70 feet to 160 feet in depth. As with the spring water samples, groundwater from all wells is a Ca-Mg-HCO 3 type water with pH values ranging from 6.56 to 6.73. Tritium in groundwater samples ranged from 0.60 to 5.45 tritium units (TU) (Figure 5). The concentration of tritium in todays rainfall (ca. 2014) is about 5.5 TU (Fanta, ISGS, personal communications, September 2014). The presence of tritium in the well water samples indicates that there is some component of modern recharge. The Ca-Mg-HCO 3 type water was expected because each well intersected the fractured and creviced Galena Dolo mite aquifer. Groundwater in equilibrium with dolomite should have a pH value of about 7.5 (e.g., Parkhurst and Plummer 1993). The Mg/Ca ratio of the samples ranged dissolved by fresh water would have a Mg/Ca ratio of 0.61. Thus, most of the samples (all but one between 0.516 and 0.623) are consistent with dissolution of do lomite. The sample with the lowest ratio of 0.417 may be affected by a thin bed of limestone. It is clear that dolomite has not been dissolved to equilibrium within ing with recent recharge. Evidence of modern recharge is also present as surfaceborne contaminants such as Na, Cl and NO 3 -N. Tritium does not decrease with depth, and it does not co-vary tion of the highly fractured and creviced nature of the Galena Dolomite aquifer and the vagaries of well drill ing. Available data from wells screened within the deep er Plattville Formation and St. Peter Sandstone show anomalously low concentrations of Cl(about 1 mg/L). This suggests that the upper part of the Galena Dolomite aquifer is more affected by surface-borne contaminants than deeper units, as would be expected. Nitrate-N concentrations range from below detection (< 3 N concentrations that exceed the background threshold concentration of 2.5 mg/L. Nitrate concentrations in the NO 3 -N concentrations were below detection (<0.04 mg/L) for TU concentrations less than 3 TU (Figure 5). Figure 5. Nitrate concentrations decrease with time within the karst aquifer, probably due to

PAGE 8

70 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE This shows a strong correlation between NO 3 N and or mixing with NO 3 -N free groundwater. Fluoride (F ) appears to be negatively correlated with tritium (Figure 6) suggesting that F is derived from rock-water interac tion and might serve as a surrogate/proxy for tritium in this area. Calcium and sulfate concentrations co-vary within the shallow aquifer (Figure 7). Given the abundance of py oxidation of pyrite is probably the main source of SO 4 2. 18 O data from precipitation in Jo Daviess County should be similar to that of precipitation in the Chicago area. Precipitation data collected in Chicago from 1960 to 1979 (IAEA 2014) provide a local meteoric water line similar to that of the Global Meteoric Water Line. The stable isotope data for the groundwater sam ples collected in Jo Daviess County fall along this line and indicate that little has occurred to the precipitation/ recharge (e.g., evaporation) prior to entering the Galena Dolomite aquifer. Combining these data with those of the ISWS database yielded a representation of the vertical distribution of selected ions. The vertical distribution of surface borne contaminants such as Cl and NO 3 -N is distinctive (Fig ures 8 and 9). Chloride concentrations are due, in a large part, to road salt contamination. The highest Cl concen Figure 6. Fluoride and tritium are negatively correlated, suggesting that F increases in con centration with time within the karst aquifer, probably due to rock-water interaction. Figure 7. Sulfate vs Ca in shallow wells reveals a linear relationship that is common in ground water. The linear relationship suggests a single source for the sulfate (i.e., the oxidation of pyrite). Figure 8. Chloride concentrations vs depth from private wells and springs (assumed depth of 1 foot) located in Jo Daviess County. Chloride concentrations are sporadic and el evated above natural background for shallow groundwater (15 mg/L) to depths of over 300 feet. This behavior is consistent with ground aquifer.

PAGE 9

71 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 trations are generally found in wells less than 240 feet concen trations are always below background levels (Figure 8). The vertical distribution of surface-borne contaminants is consistent with that of a fractured and creviced karst aquifer. Elevated NO 3 -N concentrations are probably that land use in Jo Daviess County is dominated by rowcrop agriculture. The relatively large NO 3 -N concentra tions near the surface are typical of agricultural areas, but as with the Clconcentrations, the sporadic nature karst aquifer. Overall, the chemical composition of relatively shallow groundwater within the Galena Dolomite aquifer is con tured karst aquifer whose mineralogy is dominated by dolomite. In particular, elevated NO 3 -N concentrations at depth are indicative of an open, oxygenated system Conceptual Hydrologic Model A preliminary conceptual model for the hydrogeology of the Galena Dolomite, based on previous studies in counties east of Jo Daviess County suggests that the northeast-southwest and northwest-southeast trending fractures/crevices and bedding planes constitute the greatest porosity for the Galena Dolomite aquifer (Ek berg 2008). Panno et al. (2015c) and Luman and Panno (2015) showed that these fracture/crevice systems are oriented more north-south and east-west in Jo Daviess County (Figure 10). Based on this investigation and Panno et al. (2015b), we have documented that the Galena Dolomite has its great est porosity within at least the upper 15 to 25 feet (4.6 to 7.6 m) of the dolomite with fractures and solutionFigure 9. Nitrate-N concentrations vs depth from private wells and springs (assumed depth of 1 feet) located in Jo Daviess County. These concentrations are sporadic and elevated above natural background for shallow groundwater (2.5 mg/L) to depths of slightly over 200 feet. As with Cl this behavior is con tured and creviced aquifer. Figure 10. Daviess County in September of 2012. The lines follow fractures and crevices that make up the porosity of the underlying karst aquifer (photograph from GoogleEarth).

PAGE 10

72 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE enlarged crevices ranging from < 0.4 in to 3 feet (< 1 cm found in the upper parts of crevices greater than about tigation. In general, the crevices provide a network of many crevices become narrower and range from less trending crevices tend to retain their widths of 3 feet or more with depth, thereby providing large conduits for the karst aquifer. Bedding planes may also provide path ways for groundwater movement. The effect of depth on the porosity and permeability associated with shear ed concentrations of NO 3 -N and Cl suggest the system is susceptible to surface-borne contamination to depths greater than 300 feet. It is reasonable to assume that relatively rapid recharge to the karst aquifer occurs throughout the county, but probably less so where Maquoketa Shale is present. In areas where Maquoketa Shale constitutes the bedrock surface, and where drain tiles are used to lower the wa ter table, relatively greater amount of recharge may dis charge to streams before entering the karst aquifer. Sink holes and macropores are present throughout the county and are locations of focused recharge. Sinkholes are not commonly seen due either to cultivation, which tends to obscure all but the very largest ones, or perhaps their natural scarcity. During dry periods, macropores (desic cation cracks) are ubiquitous and form easily due to the thinness of the soil and the depth of the water table (be low the soil-rock interface) (Panno et al. 2013). References American Public Health Association (APHA), the American Water Works Association (AWWA), and the Water Environment Federation (WEF). 1999. Standard methods for the examination of water and wastewater 20. Ekberg DW. 2008. Secondary porosity development in the karstic Galena/Platteville (Trenton/Black Riv er) dolomite on the Wisconsin Arch: Implications National Ground Water Association Ground Water Summit, Memphis, TN. March 30 April 3, 2008. Field MS. 1993. Karst Hydrology and Chemical Con tamination. Journal of Environmental Systems 22 (1): 1-26. Frankie WT, Nelson RS. 2002. Guide to the geology of the Apple River Canyon State Park and surround ing area of northeastern Jo Daviess County, Il linois. Illinois State Geological Survey Field Trip Guidebook. 2002B: p. 88. Green RT, Painter SL. et al. 2006. Groundwater con tamination in karst terrains. Water, Air, and Soil Pollution: Focus 6: p. 157-170. Hackley KC, Panno SV, Hwang HH, Kelly WR. 2007. Groundwater quality of springs and wells of the sinkhole plain in southwestern Illinois: Determi nation of the dominant sources of nitrate. Illinois State Geological Survey Circular 570: p. 39. Hem JD. 1985. Study and the interpretation of the chemical characteristics of natural water. U.S. Geological Survey Water-Supply Paper. 2254: p. 263. Heyl AV Jr, Agnew AF, Lyons EJ, Behre CH Jr. 1959. lead district. U.S. Geological Survey Professional Paper. 309: p. 62-63. Hwang HH, Panno SV, Hackley KC. 2015. Effects of land use change on groundwater quality of northeastern Illinois, USA: A historic perspective. Environmental and Engineering Geology, v. XXI, no. 2, p. 75-90. IEPA-IDOA. 2015. Illinois nutrient loss reduction strategy. Illinois Environmental Protection Agency and Illinois Department of Agriculture.http://www. epa.state.il.us/water/nutrient/documents/illinoisnlrs-public-comment-11-20-14.pdf IDEXX. 2013. Colilert. IDEXX Laboratories, Inc. Westbrook Maine. Accessed October 14, 2014. Available from: https://www.idexx.com/water/ products/colilert.html Kelly WR, Panno SV, Hackley KC. 2012. Concentra in waters of the Chicago, Illinois region. Journal of Environmental and Engineering Geology. Special issue on the hydrogeological effects of KA. 2010. Relations between sinkhole density and anthropogenic contaminations in selected carbon ate aquifers in the eastern United States. Environ mental Earth Science, 60: 1073-1090. Maas BJ. 2010. Investigation of spatial and temporal variations in water quality around Nora, IL [mas ters thesis]. Illinois State University, Blooming ton, IL. p. 128. McGarry CS, Riggs MH. 2000. Aquifer sensitivity map, Jo Daviess County, Illinois. Illinois State Geo logical Survey Open File Series OFS 2000-8i. 1: 62,500 scale. Nelson JW. 1995. Structural features in Illinois. Illinois State Geological Survey ISGS Bulletin 100. p. 144.

PAGE 11

73 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 ORiordan T, Bentham G. 1993, The politics of nitrate in the UK, in T.P. Burt, A.L. Heathwaite, and S.T. Trudgill, eds., NitrateProcesses, patterns, and management: New York, John Wiley and Sons. p. 403. Ostlund HG, Dorsey HG. 1977. Rapid electrolytic enrichment and hydrogen gas proportional count ing of tritium, in Low-radioactivity measurements and applications: Proceedings of the International Conference on Low-Radioactivity Measurements and Applications. October 6-10, 1975, the Hig Nakladetalstvo, Bratislava. Panno SV, Krapac IG, and Weibel CP. 1996. Groundwa ter contamination in karst terrain of southwestern Illinois. Illinois State Geological Survey Environ mental Geology Series. 151: p. 43. Panno SV, Weibel CP, and Li WB. 1997. Karst regions of Illinois. Illinois State Geological Survey Open File Series 1997-2: p. 42. Panno SV, Weibel CP, Wicks CM, and Vandike JE. 1999. Geology, hydrology, and water quality of the karst regions of southwestern Illinois and southeastern Missouri. Illinois State Geologi cal Survey Guidebook 27, 33rd Annual Meeting, Geological Society of America, April 22-23, 1999: p. 38. Panno SV, Hackley KC, Hwang, HH and Kelly WR. 2001. Determination of the sources of nitrate contamination in karst springs using isotopic and chemical indicators. Chemical Geology. 179 (1-4): 113-128. Panno SV, Hackley KC, Hwang HH, Greenberg S, Krapac IG, Landsberger S, and OKelly DJ. 2005. tion of the sources of sodium and chloride. Illinois State Geological Survey Open File Series 2005-1: p. 15. Panno SV, Hackley KC, Hwang HH, Greenberg S, Krapac IG, Landsberger S, and OKelly DJ 2006a. of Na-Cl in ground water. Ground Water. 44 (2): 176-187. Panno SV, Kelly WR. Martinsek AT, Hackley KC. 2006b. Estimating background and threshold nitrate concentrations using probability graphs. Ground Water 44 (5): 697-709. Panno SV, Kelly WR, Hackley KC, Weibel CP. 2007. Chemical and bacterial quality of discharge from on-site aeration-type wastewater treatment sys tems. Ground Water Monitoring Review 27 (2): 71-76. Panno SV, Luman DE. 2008. Assessment of the geology and hydrogeology of two sites for a proposed large dairy facility in Jo Daviess County near Nora, IL. Illinois State Geological Survey Open File Series 2008-2: p. 32. Panno SV, Luman DE, Kelly WR, Alschuler MB. 2013. The use of drought-induced crop lines as a tool of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Envionmental Impacts of Karst: p. 9. Panno SV, Millhouse PG, Nyboer RW, Watson D, Kelly WR, Anderson L, Abert CC, and Luman DE. 2015a. Guide to the geology, hydrogeology, archaeology, history and biotic ecology of the Driftless Area of northwestern Illinois, Jo Daviess County. Illinois State Geological Survey Field Trip Guidebook 42, 70 p. Panno SV, Luman DE, and Kolata DR. 2015b. Charac data in Jo Daviess County, Illinois. Illinois State Geological Survey Circular. 587. 29 p., 1 map, 1:62,500. Parkhurst DL, Plummer LN. 1993. Geochemical mod els, in Alley, W.M., editor. Regional ground-water quality, Chapter 9. New York (NY): Van Nostrand Reinhold. p. 199-225. Reed PC. 2008. Spring place names and historic data on springs, licks and selected water wells in Illinois. Schilling KE, Helmers MJ. 2008a. Effects of subsurface sheds: exploratory hydrograph analysis. Hydro logical Processes 22: 4497-4506. DOI:10.1002/ hyp: 7052. Schilling KE, Helmers MJ. 2008b. Tile drainage as drained watershed. Journal of Hydrology. 349: 291-301. Trowbridge AC, Shaw EW. 1916. Description of the Geological Survey Bulletin. 26: p. 233. USEPA 2014. Sodium in drinking water. U.S. Environ mental Protection Agency Contaminant Candidate. Available from: drinkingwater/dws/ccl/sodium. cfm.http://water.epa.gov/scitech/drinkingwater/ dws/ccl/sodium.cfm Weibel CP, Panno SV. 1997. Karst terrains and carbon ate bedrock of Illinois. Illinois State Geological Survey, Illinois Map Series 8, 1:500,000 Scale. Willman HB, Frye JC. 1970. Pleistocene stratigraphy of Illinois. Illinois State Geological Survey Bulletin. 94: p. 204.

PAGE 12

74 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

113 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 DYE TRACING THROUGH THE VADOSE ZONE ABOVE WIND CAVE, CUSTER COUNTY, SOUTH DAKOTA Abstract During the 1990s, in an attempt to better understand threats posed by surface developments overlying the cave, National Park Service staff at Wind Cave National Park in Custer County, South Dakota carried out a se overlying the cave. Wind Cave is located within the 100 m-thick Madison formation (limestone and dolomite), which in most locations is capped by varying thick nesses of the basal units of the Minnelusa formation (in termingled beds of sandstone, limestone, and shale). A variety of cave locations with dripping or pooled water tion. Transit times to the cave varied from less than six hours to as much as 4.8 years. Despite a variety of posi tive results, there appears to be little correlation between transit time and lateral or vertical distance from the in dye recovery curves in some locations, albeit stretched out over hundreds and possibly even thousands of days tion sites would quickly enter multiple sites in the cave system, and could persist for years. Introduction Wind Cave is located within Wind Cave National Park in the southern Black Hills in Custer County, South Da kota. The cave is an extensive three-dimensional laby rinth that as of this writing consists of more than 230 ki lometers of surveyed passages contained within an area of roughly 2.5 square kilometers (Ohms, 2015, personal communication). It is contained within the Madison for mation, which consists of beds of limestone and dolo mite of Mississippian age roughly 100 meters in thick ness in the Wind Cave area (Palmer and Palmer, 2008). Overlying the Madison formation in the vicinity of the cave is the Minnelusa formation, consisting of sand stone, limestone, dolomite, and shale (Strobel and oth the lower 30 to 100 meters of the Minnelusa formation, but a small number of cave passages are located beneath areas where the overlying Minnelusa formation has been largely or entirely eroded away. The only known com pletely-natural entrance to the cave is located in such an area, in a drainage known as Wind Cave Canyon (Figure 1). The contact between the Minnelusa and Madison forma tions is complex. The Madison was exposed soon after deposition, and a karst topography formed, with many caves, sinkholes, and other typical karst features. As a result, the base of the Minnelusa was deposited atop an solutional features up to tens of meters below the highest remaining reaches of the Madison (Palmer and Palmer, 2008). The area above Wind Cave is semi-arid with a meanannual precipitation of 45.8 cm/year (Ohms, 2012). Streambeds above the cave area have highly intermittent any obvious way. Wind Cave National Park was established in 1903, making it one of the oldest national parks in the United States. For the convenience of the visiting public, the park headquarters and its associated buildings, utilities, and parking lot were established near the caves en James Nepstad Figure 1. An aerial image of surface features and developments located in the vicinity of the entrance to Wind Cave.

PAGE 2

114 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE trance. During the 1980s and 1990s, park staff became increasingly concerned about the aging infrastructure ly ing above the cave. Water and sewer lines were known to be compromised, and untreated runoff from the 1-hect are asphalt (at that time) parking lot was entirely disap pearing within 15 meters of the drains, which are located in Wind Cave Canyon (Nepstad, 1997). How much of the cave was affected by these potential contaminant sources? How long did it take to reach the cave? Once there, how long was the contamination present? To answer these questions, it was necessary to trace wa ter from the surface to the cave. Although dye tracing for this purpose for decades, the circumstances at Wind Cave presented challenges. The surface overlying the cave lacks traditional karst features such as sinking streams or sinkholes, requiring ingenuity with respect and testing for it at resurgences or within cave streams, Previous Work In 1986, Marsha Davis, a graduate student at the Univer sity of Minnesota working under E. Calvin Alexander, conducted two dye traces at Wind Cave (Alexander and mine whether links existed between the parks visitor center parking lot and wet areas in the underlying cave. of Wind Cave Canyon in the parks picnic ground, along with roughly 1,000 liters of water. On August 16, 1986, 1076 grams of rhodamine WT (Acid Red 388, often re ferred to generically as rhodamine throughout this pa beneath the south parking lot drain located near the elevator building. At the time of these traces, there was no database doc umenting the location of wet areas within Wind Cave. Park staff and volunteer cavers, acting on personal recol lections, were able to inform Davis of a limited number dripping or pooled water. Frustrated at not being able to provide researchers with more information for such a critical study, in 1989 park staff instituted an aggressive inventory of cave features at Wind Cave. By the early 1990s, inventory data was already showing that Davis work had missed several key wet areas. Despite the fact that sampling was unintentionally in complete, Davis reported some positive results for both traces. Fluorescein was reported at survey stations NFP5 (Fairy Palace) and NM4 (Before Fairy Palace). Possible Following some background sampling in these areas in 1993-1995, park staff learned that Fairy Palace and Before Fairy Palace consistently produce false positives beneath the south parking lot drain, park staff suspected spills on the parking lot. All of this, combined with the scarcity of sampling sites and the short duration of sam pling (only about three weeks), created some skepticism Davis rhodamine WT trace also produced some positive results. One site, RS11 (Minnehaha Falls), produced dye tually tested positive for dye including RS13 (Mules Ears), UB10 (Caving Tour Drips), and C58 (Garden of Eden Drips). Dye may have shown up at additional sites, but once again monitoring was limited to a small number of sites in the cave and was cut off relatively early. gan showing up in the parks well, located within Wind tion site off the parking lot. Davis attributed the dye in into the Beaver Creek sink, located about 3.8 km north west of the well, but since the same dye had been used for both traces, an additional trace would be necessary to produce conclusive results. By 1992, the number of cave management staff at Wind possible to contemplate additional dye traces. Park staff, advised by Joe Meiman, a hydrologist who worked for Mammoth Cave National Park at the time, began to experiment with alternative methods of tracing wa of Davis dyes in the vicinity of her traces led them to experiment with Lycopodium spores. Given the limited if such a trace could be successfully accomplished, a nearly unlimited number of tracing elements would be available, since the spores could, in theory, be dyed with a large number of colors.

PAGE 3

115 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 An experimental trace was attempted in 1992, when roughly 200 grams of Lycopodium spores, together with cein into back in 1986. Sampling stations containing drips that were known to respond quickly to rain events on the surface, including some that Davis had believed ters were removed from the cave on an occasional basis and examined carefully with a microscope. No spores were ever detected, so the Lycopodium experiment was overlying soil and sediments in the epikarst. By coincidence, however, park staff had been back ground testing the water at many of these same sites for the presence of optical brightener, in the hopes of using that dye as a tracer element in the future. Untreated cot ton had been left in several of these locations, and these samples were observed under a handheld UV lamp after being removed from the cave. During the Lycopodium experiment, a cotton sample from NM4 (Before Fairy Palace) was found to be strongly positive for optical since all known potential sources of optical brightener (leaking sewage lines) were located down-dip from this location. Eventually it was recalled that the Lycopodium with four liters of water mixed with laundry detergent (to prevent clumping). Although the Lycopodium ex periment proved to be a failure, it is interesting that it indirectly led to a successful dye trace of sorts. By 1993, it appeared as though future traces would have tracer elements available at that time. A series of three rhodamine WT, and optical brightener. All of the traces were conducted considerably to the west of the 1986 traces. Over 60 sampling locations were set up, blanket ing nearly all of the known portions of the cave. Small samples of untreated cotton and activated charcoal bugs, consisted of a few grams of activated charcoal contained within a small screen pouch, were tested on at in order to gather background data. Once they were removed from the cave, the charcoal bugs were eluted with roughly 10cc of an elutant con hydroxide (mixed at 5:3:2), then tested with a Turner the spectra of the dyes to be used. Cotton samples were visually examined for optical brightener exposure with a handheld UV lamp. With the exception of the single cotton sample that tested positive for optical brightener described in the Lycopo cein or rhodamine in the vicinity of the 1986 work, the entire cave proved to be free of the presence of all three dyes. a drainage roughly overlying the Club Room portion of the cave, along with roughly 350 liters of water from a pumper truck. The water in the truck had been left ex posed to the atmosphere for four days to allow any re sidual chlorine from the parks water treatment system to dissipate. Roughly 125 liters went in prior to the dye, with the remaining 225 liters going in after the dye had been poured into the sink. depression above the southeast portion of the cave, along with roughly 1200 liters of chlorine-free water from a it into the system. Later that same day, 4000 grams of Tinopal CBS-X (Fluorescent Brightener 351), an optical of chlorine-free water into a small depression in a drain age overlying the southwestern portion of the cave. During the 1986 traces in Wind Cave Canyon, dyes were overlying Minnelusa formation. The dyes used in the the Minnelusa cap above the cave. The rhodamine trace above the Club Room was roughly 25 meters fur No longer dependent on visiting researchers for sam pling, these traces were sampled by park staff according to a far more rigorous schedule, and for a much longer period of time. Some sites were monitored on a daily or weekly basis for months, and the remainder of the sites were monitored at least quarterly. After two full years of monitoring and over a thousand samples being collected, no dyes were detected in the cave from any of the 1993

PAGE 4

116 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE glance, they nonetheless demonstrated a critically im of the overlying Minnelusa formation, which could ar up into the Minnelusa were either carried away laterally from the cave area, or diffused or delayed by the Min nelusa cap to such an extent as to render them undetect able for at least two years. developments its parking lots, buildings, and sewer and water lines were perched above the most hydrologi cally vulnerable portion of the cave, where the protective Minnelusa formation was thinnest. It was time to supple ment the results of the 1986 traces to get a better idea of the extent of this problem. The 1996 Traces In early 1996, additional background sampling demon strated that dye from the 1986 traces was either entirely absent or barely detectable at cave sites. park staff were able to identify over 60 locations suit able for sampling in areas including and surrounding Da of background sampling in the spring of 1996 (includ ing both grab samples and bugs), plans were made to face locations, and to better measure travel times and dye persistence. mine WT solution, representing 428 grams of pure dye, with enough water to simulate a one-inch rain event roughly 140000 liters for this particular drain. About a fortunately, due to the large quantity of water needed to accomplish this, there was no way to ensure the water was completely chlorine-free. Chlorine levels in the water were very low, however, as chlorination had been shut down for over 24 hours prior to opening the hydrant used to supply the water. As chlorine levels in the water are ordinarily very low to begin with, it likely destroyed very little of the dye. same small swallow near the park picnic ground that was used in 1986. As in the earlier experiment, roughly 1,000 liters of chlorine-free water from a pumper truck was and about 650 liters following it. As in the 1993 traces, samples were collected in 40ml Due to a limited budget and the very large number of samples taken over the course of these traces, vials were reused. To avoid contamination once a site tested posi vials in the cave (emptied and rinsed prior to each use) were kept at each sampling site to avoid cross-contami nation. Activated charcoal bugs were placed at each site ing dye arriving at quantities too low to measure in grab samples. tween instruments. The Rhodamine WT Trace It had been known for years that a site in the cave near survey station BI25 (Upper Minnehaha Falls), located almost directly below the parking lot drain, consistently Figure 2. The locations of the 1986 and 1996 dye injection points on the surface above Wind Cave. The same dyes were injected in the same locations utilizing the same methods for both traces.

PAGE 5

117 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 responded to rain events within a matter hours. It was be lieved that the dye would make its initial appearance in the cave here, so an auto-sampling device was installed to automatically collect samples on an hourly basis fol 0.07 parts per billion (ppb), and one hour later it was nearing 1 ppb. By midnight that evening, it was ap proaching 20 ppb. Already a site had been discovered that received dye several times faster than the site that BI25 would go on to become the most carefully docu mented site during the life of this dye trace. The site is easy to access, and was sampled on an hourly basis for weeks, on a nearly-daily basis for over a year, and sev thor left the park in October 1998. The rhodamine results illustrated in Figure 3. A plastic tarp was arranged beneath the site to capture rates, which proved to be highly variable. Following heavy rains during the spring and summer, steadily streaming or heavily dripping water will pour in from a dome above the site at up to 1700 ml per minute. Flow rates quickly reach a peak within a day of the rain event, lowing. The steep peaks in the drip rate in Figure 3 all relate to heavy rain or melting events on the surface. During cold, dry stretches in the winter, the drip rate slows dramatically, and on some occasions even ceases completely. Rhodamine levels rose steadily at the site for a week, rising to 87 ppb before declining to 43 ppb 11 days after tion had come and gone, and the site would continue to rapidly decline over the coming weeks. On the twelfth cant rain event on the surface occurred. As expected, drip rates quickly rose, but dye levels surged even faster, peaking at 250 ppb two days later before starting yet an other decline. Another rain event occurred on the sev enteenth day, creating another rapid surge in dye levels that peaked at 380 ppb within 48 hours. This reading on the nineteenth day proved to be the true peak at this site. Although spikes in dye concentration continued to occur Figure 3. the most carefully studied site in the 1996 traces.

PAGE 6

118 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE immediately following subsequent precipitation or melt ing events on the surface, the magnitude of the spikes They become barely noticeable after about 300 days. Despite the spikes which occurred early in the trace, the overall shape of the concentration curve strongly resembles a traditional positively skewed, bell-shaped curve that is steeper on the rising limb than on the fall ing limb, (Mull, et al, 1988), except that instead of a time scale measured in hours or days, the time scale here stretches over multiple years. The persistence of the dye in the system was rather re concentrations were still double the background levels tence perfectly illustrates the threat posed by having the parking lot perched above the cave in this area. If con taminants in parking lot runoff behave like the rhoda mine used in this trace, then relatively minor spills on the parking lot could impact cave resources for years follow ing an incident. BI25 was an interesting site, and it received more atten tion than any other sampling location during the 1996 rhodamine trace. But it was by no means the only site in the cave where positive results were found. Table 1 documents all of the sites that received measurable quan tities of rhodamine, together with the number of days it took for the leading edge of the dye to reach those sites and the number of days it took dye concentrations to reach their peak. There appears to be no clear correlation between the emergence of dye at a site, its peak concentration, or the time it took to peak and any of the other variables tive well before other sites that were nearly beneath it. Likewise, vertical depth is a poor indicator of when dye will arrive or at what concentration it will peak, as is the drip rate associated with a site. Rather, travel times and peak concentrations appear to depend on highly local but unmeasurable variables, such as the sites proximity to the site), or the mixing characteristics of other waters encountered in transit. The concentration curves for many of these other sites also resemble the classic shape of curves associated with traditional dye traces in karst aquifers. Other sites exhibit concentration curves that are anything but tra ditional. See Figure 4 for examples of both. Concentra tions at C47 (Assembly Room), for instance, rise abrupt course of an entire year before increasing once again. One cannot even be certain the peak concentration had been measured after nearly 1000 days, when sampling peak concentration was measured at either C53 (Escape Route) or C53A (Silent Lake) after more than two years of sampling. Years of additional sampling may have been necessary in order to capture an entire curve at these sites. Rhodamine Recovery trace is challenging. In a traditional dye trace in a karst Table 1. Cave sites that tested positively for rhodamine WT following the July 1996 injec tion. Also listed are the vertical and horizontal distances from the injection point, days fol the peak concentration measured, and the number of days following injection that the peak concentration occurred. Missing data for RS10, BI12, and UYA4 are due to very infre quent sampling at those sites. Site Depth (m) (m) Days to Cave Peak (ppb) Days to Peak BI25 31 12 1 380 19 RS11 34 28 2 235 35 BI9 56 15 2 290 19 RS13 31 31 17 9 56 SA1 40 87 23 23 93 SA1A 42 85 23 11 57 UB10 57 54 44 6 160 RS10 34 28 83 3.5 BI12 46 21 99 4.5 C53A 34 40 114 0.96 697 UU10 47 8 155 1.5 245 SK3 36 49 232 8.5 380 SK3F 38 43 232 2.6 629 C53 37 40 233 3.39 725 SK6 41 62 246 0.25 413 C47 57 33 414 0.41 957 UYA4 48 111 414 0.12

PAGE 7

119 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 landscape, where dye is poured into a stream at an insur gence, the researcher can at least be relatively certain that Although park staff could be certain about the amount of dye poured onto the ground in these Wind Cave dye trac es, it is impossible to know how much of that dye might have been adsorbed by soils and sediments at or near of roughly 150-200 square meters before gradually sink this was deemed acceptable, since the experiment was designed to model the fate of parking lot runoff and that was how parking lot runoff in the area behaved, it none completely resolve. We can calculate roughly how much dye reached the various sites in the cave, but the fate of the dye beyond the sum total of all sites will remain a traveled through BI25. This site experienced not only times exceeded 1700 ml/min. Other sites had far lower To arrive at an estimate for the quantity of dye captured at BI25, certain assumptions had to be made. It was Figure 4. Sampling sites such as RS11 and RS13 had traditional-shaped dye curves, albeit stretched out over periods of hundreds of days. Other sites, such as C47 and C53A had more complex curves, and may not have peaked before sampling concluded.

PAGE 8

120 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE samples dye concentration could be viewed as the mean concentration for that day. Likewise, if more than a day passed between samples, then the values for the most re cent sample were used for the day(s) not sampled. While this inevitably resulted in over-calculating the daily mass declining, the opposite surely happened on numerous oc casions as well. It is assumed that the errors thus intro duced into the estimate will cancel each other out, since the number of samples collected was quite large. Data If this assumption is correct, then roughly 12.06 grams of rhodamine WT (dye, not dye solution) traveled through illustrates the temporal distribution of the dye recovery, with the solid red line representing the running total of rhodamine received at the site by a given day. Roughly the site markedly slowed. The total stood at about 9.9 grams at the beginning of the spring of 1997, and rose to 11.9 grams by the beginning of the winter of 1997-1998. During the spring, summer, and early fall of 1998, the total slowly rose to 12.06 grams. Although dye was still arriving at low concentrations at that time, extrapolating ing above 12.5 grams. from BI25. Due to its location, and due to the fact that it tracked so closely with data from BI25 (they peaked at precisely the same time), it is strongly suspected that the through BI25. Although it peaked at a lower concentra tion, this could be due to mixing that occurs after the water leaves BI25. Unless future evidence indicates oth erwise, it is probably not appropriate to consider this site when accounting for dye recovery. Figure 5. Cumulative totals for dyes recovered at BI25 by a certain period in time. Note the rhoda does not. Graph represents total dye recovered, not total dye mixture recovered.

PAGE 9

121 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Another site where very high concentrations of dye were encountered was RS11, Minnehaha Falls. Although this ute at RS11, compared to an average 226.9 ml/min at ing the site to be measured, however. It is estimated that the total amount of water entering the site at any time is A total of 309 samples from RS11 were collected and rates and dye concentrations are the mean values for that particular day, then a total of 0.18 grams of rhodamine trace. If our assumption is correct that twice as much water enters the site as a whole, and if that additional water had the same concentration as the water that was is possible that a total of up to 0.36 grams of rhodamine WT entered the cave at RS11 during this time period. The remaining sites were sampled far less frequently, large gaps between sample collections, the author be lieves that recovery estimates are not precise enough to cite, but even considering simple order of magnitude ba sis estimates, most of these sites can account for a few hundredths of a gram total at best, with the possible ex ception of SA1, which may be able to account for a few tenths of a gram. In all, it is fairly safe to say that less than 13 grams of dye can be accounted for from the rhodamine WT trace, planation for the remaining dye was that it adsorbed to soil, sediments, or bedrock above the cave. It is also possible that the dye passed into a series of fractures un connected to the cave, or too small for humans to enter, and thus passed to the water table unobserved. Addition ally, it is possible that the dye was intercepted by lateral carried away from the sampling area, and possibly even the cave area. Finally, it is possible that the rest of the dye traveled through unexplored cave passages. This last possibility is the least likely. The sample area is located in close proximity to the elevator entrance, and as a re sult is one of the most thoroughly explored portions of the cave. More than likely, the fate of the missing rhodamine was a combination of the other three possibilities. Rhodamine WT is known to adsorb to soils and sediments (Smart, occur at a discrete point. Also, the Madison formation is highly fractured in the cave area, as evidenced by the incredibly complex nature of Wind Cave itself. It is not trace began to emerge. The Fluorescein Trace 1996, a total of 25 sites were repeatedly sampled in the northern portion of the cave. These sites included NFP5, cein trace, and NM4, which Davis also believed tested for optical brightener following the Lycopodium experi ment. Given this past history of the area, it was assumed that a number of sites would receive measurable quanti ties of dye in this trace. After 15 months of exhaustive sampling, no sign of the tion point is very near the Minnelusa/Madison contact. Due to that proximity, it was believed that the dye would travel in a mostly vertical fashion, with little if any lat eral movement. While some of it may have traveled vertically via fractures too small for humans to enter, in beneath Wind Cave Canyon. Despite our early assumptions, background sampling ground grab samples and bugs were tested for all dyes that were going to be used, regardless of how unlikely we believed a particular dye might be encountered at a particular site in the future. Between the background sampling performed in June 1996 and routine sampling that occurred during the course of the 1993 dye traces, but always hovered at levels equivalent to 10 14 ppb (Figure 6). els at BI25 and RS11 remained at familiar levels. Lulled into complacency by the initial readings and desperate to save time due to the crush of data coming in (we col

PAGE 10

122 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE tween day 19, when it was still within the background ly doubled at both sites. Levels slowly rose during the course of the fall and winter, and eventually peaked at 40 ppb at BI25 and 38 ppb at RS11 the following spring. anything we had seen. For more than two years, con cycle. Levels were highest during the winter months, as two separate components, one very local, another with a more distant origin. Such unexpected results cried out cein that was being measured? If so, could it possibly be order. After elution, charcoal bugs retrieved from the luminated with a bright beam of light, was visible even through the pink cloud produced by the rhodamine at the sites. Later, we discovered that the same could be done with raw samples of water from the site. Joe Meiman at entering the cave. ny, who, after a bit of arm twisting, revealed the rough Figure 6. neath Wind Cave Canyon to arrive at this location in the cave.

PAGE 11

123 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 have come off the parking lot. The only other location area was a considerable distance down dip from BI25 and RS11, near the southeast corner of the cave. been present at these levels in any of the samples we had tested over the course of the previous three years. In light of all this, and considering the fact that Davis also re to BI25 is 449 meters. endipitous discovery of optical brightener at NM4 dur ing our Lycopodium experiment seemed to provide inde the picnic ground and NM4. If that connection exists, it did not exhibit itself during the course of this trace, suggesting it is sporadic in nature. Results from other (Aley, 2012). An examination of local precipitation data showed that 1986 and 1993 were both much wetter than normal, with an annual total of 58.2 cm in 1986 and 69.9 cm in 1993 (Ohms, 2012). Annual precipitation for 1996 was a more average 47.8 cm. Further, precipitation in the two 30 year average for the area (7.3 cm, compared to the average 14.5 cm). It is possible the system wasnt active enough to move dye to NM4 in 1996. same time it arrived at BI25, further leading us to be lieve that most of the water arriving at that site is water that has previously traveled through BI25. The concen trations detected at BI9 consistently lagged behind the concentrations measured at BI25 by 2-3 ppb. The consis tency of the data between these two sites, together with started to arrive by that date. higher than any background sample from the site, and given the consistency of the 2-3 ppb lag behind BI25 readings, suggests that BI25 would have been reading 16-17 ppb. The increase is small and inconclusive, but it does suggest that it is at least possible that the leading Wind Cave Canyon was traveling about 15-16 meters per day. No other sites that were sampled could conclusively be readings, although a small number of sites, including SA1 and UB10, eventually rose slightly above back ground levels. Charcoal bugs left for months at these very weakly. Fluorescein Recovery ence induced by a substance naturally in the water, or both BI25 and RS11 in small but measurable levels prior ings were subtracted from readings measured after dye began to arrive in order to arrive at an estimate for the When calculated in the same manner as the rhodamine WT totals discussed above, it appears that 3.704 grams the same time period was 0.071 grams. As with the rho damine calculations for this site, if we assume that twice as much water is passing through the RS11 area as we were able to capture and measure, then this total could be as high as 0.142 grams. This means that about 3.846 cein recovery totals is that while the rhodamine total monitoring (since the sites that were still positive when sampling ceased were only weakly positive, and mostly increase had sampling continued. Figure 7, which docu this beautifully.

PAGE 12

124 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Epilogue ever emerged in measurable quantities at the park well. Grab samples were regularly collected from a spigot in the well house that draws water from the system prior to chlorination, and charcoal bugs were placed in a small piece of PVC pipe that was exposed to water continu or bugs ever tested positively for any dye. This seems to lend validity to Davis assertion that the rhodamine she found in the park well in the 1980s was earlier. During the 1990s park staff also discovered a in Beaver Creek as measured by a USGS gaging station and the rise and fall of the water table in the Lakes portion of the cave. A lag time of roughly two months was observed between a rise in Beaver Creek same amount of time that passed between Davis rhoda (which is located about the same distance from Beaver Creek as the Lakes area of the cave). tions had grown so infrequent that we could no longer retrieve any bugs that were still present in the cave in those areas. tion of the rhodamine WT in 1993, the collection site mine trace, and surprisingly, this sample turned out to be weakly positive for rhodamine roughly 0.14 ppb. Additional grab samples collected 51 and 89 days later both produced readings of 0.36 ppb. Although these are the only three samples collected during this time, it ap pears we inadvertently caught the near-arrival of the dye during the April 1998 sampling. Figure 7. two years of sampling. Graph represents dye recovered, not dye mixture recovered.

PAGE 13

125 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 preting results, and served as our lead cheerleader when positive data began to pour in. I am deeply grateful for all of his assistance. References Alexander EC, Davis MA, Alexander SC. 1989. report. National Park Service Contract CX-12005-A047. Aley TJ, Kirkland SL. 2012. Down but not straight 27: 193-198. Nepstad JA. 1997. Surface developments above Wind Cave studying the impacts. In: Stitt RR, editor. Proceedings of the 1997 Karst and Washington. p. 137-140. Ohms MJ. 2012. Water resources of Wind Cave National Park 2012 update. National Park Service report. Palmer AN, Palmer M. 2008. Field guide to the paleokarst of the southern Black Hills. In: Saskowski ID, editors. Karst Waters Institute Conference Karst From Recent to Reservoirs. Karst Waters Institute Special Publication 14. Smart PL, Laidlaw IMS. 1977. An evaluation of some Research 13 (1): 15. Strobel ML, Jarrell GJ, Sawyer JF, Schleicher JR, Fahrenbach MD. 1999. Distribution of hydrogeologic units in the Black Hills, South Dakota. U.S. Geological Survey Hydrologic Investigations Atlas HA-743. since no dye was recovered in any of the sites between liest explanation is that this represents a weak positive from the 1993 dye traces. Unlike the parks developed tion and initial recovery during a dye trace has ever been measured, this author is not aware of it. The discovery of dye at CI17 provided us with an addi tional insight. We had often wondered how long an acti vated charcoal bug could be left at a drip site in the cave and remain active. The charcoal bug that we retrieved in April 1998 had been in the cave since July 1995. If we are correct in our belief that we discovered the dye relatively soon after its initial appearance in the cave, then a charcoal bug exposed to dripping water at this site could still adsorb dye after being in the cave for nearly three years. Complete turnover of the parks cave management staff occurred during 1998, and as a result, regular sampling came to an end by late October of that year. Sampling was sporadic during 1999, and continued at some sites on an occasional basis into 2000. Following the collec tion and analysis of more than 4,400 grab samples and hundreds of bugs throughout the 1990s, this exercise in Acknowledgments It requires discipline to stick to a sampling schedule for as we learned, negative results in such experiments can produce valuable information as well, and surprising re sults can sometimes be attained years after all hope is gone. The author is grateful for the exceedingly dedi cated cave management employees both paid staff and volunteers who assisted with data collection and anal Bonnie-Ann Burnett, and Tonya McDermott. Marc Ohms continued collecting samples after the au thor left Wind Cave National Park in October 1998, and has provided very useful information and advice in the intervening years. His patience and diligence in assist bits of valuable data out of this long-term experiment. Finally, NPS hydrologist Joe Meiman provided critical advice and training in study design, and also provided out on numerous occasions when we had trouble inter

PAGE 14

126 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

289 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 EVALUATION OF CAVITY DISTRIBUTION USING POINT-PATTERN ANALYSIS Abstract City Municipality (ADM) underlain by soluble rocks such as gypsum, calcarenite, or mudstone. This is es pecially critical if they are located at a relatively shal low level and are likely to cause settlement or sudden soil collapses. The Gachsaran Formation, which is com posed of interlayered mudstone and gypsum, underlies all of the ADM and is known to be vulnerable to cav ity formation in the area. Information associated with cavities was cataloged and reviewed based on available data from an existing geotechnical borehole database maintained by the ADM. Cavity data obtained from distributions based on the following factors: lithology, cavity, and depth to bedrock. All cavities were grouped into geographic clusters and lithological clusters for point-pattern analysis. Most cavities (87 percent) occur in mudstone or gypsum, or at an interface between these two rock types, which compose part of the Gachsaran curred in the Shakhbout City area hence pattern analysis including average nearest neighbor analysis, Morans I for measuring spatial autocorrelation, and G-statistics for measuring high/low clustering were conducted in this area using spatial statistics tools in ArcGIS. Aver age nearest neighbor analysis and MoransI show that cavities are strongly clustered in this area with a high relatively high values of depth to cavity, depth to bed rock, and number of cavities per borehole. No highly clustered large cavities were detected by the General Gnineth nearest neighbors were determined for cavities in different lithological materials and geographical clus ters. Outcome of these spatial correlations and statistical analysis can be used to conduct risk assessment and the probability of occurrences of cavities in the future. Introduction City Municipality (ADM) underlain by soluble rocks such as gypsum, calcarenite, or mudstone. This is es pecially critical if they are located at a relatively shal Raghav Ramanathan RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, raghav.ramanathan@rizzoassoc.com Yongli Gao Center for Water Research, Department of Geological Sciences at The University of San Antonio, One UTSA Circle, San Antonio, TX, 78249, USA, yongli.gao@utsa.edu M. Melih Demirkan RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, melih.demirkan@rizzoassoc.com Bulent Hatipoglu RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, bulent.hatipoglu@rizzoassoc.com Mazen Elias Adib Town Planning Sector, Abu Dhabi City Municipality, P. O. Box 263, Abu Dhabi,UAE, Mazen.Adib@adm.abudhabi.ae Michael Rosenmeier RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, michael.rosenmeier@rizzoassoc.com Juan J. Gutierrez RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, juan.gutierrez@rizzoassoc.com Hesham El Ganainy RIZZO Associates, 500 Penn Center Boulevard, Pittsburgh, PA, 15235, USA, hesham.elganainy@rizzoassoc.com

PAGE 2

290 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE low level. Sometimes halite, gypsum, or carbonate-rich unconsolidated crystalline formations dissolve when the groundwater condition changes, especially if originally unsaturated (Tose and Taleb, 2000). When the sedi ment layer starts losing material due to these reasons, ing to bigger voids. Unconsolidated sediments can dis place into cavities if the roof of the cavity collapses or is punctured by human activities such as drilling. Loss of unconsolidated sediments, due to washing out or leaking into cavities, is also a common problem in the region. These types of problems are likely to cause settlement or sudden soil collapse. Likewise, unconsolidated soil sediments are weakly cemented by soluble materials and can experience settlement or collapse if this weak bond is destructed by mechanical or chemical factors, such as excessive pressure or wetting (Hausmann, 1990). Evaluation of the lithologic sections indicated that exca vations periodically intercepted open voids in the mud commonly reported on drilling logs. Borehole data in dicated that most of these cavities occur close to the top of the bedrock often at the interface between the over ing mudstone and gypsum. This formation of cavities is believed to be formed by groundwater movement along the interface of the mudstone and gypsum layers forming cavities that are more vulnerable to collapse in the vicin ity of the top of rock (Farrant et al., 2012a). Shakhbout City and Zayed City areas within the ADM. Most notably, Tose and Taleb (2000) developed a ground bout City and Zayed City areas. Although ostensibly designed to identify generally adverse subsurface condi risk with shallow, less than 20 m or 66 ft below ground surface [bgs], cavity distributions (heights) and so-called broken subsurface strata extents, as inferred from ex tensive geotechnical boring and geophysical survey data. titioners for a discontinuous 44-plot area located within map risk distribution was based solely on cavity (void) density and depth below ground surface. An overall in which voids were determined to be located more than 16 m or 52 ft below the ground surface. In contrast, (including voids and water-loss instances) were largely within any class (low, medium, or high) was in turn de low-risk conditions were assigned to individual plots with no more than three deep cavities. Moderately lowrisk conditions were similarly assigned to plots with less than 10 (more than 3) deep cavities. Slightly higher but still generally low-risk conditions were assigned to plots containing abundant (more than 10) but deep cavities. cover thickness, overburden lithology and mechanical characteristics, hydrogeologic conditions, and ground water geochemistry due to lack of data availability. Study Area Abu Dhabi is located in the stable cratonic region of the Arabian Plate. The study area covers an area of 11,000 square kilometers (4,250 square miles). It includes the mainland urban area of Abu Dhabi in addition to the coastal islands. Based on data availability the extent of study area was chosen as shown in Figure 1. The coastal proximately 35 m (115 ft) above sea level to the east and southeast across an arcuate escarpment trending from Mafraq in the south to Al Shahama in the north (Price et including many of the coastal islands is reclaimed land ly in places in an uncontrolled way over pre-existing, coastal barrier and supratidal sabkha sediments. The sedimentary sequence underlying the region consists m or 26,250 ft (Al-Jallal and Alsharhan, 2005). Above this are extensive Holocene aeolian deposits forming the to periodic inundation and evaporate depositions, domi are commonly formed in arid shallow-shelf environ ments, and are formed in response to two environmental accumulation in a lagoon, or by a combination of both processes (Evans, 1970). Most of the solution cavities occur in the Gachsaran Formation part of the Neogene system (Alsharhan and Narin, 1997). The Gachsaran Formation is a thick evaporitic basinal succession that was deposited in a shallow marine/brackish setting with input from a nearby land source indicated by plant mat

PAGE 3

291 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 ter. It is well known from offshore oil wells, but is only poorly exposed onshore in the Abu Dhabi Area where it is recorded in numerous temporary excavations and boreholes that have penetrated up to 100 m (328 ft) of in terbedded mudstone and gypsum (Farrant et al., 2012a). The Gachsaran Formation is covered by the Abu Dhabi Formation along the coast, and by younger Miocene and Quaternary sediments inland. Small exposures occur around Mafraq, Shakbout City, Shahama, Al Bahya, and along the foot of the Dam Formation escarpment around the Al Dhafra Air Base at Al Maqatrah (Farrant et al., 2012a, b). In many exposures and borehole logs the gyp sum layers have been shown to contain well-developed observed at or close to the surface of the bedrock, par and the underlying mudstone and gypsum (Farrant et al., 2012a). Figure 2 shows the extent and depth to Gachsa ran Formation within the study area. Abu Dhabi Cavity Characteristics and Distribu tion The ADM maintains a borehole database consisting of around 21,000 geotechnical borings. This borehole da tabase is called Geotechnical Information Management System (GIMS). The GIMS for Abu Dhabi City supports a consolidated geotechnical database in accordance with internationally accepted standards. For this study, the GIMS borehole dataset was queried for string drops (also loss of water, which are indicators of voids or cavities within a given boring. Since these are only indicators of the presence of subsurface cavities and voids for the purpose of this study it is assumed that these indicators are in fact subsurface cavities and voids adopting a con servative approach. A detailed geophysical and ground exploration investigation should be performed for con database was developed to manage spatial data acquired during the data collection process of this study. drop, free fall or loss of water the GIMS borehole data. However, some boreholes may encounter multiple string drop, free fall, or loss of water features. The top ard assessment for this phase of the GGHIP. Therefore, a total of 729 cavities nearest to the surface for each borehole were selected for analysis. Overburden thick Figure 1. The extent of the study area shown here was decided based on the availability and spatial distribution of data within the ADM. Figure 2. The areal extent and vertical depth of the Gachsaran Formation below ground surface level.

PAGE 4

292 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE fractures of the soluble rocks through suffosion or piping processes. Therefore, the depth to the Gachsaran Forma assessment. Other reasons for occurrence of voids in in soluble materials could be due to drilling activities due to weak material collapsing. Figure 3. Spatial distribution of cavities in the Abu Dhabi Municipality. Table 1. Cavity distribution in different litho logical materials. ness or depth to bedrock, depth to cavity, cavity den of cavities occurred included the southeastern Capital District, the Abu Dhabi International Airport, and the Al Falah areas. A small number of cavities were sparsely distributed in other areas. All cavities were grouped into geographic clusters and lithological clusters for subse quent statistical analysis. Figure 3 shows the spatial dis tribution of cavities in the ADM. The occurrence of cavities in different types of lithologi cal materials in the ADM area is shown in Table 1. Figure 4 shows a chart representation of Table 1. Most cavities between these two rock types, which compose part of the Gachsaran Formation (Ga). The Gachsaran Forma tion is a thick evaporitic basinal succession consisting of carbonates and evaporites, with marls and thin limestone (Bahroudi and Koyi, 2004). It does not form natural out crops at surface (Farrant, A.R., et al., 2012). However, the dissolution of carbonate and evaporites within this formation causes subsurface voids formed in the ADM area. Even though some voids occurred in non-soluble rocks such as siltstone and sandstone, they were most likely associated with the dissolution of Gachsaran For mation underneath. Since the Gachsaran Formation is so extensive in the ADM area soil and sediment above the Gachsaran Formation can migrate down into voids and Figure 4. Cavity distribution in different litho logical materials. Chart showing the occur rence of cavities in different types of litholo gies prevalent in the ADM.

PAGE 5

293 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Mudstone has the second highest distribution of cavities among the other lithological material in the ADM. Even though gypsum is known to be more soluble than mud stone, mudstone layers generally have low compressive strength compared to gypsum layers. The mudstone in compressive strength as low as 100 kPa which is less gypsum core samples tested. Three factors can be attrib uted to weathering and cavity formation in the mudstone: nature of the encountered mudstone, given the fact that gypsum interbedded within the mudstone (Canton et al., 2001). Point Pattern Analysis Many attempts have been made in the past to study pat terns among point data in various natural systems. Clark and Evans (1954) and Thompson (1956) developed a nearest-neighbor analysis (NNA) method which has been used in many research areas. Another study (Drake ation of sinkholes in Mendip, England by comparing the bors between the two generations of sinkholes. A comprehensive investigation of cavity distribution is cavities are clustered or randomly distributed. Depth to or thickness, and distributions of cavities in different geographic and lithological clusters help to character analysis is the study of the spatial arrangement of point features in two-dimensional space (Gao, 2002). ArcMap can be used to determine clustering or level of disper the study area. The Average Nearest Neighbor tool mea sures the distance between each feature centroid and its nearest neighbors centroid location. It then calculates the average of all these nearest neighbor distances. If the average distance is less than the average for a hypotheti cal random distribution, the distribution of the features The Spatial Autocorrelation (Global Morans I) tool measures spatial autocorrelation based on both feature locations and feature values simultaneously. Given a set of features and an associated attribute, it evaluates whether the pattern expressed is clustered, dispersed, or Low Clustering (Getis-Ord General G) tool measures the concentration of high or low values for a given study area. The High/Low Clustering tool is most appropriate when there is a fairly even distribution of values and un expected spatial spikes of high values need to be identi Results A pattern analysis usually demonstrates if a distribution pattern is random, dispersed, or clustered. In addition, a distribution pattern containing clusters of high or low section discusses the results of the pattern analysis per formed on the cavity dataset. Pattern Analysis in the Shakbout City Area Area pattern analysis, including average nearest neighbor were conducted in this area using spatial statistics tools in ArcGIS. Figures 5 through 11 illustrate results of point pattern analysis of cavities in the Shakbout Area. Aver age nearest neighbor analysis (Figure 5) and Morans I (Figure 6 ) show that cavities are strongly clustered in with relatively high values of depth to cavity, depth to Figure 5. Shakbout Area Average nearest neighbor analysis indicates a clustering pattern based on the p-value.

PAGE 6

294 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE bedrock, and number of cavities per borehole (Figures 8 and 9). No highly clustered large cavities were detected by the General G statistics (Figure 7). tion of Cavities Depth to Gachsaran Formation, depth to bedrock, depth in three geographic clusters including the Shakbout City and southeastern Capital District, the Abu Dhabi Inter national Airport, and the Al Falah areas. Most cavities occurred in areas surrounding the Shakbout City area, including the southeastern Capital District area, and these areas represent typical geological settings for the cavities within the Shakbout City and the southeastern Capital District area are discussed in this paper. Cavity Figure 6. Shakbout Area Morans I with cavity size indicates a clustering of cavities with similar size. Figure 7. Shakbout Area General G with cav ity size indicates large cavities are not clus tered. Figure 8. Shakbout Area General G with depth to cavity indicates cavities occurring at similar depths are clustered. Figure 9. Shakbout Area General G with depth to bedrock indicates high clustering of cavities at certain depths to bedrock.

PAGE 7

295 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 thickness of each cavity or the distance between the top and bottom elevations of each cavity. Depth to Gach saran Formation, depth to bedrock, and depth to cavity distribution (Figure 11) is more random similar to the Poisson distribution, which is consistent to results of the General G statistics (Figure 7). Nearest Neighbor Analysis bors were conducted for cavities in different lithological materials and geographical clusters. Figure 12 demon strates a histogram of the distance to the nearest cavity within the Gachsaran Formation. The median distance increasing within the Gachsaran Formation (Figure 12). For nearest neighbor analysis of the entire ADM area, some cavities may have a nearest neighbor that lies out side of the district boundaries or areas without detailed borehole data. This phenomenon is called edge effect. To avoid edge effects, cavities were evaluated for proximity to district boundaries or areas without enough borehole data. Some isolated cavities are very far away from the main populations. These areas have not been fully in vestigated and some cavities might exist, but may not be mapped or recorded in the database. Three kinds of cavi ties were removed for NNA: cavities that have nearest neighbors outside of the clustered area whose distance Figure 10. Cavity distribution in relation to depth to Gach saran Formation, depth to bedrock and depth to cavity follow a normal distribution. Figure 11. Cavity distribution in relation to cavity size fol lows poisson distribution indicating a random spatial distirbution.

PAGE 8

296 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE different from those in the clustered area, cavities whose DNN are greater than the distance to the boundary of borhood has not been fully investigated for cavities by boreholes. The overall DNN distribution of all cavities does not follow Poisson, Normal, or Log-Normal dis tributions. However, the distribution of the DNN for all cavities more or less follows normal distribution once DNN is greater than 160m. A decision tree model based on cavity characteristics and sessment in the ADM area (Figure 13). The decision tree includes characteristics of bedrock geology, depth to the distances to the nearest cavities in the ADM area. The primary controls on cavity development are lithostrati graphic position or bedrock geology and depth to the soluble Gachsaran Formation. Conclusions cavity formation in comparison to insoluble rock, as as mudstone can be attributed to its weak compressive strength properties (Canton et. al, 2001). Out of the various soluble bedrock formations within the ADM the Gachsaran Formation has the highest likelihood for cav ity formation. Cavities are formed either due to the dis Figure 12. (a) Distance to the Nearest Cavity and (b) Median Distance to the Nth Nearest Cavity within the Gachsaran Formation. Figure 13. The decision tree model that was developed for hazard assessment related to cavities in the ADM area.

PAGE 9

297 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 solution of mudstone or gypsum layers in the formation, between these two lithological materials. It is also evi dent that more cavities are likely to be formed in regions with shallow bedrock than in regions with relatively deeper bedrock. Based on the cavity distribution in relation to depth to Gachsaran Formation (Figure 10) it is more likely that cavities are formed in locations where the Gascharan formation occurs at a depth of less than 30 m (100 ft) below ground surface. Similarly, based on the histogram ure 11), it is statistically more likely that a cavity prone 10 ft thick) cavity population tends to form in highly concentrated important role in cavity distribution and formation. Formation, depth to cavity, cavity density and distances to the nearest cavity in the Abu Dhabi Municipality. This decision model, when compared with earlier, elemen cavities and supplements the existing cavity distribution maps when comparing the depth and resolution of evalu formation of cavities, such as groundwater chemistry and changes to landscape and groundwater, were not consid uisites for determining regions that are more susceptible to forming cavities, this decision process can only pre dict future occurrence of cavities with low accuracy as information relating to all cavities used in this study are solely collected from boring logs. This contributes to a lot of noise in the accuracy of cavity distribution. Also, in this study cavities are assumed as discontinuous 2D features, while in reality cavities tend to develop and propagate in vertical and lateral directions. Therefore, this decision tree model needs to be constantly updated formed and made available. References Al-Jallal, I.A. and A.S. Alsharhan. 2005. Arabia and the Gulf. In: R.C. Selley, L.R. Cocks, and I.R. Plimer Ed. Encyclopedia of Geology Vol. 1. Amsterdam (The Netherlands): Elsevier Science B.V. p. 140152. Alsharhan AS, Nairn AEM. 1997. Sedimentary basins and petroleum geology of the Middle East. 1st ed. Amsterdam (AE): Elsevier Science B. V. p. 441. Bahroudi, A., and Koyi, H A. 2004. Tectonosedimentary framework of the Gachsaran Formation in the Zagros foreland basin. Marine and Petroleum Geology 21: 1295. Clark PJ, and Evans, FC. 1954. Distance to nearest neighbor as a measure of spatial realtionships in populations. Ecology 35: 445-453. R. 2001. Weathering of a gypsum-calcareous mudstone under semi-arid environment at experimental approaches. Catena 44: 111-132. Drake JJ, and Ford, DC. 1972. The analysis of growth pattersns of two-generation populations: the example of karst sinkholes. Canadian geographer XVI (4): 381-384. Ebdon, David. 1985. Statistics in Geography. Blackwell. Farrant, AR, Ellison, RA, Merritt, JW, Merritt, JE, Newell, AJ, Lee, JR, Price, SJ, Thomas, RJ, and Leslie, A. 2012. Geology of the Abu Dhabi 1:100 000 map sheet. United Arab Emirates (Keyworth, Nottingham: British Geological Survey). ISBN 978 085272 725 6. Farrant, AR, Ellison, RA, Merritt, JW, Merritt, JE, Newell, AJ, Lee, JR, Price, SJ, Thomas, RJ, And Leslie, A. 2012. Geology of Al Wathba 1:100 000 map sheet. United Arab Emirates (Keyworth, Nottingham: British Geological Survey). ISBN 978 085272722 5. assessment in Minnesota using a decision tree model. Environmental Geology 54 (5): 945-956. Gao, Y. 2002. Karst Feature Distribution in Southeastern Minnesota: Extending GIS Based Database for Spatial Analysis and Resource Management [Dissertation]. Minnesota (MN): University of Minnesota. p. 210. General Geotechnical Investigation Report (Draft), Spektra Jeotek, 2012, Presidential Affairs Building, Khalifa City-B. February 16 2012. Geotechnical Risk Map, Presidential Affairs 44 Plots Khalifa City-B. Spektra Jeotek, 2011, Abu Dhabi, UAE. Getis, Arthur, and J. K. Ord. 1992. The Analysis of Spatial Association by Use of Distance Statistics. Geographical Analysis 24 (3). Primer. Resource Publications in Geography, Association of American Geographers.

PAGE 10

298 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Hausmann, M.R., 1990, Engineering Principles of Company. Mitchell, Andy, 2005. The ESRI Guide to GIS Analysis Volume 2. ESRI Press. Price SJ, Farrant AR, Leslie A, Terrington RL, Merrit J, Entwisle D, Thorpe S, Horabin C, Gow H, Self bedrock geological model of the Abu Dhabi urban area. United Arab Emirates. British Geological Survey Commercial Report CR/11/138. Thompson HR. 1956. Distribution of distance to n-th nearest neighbor in a population of randomly distributed individuals: Ecology 37: 391-394. Tose JA, Taleb A, 2000. Khalifa city b ground conditions program. In: Mohamed & Al Hosani, editors. Georngineering in Arid Lands. Balkerna (Rotterdam). p. 75-81.



PAGE 1

537 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 EVALUATION OF FIRST ORDER ERROR INDUCED BY CONSERVATIVE-TRACER TEMPERATURE APPROXIMATION FOR MIXING IN KARSTIC FLOW Philippe Machetel Gosciences Montpellier UMR5243 CNRS/UM2, Place Eugne Bataillon-CC60, 34095 Montpellier cedex5, France, philippe.machetel@laposte.net David A. Yuen Minnesota Supercomputing Institute and Department of Earth Science, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, USA, and School of Environmental Studies, China University of Geosciences, Wuhan, China, daveyuen@gmail.com formalism has been used to conduct a several orders of magnitude parametric exploration based on the Peclet and the Reynolds numbers. The final errors, between the all of the parametric range. The combination of error curves bounds a closed volume in error space that gives a first upper bound of the error made by considering the temperature as a conservative tracer. Applying the method to an illustrative example of karst allows us to reach a first order error within a few degrees C. Introduction Karst aquifers are often considered as potential solutions for meeting water needs required by agriculture, industry and human consumption. A highlight is the ability of these systems to return, during the dry season, the rainfalls of the watershed. However, exploitation by surface, or underground catchment, or pumping of balance of downstream ecological systems (Weber et resources by surface catchment, underground catchment or pumping requires careful evaluation in order not to Karst environments present broad spectra of hydrodynamic properties which range from a Porous Fractured Matrix (PFM) to thin interconnected Conduit Systems (CS) whose diameters can range from fractions of a centimeter to tens of meters. This variability induces, for the discharge and recharge phenomena, temporal responses ranging from a few minutes to several months. In addition, except for the largest caves, it is generally not possible to enter the heart of the karst system to proceed at direct measurements of temperatures, Abstract Fluid dynamics in karst systems is complex due to the heterogeneity of hydraulic networks that combine the Porous Fractured Matrix (PFM) and the interconnected drains (CS). The complex dynamics often requires to be treated as black boxes in which only input and output properties are known. In this work, we propose to assess the first-order error induced by considering the temperature as a conservative tracer for flows mixing in karst (fluvio-karsts). The fluvio-karstic system is treated as an open thermodynamic system (OTS), which exchanges water and heat with its surrounding. We propose to use a cylindrical PFM drained by a water saturated cylindrical CS, connected on one side flowing at the base level of the karstic system. The overall structure of the model is based on the conceptual model of fluviokarst developed by White (2002, 2003). This framework allows us to develop the equations of energy and mass conservation for the different parts of the OTS. Two numerical models have been written to solve these equations: 1) the so-called AW (for Adiabatic Wall) configuration that assumes a conservative tracer behavior for temperature with no conductive heat transfer, neither in the liquid, nor in the PFM or even 2), the CW (for Conductive Wall) configuration that takes into account the heat and mass transfers in water and from water to aquifer rocks both in the CS and in the PFM. Looking at the large variability of karstic system morphologic properties, dimensionless forms of the equations have been written for both AW and CW configurations. This approach allows us to gather the physical, hydrological and morphological properties of karstic systems into four dimensionless Peclet, Reynold, Prandtl and dimensionless diffusivity numbers. This

PAGE 2

538 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE less rapid fluctuations depending on the sunlight, the daily cycle, the weather trend or season. Conversely, in depth, with the notable exception of phenomena consecutive to run-off or sudden rainstorms, the thermal fluctuations are damped in the CS and the PFM. Thus, most of the studies mentioned above made the assumption of rapid temperature fluctuation damping (Sinokrot and and Wicks, 2005). Our work aims to evaluate the influence of conservative temperature approximation based on the theoretical framework provided by the OTS. For this purpose, our theoretical development takes into account the different heat transfer possible between the different elements of the OTS. However, it is impossible for a black box like approach, to account for all of the natural karstic system complexity. While early models sometimes regarded karst as continuous media, more recent studies have shown the way for the inclusion of more complex internal works we have considered a cylindrical water saturated CS, carried by an abscissa axis x, separated from a cylindrical PFM by a permeable wall. This framework allows checking the mass and energy conservations solving water and rock temperatures in the different parts of the OTS with two different configurations : 1), a conservative tracer behavior for temperature in which the mixing between CS and PFM flows occurs in the CS without heat dissipation or dispersion in the CS, in the PFM or through the wall separating the CS from PFM (in the following this configuration is called AW between CS and PFM flows but with conductive heat dissipation within the CS in the PFM and through the wall that separates them (This configuration is called CW for Conductive Wall). The overall morphological structure used for our study is that the conceptual model developed by White (2002 and 2003) for fluviokarst. This structure corresponds to the Cent-Fonts (Hrault, France) karstic system whose description is given in the Annex part and which serves as an illustrative example for a potential application of the method. In the Whites fluviokarst model (redrawn in Figure 1a) the stream that dug the valley is partially lost in a sinkholes area at the entry of the CS. However, compositions or flow rates. It is often unavoidable to treat the entire system as a black box whose physical properties are more or less well-known except at the entrance and exit. So, for several decades, studies of hydrology properties of karstic systems use time series of temperature recording and flows, often supplemented by borehole measurements and test pumping although the latter cannot take into account the local complexity of karst systems and that the results may be biased, for example, by the location of the borehole in the PFM or in the CS network. In the first case, the measures will be typical of a damped environment and mainly sensitive to long term changes while, in the second case, they will be more sensitive to abrupt and rapid events such as heavy rains or run-off. to study the karst systems as black boxes. In this framework, they can be considered as Open Thermodynamic Systems that exchange water and heat energy with the atmosphere, rocks or other surface or underground water circulations. However, it is not always possible to obtain accurate assessments of flow mixing and energy exchanges that occur because the temperature cannot be considered without caution as a conservative tracer. Heat exchanges exist in the CS water, in the PFM and, through the walls, with the porous rocks. Despite these difficulties, recent studies have used the temperatures and surface and underground flows to hydraulic conductivity and temperature. Tabbagh et al. (1999) estimated the magnitude of aquifer recharge by setting vertical values of measured temperature as boundary conditions to energy equations. Benderitter et al. (1993) used long series of several year temperature records years to assess reservoirs depths on the basis of temperature equilibrium between groundwater and rocks. The couplings between the temperature of underground streams, rainfall and flow rates were also used on an the slow drainage of the limestone lagoons. The same author also used the annual changes in temperature sources to determine the preferred path of deep rain water in karst cavities (Genthon et al., 2005). Finally, combined with the recession curves to estimate the hydraulic properties, recharge properties and hydraulic transmissivity of karst systems.

PAGE 3

539 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 velocity field in the PFM and in the CS. In this study, we will also consider that the axial velocity, v x PFM and that the radial velocity, v r Therefore, the mass conservation equation can be written in cylindrical coordinates ( x, r ) as : (1) By linearity, it is possible to apply the Ostrogradskis or Greens theorem over each CS and PFM cylindrical section (of thickness dx ) to convert the volume integral in a flux integral. The resulting Eq. 2 and Eq. 3 ensue for the PFM ( r r h ) and for the CS ( r < r h ). (2) (3) Furthermore, to describe the CW configuration, it is necessary to take into account the conservation of thermal energy during the conductive transfers in the CS, in the PFM and between water and the rocks of the aquifer. Following the work developed by Covington et al (2011), we apply a conduction-temperature advection equation in PFM, but adding a new dispersion-conduction term that takes into account the cooling effects of the diffuse infiltration. In the PFM for ( r r h [0, L]), the energy conservation becomes Eq. 4: (4) and, in the CS, for ( r < r h [ 0, L ]): (5) where v x v r and T are the velocity and temperature fields depending on the cylindrical coordinates x and r However, to simplify the presentation of Eq. 5, these dependencies are not explicitly written except if the values are taken at specific points as r = 0 or r h In nature, karst systems exhibit great variability which is mainly due to the diversity of soils, to the extent of the watershed and the degree of karstification. CS present lengths ranging from tens of meters to a few tens of kilometers, while the permeability of the PFM, when it is not completely dried, the remaining of its level of the fluviokarstic system. Along the underground path, the CS receives a diffuse infiltration from the far field through the PFM. The outlet of the system is a the base level of the karstic system. Theoretical Approach The CS of karstic systems is often impossible to explore and it is often unavoidable to consider the CS and the PFM as a black box whose thermodynamic properties will be treated as an OTS. Considering water as an incompressible fluid with constant physical parameters (heat capacity, thermal expansion, density, or viscosity, etc.), the equations of mass and energy have been written for AW and CW configurations. Under these conditions, the mass conservation equation Figure 1. Fluviokarst model. (a) Conceptual fluviokarst structure proposed by White (2002 and 2003). The karstic aquifer is composed of a sinkholes area, of a saturated porous fractured matrix (PFM) drained by a conduit system (CS) and of a spring resurging at the base level. The Whites model is restricted to recession periods with no surface stream and no run-off flow; (b) A sequence of open thermodynamic system sections is used to describe the fluvio-karstic system. Each section represents a slice of CS and PFM over which mass and energy balances will be checked.

PAGE 4

540 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE water and rocks. In the following of this work, we will consider that the uncertainties on these parameters are significantly lower than those that might come from the variability of the hydrological. The latters mainly affect the Reynolds number (through the hydraulic radius of the CS) and the Peclet number (through the CS length). We therefore have built our parametric study on these two parameters which endorse most of the variability. If the mass conservation equations (Eqs. 6 and 7) are not changed when the temperature is considered as a conservative tracer (AW case), this is not the case for the energy conservation equations (Eq. 8 and Eq. 9). In that case, it is necessary to consider the limit when the conductive heat dissipation disappears, that is to say when D w and D m -> 0. In the PFM, for ( r r h ), it comes: (10) According to the constant thermal boundary condition for the far field temperature ( ), the solution is given by Eq (11): (11) It follows as an expectable result that, in the absence of heat conduction, the PFM temperature is uniform and equal to those of the far field. Now, lets study the temperature equation in the CS for the AW case. It is then conductivity in water and rocks. In this case, if D w -> 0 Pe sides of Eq. 9 leads to: (12) That is to say: (13) Eq. 13 matches the classical expression of the thermal energy conservation for a mixing without heat dissipation. Numerical Modeling Two numerical programs were written to solve the conservation of energy in the CS and in the PFM for both AW and CW configurations. They are based on finitedifference second order accurate methods (Douglas and or the hydraulic radius of CS cover several orders of magnitude. Only a non-dimensional approach, classical for fluid mechanics studies, can describe such broad ranges of local properties through dimensionless number approach. The morphological aspect ratio of the karstic system, r h / L may display broader variations. However, hydrological and thermal properties of the various flows ( T i Q i for the intrusion at the sinkholes), but also the far field temperature and the rate of diffuse infiltration in the CS ( T qM ) also impact directly the temperature of the mixed water in the CS. Similarly, the physical parameters of rocks and water will also be gathered in these dimensionless groups as the viscosity or the ratio of water to rock thermal diffusivities. Eqs. 1 to 5 have been rewritten using L the CS length, resurgence and the intrusive flow at the sinkholes area (divided by h 2 the CS wall surface) defines a natural velocity scale for diffuse infiltration V = ( Q s -Q i ) / h 2 A time scale = L / V ensues from these two scales and, with these conditions, the relationships between dimensionless parameters of length, speed and time are respectively h 2 / ( Q s -Q i ) and The relationship that links the temperature to the dimensionless temperature is Within this framework, with = r h /L and = u rh /V the mass conservation equations in PFM and in CS become: (6) (7) and the energy equations in PFM and in CS become: (8) (9) Four groups of dimensionless numbers appear in Eqs. 6 to 9. They are combinations between the Peclet number, w the dimensionless thermal diffusivity, D m / D w the Prandtl number, w and the tubular Reynolds number h The Prandtl number and the dimensionless thermal diffusivity depend only on the physical properties of

PAGE 5

541 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 following dimensionless numbers: 8 Red 4 and Figure 2 shows the temperature field obtained for the CW configuration in PFM and CS, after resolutions of Eqs. 8 and 9. The monotonic cooling with abscissa along the CS results from the mixing of the cold diffuse flow with the hot water that enters the CS at the sinkholes area. However, in the PFM around the CS, close to the sinkholes area, the heat transmitted by conduction in the PFM, induces thermal boundary layers that encounters the cooling flow from the far field. At a radial distance of a few tens of meters away from the CS wall, the thermal boundary layers attenuate rapidly and fall to a temperature close those of the far field. On the other hand, solving Eq.13 provides the temperature of CS water with the conservative tracer hypothesis (AW). In Figure 3, the Comparison of the blue curve (AW configuration) with the red curve (CW configuration) shows that the CS cooling is overestimated by the conservative temperature hypothesis. Due to the Rachford, 1956). Both codes strictly comply with the same boundary conditions given by the flows entering the sinkholes, the rate of diffuse flow from PFM to CS and the intrusive and far field temperatures. The dimensionless temperature at the CS input is T i = 0, while the CS temperature at the CS resurgence remains free. The thermal boundary condition at the outer edges of the calculation grid ( ) is the far field dimensionless temperature, For AW cases, a numerical 1-D code, solves Eq. 13 in the CS to calculates the temperature as a function of the abscissa x On an other hand, for the CW configurations, Alternate Direction Implicit (ADI) finite-difference methods are used to successively solve the temperature equations in the radial and axial directions. The equations direction and 500 points in the longitudinal direction. A convergence test stops the iterative process when a steady state solution is achieved (in effect when the relative changes of the temperature are less than 10 -8 for all the points of the computation grid). Continuity and coupling of the temperature between the CS and the PFM are ensured by sharing the thermal boundary conditions along the CS wall: the temperature calculated in the CS serves as radial boundary condition for solving the temperature in the PFM. This method allows coupling the resolution in both parts of the solution while taking into account the heat dispersion in the CS and heat spreads within the CS and between CS and PFM. A first comparison of the AW and CW hypotheses was conducted by introducing the values of the morphological, hydrological, and thermal main data observed on the Cent-Fonts fluviokarst site in summer 2005 (see annex part, and Ladouche et al., 2005 for report). Thus, a CS length (L = 5,000 m) is the distance, as the crow flies, between the sinkholes area of the Buges stream and the Cent-Fonts resurgence near for the CS ( r h ) corresponds to the speleological flow rate at the sinkhole is Q i 3 /s stream at the resurgence is Q s 3 /s at sinkhole is temperature is (Ladouche et al., 2005). Further, we assumed a water thermal diffusivity D w -7 m 2 /s and for PFM D m 10 -6 m 2 /s (Pechnig et al., 2007). These values lead to the Figure 2. CS and PFM temperature field. Illustrative example of the temperature field obtained in the CS and in the PFM with CW model. The morphological, hydrological, and physical parameters are characteristic of the Cent-Fonts (Hrault, France) fluviokarst system (L = 5,000 m, r h = 5 m, D m = 1.42 10 6 m 2 /s, D w = 1.4310 -7 m 2 /s, T i = 295.25 K (22.1 i =0.055 m 3 /s, Q s = 0.392 m 3 /s). These values lead to Pe = 1.50 10 8 Red = 4.29 10 4 Pr = 6.99) and D m /D w = 9.93. The computation has been done in the dimensionless space. The temperature field has been rescaled in the physical space for drawing.

PAGE 6

542 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE (14) (15) As mentioned above we have restrained our attention in this section, to the study of the effects of Reynolds numbers and Peclet numbers changes. The Peclet number ( w ) measures the ratio of advection to conduction characteristic times. Therefore, the higher Pe the preeminent effect of conduction compared to length, which is one of the most variable morphological properties of karst systems. The second entry of this study is the Reynolds number ( h ), which is characteristic of a second highly variable morphological parameter of the karstic system : the CS hydraulic radius. These two quantities are mathematically linked through the Prandtl number h However, it should be noted that in the energy equation (Eq. 9), Pe is present at the numerator of the Eq. 9s terms while Red is present at the denominator. According to their increasing or decreasing values, they will induce opposite effects on conduction, dispersion and advection. To better understand the cross-influence of these two parameters we have, initially studied the behavior of the depending on Pe (ranging from 10 6 to 10 9 ) at constant 4 the value previously second step, we studied the behavior of by varying Red from 10 3 to 10 7 at constant 8 ). The results of these two steps are shown respectively on the upper and lower parts of Figure 4. The left panels show the evolution as a function of the abscissa x in the CS, while the right panels give an overview of the final errors at the system output. In any case, at the sinkholes area location ( ), the error stage, no energy transfer by conduction has occurred between hot intrusive water at sinkhole and cold intrusive water from PFM diffuse infiltration. Conversely, as soon as the offset x of the sinkhole increases, increases especially as the boundary layers observed in Figure 2 are stronger. For higher abscissa x the rate of variation of decreases. In fact, two distinct types of behavior are displayed for on the left panels of Figure 4. effect of heat conduction in the CS, in the PFM and from the CS to PFM, a significant amount of thermal energy is consumed in the PFM to hold the hot conductive boundary layers against the diffuse cold water advection from the far field. Both curves diverge quickly where the maximum of energy is required to maintain the higher boundary layers (abscissa close to the sinkholes area). For larger abscissa, the decrease of the boundary layer difference between the two AW and CW assumptions. The final temperature difference at the CS output corresponds to the thermal energy used to maintain the equilibrium of water and rocks temperature in the PFM in spite of antagonistic effects between conductive and advective heat flows while the thermal diffusivity of the rock is about one order of magnitude higher than those of water. This final temperature difference bears a firstorder information on the error due to the conservative tracer assumption in the PFM. Beyond this simple illustrative example, the next section will describe the general quantification of this error. Parametric Study The high variability of morphological, hydrological and thermal properties of karst systems stresses the need to conduct a parametric study to cover the wide ranges of dimensionless parameters that describe them. In this context, we conducted a systematic study of the temperature relative deviation (Eq. 14) obtained in the CS for both CW and AW configurations. Its dimensionless expression is given by Eq. 15. Figure 3. AW versus CW, CS temperature. Comparison of CS temperature obtained with the CW model (red curve) and the AW model (blue curve).

PAGE 7

543 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 (bottom panel)). The maxima of error are encountered for intermediate values of both parameters. They remain are respectively achieved for 7 and 5 ). It for the illustrative example of Cent-Fonts fluviokarstic system naturally fall within this bounded error range. Summary and Discussion The purpose of this work is to try to assess a first order of the error done by considering temperature as a conservative tracer in fluvio-karstic systems. For that, we developed and solved the energy and mass conservation equations, leaned against the Whites conceptual model for fluviokarst, and within the theoretical background of OTS. We applied theses equations to a cylindrical CS, which receives hot intrusive water from a sinkholes area and, through a PFM, a cold diffuse flow from the far field all along its underground path. This set forms an OTS in which we studied two configurations. The first (AW) assumes that no conductive heat is lost in the CS water, neither between the PFM water and the aquifer permeable separation wall. The second (CW) takes into account the conductive heat dissipation in CS water, in PFM and the dispersion of heat by conduction through the PFM wall. These equations have been rescaled that leads to a new system of equations where four groups of dimensionless numbers measure the relative magnitudes of the various conductive and advective terms. Solving these equations in both configurations with strictly similar thermal and dynamic, boundary conditions allows assessing a first order of the errors induced by the conservative tracer assumption for temperature. However, it is clear that our results lead only to a firstorder information about this error because the method cannot completely eliminate or estimate other sources of error. Indeed, in order to proceed to the numerical solving of the mass and energy equations we need to consider laminar flows, in a saturated CS. Furthermore, we must keep in mind that the method is better applied during the recession periods when the hydraulic regime of the fluviokarst is as close as possible of a steady state. Further studies are needed, to go beyond the first results presented in this paper. However, some of our results seem encouraging since whereas the variability of karst systems has largely been accounted by the range extent of the parametric exploration, errors in the or less slowly for higher abscissa. The seconds shows monotonic growths of The transitions between the two regimes are progressive and occur with the increases of the test parameters ( Pe or Red). However, it should be noted that, in all the cases, the error curves remain Lets now examine in more detail how evolves with Pe (at constant Red) (Figure 4, top panels). According to the physical meaning of the Peclet number, its decreasing corresponds to a decrease of the relative importance of conductive transfers face to the advective ones, homogeneously in the whole system. In fact, for the lowest values of Pe (10 6 and 5 10 6 left panel, light brown and red curves) increases monotonically with x However it converges uniformly Pe Conversely, for the highest values of Pe (5 10 8 and 10 9 Figure 4, top left panel, dark blue and purple curves) reaches maxima for abscissa close to the sinkhole area (low values of x ). In these cases, the amplitudes of these maxima decrease with increasing Pe the value of the abscissa. In any cases, the maximum of the relative errors are achieved for intermediate values of the dimensionless parameter Pe They remain less than 0.01 (actually 0.0092 for Pe = 710 7 ). For the highest values of Pe D w -> 0 described in Section 2 since in that case the assimilation of temperature to a conservative tracer is equivalent to consider the limit that induces that which measures the difference between the two AW and CW Lets now consider the physical situation that prevails when Red increases at constant Pe This scenario allows to assess the influence of the CS hydraulic radius ( h ). Since Red and Pe perform in antagonistic manners n Eq. 9, it is now for the highest values of Red (10 6 and 10 7 Figure 4, bottom left panel, orange and red curves) that Conversely, this I for the lowest values of Red (10 3 and 10 4 Figure 4, bottom, left panel curves purple and dark blue) that reaches maxima, for low values of the abscissa, in the sinkholes area. The right panels of Figure 4 show synoptic representations of the relative errors at the output of the CS (following the variation of Pe (top panel) and of Red

PAGE 8

544 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE temperature assumption for particular karstic system by reversing the scaling scheme. This quite simple calculation allows quick retrieving of the error in the physical space thank to the morphological and hydrological properties of a particular fluviokarstic system. The comparison of model results with field data open the possibility of critical analyses and may offer a decision-making support for its applicability to local cases. If we focus our attention on the illustrative example given by the Cent-Fonts fluviokarstic system, Figure 4 shows that the relative error 0.0092 to 0.007). The error volume formed by the error curves, due to the variations of these two parameters, the maximum of errors are reached for median parameter values that are characteristic of realistic karstic system in both morphology and hydrological considerations. From the results obtained in the non-dimensional space, it is possible to evaluate an upper bound of the first order of the error induced by the conservative Figure 4. Parametric study of the error e(x) versus Pe and Red numbers. Evolution of the give a synthetic view of the errors reached at the output of the CS (x=1) versus the dimensionless parameters.

PAGE 9

545 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 confluence with the Lamalou. On the other hand, the underground pathway from Buges sinkhole to the Cent Fonts resurgence was established by tracing (Dubois, reaches, 0.00613 at the exit of the resurgence in the dimensionless field ( 8 and 10 4 ). When rescaled in the physical domain, the error indicates a temperature disparity T CW T AW = 1.77 C ( 0.00613 x 288.50 ). It is clear that this information can be used to infer potential propagation of uncertainties in calculations or cooling rates. It can also be used to propose a calibration of the effects of the conservative temperature approximation depending on the equations and on the dimensionless properties of the karstic system (Machetel and Yuen, in preparation). In these next works, we will focus on examining that the theoretical evaluation of the error proposed in this work is compatible with the thermal data available on other karst systems. Indeed, at the sight of this study, it seems possible to use the results from a theoretical analysis to coerce information on internal thermal conditions of the karstic system. The results seem open interesting research opportunities that may be applicable to other systems whose workings are often described in terms of black box, as geothermal or hydrothermal studies. Appendix : Cent-Fonts resurgence (Hrault, France) The Cent-Fonts resurgence is the base level outlet of a fluviokarstic system, which watershed covers an area of 40-60 km 2 (Figure 5). This basin is located north of Montpellier, on the right bank of the Hrault River within a thick dolomitic Middle and Late Jurassic limestone sequence. Several structural, geological, geochemical and hydrological studies were devoted to this karstic system Drfliger et al., 2009). The watershed is bounded to the north and northeast by the Cevennes fault, the surface course of the intermittent Buges stream, and southeast by the Hrault River, which drains its base level. The watershed encompasses the upper course of the Bueges stream, which flows on an impermeable Triassic outcrop until it reaches a sinkholes area crossing a batonian dolomitic area a few kilometers downstream from SaintJean-de-Buges (Figure 5). From that point, the Buges stream surface course forms a valley mostly dried up that Figure 5. Cent-Fonts fluviokarst watershed. Redrawn on the Hrault geological map 1/50000 France BRGM (J2: Bajocian; J35: Bathonian, Callovian, Oxfordian; J6: Kimmeridgian; J7: Tithonian). The Cent-Fonts fluviokarst watershed covers about 60 km 2 including 10 km 2 for the Buges resurgence watershed (Schoen et al, 1999). Figure 6. CS map of the Cent-Fonts resurgence. Unrolled 3-D speleological map of the Cent-Fonts CS near the resurgence. Vertical distances (m) are conserved. Depths below the main cave entry are given in m (italic). During the summer 2005 pumping tests, piezometric heads and temperatures have been measured in the so-called CGE Reco and F3 boreholes. Temperature and discharges were also recorded at the output of the pumping device located in the F3 borehole, at the entry of the Buges Stream swallow zone and in the Hrault River near the resurgence (Ladouche et al., 2005).

PAGE 10

546 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The Cent-Fonts karstic system takes root in a dolomite layer Bathonian of 150 to 300 m thick and probably extends into the layer of Aalnien catching the Buges Stream loss at the Sint-Jeande-Buges sinkhole and by draining the watershed rainfalls which percolate through an upper Jurassic epikarstic layer (Petelet-Giraud et al., 2000). Therefore, the Cent-Fonts karstic system is similar to the conceptual model of White (Figure 1a) with a CS collecting sinkholes losses and a diffuse infiltration flowing from a PFM into the CS. After 5 km of underground path (as the crow flies), the CS poors through the Cent-Fonts resurgence in the Herault River which drains the base level of the karstic system. The resurgence discharges through a shallow network of springs that flow a few tens of centimeters above the Hrault (Schoen et al, 1999). During the dry season, the resurgence discharge ranges from 0.250 to 0.340 m 3 /s summer (Marchal et al., 2008). The detailed structure of the CS near the resurgence output (Figure 6) has been explored by divers (Vasseur, 1993). In the maped area, the CS cross section ranges between 4 to 16 m 2 (Drfliger et al., 2009). The Cent-Fonts resurgence has undergone numerous field observations since 1997. Several years of flow measurements have been recorded to calibrate the base flow of the Cent-Fonts resurgence, of the Buges stream and the losses in the sinkhole area. A pumping test campaign was conducted during the summer of 2005 by BRGM under contracting authority of Conseil Gnral de lHrault. This campaign provides to scientists many temperature and flow records from surface and deep holes measurements (Ladouche et al., 2005). Acknowledgments The authors thank the Conseil General de lHrault for the electronic transmission of the whole set of data of the 2005, Cent-Fonts pumping test campaign. We also thank Alexander Calvin, Martin Saar, and Andrew Luhmann for friendly and useful discussions about heat transfer in karstic systems. Research received funds from geochemistry program from the National Science Foundation. This work received a financial support from the Organising Committee of the International Workshop of Deep Geothermal Systems Wuhan, China, June 29-30, 2012. Table Acronyms AW Adiabatic Wall CS Conduit System CW Conductive Wall CV Control volume OTS Open Thermodynamic System PFM Porous Fractured Matrix Not. Units CF Description D m (m 2 /s) 1.42 10 -6 Matrix thermal diffusivity D w (m 2 /s) 1.43 10 -7 Water thermal diffusivity (m 2 /s) 10 -6 Kinematic viscosity L (m) 5 10 3 CS Lengh q m (m/s) surfacic discharge of PFM Q i (m 3 /s) 0.055 discharge of stream sink Q s (m 3 /s) 0.392 discharge of spring r (m) radial coordinate r h (m) 5 half hydraulic radius t (s) time T(x,r) (K) temperature (K) 285.35 far field temperature T i (K) 295.25 Temperature of stream sink v (m/s) Fluid velocity vector v r (x,r) (m/s) Velocity r component v x (x,r) (m/s) Velocity x component V (m/s) 4.29 10 -3 Velocity scale x (m) abscisse coordinate Dimensionless numbers D D m /D w 9.93 Thermal diffusivity ratio Pe LV/D w 1.50 10 8 Conduit Peclet number Pr w 6.99 Prandtl number Red (2V r h 4 Conduit Reynolds number

PAGE 11

547 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Dubois P. 1962. Etude des rseaux souterrains des rivires Buges et Virenque (Bas Languedoc), paper presented au 2e congrs international de splologie, Salerne, 2-12 Octobre 1958. Genthon P, Wirrmann D, Hoibian T, Allenbach M. 2008. Steady water level and temperature in a karstic system: the case of the coral Lifou Island (SW Pacific). C R Geoscience. 340: 513-522. Genthon P, Bataille A, Fromant A, DHulst D, Bourges F. 2005. Temperature as a marker for karstic waters hydrodynamics. Inferences from 1 year recording at La Peyrere cave (Ariege, France). Journal of Hydrology. 311 (1-4): 157-171. Jemcov I. 2007. Water supply potential and optimal exploitation capacity of karst aquifer systems. Environmental Geology. 51: 767-773. hydrograph and water temperature analysis Hydrology. 45 (3-4): 203-217. Ladouche B, Drfliger N, Pouget R, Petit V, Thiery D, des systmes karstiques nord-montpellirains, rapport du programme 1999-2001-Buges, RP 51584-FR., BRGM, 2002. Ladouche B, Marchal J C, Drfliger N, Lachassagne P, Lanini S, Le Strat P. 2005. Pompage dessai sur le systme karstique des Cent-Fonts (Commune de Causse de la Selle, Hrault), Prsentation et interprtation des donnes recueillies, BRGM RP54426-FR, 82 ill., 45 tabl., 9 ann., 245 pp. Luetscher M, Jeannin PY. 2004. Temperature distribution in karst systems: the role of air and water fluxes. Terra Nova. 16: 344-350. Luhmann AJ, Covington MD, Peters AJ, Alexander SC, Anger CT, Green JA, Runkel AC, Alexander EC. 2011. Classification of thermal patterns at karst spring and cave streams. Ground Water. 49 (3): 324-335. Marechal JC, Ladouche B, Drfliger N. 2008. Interpretation of pumping tests in a mixed flow karst system. Wat Res Res. 44: W05401. http:// dx.doi.org/10.1029/2007WR006288 Paloc H. 1967. Carte hydrogologique de la France, rgion karstique nord-montpelliraine, notice explicative. Mmoire BRGM. n. Pechnig R, Mottaghy D, Koch A, Jorand R, Clauser C. 2007. Prediction of thermal Germany, European Geothermal Conference, GeothermischeVereinigunge.V. Bundesverband Geothermi. References Aquilina L, Ladouche B, Drfliger N. 2005. Recharge processes in karstic systems investigated through the correlation of chemical and isotopic composition of rain and spring-waters, Applied Geochemistry. 20 (12): 2189-2206. Aquilina L, Ladouche B, Drfliger N. 2006. Water storage and transfer in the epikarst of karstic systems during high flow periods. Journal of Hydrology. 327 (3-4): 472-485. Benderitter Y, Roy B, Tabbagh A. 1993. Flow in a carbonate fractured medium: first approach. Water Res Res. 29: 3741-3747. Camus H. 1997. Formation des rseaux karstiques et mridionnal, Hrault, France. Karstologia. 29: 23-42. temperature, stream flow, and groundwater exchanges in Alpine streams. Water Resources Research. 34 (7): 1609-1615. streambed water exchanges. Water Resources Research. 44: W00D10. http://dx.doi. org/10.1029/2008WR006996 Covington MD, Wicks CM, Saar MO. 2009. A dimensionless number describing the effects of recharge and geometry on discharge from simple karstic aquifers. Water Res Res. 45: W11410. http://dx.doi.org/10.1029/2009WR008004 Covington MD, Luhmann AJ, Gabrovsek F, Saar MO, Wicks CM. 2011. Mechanisms of heat exchange between water and rock in karst conduit. Water Resources Research. 47: W10514. http://dx.doi. org/10.1029/2011WR010683 Covington MD, Luhmann AJ, Wicks CM, Saar MO. 2012. Process length scales and longitudinal damping in kart conduits, Mechanisms of heat exchange between water and rock in karst conduit. J Geophys Res. 117: P01025. http://dx.doi. org/10.1029/2011/JF002212 Dogwiller T, Wicks C. 2005. Thermal variations in the and Evolution of Karst Aquifers. 3 (1): 1-11. Drfliger N, Fleury P, Ladouche B. 2009. Inverse modeling approach to allogenic karst system 414-426. 6584.2008.00517.x Douglas J, Rachford HH. 1956. On the numerical solution of heat conduction problems in two and three space variables. Trans Am Math Soc. 82: 966-968.

PAGE 12

548 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Petelet E, Luck JM, Ben Ohtman D, Negrel P, Aquilina L. 1998. Geochemistry and water dynamics of a and trace elements. Chemical Geology. 150 (1-2): 63-83. Petelet-Giraud E, Drfliger N, Crochet P. 2000. RISK : mthode dvaluation multicritre de la vulnrabilit des aquifres karstiques. Application aux systmes des Fontanilles et cent-font (Hrault, sud de la France). Hydrogologie. 4: pp.71-88. Petelet-Giraud E. 2003. Dynamic scheme of water circulation in karstic aquifers as constrained by Sr and Pb isotopes. Application to the Herault watershed, Southern France. Hydrogeology Journal. 11 (5): 560-573. 1999. Caractrisation du fonctionnement des systmes karstiques nord-montpellirains. Rapport BRGM R 40939, volume III, 91 p. Sinokrot BA, Stefan HG. 1993, Stream temperature dynamics measurements and modeling. Wat Res Res. 29: 2299-2312. Determination of recharge in unsaturated soils using temperature monitoring. Water Resources Research. 35 (8): 2439-2446. Vasseur F. 1993. Les explorations de lAssociation Celadon. Info-plonge. 63: 17-21. Weber KA, Perry RG. 2006. Groundwater abstraction impacts on spring flow and base flow in the Hillsborough River Basin, Florida, USA. Hydrogeology Journal. 14 (7): 1252-1264. White WB. 2002. Karst hydrology: recent developments and open questions. Engineering Geology. 65 (2-3): 85-105. White WB. 2003. Conceptual models for karstic aquifers. Speleogenesis and evolution of karst aquifers. 1 (1): 6.



PAGE 1

313 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 EVALUATION OF VETERINARY PHARMACEUTICAL AND IODINE FOR USE AS A GROUNDWATER TRACER IN HYDROLOGIC INVESTIGATION OF CONTAMINATION RELATED TO DAIRY CATTLE OPERATIONS Larry Boot Pierce Missouri Geological Survey, 111 Fairgrounds Road. Rolla, MO 65401, USA, larry.pierce@dnr.mo.gov Honglin Shi Missouri University of Science and Technology, 1870 Miner Circle, Rolla, MO 65409, USA, honglin@mst.edu a long term groundwater tracer does not seem to be an immediate option. Further research may reveal that its degradation products are potentially useful as a tracer. In some instances, such as catastrophic discharges of large volumes of milk when samples can be collected environmental tracer may prove possible. The validity of pharmaceutical iodine as a groundwater tracer appears to be much greater than that of cephapirin. Iodine was detected in all of the environmental samples including the highly organic and anaerobic environment of the dairy wastewater lagoon. This study concludes that iodine is capable of surviving the hostile wastewater environment. If sufficient data is collected to determine natural background levels, iodine may prove useful in determining hydrological connections between iodine laden dairy effluent and the underlying groundwater. Introduction Investigation of the hydrologic environment in the vicinity of a fitness center in Southeast Missouri began in the summer of 2010. This investigation included a dyes such as fluorescein and Rhodamine WT to evaluate potential sources of bacteria entering the centers well. The sources included surface water infiltration surrounding the wellhead, onsite septic system failures and a nearby dairy operation (Pierce, 2012). After reviewing operations of the nearby dairy it was determined that additional groundwater tracers may be present in the form of veterinary pharmaceuticals then initiated. The first phase of the investigation was were detectable in a laboratory environment. Once the analysis methodology was proven to be successful, Abstract Standard groundwater tracers such as Rhodamine WT, Fluorescein, Eosin and Tinopal CBX effectively provide a snapshot of hydrological conditions over a brief period of time and in a tightly controlled setting. However, in complex environmental situations with multiple potential sources, groundwater hydrologists are often seeking groundwater tracers that have extended longevity in the natural environment and the ability to directly pinpoint source locations. After reviewing operations of the nearby dairy it was determined that emerging contaminants, specifically two bovine veterinary pharmaceuticals (antibiotics), cephapirin and a sanitation agent, elemental Iodine (I) may have potential as extended longevity groundwater tracers if analytical methodology could be established. Initially, sample analysis indicated that cephapirin is undetectable in unconcentrated samples of lagoon wastewater at for better detection at part per trillion level. Concentrated samples from one of the two lagoon cells sampled (cell #3), detected cephapirin at 13.14 ppt level, while cell #1 failed to detect any cephapirin present. Controlled laboratory testing later indicated that in a wastewater of initial concentrations within 4 days, with complete degradation within 6 days. Degradation patterns in surface water and groundwater samples were less dramatic and at slower rates. Degradation curves of the surface and groundwater samples indicate that concentrations of cephapirin are still detectable for approximately 25 days. Unconcentrated Iodine samples collected in lagoon cells ranged from 50.896 ppb and 1,704.55 ppb with variations determined to be a result of the primary inflow of the lagoon. Cephapirins use as

PAGE 2

314 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE If a hydrologic connection between the CAFFC well and the nearby dairy operation exist then testing should reveal cephapirin, used in the dairy operation, to be present in CAFFC water samples. Additionally, iodine levels should be elevated in water samples when compared to natural background iodine concentrations. To determine if Cephapirin was readily detectable Dr. Honglan Shi at Missouri University of Science and Technology was contracted to develop liquid chromatography-tandem mass spectrometry (LC-MS/ antibiotic. In addition Dr. Shi developed protocol for inductively coupled plasma mass spectrometry (ICPcompletion, the methodological process was put into samples associated with the CAFFC well investigation. Cephapirin and Iodine Analysis Development The results of the cephapirin analysis methodology development were mixed. Dr. Shi was able to successfully develop a reliable, fast and simple method for detecting the base Cephapirin (CEP) molecule in water samples to the part per trillion (ppt) level. However difficulties were also observed in the analysis of field samples, stability studies showed that the Cephapirin degrades to form a degradant (DACEP) in all sample matrices (water and lagoon effluent). In a series of further test conducted on spiked samples the degradation was shown to be highest in the lagoon samples with complete degradation of CEP to DACEP within six days when left at room temperature. Figure 1 highlights these degradation rates. Analysis of samples in both groundwater and surface water matrices also show similar degradation patterns, but at a slower rate. Testing revealed that the currently degraded to undetectable levels. The ICP-MS analysis methodology for total iodine in water samples was previously developed by Dr. Shi and the MS&T laboratory in 2009 and would be used for the investigation (Shi and Adams, 2009). Field Sampling and Investigation As part of the initial dye tracing, a well survey was conducted by MGS to locate and inventory any groundwater wells servicing residence, farms, or the potential for their survival in natural environments or highly organic and anaerobic conditions of dairy wastewater lagoons needed to be determined. The second phase was to then test the applicability of the groundwater tracers in the natural environment. Background Literature reviews and research revealed that the potential for water tracing using emerging contaminants such as bovine pharmaceutical compounds, trace elements and Jones, 2009). Emerging contaminants, specifically two bovine veterinary pharmaceuticals (antibiotics), cephapirin and a sanitation agent, elemental Iodine (I) are detectable potential for use as long-term groundwater tracers for hydrogeologic investigations where dairy operations are Through the course of several conversations the State of Missouris dairy inspector indicated that the pharmaceutical products cephapirin sodium (CEPNa) used for treatment of dairy cows at the nearby dairy cephapirin is excreted unchanged in the urine. Most excretion occurs within 8 hours of application and the process of excretion completed within 72 hours (Jones, 2009). Additionally, the inspector indicated that iodine both before and after milking at the dairy operation. It too is washed to the lagoon and thus has the potential to reach the natural environment in a similar manner to the cephapirin. Further literature research indicated that some pharmaceuticals and iodine are detectable in groundwater samples. (Santschi et al., 1999) Therefore, CEPNa, CEPB, and iodine may potentially be used as tracers to investigate the groundwater hydrology of the CAFFC area to determine whether the contamination in the affected well is related to the dairy operations.

PAGE 3

315 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 collected from each cell of dairys two-cell wastewater lagoons. A representative population of inventoried wells was selected for sampling and analysis to establish natural background levels of iodine. As well as a small nearby spring and surface water samples from Williams Creek and the small tributary below the dairy operation. A total of nineteen water samples were initially collected for analysis. Two additional samples were collected at a later date to allow for the SPE and pre-concentration analysis. Figure 2 is a map showing all sample locations. Samples were collected using procedures to meet MDNR standards and the methodology needs for Dr. Shis analysis. Water samples were collected in pre-cleaned, 4-ounce, Tefloncapped, wide-mouth amber bottles. For well water collection at homes, the faucet aerator was first removed (if present) and the water allowed to flow for approximately 5 minutes. Sample bottles were then filled. Surface water sample bottles were submersed and filled directly in the sampling bottle. For wastewater businesses in the vicinity of the CAFFC. This survey was expanded to additional locations to help establish a more representative natural Iodine background for the area. Following the survey, wells in the vicinity of the of cephapirin and iodine. Wastewater samples were Figure 1. Cephapirin degradation patterns by matrix. Figure 2. Sample location map.

PAGE 4

316 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE some barnyard runoff similar to cell #1. However, the influent of wash water from the milking parlor was also observed discharging into the cell. Since iodine is the daily cleaning effluent is not uncommon and likely explains the large difference in iodine concentration between cell #1 and cell #3. Conclusion Due to the instability of cephapirin its use as a long term groundwater tracer does not seem to be an immediate option. Future research may at some point identify a specific cephapirin degradant product that has the stability and environmental longevity required for groundwater tracing, but at this time those degradants have yet to be identified. In some instances such as direct runoff, catastrophic lagoon failures or discharges of large volumes of milk, where samples can be collected cephapirin or its degradants as an environmental tracer may prove possible. subsequently cleaned and decontaminated between uses to avoid cross-contamination. Field blanks, trip blanks, and duplicate samples were collected throughout the sampling event as a quality assurance measure. The collected samples were sealed and placed in an iced cooler then transferred to the University of Missouri Science and Technology laboratory using standard MDNR chain of custody protocol. Results Analysis of collected samples initially indicated no detection of cephapirin in the wastewater lagoons at part per billion (ppb) levels. A second round of sampling was conducted on the wastewater lagoons. At this time larger volumes of sample were collected so that a solid phase extraction (SPE) could be used to pre-concentrate the cephapirin and allow for better detection sensitivity to the PPT level. In this analysis large volumes of original amount while still retaining the concentrations of cephapirin. In this second analysis a two liter sample taken from the taken from cell #3 of the dairy lagoon treatment system, detected 13.14 ppt of Cephapirin. Analysis of a sample from cell #1 failed to detect any cephapirin. Iodine levels collected from the study area ranged from 4.939 ppb to 1704.55 ppb and are shown in Table 1. Iodine was detected in all eight drinking water wells with concentrations ranging from 4.939 to 24.712 ppb with a mean concentration of 12.532 ppb and a standard deviation of 6.672 ppb. Six surface water samples were 6.671 to 21.247 ppb. The mean concentration of iodine in surface water samples was 14.732 ppb with a standard deviation of 5.489 ppb. A single sample from a shallow spring was collected and determined to have 5.855 ppb iodine present. and lagoon cell #3 contained 50.896 ppb and 1,704.55 ppb iodine, respectively. A considerable difference in iodine concentrations is obvious in the lagoon samples. Further investigation revealed that the lagoons are not successive and that the source area and influents of waste were different for each. The primary waste source for lagoon cell #1 is runoff from the animal feed lot and loafing areas. Lagoon cell #3 received Sample ID Sample Name Concentration PPB FB001 Field Blank <0.2 TB001 Trip Blank <0.2 WW001 Class Act Family Fitness Center 16.178 WW002 Koehler Engineering Well 11.472 WW003 University Farm Well 11.93 WW004 Volkerding Well 4.939 WW005 New Midway Dairy Well 24.712 WW006 Old Midway Dairy Well 6.255 WW007 Dave Brown Well 19.31 WW008 Wayne Eakins Well 5.457 SS001 Upper Williams Creek 6.671 SS002 Lower Williams Creek 9.216 SS003 Tributary at Shale Lane 21.247 SS004 confluence with Williams Creek 20.23 SS005 outfall of dairy lagoon #3 13.117 SS006 Tributary above dairy lagoon outfall (Ponco Lane) 17.913 SP001 Spring on University Farm 5.855 LG001 Lagoon Cell #1 50.896 LG002 Lagoon Cell #3 1704.55 Table 1. Sample locations and iodine concentrations.

PAGE 5

317 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 The validity of iodine as an extended-longevity groundwater tracer appears to be much higher than the cephapirin. Iodine was detected in all of the environmental samples including the highly organic and anaerobic environment of the dairy wastewater lagoon. Iodine levels within dairy lagoons #1 and #3 were found This study concludes that iodine is capable of surviving a hostile wastewater environment. If sufficient data is collected to determine natural background levels, the use of iodine may prove useful in determining hydrological connections between iodine laden dairy effluent and the underlying groundwater. References Boxall, ABA, Kolpin DW, Halling-Srensen B, Tolls J. 2003. Are Veterinary Medicines Causing Environmental Risks? Environmental Science and Technology. 37 (15): 286AA. Drewes JE, Heberer T, Rauch T, Reddersen K. 2003. Fate of Pharmaceuticals During Ground Water Recharge. Ground Water Monitoring & Remediation. 23: 64. Jones GM. 2009. On-farm Test for Drug Residuals in Milk. Virginia Cooperative Extension Publication number 404-401. Virginia Polytechnic Institute and State University: [cited 2015 Jul 20]. 5 p. Available online at: https://pubs.ext. vt.edu/404/404-401/404-401.html Pierce LD. 2012. Report on the Methodological Development and Analysis of Veterinary Pharmaceutical and Iodine for use as a Tracer in an Investigation of Groundwater Contamination Related to a Dairy Cattle Operation in Southeast Missouri. Missouri Geological Survey. Santschi PH, Moran JE, Oktay S, Hoehn E. 1999. 129Iodine: A new tracer for surface water/ groundwater interaction. Isotope techniques in water resources development and management Proceedings, (Austria): 10-14 May. International Atomic Energy Agency (IAEA),Vienna. Available on CD-ROM. Shi H, Adams C. 2009. Rapid IC-ICP/MS method for simultaneous analysis of iodoacetic acids, bromoacetic acids, bromate, and other related halogenated compounds in water: Talanta. 79: 523-527.

PAGE 6

318 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

465 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 EVAPORITE GEO-HAZARD IN THE SAURIS AREA (FRIULI VENEZIA GIULIA REGION NORTHEAST ITALY) Chiara Calligaris Dipartimento di Matematica e Geoscienze, Universit degli Studi di Trieste, Via Weiss, 2, Trieste, 34128, Italy, calligar@units.it Stefano Devoto Dipartimento di Matematica e Geoscienze, Universit degli Studi di Trieste, Via Weiss, 2, Trieste, 34128, Italy, sdevoto@units.it Luca Zini Dipartimento di Matematica e Geoscienze, Universit degli Studi di Trieste, Via Weiss, 2, Trieste, 34128, Italy, zini@units.it Franco Cucchi Dipartimento di Matematica e Geoscienze, Universit degli Studi di Trieste, Via Weiss, 2, Trieste, 34128, Italy, cucchi@units.it associated to evaporite rocks. Introduction Subsidence phenomena associated to the presence of evaporite rocks are common in Europe. Evaporite sinkholes affect the central and northern part of England (Cooper, 2008), Lithuania (Taminskas and Parise et al., 2008). As reported by Nisio (2008) and by Caramanna et al. (2008), sinkholes occur also in Italy where are distributed along the whole peninsula, especially in some regions, part of Italy and covers an area of 7,858 km 2 Here are present 221 municipalities of which approximately 40 existence of outcropping or mantled karstifiable rocks. Limestones and dolostones represent approximately where evaporites and limestones outcrop or are mantled by quaternary deposits or other rock types, sinkholes Abstract Evaporite sinkholes represent a severe threat to many European countries, including Italy. Among the Italian regions, of the area most affected is the northern sector had two main depositional periods first in the Late Permian and then during the Late Carnian (Late Triassic). Evaporites outcrop mainly in the Alpine valleys or are partially mantled by Quaternary deposits, as occur along the Tagliamento River Valley. Furthermore, evaporites make up some portions of mountains and Alpine slopes, generating hundreds of karst depressions. This paper presents the preliminary results of the research activities carried out in Sauris Municipality where sinkhole phenomena related to the presence of gypsum are very common. Field investigations were devoted to recognition, mapping and classification of evaporite sinkholes. To included the analysis of historical documents collected in archives, the analysis of aerial photos and Airborne Laser Scanning (ALS) surveys. The integration of the abovecited activities allowed a preliminary identification of the phenomena, which were later validated by detailed field surveys. All the collected data populate a geo-database

PAGE 2

466 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE ones, belonging to the Bellerophon Formation, are Over time, water infiltration and weathering led to the solution of the Permian evaporites and to the progressive failure of the overlying terrigenous rocks, less soluble but more plastic. The result is the genesis and evolution of depressed landforms classifiable as sinkholes giving to the Alpine environment peculiar morphologies. This paper illustrates the results of the research activities performed in the Sauris Municipality, which is the area of the NE Italy most affected by sinkhole phenomena related to the presence of evaporite rocks and clearly visible on the topographical surface. means of traditional activities such as desk analysis and field surveys. Study Area The Sauris territory has an area of about 42 km 2 and a mean elevation of 1,212 m with an inhabitant density of 10 persons/km 2 Even if the average elevation is not so high, in the northern side of Sauris, some peaks as are widespread. Despite the low percentage of evaporite rocks, sinkholes associated with these lithologies reported by Zini et al. (in press), the northern portion of Tagliamento River Valley is affected by several sinkhole phenomena, which caused and are still causing severe damages to the existing infrastructures. This area has been historically hit by subsidence phenomena since 1960s, as reported by Gortani (1965). Giulia Region are caused by evaporites originated as the result of two main depositional episodes occurred respectively at the end of the Permian (Bellerophon Formation) and in the Carnian (Raibl Formation) (Venturini et al., 2006). In this context, evaporites, in role during the Alpine tectonic compressions, during which they may be treated as tectonic lubricant. Evaporites mainly outcrop in the valley floors, but can be identified also in the mountains, and this is the case of siltstones belonging to the Werfen Formation (Triassic) are widely present. These rocks, capping evaporitic Figure 1. Location of the study area (pale blue), which is situated in the NE Italy. Pink and brown areas indicate the main evaporite outcrops.

PAGE 3

467 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 regional structural trend. This fault is 40 km long from West to the East of the Region reaching the Tagliamento River where is present its further eastern portion. The above-cited fault permits Permian units to overlap the Triassic rocks. The presence of gypsum at the base of the northern overthrusted units facilitated the process. The Sauris Fault is also crossed by secondary faults. The Bellerophon Formation incorporates two different members: at the base gypsum alternates with black dolostone (thickness of about 60 m) whereas the upper member consist of dolostones and black limestones (200 m). The plastic behavior causes strong deformations of the evaporite member that for this reason seldom outcrops. Conversely, the upper Member widely outcrops In the northern part of the study area, the Bellerophon Formation is capped by Werfen Formation (Early Triassic). Werfen Formation incorporates six different members. The lower member consists of an oolitic limestone with an average thickness less than 7 m. Middle members are made up by limestones and marls, dolomitic limestones, marls and pelites, alternating calcareous sand and mud with a thickness of approximately 200 m. The Formation ends up with a member of fine-grained violet sandstones and pelites, reaching a thickness of about 200 Moraine deposits partially mantle the valley bottoms and the flat parts of the highlands. Bivera Mountain reach an altitude of 2,474 m. In this municipality, evaporite rocks are not so common at the surface because they are overlyed by Werfen Formation or mantled by Quaternary deposits. The presence of sinkhole phenomena is historically known in this area (Calligaris et al., 2009). Tens of sinkholes were ridges and in the middle of the slopes. Their presence partially compromises urban expansion and consequently affects land use planning. In the villages and settlements, facilities and inhabitants (Figure 2). Geological Overview of the Area From a geological viewpoint, Bellerophon Formation and Werfen Formation are the most important. The Bellerophon Formation is a regionally extensive unit, outcropping from Slovenia to Veneto Region. The intensity of the alpine tectonic deformations strongly affected its stratigraphic continuity. In fact, it is difficult to find a comprehensive section and it frequently appears cataclastic (Venturini et al., 2006). It is difficult to define the original thickness of this unit due to the presence of several thrusts and high rates of gypsum dissolution (Buggisch and No, 1986). The unit is mainly outcropping in a wide belt coincident with the valley bottoms. In the study area, the main discontinuity is the E-W oriented Sauris Fault (Figure 3), which follows the Figure 2. Aerial (A) and oblique (B) views of a sinkhole located near Sauris Village.

PAGE 4

468 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The above-mentioned activities allowed us to identify sinkholes. All the desk data were validated by field surveys (Figure 2). Particular attention was devoted to the selection of the sinkhole classification. In the present paper, we used the methodology developed by the encountered evaporitic phenomena. and are common in the northern part of the study area Methods and Results steps have included the analysis of historical documents collected in archives and the study of scientific papers and technical reports. These investigations have permitted us to define the geological setting of the Sauris area and to outline the possible locations of sinkholes associated with the carbonate and evaporite rocks. These preliminary stages were integrated with the interpretation of aerial photos and ALS surveys acquired recently by Regional Civil Protection. Figure 3. Spatial distribution of sinkholes affecting Sauris Municipality. The different colors indicate the types of sinkholes. 28= Quaternary deposits; 23= Moraine deposits Late Pleistocene; 12a= Carbonates alternated with marls Late Trias; 11= Carbonates alternated with marls Late Trias; 10b= Vulcanites Trias Middle; 10a= Sandstones and shales (flysch) Middle Late Trias; 9= Massive carbonates Middle Late Trias; 8a= Massive carbonates Middle Trias; 8b= Sandstones and shales (flysch) Middle Trias; 7= Sandstones and shales (flysch) Werfen Fm. Early Trias; 6c= Carbonates alternated with marls (Bellerophon Fm.) Late Permian; 6b= Evaporites (Bellerophon Fm.) Late Permian. In red the main thrusts (after Carulli, 2006); SF= Sauris fault.

PAGE 5

469 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 are associated with heavy rainfall, which often exceeds 2,000 mm per year in northern Friuli, and mainly by the presence of a torrent, which accelerates the dissolution processes and the gypsum erosion. The cover suffosion sinkholes (5) and cover collapse sinkholes (2) are mainly located near Sauris di Sotto and involve the Quaternary glacial deposits. Conclusions The results of the integrated analysis has produced an inventory of sinkholes associated with the evaporite rocks in the Sauris Municipality. Here, different types of karst phenomena occurred and involve mainly the Werfen Formation, which overlies the evaporites of the Bellerophon Formation. The sinkholes located in the surroundings of the top of mountains situated in the northern part of Sauris Municipality, affect a 200 m thick of poorlykarstifiable formation. This is due to the different mechanical properties of the Werfen and Bellerophon Formations. Conversely, (Figure 4). Caprock sagging sinkholes involve the Werfen Formation, which overlies the Bellerophon Formation. For this type of sinkhole the circular or subsinkholes. The left sinkhole (A) is a caprock collapse right (B) caprock sagging sinkhole is in an early stage of evolution. Although bedrock collapse sinkholes are extremely rare, two spectacular features of this type were observed 500 meters East of Sauris di Sotto Village. These phenomena involve Bellerophon Formation rocks and are limited by steep slopes (Figure 6). The diameters exceed 100 m and bottom. These karst depressions can reach depths of tens of meters. The sinkhole slopes are affected by abundant falls of trees and debris. These bedrock collapse sinkholes Figure 4. The bars indicate the different type of sinkholes. Caprock sagging sinkholes are dominant. Figure 5. Oblique view of two sinkholes affecting Werfen Formation. Figure 6. View of a 30 m deep bedrock collapse sinkhole.

PAGE 6

470 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE subsidence: Effects on alluvial systems and Range, Spain). Geomorphology 16: 277-293. subsidence in the historical city of Calatayud, Spain: damage appraisal and prevention. Natural classification of sinkholes illustrated from evaporite paleokarst exposures in Spain. Environmental Geology 53: 993-1006. Nisio S. 2008. I sinkholes nelle altre regioni. Memorie Descrittive della Carta Geologica dItalia 85: 419-426. 569-581. Albania: main feature and cases of environmental degradation. Environmental Geology 53: 967-974. Taminskas J, Marcinkevicius V. 2002. Karst geoindicators of environmental change: The case of Lithuania. Environmental Geology. 42: 757-766. Venturini C, Spalletta C, Vai GB, Pondrelli M, Fontana Geologica dItalia alla scala 1:50.000, Foglio 31 Zini L, Calligaris C, Devoto S, Zavagno E, Forte E, Boccali C, Petronio L, Cucchi F. In press. Fenomeni di sprofondamento nella piana di cavit sotterranee: ricerca storica metodi di studio ed intervento. 8 May 2014 Roma. bedrock collapse sinkholes are scarce but are depth exceeds 35 m. The spatial distribution of sinkholes does not coincide with any particular structural alignment even if the investigated area is crossed by several regional faults, which are oriented approximately E-W. Thanks to the financial support of Regional Geological Survey, the research activities here presented are not limited to Sauris area but include all the municipalities inventoried approximately 200 sinkholes, spreaded characteristics of each phenomena populate a Geodatabase. The latter is crucial to assist local associated to evaporite rocks and can be used for landplanning purposes. References Buggisch W, No S. 1986. Upper Permian and PermianTriassic boundary of the Carnia (Bellerophon Memorie Societ Geologica Italiana 34: 91-106. Calligaris C, Zini L, Cucchi F, Stefanelli N. 2009. sinkholes. Proceedings of the 2Seminario I sinkholes: gli sprofondamenti catastrofici ISPRA. Roma, 213-221. Caramanna G, Ciotoli G, Nisio S. 2008. A review of natural sinkhole phenomena in Italian plain areas. Giulia alla scala 1:150.000 e Note Illustrative. Cooper AH. 2008. The GIS approach to evaporiteGeology 53: 981-992. Gortani M. 1965. Le doline alluvionali. Natura e montagna 3: 120-128. and active subsidence due to evaporite dissolution processes, spatial distribution and protection measures for transport routes. Engineering Geology 72: 309-329.



PAGE 1

501 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 EXPERIMENTAL AND NUMERICAL INVESTIGATION OF SINKHOLE DEVELOPMENT AND COLLAPSE IN CENTRAL FLORIDA Abstract The mechanisms of sinkhole formation, development, and collapse are investigated in this study using experi mental and numerical methods. Sandbox experiments are conducted to understand how excessive groundwater pumping triggers sinkholes formation. The experimental results indicate that the change of hydrologic conditions is critical to sinkhole development. When seepage force increases due to increase of hydraulic gradient, clay and sand particles start moving downward to form a cavity. ceptual model developed from the sandbox experiments, the Fast Lagrangian Analysis of Continua (FLAC) code and Particle Flow Code (PFC) are coupled to simulate the sandbox experiments. PFC was used to simulate par ticle movement in the sinkhole area, and FLAC is used for other areas. While the current numerical simulation the cavity and the sinkhole, the simulation capability is limited by the computing cost of PFC. More effort of model development is necessary in the future study. Introduction Sinkholes are a common geological feature of karst landscape in Florida, southeastern United States, and worldwide. In particular, cover-collapse sinkholes oc cur abruptly and can cause catastrophic damages such as resident vanished into a sinkhole that opened under his bedroom on a night in March, 2013. In the last several years, sinkholes have become Floridas insurance di saster due to sinkhole collapse in urban areas. Covercollapse sinkholes also do severely damage buildings, drain farm ponds, damage roads, and wreck farming equipment, and lead to engineering and environmental problems (Beck, 1988). There is an urgent need to un derstand the mechanisms of sinkhole development and catastrophic collapse. Cover-collapse sinkholes occur in the soil or other loose material overlying soluble bedrock. The thickness and cover-collapse sinkhole. Figure 1 shows a typical pro cess of cover-collapse sinkholes formation caused by excessive groundwater pumping. A karst aquifer is the Xiaohu Tao Geophysical Fluid Dynamics Institute, Florida State University, Tallahassee, FL, 32306 College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing, Jiangsu, 210098, P.R. Chi Ming Ye FL 32306, U.S.A, mye@fsu.edu Dangliang Wang FL 32306, U.S.A, dlw9800@163.com Roger Pacheco Castro FL 32306, U.S.A, matbnt@gmail.com Xiaoming Wang Department of Mathematics and Geophysical Fluid Dynamics Institute, Florida State University, Tallahassee, FL 32306, U.S.A, xwang@fsu.edu Jian Zhao College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing, Jiangsu, 210098, P.R. Chi na, zhaojian@hhu.edu.cn

PAGE 2

502 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE pre-requisite, and sinkhole development always starts from dissolution of soluble rocks or fractures and con duits to create an opening, which provides a passage for soil transport downward. Groundwater is one of the pri mary triggering mechanisms for sinkhole development and collapse, because seepage force due to groundwa 1(A) shows two layers with cohesive and non-cohesive and is beneath the non-cohesive soil layer, which is com mon in central Florida. In the initial stage of sinkhole development, an opening forms in the rock at the inter face between the bed rock and cohesive soil. While the metric surface decreases (Figure 1(B)) and downward the cohesive soil layer starts falling down due to gravi tational force and seepage force applied on the particles (Figure 1(B)). This forms a cavity in the cohesive layer. The cavity gradually increases, and it in turn increases hydraulic gradient and thus seepage force. Because of these, the cavity expansion accelerates (Figure 1(C)). Once the cavity expands to the non-cohesive soil layer, sand movement becomes dramatic. When the non-co hesive soil layer cannot support the overlying material, collapse occurs and a sinkhole propagates to land surface (Figure 1(D)). While the process of cover-collapse sinkhole formation triggered by groundwater pumping has been understood, mathematical models and numerical modeling tools have not been available for predictive understanding. A coupled model based on FLAC and PFC was used to simulate the soil-structure interactions during a sinkhole event (Caudron et al. element analyses to detect three-dimensional (3-D) de formations due to submerged cavities that lead to sink hole. Tharp (2003) employed an elastic-plastic model to demonstrate the development of a sinkhole above a karst cavity. Shalev (2012) adopted a two-dimensional (2-D) visco-elastic model to simulate the sinkhole formation to take into account of the brittle and ductile aspects of sinkhole collapse. Baryakh et al. (2009) established a nu merical model that uses the discrete element method to simulate the evolution of the stress-strain state of a rock mass containing a karst cavity. Baryakh and Fedoseev cavern to describe possible scenarios of sinkholes devel opment in the karstic areas, to determine formation cri teria for ground surface sinkholes and underground cav et al. (2006) simulated the dissolution of salt layer and The numerical simulation showed the growth of cavities from the bottom to the top of the salt layer, and suggest ed that sinkhole collapses shortly after the cavities reach the top of the salt layer. These modeling effort suggests that, while the continuum theory can estimate the stressaccount the change of cavity geometry such as enlarge ment of the cavity in the cohesive soil layer. While discontinuum theories can be used to resolve this problem, the dis-continuum theories are computationally intensive and not always practically affordable. Sandbox Experiment Sandbox experiments are conducted to better understand the process of sinkhole development and collapse. Fig ure 2 shows the schematic view of experiments. A sand box of 150 cm 120 cm 20 cm was constructed with plastic material. There are four tanks to control the water hydrogeologic al materials in three layers. The bottom one (in black) represented a karst aquifer with void space. A clay layer (in yellow) overlaid the karst layer three opening were designed, but only the one of 1 cm in the middle was used in this study to create a sinkhole in the middle of the sandbox. Above the clay layer was Figure1. Model for cover-collapse sinkhole.

PAGE 3

503 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 by the inner reservoirs at the both sides of the sandbox. The outer reservoirs were used to control hydraulic head The sandbox experiments are designed to understand impacts of groundwater pumping on sinkhole develop ment and collapse. The impacts are believe to be the City area during the winter of 2010, when more than 100 sinkhole collapses were triggered by excessive ground water pumping for irrigation to prevent crops from be mains constant. After the drop of hydraulic head in the downward through the opening due to the seepage force and slowly expands upward. Once the cavity reaches the sand layer, sinkhole development is accelerated, and sinkhole collapse occurs shortly because of small cohe sion of the wet sand. Figure 4 shows the picture after sinkhole collapse. FLAC/PFC coupling approach In this study, we use the continuum and dis-continuum FLAC, based on the continuum theories with the discrete element code, PFC, based on the dis-continuum theo ries (both FLAC and PFC are developed by the Itasca Consulting Group, Inc.). Since PFC is computationally demanding, it is only used for the small area of exces sive displacement above the opening. FLAC is less com putationally demanding, and thus used to simulate the larger area of small deformation away from the opening. computational requirement for simulating the process of sinkhole development and collapse. Coupling FLAC and PFC ing displacements, velocities, and forces at each model ing step. The data exchange is made possible by the I/O socket connection ability to pass data rapidly between Figure 2. Schematic of sandbox experiments. Figure 3. Photo of sandbox experiments. Figure 4. Sinkhole collapse in a sandbox experi ment.

PAGE 4

504 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the two codes running on the same machine or on sepa rate machines with a network connection. As shown in Figure 5, the data exchange is two-directional between FLAC and PFC. In each step of numerical simulation, the velocity at the interface between FLAC and PFC FLAC run and then sent to PFC via the I/O socket. Af ter receiving the data from FLAC, PFC starts to update the forces at the interface and then send the results of forces back to FLAC via the I/O socket. Afterward, the simulation moves to the next time step, and the iteration continues until the end of simulation time. Numerical simulation Table 1 lists the values of the parameters used for the nu merical simulation. While the clay/sand particle move ments change hydraulic conductivity, to simplify the numerical simulation, it is assumed that the deformation and the particle transport have negligible effect on the hydraulic conductivity and that hydraulic conductivity is constant during the process of sinkhole development. using the Darcys Law and heat equation. The FLAC/PFC simulation is set up as shown in the placements at the bottom boundary in the FLAC model ing area. For the PFC modeling area, the bottom bound aries are the two walls (Figure 6) with the distance of 1cm. At the initial time, the model is in the steady state with the hydraulic head of 0.45m (the datum is at the bot hydraulic pressure and pressure gradient is calculated and then passed to the FLAC-PFC-based mechanical Figure 5. Coupling of FLAC, PFC, and ground Figure 6. Illustration of modeling domain. PFC is used for the black area, and FLAC is used for rest of the area. Variables Clay Sand Density (Kg/m3) 2200 2600 Bulk modulus (Pa) 7.00E+05 1.30E+07 Shear modulus (Pa) 4.00E+05 8.00E+06 Cohesion (Pa) 8.00E+05 0 Friction angle () 25 35 Hydraulic conductivity (m/s) 1.00E-08 5.00E-05 Table 1. Parameter values of soil properties

PAGE 5

505 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 modeling (Figure 5). The time step used in the mechani cal modeling is smaller than the time step used in the cavity geometry due to particle movements. The new cavity geometry (i.e., the cavity boundary) is passed to Results of Numerical Simulation lected, and the time step of the mechanical modeling is A total of 20,000 particles are used for simulating the clay layer and 8,000 particle for the sand layer. The simulation results at the time of 1s, 3s, 5s, 8s, and 15s hydraulic head distribution. Cavity expansion Figure 7 shows the cavity in the clay layer obtained in a sandbox experiment and at t = 3s, 8s, and 15s of the numerical simulation. The numerical modeling is able to simulate expansion of the cavity. The shape of the simu lated cavity is similar to that of the experimental cavity. For example, the angles between the experimental cav 49~50 (Figure 7), and the corresponding angles of the simulated cavity are about 48~54. Hydraulic pressure and head along the vertical line perpendicular to the opening at the in particular in the early time when the sinkhole starts forming. The pressure change has substantial impacts on cavity geometry and expansion. At the beginning of cav ity formation the largest hydraulic gradient occurs at the point of the opening, and it induces a large seepage force on the particles and causes downward movement of clay particles. As a result, the cavity will expand upward to the sand layer until sinkhole collapse. The pressure opening but at its vicinity, as shown in Figure 9 that plots the spatial distribution of hydraulic head in the entire These results indicate that, during the process of sinkhole formation, monitoring hydraulic pressure and hydraulic two quantities do not change with time. The reason is that the clay layer isolates the pressure propagation to Figure 7. Cavity in the clay layer obtained in a sandbox experiment (top) and numerical simu lation at t = 3s, 8s, and 15s. Figure 8. at different simulation times.

PAGE 6

506 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE at this stage, the sinkhole has been largely formed, and lenge is to determine where to install monitoring well in Conclusions This paper presents a laboratory study for understand ing the processes of sinkhole formation, development, and collapse. The experiments are useful not only to illustrate the sinkhole processes but also to develop a conceptual model for mathematical and numerical mod of clay is important to sinkhole formation and develop ment. This paper also demonstrates an approach that uses FLAC and PFC to simulate the laboratory experi ments. The results of the numerical simulation are con sistent with the phenomena observed in the experiments, in particular the expansion of the cavity due to particle movement caused by the increase of seepage force after ing layer is of particular importance, because the cavity sinkhole in its early formation. More effort is warranted to develop more robust numerical models for simulating sinkhole events in the future research. Acknowledgements This research is funded by FSU MultiDisciplinary Sup the China Scholarship Council for their study at FSU. The fourth author was supported by the Fulbright Schol arship of the U.S. Department of State. References Ahmed M. 2013. Experimental and Numerical Modeling of Sinkhole Collapse. In: Proceedings of the Transportation Research Board 92nd Annual Meeting, 13-17 Jan 2013. Washington, D.C. strain state of karst areas. J Min Sci 45 (6): 517524. Baryakh AA, Fedoseev AK. 2011. Sinkhole formation mechanism. J Min Sci 47 (4): 404-412. Beck BF. 1988. Environmental and engineering effects of sinkholesthe processes behind the problems. Environ Geol Water Sci. 12 (2): 71-8. Caudron M, Emeriault F, Kastner R. 2006. Collapses of underground cavities and soil-structure interactions: of the Proceedings of the 1st Euro mediterranean symposium on advances on geomaterials and structures, 3-5 May 2006, Hammamet, Tunisia. Shalev E, Lyakhovsky V. 2012. Viscoelastic damage modeling of sinkhole formation. Journal of Structural Geology. 42 (0): 16370. Shalev E, Lyakhovsky V, Yechieli Y. 2006. Salt dissolution and sinkhole formation along the Dead Sea shore. Journal of Geophysical Research: Solid Earth 111(B3): B03102. Tharp T. 2003. Cover-collapse sinkhole formation and soil plasticity. In: Beck BF, editor. Sinkholes and the engineering and environmental impacts of Publishing. p. 110-123. Figure 9. Spatial distribution of simulated hy draulic head (m) at t = 1s, 8s, and 15s.



PAGE 1

Ordovician Karst of Southeast Minnesota Field Trip Guidebook 6 October 2015 Field Trip 14th Sinkhole Conference Rochester, Minnesota, U.S.A. Calvin Alexander & Jeff Green, Leaders Tony Runkel & John Barry, Co Leaders Fountain Big Spring 1

PAGE 2

Sinkhole Conference Field Trip Outline and Timeline 7:30 Leave Civic Center both buses together 7:45 Stop 1 Paragon Chateau 14 Theater, St. Peter Sandstone Formation; 30 min. 8:15 Leave Stop 1 8:30 Stop 2 (Rolling) Golden Hills road cut, Platteville, Decorah and Cummingsville Fo rmations. 8:45 Stop 3 Orion Township Dye Traces/CAFOs; 30 min. 9:15 Leave Stop 3 9:45 Stop 4 Lower Shakopee Formation, Prairie du Chien Group road cut south of Chatfield at Middle Branch Root River; 30 min. 10:15 Leave Stop 3 Buses separate Bus 1 (low stair climbing) turns hard right on Co. 8 at the northwest corner of Fountain and then right on Keeper Road to Fountain Big Spring. This group will be led by Jeff Green and Tony Runkel 10:30 Stop 5 Fountain Big Spring Group; 30 min. 11:00 Leave Stop 5 11:15 Stop 6 Root River Trail Sinkhole Kiosk; 30 min. 11:45 Leave Stop 6 12:15 Stop 8 Mystery Cave Visitor's Center; 2 hr. and 15 min. Split into two groups. One group eats box lunch while other group tours cave & vice versa. 14:30 Leave Stop 8 Bus 2 (lots of stair climbing) turns left on Co. 8 at the south edge of Fountain and goes to the Fountain sinkhole kiosk parking lot. This group will be led by Calvin Alexander and John Barry. 10:30 Stop 6 Root River Trail Sinkhole Kiosk; 30 min. 11:00 Leave Stop 6. 11:15 Stop 5 Fountain Big Spring Group; 30 min. 11:45 Leave Stop 5 12:15 Stop 7 Niagara Cave; 2 hr. and 15 min. Split into two groups. One group eats box lunch while other group tours cave & vice versa. 14:30 Leave Stop 7 2

PAGE 3

Buses rejoin 14:45 Stop 9 Cherry Grove Blind Valley SNA/Goliath's Cave; 1 hour 15:45 Leave Stop 9 16:00 Stop 10 Wykoff Storm Water Treatment System, 30 min. 16:30 Leave Stop 10 17:30 Rochester Civic Center, End of field trip. Figure 1. Field trip route 3

PAGE 4

Introduction This field trip is designed to acquaint you with the Paleozoic bedrock aquifers of southeastern Minnesota. A major portion of Minnesota's population live s on top of those aquifers and depend s on them for their domestic, agricultural and industrial water su pplies T hese aquifers provide the base flow for the rivers and streams in the area. S ociety's ability to manage this critical groundwater resou rce becomes ever more important as p opulation grow s and water usage steadily increases One major challenge is t hat these Paleozoic aquifers are dual or triple porosity fractured and karst aquifers. T he movement of water in these aquifers is not adequately described by isotropic, homogeneous porous media mechanics. Unfortunately, those mechanics are built into the s cientific and public culture of ground water management. Darcy's equations seem inextricably imbedded in ground water management efforts at every level. This field trip will focus on the heterogeneities in the routes that water uses as it moves from rechar ge areas to discharge areas though the Paleozoic aquifers. We will look at a variety of conduits, pipes, high transmissivity zones and other ground water flow paths. We will see porosity that ranges from the resolution of our vision to voids that houses co uld be built in. 4

PAGE 5

Fi gure 2 Minnesota Karst Lands Map The colored portions of this map are underlain by carbonate bedrock. In the green and yellow areas there is >100 feet and 50 to 100 feet of glacial cover. Almost all of the surface karst features are in the < 50 feet of cover, red areas. Figure 2 is a cross between the bedrock geology and depth to bedrock maps in southeast Minnesota In t he red areas, active karst occur s where the cover over the bedrock is less than about 50 feet thick. The images on the next two pages show the bed rock lithostratigraphy in more detail. Figure 3 is a conventional stratigraphic column for the Paleozoic bedrock of southeast Minnesota. The vertical scale is proportional to the thickness of the individual rock units. Uncon formities are shown but represent gaps in deposition and or erosional removal of previously deposited rocks. Figure 4 is the same information plotted on a chronostratigaphic column where the vertical scale is linear time. This figure illustrates that the u nconformities represent large gaps during which karst processes were operating. Many of the karst features we will see today may be traced back to syndepositional features. Many of the karst features may be very old. 5

PAGE 6

Figure 3. Paleozoic l ithostratigraphic column of SE Minnesota bedrock The vertical scale is linear unit thickness. U nconformities accented in bla ck (Alexander et al., 2013) 6

PAGE 7

Figure 4. Chronostratigraphic column of SE Minnesota Paleozoic rocks The vertical scale is linear time (Alexander and others 2013) 7

PAGE 8

Stop 1 St. Peter Sandstone and Glenwood Shale Parking lot, Paragon Chateau 14 Theater, 37th St. and US 63 in Rochester, Olmsted Co., Minnesota 44 ¡ 03' 33.82" N; 92¡ 26 58.93" W 1039 ft. elevation The St. Peter Sandstone is fine to medium grained, friable, quartzose sandstone, which is typically about 100 feet thick in Olmsted County. Some exposures are silica cemented but much of the Formation is so friable that it can be dug with a shovel or by hand. Freshly exposed surfaces of the St. Peter are typically bright white to light gray and often display bright yellow to red to purple Liesegang banding near its top. Newly exposed surfaces develop a hardened gray to black surface within a year or two. The St. Peter is widely exposed in both natural and artificial faces. The basal con tact with the underlying Shakopee is rarely exposed (a n erosional unconformity and interstratal karst surface ) The St. Peter should be close to an ideal porous medium However, as you look at this face note the large, systematic, open, near vertical joint s that are exposed. We will pass several exposures of St. Pete in road cuts on this trip. Watch for the big open joints. They are always there. The St. Peter is at least a dual porosity aquifer. In addition this face has exposed a large filled crevice in the St. Peter. This cryptic feature is not well understood. The fill does not appear to contain any glacial materials and so is probably pre Pleistocene. Note the complex smaller joints near the edges of the fill. Is this a filled feature formed by collaps e into the underlying Prairie du Chien? Engineering boreholes in the Rochester area have encountered a significant number of open voids in the St. Peter. The foundations of several of the buildings in Rochester had to be modified and upgraded when such voi ds were found. There are several small caves known in the St. Peter in Minnesota and sinkholes have been mapped where the St. Peter is first bedrock. Figure 5 Cut face of the St. Peter Sandstone at the Paragon Chateau 14 Theater parking lot Rochester, Minnesota. Both systematic joints and a large filled fracture are visible. 8

PAGE 9

In October 2005 a stormwater retention pond in a new housing development in Woodbury, Minnesota was filled for the first time by a heavy rain fall. Within a few days about a dozen sinkholes developed in the sides and bottom of the pond and drained 60 acre feet of water into the underlying St. Peter and Prairie du Chien aquifers. The largest sinkhole developed on top of a feature that probably resembled the filled crevice visible at this stop. Figure 6 Catastrophic collapse sinkholes in the St. Peter Sandstone These drained a storm water retention pond in the Dancing Waters development, Woodbury, Minnesota in the southern Twin Cities metropolitan area 9

PAGE 10

Figure 7 Close up photo o f the largest sinkhole that drained the Dancing Waters storm water retention pond There is a thin delta of recent sediments over an 18 inch clay liner, over a varying wedge of glacial deposits over the St. Peter Sandstone. 10

PAGE 11

Stop 2 (Rolling) Golden Hill Cut Western Face Road cut on US 52 between the Zumbro River and US 63 in southwestern Rochester Rochester Olmsted Co., Minnesota 43¡ 59' 18" N; 92¡ 28' 47" W The Golden Hill Road Cut on US 52 has long been a favorite stop with geologists and hydrogeo logists. T he road cut was expanded to the west i n 2003 and 2004 as part of a M innesota Dep artment of Transportation project to widen and rebuil d US 52 through Rochester This new excavation work exposed a beautiful section from the top of the St. Peter San dstone through the Glenwood Formation the Platteville Formation, the Decorah Shale and the bottom section of the Cummingsville Formation of the Galena Group. In addition to excellent fossil hunting in the exposed Decorah Shale, the road cut exhibits seve ral well exposed hydrogeologically significant features. As noted above, however, we may not be able to stop. Numerous solutionally enlarged joints cut the Platteville Formation, which is exposed as a vertical face. Secondary calcite flowstone is visible o n several of the exposed joint faces along with iron staining showing where groundwater has moved through the bedrock. The Decorah Shale is exposed as a cut slope for stability and to help control land slips. Water is normally visible in the roadside ditch accumulating from flow in and at the top of the Decorah. Note that the Decorah Shale has a distinctly rusty orange "oxidized" tint in the northern 100 meters of the exposure while the rest of the Decorah Shale is a gray/green "reduced" color. The alterna ting bands of carbonate and shale are clearly exposed in the vertical Cummingsville wall at the top of the road cut. There is considerably more topography on the eroded top of the Cummingsville than there is on the overlying land surface. In addition to nu merous joints, several small conduits intersect the cut face. Some of these are currently actively carrying water. Others of them are currently dry. This difference is most obvious in the winter when large accumulations of ice form on the cliff faces where the groundwater emerges into the subfreezing air. The movement of water through the Cummingsville Formation i s neither homogeneous nor isotropic. 11

PAGE 12

Figure 8 West Side of Golden Hill Road Cut looking north March 2005, photo by Terry Lee. Figure 9. We st side of Golden Hill Road cut, looking west March 2005, p hoto by Terry Lee. 12

PAGE 13

Stop 3 Schoenfelder Farms Blue Ridge CAFO Section 5, Orion Township, Olmsted Co. MN 43¡ 55' 30.24" N, 92¡ 17' 18.71" W 1273 ft. elevation Schoenfelder Farms owns a large cattl e feedlot operation (Blue Ridge) centrally located in the Orion sinkhole plain. The feedlot was initially permitted by the MPCA in 1991. The Schoenfelders' have been researching solutions to update the site and meet current pollution control laws. Accordin g to federal feedlot laws, the facility must comply with a "zero runoff" status and therefore must eliminate or collect all water leaving the feedlot. Many proposals have been presented by Schoenfelder Farms and their private engineers to the MPCA, but cha llenges have set back construction for three seasons. In attempts to find a suitable site for a large manure storage structure, many locations have been evaluated at Blue Ridge. Three primary challenges exist for locating the proposed storage area. Minnes ota feedlot rules require the following. 1. Due to the proximity of sinkholes within the plain, a liquid manure storage area (LMSA) cannot be constructed on most of the Schoenfelder property. According to Minnesota Feedlot rules LMSAs must not be located wit hin 91 meters (300 feet) to any sinkhole. ( Minn. R. 7020.2005 Subp. 1.) 2. The total size of the LMSA may not be larger than 950,000 liters (250,000 gallons) when four or more sinkholes exist within 305 meters (1,000 feet.) The required volume for the facilit y is significantly more than 960,000 liters, and many sites would not allow for the 22 million liter (6 million gallon) or larger structure. ( Minn. R. 7020.2100 Subp. 2. A.) 3. Minnesota feedlot rules require a minimum vertical separation to bedrock of 3 mete rs (10 feet) when the site has a capacity of 1,000 or more animal units (1 slaughter steer is equal to 1 animal unit). Most of the sinkhole plain has soil cover of less than 1.5 meters. ( Minn. R. 7020.2100 Subp. 2. B. [ 3 ] ) Since the feedlot is located at t he northern edge of the plain, the only potentially suitable location is 1000 feet north in an open field. One small corner of the Schoenfelder property appears to allow for separation to sinkholes and vertical separation to bedrock. In the fall of 2014, t est pits were dug with an excavator to determine separation to bedrock. As of March, 2015 research is continuing to find a solution that meets Minnesota Feedlot Rules and is a viable option for the Schoenfelders'. Additional geophysics exploration may inc lude Electrical Resistivity Imaging in the spring of 2015. This stop and surrounding issues will be discussed by Larsen (2015) during the Conference. 13

PAGE 14

Figure 10. A map of the Orion Sinkhole Plain (Larsen, 2015) Figure 11. A pail of foaming water col lected from the contaminated well Photo by Steven Schmidt, 2013 14

PAGE 15

Stop 4 Shakopee Formation, Prairie du Chien Group Road cut on US 52 just north of the Middle Branch of Root River, Chatfield Township, Fillmore Co., Minnesota Stop 4 is stratigraphically t he lowest stop on this field trip near the middle of the Ordovician Prairie du Chien Group The top of the New Richmond Sandstone is just above river level. It was exposed during construction of this road cut and bridge but was subsequently covered. The ve rtical road cuts are in the Willow River Member, Shakopee Formation, and Prairie du Chien Group. The thin to medium bedded dolomite of the Shakopee Formation contain s thin interbeds of quartzose sandstone and shale. There is a karst erosional unconformity between the Shakopee and lower Oneota Formation. The paleo epikarst developed on this contact is proving to be a significant horizontal high transmissivity zone in Prairie du Chien. The most obvious solution features in these road cuts are the large soluti onally enlarged vertical joints that are filled with rusty brown sediments. The spacing of these systematic joints is similar to those observed in the St. Peter outcrops. There is a lot evidence for bioherms and complex, small scale deformation in these ou tcrops. Many of the Prairie du Chien outcrops show evidence of solution brecciation and many contain drusy quartz and sulfide minerals from low grade hydrothermal solutions. The Prairie du Chien is one of the most heavily used aquifers throughout southeas tern Minnesota. Downhole flow logging by the M innesota Geological Survey has shown that most of the water movement in the aquifer typically takes place along thin "high transmissivity" zones. The Prairie du Chien is also commonly quarried for road aggregat e. The massive bed of carbonate that makes up the bottom of this outcrop is one of the most valuable beds in the entire Prairie du Chien section. This bed, when quarried and crushed yields the highest quality gravel for highway construction. 15

PAGE 16

Figure 12 Low er Shakopee Formation road cut Both major solutionally enlarged vertical joints and a solutionally enl arged subhorizontal features. Note also the thin soil cover over the bedrock but no sinkholes or other surface karst features. 16

PAGE 17

Stop 5 Fountain Big Sp rings, MN23:A0037 Cummingsville Formation, Galena Group Fountain Township, Fillmore Co., Minnesota As you dr i ve up onto the upper Galena Group north of the town of Fountain, MN, you enter a sinkhole plain geomorphic unit. R ed "x"s are s uperimposed on the LiDAR image in Figure 13 of the Fountain area marking many, but not all, of the sinkholes. These sinkholes quickly route any surface runoff into the underlying karst aquifer. The light blue dots mark many, but again not all, of the springs where the groun dwater returns to the surface. The dozens of dye traces that have defined the three springsheds shown in Figure 13 have been omitted for clarity. Figure 13. Fountain Sinkhole Plain. Stops 5 and 6. Stop 5 is a t the major spring complex that drains much of the area. Many of the sinkholes in and around Fountain drain to the spring complex at Stop 5 with a time scale of hours A map of the largest of these springs, Fountain Big Spring MN23:A00037, is shown as Figure 14 detailing the current and former o utlet points of Fountain Big Spring's orifices which highlights the fact that springs move. There are several "Big Springs" scattered around SE Minnesota, hence the unique identification number assigned to each specific spring and other karst feature. 17

PAGE 18

Figure 14. Fountain Big Spring MN23:A00037 Diagram by Holly Johnson, Minnesota DNR 18

PAGE 19

Figure 15. Fountain S pring Group Diagram by Holly Johnson, Minnesota DNR These four springs are the head of Rice Creek. One of the conduits feeding th e Fountain S pring Group can be accessed, via open crevices, at least four places in the quarry just south of the springs. The conduit is oval in cross section about 4 feet high and about 10 feet wide. It has smooth, sinuous walls. Stand up spaces develop when the conduit c rosses one of the major systematic joints cuts. Figure 15 shows t he four distinct but interacting springs emerg ing within a few hundred feet of each other near Stop 5 All four springs emerge near the base of the Cummingsville Formation in the Galena Group just above the Decorah Shale. These are typical conduit springs. Their flow often increases by an order of magnitude within a couple of hours after a rainfall and the springs become very sediment laden. 19

PAGE 20

Springs can become plugged by colluvium or debris or the spring can develop a new lower elevation outlet either naturally or through human intervention. Fountain Big Spring has moved in the not too distant past. The previous outlet is a few feet down the valley and is now acting as a high flow overflow spr ing. That this was the main spring outlet for a long period is evident from the size of the notch or "steep head" that was eroded into the valley wall by the old, higher spring outlet. Construction of the road down the ravine may have caused the new openin g of the spring to form. Little Quarry Spring, MN23:A00045, emerges from a joint/bedding plane intersection across the road from Fountain Big Spring. Under some flow conditions water entering sinkholes in Fountain emerges from the springs on both sides of the road. Under other conditions it emerges from one or the other depending on where it sinks. Stage dependent changes in the flow systems feeding karst springs are common. Quarry Spring, MN23:A 0 0051, emerges around the point of the ridge east of Little Qu arry Spring in an old limestone quarry. Little Falls Spring, MN23:A0044, emerges on the south side of the ridge north of Quarry and Little Quarry spring. The spring runs from all four springs converge at the culvert just north of Fountain Big Springs to fo rm the headwaters of Rice Creek which flows north and then northeast to the Middle Branch of the Root River, downstream from Stop 5 Dye traces from many of the sinkholes in the Fountain Sinkhole Plain have defined two major subsurface springsheds that dr ain the Fountain Sinkhole Plain. Most of the sinkholes drain to the four springs at Stop 5 The subsurface drainage cuts across several first order surface drainage basins. A smaller set of the sinkholes in the southeastern corner of the Sinkhole Plain dra in to a spring that is the headwaters of Mahoney Creek which flows east to the Middle Branch of the Root River. The question of why so much of the groundwater drains to the four springs at Stop 5 is beginning to be understood as the result of detailed stru ctural contour mapping on the area bedrock by the Minnesota Geological Survey (Tony Runkel, written communication, 2012 ; see also Runkel and others 2013 ). These four springs drain from the axis of an asymmetric syncline that is plunging to the northwest. Figure 16 below is a preliminary version of Tony Runkel's work showing the relationship between the bedrock geology, the sinkholes, dye traces, springs, and the synform. In the lower left you can see a couple of other nearby springsheds near the community of Wykoff. We are actively conducting additional dye traces and stratigraphic field work to refine this map. 20

PAGE 21

Figure 16. Structural c ontour map of central Fillmore C ounty This shows structural control of spring locations (Runkel, written communications 2012) 21

PAGE 22

Stop 6 DNR Sinkhole Kiosk, Fountain. From Root River Trail parking lot on south side of Co. 8 in southeast edge of Fountain proceed 0.4 miles along the trail to Kiosk. 43¡ 44' 08.97" N; 92¡ 07' 35.95" W 1295 ft. elevation At this stop, we will w alk on the Root River State Trail (which is built on an abandoned railroad right of way) to a sinkhole where the Minnesota Department of Natural Resources Parks and Trails Division installed a karst educational display. The display is designed to educate t rail users about karst land & water resources. The first bedrock here is Prosser Formation limestone of the middle Ordovician Galena Group. Unconsolidated material depth over the bedrock is less than 15 meters. This area is representative of the high densi ty sinkhole plains that occur on broad ridge tops underlain by the Prosser limestone or Stewartville limestone of the Galena in southeastern Minnesota and northeastern Iowa. The density of the sinkholes can be seen on the 2009 LiDAR imagery. This is the no rthern end of a broad arc of Galena that runs from southeast Minnesota through Iowa into Wisconsin and Illinois. As part of the display design, a dye trace was conducted from this sinkhole. Liquid fluorescein (206 gm.) was poured into the sinkhole's swall ow hole and was flushed with 4500 liters of water from a fire truck. The dye was detected at the Fountain Big Spring which lies to the northwest. Breakthrough flow velocity was 1 2 miles/day. The area around Fountain is in the groundwater springshed which feeds the Fountain Big Spring and three other smaller springs that are associated with it. 22

PAGE 23

Stop 7 Niagara Cave 2.5 miles south of Harmony on Minnesota 139, 2.3 miles west on Fillmore Co. 30 Harmony Township, Fillmore Co., Minnesota 43¡ 30' 50.16" N; 92¡ 03' 18.06" W 1265 ft. elevation http://www.niagaracave.com Niagara Cave will be one of the lunch stop s on the field trip. The group will split into two groups. One group will tour the cave while the other eats lunch. Then the two groups will reverse. Our hosts, the Bishop Family, will be happy to sell you a wide variety of interesting, geologically oriented souvenirs from their gift shop. Figure 17 is the map of Niagara Cave. Niagara Cave is an unusual upper Mis sissippi Valley cave. Most of the large caves in the southeast Minnesota/northeast Iowa karst are maze caves developed near the Galena/Dubuque contact (Mystery Cave, for example) or dendritic stream conduit caves developed near the bottom of the Galena in the Cummingsville Formation (Cold Water Cave, Tyson's Spring Cave, etc.) Niagara's entrance is a sinkhole through the Dubuque Formation. The cave cuts down through the Stewartville Formation in several dome pits and vadose canyons to a stream passage in th e Prosser Formation. 23

PAGE 24

Figure 17. Niagara Cave Map 24

PAGE 25

Dye tracing has shown that the cave stream returns to the surface in Hawkeye Spring on the edge of the Upper Iowa River. Niagara Cave contains the best, most accessible example of a cave stream, active vadose canyon and waterfall of any cave in the area. The landscape we have been driving across south of Rochester consists of an ancient, erosion surface that cuts across the Paleozoic stratigraphy. That ancient erosion surface probably dates back to at least Cretaceous time. It has been glaciated several times during the Pleistocene It was not covered by ice during the Wis consina n Glacial but did receive an agriculturally significant layer of lo ess at the end of the Wisconsina n. That flat surface is b eing dissected by modern streams (the Cannon, Zumbro and Root Rivers) which drain eastward to the Mississippi River. The landscape consists of flat topped interfluves (which are in intensive row crop agriculture) cut by steep sided recent valleys. Many of the springs emerge from near the bottoms of the steep valley sides and drain to the base flow rivers. The water tables under the interiors of the interfluves are often shallow and low gradient. Many of the base level springs emerge, at a significantly lowe r level, from low gradient stream caves which can extend for miles back into the interfluves. Niagara Cave shows how the groundwater moves from the high flat water tables to the low flat water tables. The ground water moves in high gradient conduits and wa terfalls between the two parts of the groundwater system. A fascinating system of up gradient stream piracies via dome pits is evident in Niagara Cave. 25

PAGE 26

Stop 8 Mystery Cave I Visitor Center Forestville /Mystery Cave State Park Forestville Township, Fillmo re Co., Minnesota 43¡ 37 02.25 N; 92¡ 18 42.58 W 1243 ft. elevation http://www.dnr.state.mn.us/state_parks/forestville_mystery_cave/index.html Mystery Ca ve is a n over 12 mile long maze cave developed in the Dubuque and Stewartville formations and is the longest mapped cave in Minnesota. The goal of this stop is to see the Mystery I Park and visit the Visitors Center. It was operated as a private commercial cave from 1947 to 1984 and then the cave was purchased in 1984 by the State of Minnesota and made part of the pre existing Forestville State Park. Guided tours are conducted in Mystery I along lighted, handicap accessible paths by D epartment of N atural R esour ces staff cave guides. These tours leave from the Mystery Cave Visitor Center. A plan view map of Mystery Cave is shown in Figure 18 The strong joint control on the cave's development is evident in the preferred directions of the cave passages. Mystery C ave is divided into three parts, Mystery I, II, and III, based on the history of the discovery of the sections of the cave. Mystery Cave contains several types of passages. Solutionally enlarged joints in the Stewartville Formation are typically straight, narrow passages, several 10s of feet high and often with a solution tube at the contact with the overlying Dubuque Formation. The straight crevices with tubes at their tops form a characteristic "key hole" cross section. Differential weathering of trace f ossils in the Stewartville produce characteristic "knobblies" on the surface on the sides of the crevices. The resulting passages are called Stewart ville Crevices and range from just wide enough to get one's body through to not quite wide enough to get o ne's body through. A second major passage type are "big Dubuque passages" which are rectangular shape passages formed by ceiling collapse of the Dubuque layers into lower solution voids These passages are often 10 to 30 feet wide with flat ceiling and s traight walls. A third significant passage type is large, solutionally sculpted passages in the Stewartville. These are only present in Mystery II as the east part of 5th Avenue and its continuation down the Garden of the Gods at the east end of the known cave. A fourth type of passage is the currently active stream passages in the Stewartville that make up the lower levels of the cave. The cave currently functions as a meander cut off for a series of entrenched bedrock meanders on the South Branch of the Root River ( cave cross section, Figure 19 ). The South Branch sinks into the cave, flows underground through the lower level passages of the cave, and then re emerges in the Rise of the South Branch at Seven, Crayfish and Saxifrage Springs. During all but flood flows a several mile long stretch of the River is completely dry. The South Branch is a 26

PAGE 27

low grade warm water fishery above its sinks. It becomes a h igh grade trout fishery down stream from the springs. Fishing for anything is lousy in the dry stretch Air filled passages in Mystery I extend under the Root River. The river is perched about 20 feet above the local water table. The sinking points move around the riverbed. A sink ing point is an ephemeral feature. As water flows into the sink ing point any limbs, leaves or other trash floating in the river get sucked into the sink and plug it. Conversely, floods scour the bottom and sides of the channel and open new connections to the underlying conduit system. Depending on flow conditions the terminal sin k moves up and down the River. The Underground R ivers of Mystery Cave map summarizes what is currently known from tracing experiments about the various flow connections between the sink points, through the accessible stream passages in the cave, to the spr ings. Sink points near the top of the meander cut off feed both the R ise of the Root River springs and the two large springs that form the head of Forestville Creek several miles to the north east. Note that the river passages in the cave run entirely in j oint segments but the overall trend of the underground stream is from the sink points to the springs and the overall trend does not follow the preferred joint direction. 27

PAGE 28

Figure 18. 1984 Minnesota Speleological Survey Map of Mystery Cave 28

PAGE 29

Figure 19. Simplified cross section of Mystery Cave to Seven Springs Fig. 3, Milske et al., 1983. 29

PAGE 30

Figure 20. The underground rivers and dye connections of the Mystery Cave The meander cut off of the South Branch of the Root River. 30

PAGE 31

Figure 21. The resurgence of th e South Branch of the Root River Located at Seven, Crayfish and Saxifrage Springs. 31

PAGE 32

Stop 9 Cherry Grove Blind Valley Scientific & Natural Area/ Goliath 's Cave. 43 ¡ 34 54.88 N; 92 ¡ 15 51.07 W 1290 ft. elevation http://www.dnr.state.mn.us/snas/detail.htm l?id=sna02022 http://www.cavepreserve.com/goliath.html The Cherry Grove Blind Valley Scientific and Natural Area (CGBV SNA) is a 40 acre parcel owned by the Minnesota Department of Natural Resource s (DNR) and is part of the state Scientific and Natural Areas program. The CGBV SNA is an excellent example of a blind valley in the Galena Group karst. It was purchased in order to preserve and protect the surface and subsurface karst features. Dye tracin g from the blind valley demonstrated that the water sinking there flowed through a large conduit (Goliath's Cave) to the Canfield Big Spring This is the headwater spring of Canfield Creek, a state designated trout stream. The former owners of the property had submitted an application to Fillmore County to convert the property into a limestone quarry. The county preferred that the property be purchased by the state to provide for its permanent protection. The property was purchased by the DNR in 1999. The Minnesota caving community had been aware of what was then called Jesse's Grove since 1950s. The area was named for Jesse McCracken an early settler in the area who had lived her entire life adjacent to the property. The cavers knew that the woods contain ed numerous sinkholes, several sinking streams and at least three caves. One small cave was dry. One cave was a stream sink that led to tight, partially water filled passages. The third cave had an entrance crawl that was often sumped or clogged with debr is washed from the large sinkhole/ephemeral stream sink entrance. Cavers found that beyond the dangerous entrance crawl was a maze of walking passages in the Dubuque Formation. At the east side of the maze section of the cave a small and wet passage led to walking passage in the top of the Stewartville Formation. That passage extended for half a mile to the east before dropping down into a major flooded conduit. The cavers named this entire cave system Goliath's Cave after a large column found at the end of the entrance crawl. The SNA program gated the small dry cave and the entrance to Goliath's Cave; access to the cave is only allowed for research, survey and management purposes. However, most of the cave is east of the CGBV SNA property. John Ackerman pur chased a few acres of land immediately east of the CGBV SNA and the underground rights to the entire 80 acres east of the SNA and made the property part of his Minnesota Cave Preserve. He drilled a 30 inch diameter entrance shaft (named David's Entrance) i nto the main stream passage and opened his section of the cave to exploration and scientific study. The "interactions" between the SNA program and John Ackerman were not friendly and have been chronicled in the book Opening Goliath by Cary J. Griffith (20 09). Ongoing hydrogeological studies have focused on the Cave Preserve part of Goliath's Cave due in large part to the safe, rapid access to the bulk of the cave via David's Entrance. The work reported by Alexander and others (2015) at this meeting was onl y possible 32

PAGE 33

because of the availability of David's Entrance. Goliath's Cave appears to contain both old hypogenetic passages connected and enlarged by new epigenetic speleogenesis. This field trip stop will focus on the surface features of the CGBV SNA. Dep ending on the weather conditions, we will visit one or more sinking streams, several sinkholes, the natural entrance (Goliath's Sinkhole) and one of several agricultural tile drain outlets that empty into the blind valley. This stop will involve about half a mile of walking through the woods of the blind valley. Figures 22 and 23 show the relationships between Goliath's Cave and the surface karst features in the CGBV SNA. Note that there are no surface karst features over the Minnesota Cave Preserve section s of Goliath's Cave. If a few people are interested, a visit to the Cave Preserve section of Goliath's can be arranged during or immediately after the Sinkhole Conference. Figure 22. Cherry Grove Blind Valley Scientific and Natural Area and the adjacen t Minnesota Cave Preserve 33

PAGE 34

Figure 23. Flow system F rom Cherry Grove Blind Valley Scientific and Natural Area, the Minnesota Cave Preserve and Goliath's Cave to Canfield Big Spring at the south edge of Forestville/Mystery Cave State Park. 34

PAGE 35

Stop 10 City of Wykoff Storm Water System Wykoff, junction of State 80 wit h County 5, Fillmore County, MN 43¡ 42' 28.09" N; 92¡ 16' 05.72" W 1322 ft. elevation The City of Wykoff (pop. approx. 800) is located in western Fillmore County. It is in an area underlain by St ewartville Formation (fine grained dolomitic limestone and dolostone) and Dubuque Formation (fine grained limestone with thin shale beds) bedrock of the middle Ordovician Galena Group. Unconsolidated material depth over the bedrock is less than 15 meters. Wykoff lies on a sinkhole plain; surface runoff in this landscape typically flows into sinkholes. Historically, the city used sinkholes as dumping points for its storm water. Four sinkholes had been modified to receive storm water one of which was covered up and was only found during a road and storm sewer upgrade Both street surface runoff and storm sewer flow were routed into these sinkholes. The buried sinkhole was in the center of the city; the other three were at its west, east, and south boundaries. During 1992 to 1994, a County Geologic Atlas was produced by the Minnesota Department of Natural Resources (DNR) Division of Waters and the Minnesota Geological Survey (MGS). The MGS mapped & characterized the geology and the DNR mapped and characterized the hydrogeology. As part of that project, the DNR and the University of Minnesota Geology & Geophysics Department staff performed multiple dye traces to produce a springshed map. One of the dye traces was from a sinkhole developed above the Dubuque Forma tion subcrop on the west edge of the city. This sinkhole had a storm water grate over it and had been modified to serve as a storm water receptor. The dye traveled rapidly west (over two miles in 24 hours) underground to a spring on Mahoods Creek, emanatin g from Galena Group strata that approximate the contact between the Cummingsville and Prosser Formations. Mahoods Creek is an important coldwater tributary to Spring Valley Creek, a state designated trout stream. In the late 1990's, the City began to move forward with the idea of upgrading their water, sewer, and storm water systems. In order to demonstrate the importance of modifying the storm water system and make a case for state and federal grants, the Fillmore County local water planning coordinator f unded two more dye traces from the storm water sinkholes on the east and south borders of the city. DNR Waters and the University of Minnesota Geology & Geophysics Department staff performed these traces which also went to Mahoods Creek. This work undersc ored the importance of modifying the cities' storm water system. The city was denied a federal grant for a total upgrade of all of its street and storm sewer system s City funds were available to reconstruct the storm sewer system and upgrade the streets i n the west part of the town. This reconstruction was engineered so that only local surface flow was routed into the sinkhole on the south edge of town. The buried sinkhole also had storm water diverted from it. This left the storm water sinkholes on the ea st and west sides of town. To remediate the east side sinkhole, the Fillmore County Soil and Water Conservation District received funds from the Minnesota Board of Soil and Water Resources. The sinkhole was 35

PAGE 36

excavated and then sealed with concrete and rock. The storm water now flows across the sinkhole down a grass waterway. For the west side sinkhole, a variety of options were explored. When it appeared that there was no way to divert flow from it, the plan was to install a peat filter to treat the storm w ater That idea was discarded a fter reviewing the literature about peat filters and discussing the long term maintenance required with city staff. The city's engineering firm, WHKS, was able to design the storm water system to reroute the storm water away from this sinkhole into a swale that flows out to the west of the city. Since the sinkhole still received overland flow, some treatment was necessary. The sinkhole was excavated down to the opening in the bedrock. A perforated pipe wrapped with filter clot h was positioned over the bedrock opening with large diameter rock placed around it. The excavation was then backfilled with smaller diameter rock and pea gravel. The pipe was in place to allow water that flowed to the sinkhole to be routed into the bedroc k opening after it passed through the rock filter and filter cloth. The project cost approximately $20,000. Half of that cost was born by the city, the other half came from a Department of Natural Resources Conservation Partners Grant. The DNR was interest ed in this site due to the impact on the groundwater system and on water quality in Mahoods Creek. Figure 2 4 City of Wykoff and mapped springsheds 36

PAGE 37

References Alexander, E. C., Jr., Runkel, A.C., Tipping R .G., and Jeffrey A. Green J.A., 2013 Deep time origins of sinkhole collapse failures in sewage lagoons in SE Minnesota in Land, L., Doctor, D.H., and Stephenson, J. B., eds., NCKRI symposium 2 p roceedings of the 13th multidisciplinary conference on sinkholes and the e ngineering and e nvironmental i m pacts of k arst : Carlsbad, New Mexico, published on line by NCKRI, Carlsbad, NM, p. 285 292. Alexander, E.C ., Jr., Alexander, S.C., Barr, K .D.L. Luhmann, A .J., and Anger, C .T., 2015 Goliath's c ave, Minnesota e pigenic m odification and e xtension of p reexis ting h ypogenic c onduits : 14 Sinkhole Conference Proceedings, 8 p. Griffith, C J. 2009 Opening Goliath d anger and d iscovery in c aving : Minnesota Historical Society Press, St. Paul, MN, 294 p. ISBN 978 0 87351 649 5 Larsen, M ., 2015 Karst topography and animal feedlots in Olmsted County : 14th Sinkhole Conference Proceedings, 10 p. Milske, J A., Alexander, E. C., Jr. and Lively R.S., 1983 Clastic s ediments in Mystery Cave, s outheastern Minnesota : NSS Bulletin vol. 45, p 55 75. Runkel, A C., Steenberg, J R. Tipping R G. and Retzler A J. 2013 Geologic controls on groundwater and surface water flow in southeastern Minnesota and its impact on nitrate concentrations in streams : Report prepared for the Minnesota Pollution Control Agency by the Minnesota G eological Survey, St. Paul, MN, 154 p. 37



PAGE 1

Living with Karst in Rochester, MN The Sinkhole Conference Oct 5 9, 2015 Rochester, MN by Jeffrey S. Broberg, LPG, REM: WSB Associates, Inc.

PAGE 2

Introductions I am Jeffrey S. Broberg, Minnesota Licensed Professional Geologist #30019 I graduated from the University of Minnesota in 1977 and spent the next 10 years in Lafayette Louisiana, riding the boom and bust of the second OPEC Oil crisis. I moved back to Minnesota and have practiced as a geologist and environmental/geotechnical consultant in Rochester since 1990 with McGhie & Betts Environmental Services, now WSB Associates, Inc.. After 25 years in SE Minnesota I have worked professionally on nearly every type of karst hazard the regional has to offer and I was happy to have the opportunity to share may experiences on a short filed trip around Rochester. Having been acquainted with Dr E. Calvin Alexander for almost 40 years we have shared many sinkhole and karst occurrences ranging from the collapse of the Bellchester Sewage lagoon, to the current crisis of nitrate contamination of our groundwater. I owe Calvin a debt of gratitude and would like to dedicate this trip to him and his revealing work, his willingness to share, teach and advance Karst Science There are many others who deserve acknowledgemaent and credit for advancing the understanding of karst I n the Rochester, MN area. First Olmsted County, City of Rochester and the surrounding Townships have always embraced a science based approach to land use and land development, where karst meets local citizens. Rochester Public Utilities, and the water mangers there have been leaders in developing the science and funding the studies of Hydrostratigraphy and water quality in a region where everyone relies on healthy drinking water. Others worthy of mention in local government service are Barb Huberty, formerly of the County and City of Rochester who has worked to make sure the public and local decision makers understood our hydrology when facing development decision ranging from the closure of old dumps, citing of the of the art landfill or stormwater management standards. Also Terry Lee, Olmsted County Environmental Coordinator raised community awareness about the Decorah Edge and the risks and benefits of groundwater discahge on local hillsides. Our community has greatly benefited from the scientific revelations made by the Minnesota Geologic Survey and United States Geologic Survey. Both agencies have allowed devoted scientists to develop a large body of work that makes Rochester a safer and more prosperous place. Jeff Delin, Richard Lindgren and Perry Jones of the USGE deserve particular mention. At the MGS Tony Runkel, bob Tipping and Julie Steenberg also deserve credit. I have borrowed extensively from their work in preparing his guidebook. The Minnesota Department of Natural Resources and the university of Minnesota are important contributors to our knowledge of the local karst. Jeff Green, DNR Karst hydrologist and Scott Alexander, UofM researcher have both made great contributions. I want to thank all of my co workers and friends at McGhie & Betts and WSB Associates, Inc. Bill Tointon, Dave Morrill, Luke Lunde and many others have helped shape Rochester using the best available information and resources. I also want to thank WSB Associates, Inc. For their growing commitment to Rochester and their contribution to the production and printing of this guidebook. Finally I want to thank the Minnesota Groundwater Association (MGWA), a group of dedicated professionals who are constantly groundwater resources Introduction and Acknowledgements 2

PAGE 3

Introduction: Local Field Trip 3 Rochester, MN sits in the Zumbro River Valley at the confluence of nine sub watersheds that drain the glaciated plain to the west and drains the bedrock controlled karst landscape to the east. Carved into the Paleozoic Plateau the City of Rochester is dissected by the eastern edge of the Wisconsin age glaciation where the DesMoines Lobe glacial ice ran past Rochester to the west and left the area to the east largely untouched. Rochester defines the eastern side of the the Driftless Area, the lands to the east of Rochester are principally a bedrock controlled terrain covered by windblown loess soils ,while the the areas to the west are a glaciated terrain dominated by glacial till and glacial outwash features. Like lateral moraines and eskers. The bedrock is a flat lying sequence of Cambro Ordovicain sedimentary rocks. The landscape from the beginning of the Devonian to the Cretaceous was a continental upland subjected to constant erosion and corrosion of the Paleozoic bedrock leaving an ancient karst landscape millions of years before the glacial advance To the south and east of Rochester the first encountered bedrock under the thin till and loess are the carbonates and shale of the Galena Group The Platteville Limestone, Glenwood Shale, St. Peter Sandstone lie conformably under the Galena, but are separated by a 10my unconformity above the Prairie du Chein Zumbro Valley. These rock units were all deposited in the shallow epiric seas near the middle of the Hollandale Embayment. The Jordan sandstone, the youngest of the local Cambrian blanket sands ,is present in the deep subsurface and serves as the local aquifer, but, is only exposed far to the north east and north end of Olmsted County, beyond the area covered in this tour. During the glacial epoch the eroded surface of the carbonate karst was draped with wind blown silty loess pushed out of the valleys in dunes across the highlands. The combination of soils with a high water capacity overlying and the Fissile carbonates resulted in a fractured and dissolved subterranean aquifer system dominated by rapid conduit flow with sinkholes and fractures near the surface, funneling as much as 50% of the local precipitation into the sensitive aquifers ,making the karst aquifers highly susceptible to pollution. Underlying the Galena dolomitic limestone the 90 foot thick Decorah Shale is largely impervious to water, creating an aquitard that prevented the downward percolation of groundwater. In the Driftless Area side hill springs and seeps ring the valleys where the fractured karst aquifers meet the impervious Decorah Shale aquitard. First Nation Paleo Indian hunters found chert for stone tools, fresh water springs and shelter in the bedrock valleys and found abundant game in the wetlands, prairies and oak savannah hillsides. The combination of glaciated prairie and wetland terrain on the west, and a bedrock controlled landscape on the east, with sinkholes and springs was a benefit for Early settlers who first occupied sand caves and sod houses near perennial springs. Their first homes and settlements used resilient flaggy fieldstone and rough hewn timbers and the rivers and sinkholes were convenient for waste disposal. As Rochester developed the karst features clearly provided benefits and risks. Now geologists, engineers and environmental managers are developing land use policy and defining development and building codes based on our growing knowledge of karst. Developers, designers, engineers and geoscientists have learned the importance of hydrostratigraphy and how interbedded aquifers and confining layers define the opportunities and risk for a rapidly growing, high tech community. Our 100 mile trip around Rochester will look at the historic benefits and the risks and the view the current state of the art of living on the karst. Jeffrey S. Broberg, Minnesota Licensed Professional Geologist WSB Associates, Inc. Rochester, MN Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 4

Rochester Karst Fieldtrip Index Page 2 15: Introduction: Geography, climate and geology Page 15 20: Mile 0 6 Groundwater/Surface Water relationships. The St. Peter Recharge Plain Decorah Edge Galena Karst Plateau of Eastern Olmsted Page 21 23: Mile 6 10 Page 24 26: Mile 10 26 The Eyota Spring fields and Orion Sinkhole Plain Page 27 28: Mile 26 39 The glacial edge of the Upper Root River Valley. Glacial seeps and calcareous fens. Page 29 35: Mile 39 53 SW Rochester Karst Plateau. Mayowood sinkhole plan and Galena Quarry Page 36 37: Mile 53 67 Salem glacial valley aggregates and Kalmar buried valley. Sand pits and Kalmar Landfill Page 38 42: Mile 67 79 Rochester sandstone collapse sinkholes and Prairie du Chein karst: The St. Peter/PdC unconformity Page 43 51: Mile 79 92 Prairie du Chein karst Page 52 53: Mile 92 104 historic riverside dumps. Index 4 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 5

Introduction: Google Earth Field Trip Route 5 Sinkhole conference Route Map Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 6

Introduction: Zumbro Watershed and surrounding Counties 6 Shaded relief from LiDAR DEM by Ben Ogren, WSB Associates, Inc. Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 7

The Zumbro River Partnership, a non profit, non government Watershed partnership, is prioritizing critical restoration sites to raise awareness and funding to sustain an interest in Zumbro River water quality. The Zumbro River in Rochester is impaired for suspended sediments, nutrient enrichment, bacteria and mercury affecting aquatic life, aquatic recreation and aquatic consumption. The land use in this karst landscape is largely responsible for the impairments. Introduction: The Zumbro Watershed 7 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 8

Introduction: Local climate and surficial hydrology 8 Edge of field studies recently completed in a Fillmore County partnership with the Minnesota Department of Agriculture measur ing field to stream hydrology on cropland found cip itation infiltrated into the soil and bedrock, evaporated or evapotranspirated. Over the 5 46% of runoff occurs when the ground is frozen, typically in February and March. Annual nitrate leaching losses varies from 6# 50 #/ acre/year http ://www.mda.state.mn.us/protecting/cleanwaterfund/onfarmprojects/rootriverpartnership.aspx Living with Karst in Rochester: The Sinkhole Conference 2015 Located at 44 North Latitude Rochester ahs a temperate climate. The HUC 8 Zumbro Watershed is 990,000 acres. The HUC 8 Root River Watershed is 1,064,961 acres and has 15 major subwatersheds The land was originally prairie, oak savannah and floodplain forests. Starting was converted to agricultural production and is now dominated by row crop production

PAGE 9

Introduction: Region relies on groundwater 9 Living with Karst in Rochester: The Sinkhole Conference 2015 Hydrochemical Survey of the Groundwater Flow in the Rochester Metropolitan Area, Minnesota, R. Tipping MGS OFR14 05, June 2014

PAGE 10

East side Escarpment Edge of Upper Carbonate Plateau and Decorah Edge: Marion Township Introduction: Layer cake hydrogeology 10 Living with Karst in Rochester: The Sinkhole Conference 2015 R unkel, Steenberg, Tipping, Retzler MGS OFR 14 02

PAGE 11

Introduction: Hydrostratigraphy of interlayered confining beds MGS OFR 14 02 11 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 12

Introduction: Unconformities in SE Minn 12 Unconformities and paleokarst horizons are interlayered with confining beds defining local and regional hydrostratigraphy Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 13

SE Minnesota Karst features: Alexander and Gao 2002 Introduction: Karst has been well defined in SE MN since before 2002 13 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 14

Introduction: 2015 Watershed Management Efforts rely on karst information 14 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 15

Rochester Wellhead Protection Plan 15 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 16

Decorah Edge Seeps, Land use and groundwater hazards. Introduction: Decorah Edge Studies 2001 16 Living with Karst in Rochester: The Sinkhole Conference 2015 Groundwater Recharge and Flow Paths Near the Edge of the Decorah Shale Platteville Glenwood Confining Unit, Rochester, MN. USGS Water Resources Report 00 4215, Lindgren, 2001

PAGE 17

federal agencies who have developed a sound understanding of local hydrostratigraphy and groundwater condition. Ongoing efforts of RPU, MGS, DNR and USGS are detailing this karst sequence layered with confining Rochester Groundwater susceptibility 17 Living with Karst in Rochester: The Sinkhole Conference 2015 Rochester Public Utilities: Rochester Wellhead Protection Plan 2007

PAGE 18

The Galena Plateau and Decorah Edge rims the City of Rochester The Decora Edge: Rochester East side 18 Living with Karst in Rochester: The Sinkhole Conference 2015 Isolated islands of Galena/Decorah have low water storage capacity with few side hill wetlands and Decorah Edge risks, however the edge of the Galena/Decorah to the east, south and west of Rochester store and discharge high volumes of groundwater resulting in persistent springs, wetlands and headwater streams Olmsted County Geologic Atlas. Bedrock

PAGE 19

Rochester/Olmsted Decorah Edge Ordinance 19 Both Rochester and Olmsted County adopted Decorah Edge protection ordinances in 2006. Decorah Edge features with active hydrologic elements, including wetlands, springs, seeps, fens and associated soil types cannot be disturbed by development except for Collector Streets and essential public utilities. The combination of the risk of flooded basements fro m groundwater flow, shrink/swell of clays, frost heaving and unstable slopes on the Decorah Edge and the findings of the USGS and Olmsted County on the ability of the Decorah edge features to clean nitrate from groundwater led to wide community support for Ordinances to protect Decorah Edge features. Developers and builders welcomed clearly defined risk features and the new avoidance maps stating that for too long they have been building homes that later developed structural damage, wet basements and mold. Public water supply mangers saw the Decorah Edge regulations as another groundwater protection tool. Today developers and their consulting planners, and engineers refer to these maps and conduct field studies to refine the mapping in high risk areas. The dynamic nature of the seasonal groundwater flows make it important to recognize the geologic and soil features that indicate groundwater movement through the Decorah slopes. Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 20

A high density of homes served with individual septic systems (ISTS) 0n the Decorah Edge needed to address multiple ISTS failures and imminent health threats from septate discharge. A subordinate sewer district, funded with grants and assessments to property owners allowed this community to run Chester Heights sewer district 20 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 21

Flood Risk 1987 Flood Introduction: Zumbro River Flood Control Projects Save Rochester in 2010 21 Flood control project proves successful under heavy rain Posted: Jun 17, 2014 6:42 PM CST Jun 17, 2014 6:42 PM CST By Devin Bartolotta, Anchor/Reporter ROCHESTER, Minn. (KTTC) -The Rochester flood control system is holding up after being put to the test by heavy rains Monday night. Officials said this system has paid for itself time and time again, and that things could have been much worse Monday night without a sophisticated flood control system in place. "It goes down to the Silver Lake Dam and then under the 37th Street bridge and out of Rochester," said Assistant City Administrator Gary Neumann. "Right now we've got quite a bit of capacity left in the system." The Zumbro River charging through town after heavy rainfall is a sign that the Rochester flood control project is working. It's a project that took nearly 40 years from start to finish, and it's the reason inches of rainfall no longer leave Rochester totally underwater. "It had a larger design than is typical because of the '78 flood. Most projects are just designed to handle 100 year flood," said Neumann. Some might remember the Rochester flood of 1978. Rains caused the deaths of five people, tens of millions of dollars in damage, and waist deep water in some parts of the city. It was the push needed to get federal funding for the project, and days like Monday show it's worth every penny of the nearly $100 million cost. Right now, water levels in reservoirs are holding back the water, just as they are designed to. "We have seven reservoirs that hold back a lot of the water and then we have this channel project in the city. All of our reservoirs have more space to hold back a lot of water," said Neumann. The highest flood level on the Zumbro River since the flood control project was back in 2010, and Neumann is hoping rainfall this week won't bring Rochester back to that level. "If we don't get major rain, more than several inches, we should be alright," he said. Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 22

Introduction: Rochester Water Primer 2014 22 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 23

Introduction; Flash Flooding Is a Frequent Occurrence in Olmsted County 23 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 24

Stop 1: Springs defined early settlement patterns Perennial cold water springs like Bear Spring were major resources for Native American and early settlers who relied on these resources of clean, cold water. Some of the earliest know Paleo Indian artifacts indicate that the nomadic tribes seasonally settled around the springs. The earliest European settlers sought land with springs and homesteaded these sites when land grant first became available in the Minnesota Territories in 1854 55. This site, the Bear Spring, is named for the first settlers who constructed a stone spring house to capture the spring for potable water and later for refrigeration to support dairy operations. Stone ruins around springs are common around local headwater springs. Stop 1: Bear Springs 24 Living with Karst in Rochester: The Sinkhole Conference 2015 Jeff Broberg

PAGE 25

Stop 2 Orion Sinkhole Plain: The Orion Township sinkhole plain is underlain by the Stewartville Limestone karst. Very thin loess soils, averaging less than five feet, cover a mature karst plain with sinkholes that appear to be pre glacial or even Cretaceous age. There are no streams or stormwater draingeways in this landscape; all excess precipitation infiltrates into the subsurface and rapidly mixes with the Upper Carbonate groundwater which is universally contaminated above the drinking water standard with more than 10 ppm nitrates. Landowners and farm managers largely ignore Best Management Practices for 50 foot cropland sinkhole setbacks from row crops and farmers have been slow to adopt nutrient management practices that would minimize the risk of groundwater contamination by applying nutrients with split applications when plant uptake is most effective. The contaminated groundwater discharges to the headwater springs where nitrate levels commonly exceed 10ppm and have been noted as high as 40ppm. Feedlots for cattle are a mixed blessing. Pastures with perennial vegetation reduces the need for fertilizers on row crops, but, manure from confined feedlots can contaminate the groundwater with both nitrates and bacteria. Orion sinkhole plain Stewartville Limestone bedrock 25 Olmsted Geologic Atlas, MGS 1988

PAGE 26

Stop 2 Orion sinkholes: Google Earth Image 26 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 27

SW Olmsted County: Edge of till and loess drape Glacial till edge and loess draped landscape 27 Living with Karst in Rochester: The Sinkhole Conference 2015 Olmsted County Geological Atlas, Surficial Geology

PAGE 28

Nitrates in Groundwater In 2014 the Minnesota Geological Survey Published Open File Reports 14 02 and 14 02 the authors synthesized LiDAR topography with the stratigraphy, hydrology, dye trace studies, water quality from wells, and base flow in streams and landuse data. The studies revealed the nitrate contamination risk of intensive row crop production over the karst of SE Minnesota. In general where row crops cover more than 60% of the landscape stream base flow nitrates exceed the drinking water standard of 10 ppm. The synthesis of topography, stratigraphic and hydrologic data, and the role of confining layers interbedded with fractured carbonates were shown to define the risk of aquifer contamination from a leaky system that relies on chemical crop nutrients and manure. The local farm operators seem to believe that it is acceptable to contaminate the water to grow crops and programs to change the trend are opposed or ignored by farmers, crop consultants and farm commodity groups. Stream Base flow Nitrate versus row crop production 28 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 29

SW Rochester bedrock controlled plateaus and valleys Sinkhole risks: Stewartville/Prosser Plateau Sinkholes Solution enlarged karst conduits allow wet loess soils to collapse into the bedrock cavities. These types of sinkholes are common in suburban areas at the top of the ridges to the east, south and west of Rochester where the vertical karst conduits impressive soil collapse features that appears to be a deep vertical shaft into the subsurface. Basal St. Peter/Upper Shakopee catastrophic collapse sinkholes : The basal St. Peter, at the unconformity with the underlying Shakopee formation is prone to catastrophic sandstone sinkhole formation, especially in drainageway settings. Geotechnical borings often find low blow counts and sand cavities just above limestone residuum at the unconformity. The sand cavities can fail leaving impressive sinkholes at the surface. St. Mary's Mary Breigh Bldg. and the Prairie Crossing sinkholes represent these sinkhole risks.. Bedrock Plateaus in SW Rochester capped with Galena Limestone 29 Olmsted Geologic Atlas, MGS 1988 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 30

Mayo Estate 30 In 1990 the Mayo heirs Ned Mayo, Maria Donovan and Dr. Charles Mayo Jr. decided that the time was right to develop the 200 acre ridge top overlooking Rochester that once was part of Dr. Charles and Dorothy Mayo estate, a sprawling 1666 acre farm complex surrounding the Mayowood Mansion on the southwest side of Rochester. A large parcel of river bluffs and rolling wooded hillsides once was home to the Mayo menagerie of deer, elk and buffalo, and award winning dairy cows, beef cattle and Friesian horses. A working stone quarry on the east and cropland and forests on the west were pockmarked with sinkholes. The Mayowood Mansion, stables, dairy and other homes were at the toe of the bluff where seeps and springs provide cold fresh water year round. The City of Rochester Land Use Plan had designated the entire Mayo property for urban development served by municipal water and sewer, but, style suburban homes served with on site septic treatment and private wells. A land use battle over the City utility extension plans focused on the risks inherent with building on karst. What was the appropriate density, disturbance and development? Sewer or septic systems? The City favored development with small lots served by sewer and water and opposed suburban development pitting the Mayo family and the Township against the City of Rochester. The suburban vision was approved by the County and the limits of City growth were set once the County and the City were convinced it was safe to develop homes on the karst. Geologic field investigations identified 39 sinkholes on the ridgetop and five thick loess dunes were identified as suitable areas for Community Waste Water Treatment drainfields. A Special District Plat was developed that identifies the sinkholes Rochester Township adopted a sinkhole Ordinance that defined sinkhole risks, setbacks and specified the need for engineering and geologic studies for each building site and septic system. Clusters of homes on the bluff top ridges were served by Community Waste Water Systems for the clustered homes and pressurized mound systems for the estate lots and a plan was developed to develop 125 lots and to preserve over 200 acres of wooded slopes with the Decorah Edge, seeps and springs. Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 31

1991 Mayo Woodlands Sinkhole Inventory 31 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 32

Mayowoodlands karst: Suburban development 1990 2015 1991: note the sinkholes and liniments north of Meadow Crossing Road 2015: The sinkhole plain has largely been developed with 125 high value homes Mayo Woodlands: 120 suburban lots on a sinkhole plain 32 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 33

Mayowoods Sinkhole investigations and mitigations. The original sinkhole investigation was conducted in the Fall and Winter of 2001. Before LiDAR McGhie & Betts Geologists found that winter weather with a thin snow cover and low barometric pressure greatly assisted the sinkhole inventory. The closed landscape depressions in the Galena Group Cummingsville Fm. often opened in the bottom to a solution enlarged cavity that would blow warm moist air from the earth during falling pressure. The typical enlarged cavity. Some were large enough to present a safety hazard and were immediately sealed with grouted rip rap. The Rochester Township Sinkhole Ordinance prescribed that a geologist or civil engineer evaluate the sinkhole risk and either require a structural and septic setback or design mitigation measures that are protective of the development and groundwater resources. All home foundations are reinforced concrete capable of spanning a 2 foot void. All the wells were drilled through the Cummingsville, Decorah, Platteville, Glenwood and St. Peter and were developed in the Prairie du Chein. Community Septic systems serve as many as 9 homes and individual septic systems are typically pressurized mounds. Mayo Woodlands Home building and sinkhole mitigation 33 Living with Karst in Rochester: The Sinkhole Conference 2015 Jeff Broberg, WSB Associates, Inc J eff Broberg

PAGE 34

Stop 3: Mayowood Quarry and was grandfathered into the County and Township Zoning Ordinance. Primarily producing crushed rock products for construction and road building the operator would blade off 8 to 20 feet of loess and the insoluble residue in the epikarst residuum before blasting an 85 foot working face the pot run rock was crushed and stockpiled on the quarry floor. In 1990 the quarry was confined by the platting of Mayowoods South subdivision and the operator blasted deeper until the encountered the fat clay of the Decorah Formation rending the crushed rock unusable due to the high clay content. The Quarry operations ceased in 2005 and the property developer who had purchased the Mayo properties retained planners to devise restoration plans that ranged from plans to develop multi family housing, a park or a solar garden; none of the plans were adopted. In 2014 the Township required a closure and restoration and a plan was developed to place as much as hopes of one day developing a higher use. Stop 3: Mayowood Quarry: Prosser/Cummingsville/Decorah 34 Living with Karst in Rochester: The Sinkhole Conference 2015 Jeff Broberg, WSB Associates, Inc

PAGE 35

Mayo Cummingsville Quarry Cummingsville Top of Decorah Mayowoods 35 McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates

PAGE 36

Olmsted County Aggregate Resources Almost all of the commercial, high quality sand and gravel aggregates in Olmsted county are found in the buried river and stream valleys draining the glaciated terrain and flow from the west where the glacail till and alluvium represented a high qualty source of sand, gravel and cobbles. The streams draining the Driftless Areas to the west have only Paleozoic carbonate, shale and sandstone, not suitable source materials except for very fine grained sand. Upland sand and gravels are found in NW Olmsted County and are derived from eskers or stream terraces in the glacial outwash channels. Salem/Rochester Township Aggregates 36 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 37

In the last leg of the tour we will be passing by most of the existing and historic solid waste facilities that have served Rochester and Olmsted County. Currently the Olmsted County Kalmar Landfill, located over a deep buried bedrock valley filled with dense and impervious glacial till ,is a state of the art landfill with a double composite liner and leachate collection system. Detailed geologic investigations following the publication of the Olmsted not prone to the typical karst risks. Prior to Olmsted County committing to the Solid Waste to Energy (WTE) Incinerator and Kalmar Landfill the areas solid waste was buried in burn dumps and landfills along the Zumbro river and in the Oronoco Sinkhole Plain. Olmsted County Kalmar Landfill 37 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 38

Mary Breigh Foundation In 2012 the Mayo Clinic decided to fill in the courtyard between the new Emergency Room and the historic Alfred building Covenant Located in what was a pre development drainageway the first encountered bedrock was the basal St. Peter Sandstone near the top of the Shakopee formation, a known unconformity with an Ordovician paleo karst. Soil borings conducted for the foundation design found sandstone voids at a depth of 25 30 feet below the planned lowest floor elevation. Additional geotechnical investigations with borings and cores identified a clay filled sinkhole extended deep into the sandstone/limestone unconformity. Numerous borings taking blow counts found consolidated sand with 50+ blows.1.5 feet above loose sand that would have between 0 and 6 blows/1.5 feet. Engineers reinforced concrete pillars extended to the top of competent bedrock. Along the east wall, pictured to the left, caissons placed on 20 foot centers encounter the Pleistocene, clay filled, sinkhole and the sand cavities. On one line of caissons the depth to the hard bedrock varied from 22 feet to 108 feet deep. Excavations for the foundation to the lowest floor level of the abutting Alfred had collapsed under the structural floor slab that was supported by caissons. We will view the cores and the drilling logs at our stop at Mayowoodlands. Breigh Building paleokarst challenges foundation engineers 38 Living with Karst in Rochester: The Sinkhole Conference 2015 McGhie & Betts (WSB Associates,, Inc.)

PAGE 39

Soil borings through sand have low blow counts just above the limestone residuum at the unconformity 39 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 40

Mary Breigh Soil Boring and Caisson 40 Living with Karst in Rochester: The Sinkhole Conference 2015 Vertical fractures intersecting the side of the caisson boring extend to large sand voids at the water table just above the limestone. Boring logs and photos by McGhie & Betts (WSB Associates, Inc.) McGhie & Betts/WSB Associates

PAGE 41

Sand cavities at the top of the Shakopee. 41 Living with Karst in Rochester: The Sinkhole Conference 2015 McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates

PAGE 42

Mary Breigh sandstone voids. Photos by McGhie & Betts (WSB Associates, Inc. 42 Photo 3 taken 10 17 2012 This photo was taken from within the void under the tunnel on the south end of the Alfred Building facing east. In this photo you can see the poly vapor barrier above, the cemented sandstone on the east wall of the void and the partially exposed face of a concrete pier. Multiple cassions exposed voids at the St. Peter/Shakopee unconformity Living with Karst in Rochester: The Sinkhole Conference 2015 McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates

PAGE 43

Upland esker aggregates Terraces along glacial outwash valleys and eskers are concentrated in the drainageways that flow off the glacial edge from the west. Aggregate resources do not exist in the valleys that drain off the bedrock terrain. The eskers include aggregates from the Superior Lobe and agates as well as traces of gold have been found in the Oronoco area giving rise to the Annual Oronoco Gold Rush celebration NW Olmsted Upland Aggregates 43 Living with Karst in Rochester: The Sinkhole Conference 2015 Shaded relief LiDAR DEM by Ben Ogren, WSB Associates, Inc.

PAGE 44

Prairie Crossing sinkhole collapse In 2006 a mixed use development on the east side of US 52, north of 65 th Street was approved. A grading plan called for a future frontage road to parallel the highway with commercial lots on both sides of the road and called for the construction of a MNDOT stormwater pond fed by the storm sewers in the street. The grading filled in across a drianageway, down gradient of culverts under US52, that drained almost 1000 acres at the edge of the glacial advance. The alluvial sands of the drianageway overlay the basal St. Peter Sandstone, Geotechnical soil borings to the standard depth of 25 feet did not reveal any karst risks. Following a heavy rain in the fall of 2006 nine St. Peter sinkholes opened and drained the ponded water into the subsurface. The sinkholes were filled with over 2500 cy of fill. The development never came to fruition due to the recession. Three sinkholes collapsed again during a snow melt in the winter of 2008. In 2012 as part of the new 65 th St NW highway interchange the sinkhole mitigation design by McGhie & Betts (now WSB Associates) uncovered 11 soil breccia pipes. Concrete taken from the 2 nd St Reconstruction was used to fill a mushroom cap over the breccia pipes and the concrete slabs were grouted, then covered with sand followed by a foot of impervious clay. The drainage was rerouted so tat no ponding would occur in the old drainageways Prairie Crossing Plat 44 Living with Karst in Rochester: The Sinkhole Conference 2015 Blocked drianageway

PAGE 45

Prairie Crossing Sinkholes. 45 2006 Sinkhole 5. Original Collapse 2008 Sinkhole 5. Second collapse Living with Karst in Rochester: The Sinkhole Conference 2015 McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates

PAGE 46

Prairie Crossing Sinkholes 46 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 47

Sinkhole excavation and mitigation 47 Living with Karst in Rochester: The Sinkhole Conference 2015 McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates McGhie & Betts/WSB Associates

PAGE 48

Oronoco Landfill 1989 90 Oronoco Landfill Dye Trace, Alexander, Alexander, Huberty and Quinlin 48 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 49

1989 90 Oronoco Landfill Dye Trace, Alexander, Alexander, Huberty and Quinlin 49 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 50

Bedrock draped with loess east of Lake Zumbro 50 Living with Karst in Rochester: The Sinkhole Conference 2015 Shaded relief LiDAR DEM by Ben Ogren, WSB Associates, Inc.

PAGE 51

MGS OFR 14 02 Zumbro River Hydrostratigraphy 51 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 52

Basal St. Peter sand collapse Excavation in north Rochester often encounters the base of the St. Peter, just above the unconformity. In 2002 the Chateau Theater sought a grading permit to carve out a flat site. At the time City Ordinance required maximum 4:1 slopes for excavations. Geotechnical evidence was presented to prove that the round sand grains of the St. Peter Sandstone were not stable at 4:1 slopes showing numerous local examples of landslides where soils and vegetation on these shallow slopes has suffered translational failures. New rules were adopted to allow near vertical sandstone slopes with terraces. During the excavation a large breccia pipe was revealed in the middle of the slope. The breccia was not cemented and immediately began sloughing. The remainder of the slope has been stable for St. Peter breccia pipe chateau theater 52 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 53

Waste Disposal History History of Olmsted County Solid Waste Disposal: having a single place to dispose solid waste. In an era when most people burned their rubbish in burn barrels and businesses dumped their refuse in the Zumbro River Rochester Opened the Rochester Burn Dump on North Broadway at what is now Elton Hills Driver. The Burn Dump filled a bedrock ravine in the Zumbro River. A full time acrid smoke upset neighbors in Indian Heights and the City closed the dump in 1958 In 1958, the year IBM came to Rochester, until 1972 the City Dump filled the floodplain and banks of the Zumbro River, north of the then new Waste Water Treatment Plan on West River Road and 41 st NW The dump operators dug sand pits in the river alluvium below the water level and used the sand for cover filling the resulting ponds with wastes from IBM and burned the municipal waste. Recycling was a forgotten practice. Due to the flood risk, the growth of Rochester to the NW, the limited space and the regulatory responsibility of the County to be the waste authority the dump was closed in 1972. In 1972 the City opened the Oronoco Landfill in a karst drainageway in the Oronoco Sinkhole Plain where numerous Prairie du Chein sinkholes had already been filled with refuse from the surrounding Township. Employing the practices of the day the soils were stripped from the site for use as landfill cover and the waste was packed and covered. Landfill leachate rapidly entered the groundwater that was the drinking water supply for the surrounding homes and the town of Oronoco. One of the first major dye trace studies in Minnesota proved the leachate risk. The landfill was transferred to Olmsted County and was closed in 1990 when the Kalmar Landfill and Waste to Energy incinerator were complete. Currently the WTE, located near the beginning of our trip, burns the waste and the ash is disposed in the Kalmar landfill which also accepts overflow, industrial wastes and demolition debris Former Rochester City Dump 1920 58 53 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 54

Road Log: Zumbro River to Bear Spring Start: 0.0 mile Rochester Civic Center South Branch Zumbro River. (East on Center St. To CR9) Mile 1.2 Olmsted County Campus including Waste to Energy steam and electric generating facility (East on CR9) Mile 2.1St.Peter Sandstone, Glenwood Shale, Platteville Limestone, Decorah Shale CR22 road cut Mile 3.3 East Side Wildlife Management Area, Rochester groundwater mound (CR9) Mile 4.4 Decorah Edge Seeps. USGS/Rochester Public Utility (RPU) monitor well sites (South on CR 11) Mile 5.0 Groundwater flooded basements (East on US14) Mile 6.7 Chester Subordinate Sanitary Sewer District (US 14 and CR119) Mile 8.4 Olmsted County Chester Woods Park and Rochester Flood Control Reservoir Mile 10.4 Milestone US14 Galena Quarry Mile 11.1 Bear Creek headwater former trout stream drained and converted to croplands (South on 110 th Ave) Mile 11.8 Wetland seeps and fens Stop 1: Mile 12.4 Bear Spring Historic Spring fed milk house and trout pond (West on 30 th St SE to CR23) Road Log:: Start to mile 12.4 54 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 55

Road Log: stop 1,Bear Spring to Cummingsville North Branch Root River Mile 14.3 Galena Cummingsville Formation Springs elev. 1230 40 (south on CR23 to CR19 South) Mile 15.0 Feedlot Springs elev. 1220 1230 (CR19 and I90) Mile 16.0 Zumbro River/Root River Watershed divide elev. 1287 (CR 19 to US52 west) Mile 17.0 Orion sinkhole plain surrounded by cropland (82 nd Ave SE Mile 17.5 Orion Sealed sinkhole (South on 82 nd ,Ave SE) Stop 2 Mile 17.9 32 nd Ave Orion Twn. Roadway dissecting sinkhole Mile 18.1 Orion Sinkhole Plain surrounded by cattle pasture and cropland (East on 70 th St SE) Mile 19.2 Shoenfelder Cattle Feedlot (South on CR129) Mile 21.1 Hog confinements Mile 21.8 Road cut Stewartville Fm. to St. Peter Fm (East on North Branch Rd SE) Mile 23.2 North Branch Root River Cummingsville (south on CR7SE then West on MN 30) Road Log: Mile 14.3 23.2 55 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 56

Road Log: Cummingsville to Mayowood Mile 25.5 Dairy Confinement and manure pit (continue West on Mn30) Mile 26.2 Loess Plain over Galena Karst/edge of Wisconsin till Mile 35.3 Maple Brook Subdivision springs. Golf course converted to urban subdivision (North on US63) Mile 36.2 Stewartville North Branch Root River (US 63 North) Mile 39.3 Airport Calcareous Fen. Mile 41.3 Cummingsville Fm. Road cut. Distinctive saw toothed exposure Mile 42.3 Decorah road cut and slope failures mitigated with rip rap buttress Mile 43.5 Willow Creek aggregates. (SU63 and 40 th St SW) Mile 43.9 St. Peter Castellated Dome Mile 44.4 Willow Creek Groundwater Mound Mile 45.9 US52 Decorah Shale Road cut (North on US52) Mile 46.9 South Branch Zumbro River Aggregates (West on 16 th St. SW then North on CR8) Mile 48.6 Platteville fountains and Decorah seeps and springs (CR 8 and Maryhills Dr .) Mile 50.0 Mayowoodlands Sinkhole Management District (West on Meadow Crossing Road) Stop 3. Mile 53.6 Maywood Galena Quarry Prosser/Cummingsville/Decorah (North on CR8 then West on CR117) Road Log Mile 25.5 53.6 56 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 57

Road Log: Mayowood to Prairie Crossing Sinkhole Collapse Mile 55.7 Heritage Hills Seeps and Springs Mile 56.7 Flooded basement from Decorah Seeps. Mold and property damage (North on 104) Mile 57.3 Salem buried valley aggregates Mile 57.8 South Fork Zumbro River (West on CR 25 then North on 70 th Ave SW Mile 59.4 Glacial Till edge (West on CR34 ) Mile 62.6 Buffalo Ranch (North on CR3, cross US14 and North to 19 th St NW) Mile 66.1 Olmsted Kalmar Mile 67.0 Kalmar Flood Control Reservoirs (Continue East to CR104 go south) Mile 69.3 Kalmar Decorah Edge City of Rochester Urban Service Limits (cross US14 then go East on CR34 (2 nd Ave SW)) Mile 71.1 St. Peter road cut Mile 71.5 Cascade Creek Flood Plain. Golf course conversion to urban subdivisions Mile 71.7 Cascade Creek Aggregate Pits. Cascade Lake Park (2 nd Ave SW and Avalone Cove Dr) sinkholes (North on 11 th St NW) Mile 73.8 Cascade Creek (West on Civic Center Dr) Mile 74.4 US 52 with Assissi Heights on hill to the right (North on US52) Mile 76.3 IBM Rochester (US52 at 41 st St NW) Mile 77.3 Car lot St. Peter sinkholes along Kings Run Mile 79.2 Prairie Crossing Blocked Drainageway Catastrophic sinkhole collapse (US 52 between 65 th and 75 th St exists) Road Log Mile 55.7 79.2 57 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 58

Road Log: Mile Prairie Crossing to US 63 Castellated Domes Mile 80 Glacail Outwash valleys and terraces (East on 75 th then North on 18 th Ave NW) Mile 88.2 Oronoco Prairie du Chein sinkholes. (18 th St NW and 85 th St NW (CR154)) Closed Oronoco Landfill 1972 1990 beyond trees to the west Mile 83.3 Oronoco Prairie SNA native short grass prairie (West on 100 th St NW (CR112)) Mile 84.5 Esker pits and mines (North on 2 nd Ave SE) Mile 85.86 Oronoco Wells and Septic. New subdivisions have community waste treatment Mile 86.6 Original small lots have no septic reserve. New Oronoco water system protects drinking water Mile 86.9 Oronoco Shady Lake Dam: Destroyed by historic 2010 Flood. 30 foot flood (East on 5 th St SE) Mile 87.8 CAPX2020 (East on CR 12) Mile 89 Rucker Hog Lot: Sinkholes identified in 2010 EAW stopped feedlot expansion Stop 4 Mile 91.1 White Bridge Road Cut: Prairie du Chein Karst Mile 92.3 Loess Dunes (CR 12 and US63 South) Mile 95.1 US 63 St. Peter/Glenwood/Platteville Castellated Domes Road Log Mile 88.2 95.1 58 Living with Karst in Rochester: The Sinkhole Conference 2015

PAGE 59

Road Log: US63 Castellated Domes to Civic Center Mile 96.9 St. Peter Road cut Mile 100.4 Hadley Creek Valley Mile 101 Historic Rochester City Dump in floodplain of Zumbro River 1 mile west 1958 1972 Stop 5 Mile 101.4 St. Peter breccia pipe. Paragon Chateau 14 Theater slope excavation (Left into Shopko Lot) Mile 101.8 Sealed St. Peter caves (6 th Ave NE to Wellner Dr, South on Broadway (US63) Mile 102.9 Rochester Rec Center/Senior Center. Historic Rochester City Dump (Broadway and Elton Hills Drive NE) Mile 102.6 Historic Rochester City Dump 1920 1958 Mile 103.3 Silver Lake Dam Mile 104.3 Broadway and Center Street. Mayo Clinic 2 blocks west Mile 104.4 Return to Civic Center Road Log Mile 96.9 104.4 59 Living with Karst in Rochester: The Sinkhole Conference 2015



PAGE 1

307 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 FINDINGS SPRINGS IN THE FILE CABINET Mason Johnson Minnesota Pollution Control Agency, 18 Woodlake Drive Southeast, Rochester, Minnesota, 55904 USA, Mjohn206@uwsp.edu Ashley Ignatius Minnesota Pollution Control Agency, 18 Woodlake Drive Southeast, Rochester, Minnesota, 55904 USA, Ashley.Ignatius@state.mn.us Stream assessment is conducted to determine if the stream supports its designated use, such as supporting aquatic life. Assessment is part of the Minnesota Water Quality Frameworks Watershed Restoration and Protection Strategies (WRAPS) process the comprehensive watersheds (United States Geological Survey Watershed Boundary Dataset subbasins (HUC08)) on a ten-year cycle. This process includes identifying water quality impairments, sources of pollution, areas in need of protection and restoration, and strategies to achieve and maintain water quality standards and goals. Stream assessment requires the review of existing information as well as the collection of new data. Of particular importance is the identification of karst features, because karst features can play a controlling role with respect to a streams physical, chemical and biological properties. Existing information may come from previous work by the MPCA or, likely, from other agencies. These may be other state agencies, or may be federal, local or non-profit agencies. Existing information may also come from landowners or local land users (e.g., hunters, fishers, birders, fungi and plant collectors). The opportunity for collaboration among these agencies proactive communication and collaboration. Simply stated, one agency may not know what data and research not proactively shared or accessible and searchable, it is not practically usable. Such was the case with the Minnesota Department of Natural Resources (DNR) stream documents of southeast Minnesota. The MPCA and DNR have a strong history of collaboration, but the challenge was that the paper Abstract The Minnesota Pollution Control Agency (MPCA), in partnership with other agencies, is currently undertaking comprehensive sub-basin assessments statewide over a ten-year period. Southeast Minnesota has over 17,500 kilometers of perennial and intermittent streams, making the the task is further complicated by karst geology. In the and Minnesota Department of Natural Resources (DNR) to digitally preserve paper documents which capture qualitative and quantitative data about the hydrology, water chemistry, geomorphology, biology, land use and karst features of southeast Minnesota streams. The paper documents in file as such, they were in a data silo. The task was to preserve the documents so as to make the data usable by converting the documents into a digital format (Adobe PDF, GeoTIFF, ESRI Feature Class). To date, more than 4,000 documents (of an estimated more than 12,000) have been converted, made text-searchable, prepared for storage in a document management system, and made more accessible through a geographic information system (GIS). This previously inaccessible data is an important piece in understanding the karst region of southeast Minnesota. Within the documents scanned thus far, over 400 springs and other karst features have been identified, which are not currently recorded in Minnesotas Karst Feature GIS Database. Introduction The karst region of the Driftless Area in southeast Minnesota proves to be challenging for the Minnesota Pollution Control Agency (MPCA), which is tasked with assessing the streams of that area. Karst features, such as sinkholes and springs, can change the properties of a stream. For example, the presence of sinkholes can modify the contributing drainage area to a stream and a groundwater-fed spring can change temperature, chemistry and flow of a stream.

PAGE 2

308 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE the digital documents to be accessed through ESRI ArcGIS. Once scanned, Adobe Acrobat Professional software was used to make the digital documents text searchable and to create a searchable index. The folders by document name (e.g., stream assessments, survey reports, survey summaries, etc.). To incorporate the digital documents into a GIS format, many steps were taken to ensure all documents were linked with their related stream feature class. After the scanned documents were indexed with Adobe Acrobat Pro, a log file was created of all the file paths and document names. The log file was imported into a Microsoft Excel Workbook where file paths were Worksheet (e.g., assessments, survey reports, summaries, etc.). Worksheets were then converted to an ESRI geodatabase table. In ESRI ArcMap, each theme table was related by kittle number (unique stream numbering system developed by the MN DNR) to a stream feature. As a result, users are able to use the Identify Tool to view PDF documents associated with a stream reach, simply by clicking on the feature in the ArcMap software. In addition to the stream documents (which were generally typed or hand-written text documents including a few illustrations or maps), there were 147 large (greater than 21.59 cm x 27.95 cm) maps. These maps were generally USGS 7.5 minute series topographic map quadrangles, with hand-written annotation. On these maps were identified springs, seeps, electrofishing stations and other notes, such as water temperature and qualitative land cover descriptions, such as good pheasant habitat. These maps were sent to a private contractor who was returned to the MPCA as digital images (in Tagged Image File Format [TIFF]). Within the ArcMap environment, the digital images were georectified, or geographically referenced. Once georectified, the hand annotation could springs in the digital image were made into a GIS feature class. Discussion Thus far, over 4,000 paper documents have been made digital, text searchable and indexed. In practical terms, a MPCA staff member could search for a certain word within a digital document or within the whole collection of digital documents, enabling a stream assessment team documents in file cabinets were not in an accessible or there was an immediate business need for the DNRs stream documents to be in a digital format. Therefore, digitally preserve the stream documents which capture qualitative and quantitative data about the hydrology, water chemistry, geomorphology, biology, land use and karst features of southeast Minnesota streams. could be an important resource for stream assessment (e.g., identifying interrelated factors impacting a streams biologic community), especially because these documents contained information about the location and characteristics of karst features. become: Protected from physical damage or loss. Easily accessible, searchable, shareable. A collective resource enhancing one another, thus producing a more spatially and chronologically complete story of a particular stream. Process The task was to preserve the data contained in the documents so as to make the data usable. This was accomplished by converting the documents into a digital format (Adobe PDF, GeoTIFF, ESRI Feature Class). Proceeding by watershed, MPCA staff removed folders from the file cabinets one at a time. Each document within a folder was carefully prepared by removing staples and repairing any damage prior to scanning each to a computer in Adobe PDF format (these are herein referred to as the digital documents). In addition to the scanned copy saved on a computer, a backup copy was saved to an external hard drive. Documents were then reassembled and returned to their respective folders in the file cabinets. Finally, folders were marked to indicate that the scanning of all of the documents in each folder had been completed. A key part of this process was the document-naming scheme. Each digital document was named in a way to effectively tag it with keywords such that it could be seamlessly loaded into the MPCAs OnBase document management system for storage and retrieval. The digital document-naming scheme also allowed for

PAGE 3

309 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Geological Survey and University of Minnesota. The comparison resulted in the determination that this process had identified 400 springs and other karst features not already in the Karst Feature Database. These 400 new features were shared with DNR, Minnesota Geological Survey and University of Minnesota staff in hopes of field verification and eventual inclusion within the database. Minnesota Pollution Control Agency staff access to robust existing information for their stream assessment the complex karst region of southeast Minnesota. This was accomplished by liberating data about southeast Minnesotas streams and karst features from the data silo that is a file cabinet of paper documents. The resulting digital documents, digital images, and GIS feature classes has made the data more manageable, accessible, and shareable. to easily search for karst features mentioned within a streams digital documents. Furthermore, in OnBase the digital documents serve as a back up to the paper document in the DNRs file cabinet. Within OnBase, documents can be found using keywords such as the DNR Stream Kittle Number, year (approximate date of creation of the original paper document) or document type (stream survey, stocking record, stream assessment, etc.). The storage of these digital documents in OnBase should be a more efficient way (rather than researching paper documents while physically sitting in the DNR office) to manage and share this data within the agency as well as outside the agency. Furthermore, the text searchable digital documents allow for copying and pasting text from the digital document to another type of document, such as a word processing document. In ArcMap, the relationship between the stream feature class and the tables (tables that include the file paths to the digital documents), allow MPCA staff to use ESRI ArcMaps Identify tool to select a stream of interest and be provided with a list of digital documents associated with that stream. Using ESRI ArcMaps Hyperlink functionality, MPCA staff can click on a document listed within the Identify Tool window and open the digital document in Adobe Acrobat Reader effectively digitally opening up the DNR folder, formerly only possible by visiting the physical office. The combination of accessing the digital documents from ArcMap and the text searchability of the digital documents allow MPCA staff to not only find springs mentioned within as a spring feature (that is geographically referenced) in ArcMap, based on the information in the text. For example, a digital documents text that reads spring, 0.5 cfs, 200 meters upstream from Bridge 2 may be however, when that spring location is captured as a geographically-referenced feature in a GIS feature class, it can be examined in context with other features on a sub-basin-wide scale. As these springs and other karst appended to the respective feature class. to create a GIS feature class of springs from the handannotated locations on the maps. Next, the feature class was compared to the existing spring features within the Karst Feature Database created by the Minnesota

PAGE 4

310 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE



PAGE 1

NCKRI SYMPOSIUM 5 Proceedings of the 14th Multidisciplary Conference on Sinkholes and the Engineering and Environmenal Impacts of Karst National Cave and Karst Research Institute 400-1 Cascades Avenue Carlsbad, New Mexico 88220 USA

PAGE 2

NCKRI SYMPOSIUM 5 Proceedings of the 14th Multidisciplary Conference on Sinkholes and the Engineering and Environmenal Impacts of Karst 2015

PAGE 3

NCKRI SYMPOSIUM 5 Proceedings of the 14th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst Edited by: Daniel H. Doctor, Lewis Land, and J. Brad Stephenson www.nckri.org

PAGE 4

NATIONAL CAVE AND KARST RESEARCH INSTITUTE SYMPOSIUM 5 SINKHOLES AND THE ENGINEERING AND ENVIRONMENTAL IMPACTS OF KARST PROCEEDINGS OF THE FOURTEENTH MULTIDISCIPLINARY CONFERENCE October 5 through 9, 2015 Rochester, Minnesota EDITORS: Daniel H. Doctor United States Geological Survey Reston, Virginia, USA Lewis Land National Cave and Karst Research Institute Carlsbad, New Mexico, USA J. Brad Stephenson CB&I Federal Services Knoxville, Tennessee, USA Co-organized by:

PAGE 5

Published and distributed by National Cave and Karst Research Institute Dr. George Veni, Executive Director 400-1 Cascades Ave. Carlsbad, NM 88220 USA www.nckri.org Peer-review administered by the Editors and Associate Editors of the Proceedings of the Fourteenth Multidisci plinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst. The citation information: Doctor D.H., Land L., Stephenson J.B. editors. 2015. Sinkholes and the Engineering and Environmental Impacts of Karst: Proceedings of the Fourteenth Multidisciplinary Conference, October 5-9, Rochester, Minnesota: NCKRI Symposium 5. Carlsbad, New Mexico: National Cave and Karst Research Institute. ISBN 978-0-9910009-5-1 ASSOCIATE EDITORS: Gregory A. Brick Minnesota Department of Natural Resources St. Paul, Minnesota, USA James E. Kaufmann U.S. Geological Survey Rolla, Missouri, USA Mustafa Saribudak Environmental Geophysics Associates Austin, Texas, USA Samuel V. Panno Illinois State Geological Survey Champaign, Illinois, USA Jason S. Polk Western Kentucky University Bowling Green, Kentucky, USA David J. Weary U.S. Geological Survey Reston, Virginia, USA Ming Ye Florida State University Tallahassee, Florida, USA Lynn B. Yuhr Technos Inc. Miami, Florida, USA Cover Photo: lapse. The museum Skydome was built over a large cave passage which suffered a catastrophic roof collapsed on the morning of February 12, 2015, causing more than $3 million in damage (see paper by Polk et al., p. 477-482). Photos provided courtesy of Jason Polk and Western Kentucky University.

PAGE 6

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 III CONTENTS Organizing Committee ..................................................................................... X-XI Foreword ............................................................................................................... XII Invited Speaker Hales Bar and the Pitfalls of Constructing Dams on Karst (abstract) J. David Rogers ................................................................................................... XIV Upper Mississippi Valley Karst Aquifers Karst Hydrogeologic Investigation of Trout Brook Joel T. Groten and E. Calvin Alexander, Jr. ..................................................... 1-8 Human Impacts on Water Quality in Coldwater Spring, Minneapolis, Minnesota Sophie M. Kasahara, Scott C. Alexander and E. Calvin Alexander, Jr. ..... 9-18 Hydrologic and Geochemical Dynamics of Vadose Zone Recharge in a Mantled Karst Aquifer: Results of Monitoring Drip Waters in Mystery Cave, Minnesota Daniel H. Doctor, E. Calvin Alexander, Jr., Roy Jameson and Scott C. Alexander ......................................................................................... 19-30 Conduit Flow in the Cambrian Lone Rock Formation, Southeast Minnesota, USA John D. Barry, Jeffrey A. Green and Julia R. Steenberg ............................ 31-42 Using Nitrate, Chloride, Sodium, and Sulfate to Calculate Groundwater Age Kimm Crawford and Terry Lee ...................................................................... 43-52 Karst Spring Cutoffs, Cave Tiers, and Sinking Stream Basins Correlated to Fluvial Base Level Decline in South-Central Indiana Garre A. Conner ............................................................................................. 53-62 Driftless Area Karst of Northwestern Illinois and its Effects on Groundwater Quality Samuel V. Panno and Walton R. Kelly .......................................................... 63-74 Seeps and Springs at a Platteville Observatory on the River Bluffs (abstract) BJ Bonin, Greg Brick and Julia Steenberg ................................................... 75-76

PAGE 7

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE IV Karst Hydrology Hydrogeological Dynamic Variability in the Lomme Karst System (Belgium) as Evidenced by Tracer Tests Results (KARAG project) Amal Poulain, Gatan Rochez and Vincent Hallet ................................. 77-84 Recharge Area of Selected Large Springs in the Ozarks James W. Duley, Cecil Boswell and Jerry Prewett ...................................... 85-92 Hydrological and Hydrogeological Characteristics of the Platform Karst (Zemo Imereti Plateau, Georgia) Zaza Lezhava, Nana Bolashvili, Kukuri Tsikarishvili, Lasha Asanidze and Nino Chikhradze .................................................................................................... 93-100 Tracer Studies Conducted Nearly Two Decades Apart Elucidate Groundwater Movement Through a Karst Aquifer in the Frederick Valley of Maryland Keith A White, Thomas J. Aley, Michael K. Cobb, Ethan O. Weikel and Shiloh L. Beeman ......................................................................................... 101-112 Dye Tracing Through the Vadose Zone Above Wind Cave, Custer County, South Dakota James Nepstad ........................................................................................... 113-126 A Comparative Study Between the Karst of Hoa Quang, Cao Bang Province, Vietnam and Tuscumbia, Alabama, USA Gheorghe M. Ponta, Nguyen Xuan Nam, Ferenc L Forray, Florentin Stoiciu, Viorel Badalita, Lenuta J. Enache and Ioan A. Tudor ............................ 127-138 Hydrochemical Characteristics and Formation Mechanism of Groundwater in the Liulin Karst System, Northwestern China Min Yang, Fenge Zhang, Sheng Zhang, Miying Yin and Guoqing Wu ................................................................................................ 139-146 Karst Geology Karst Paleo-Collapses and Their Impacts on Mining and the Environment in Northern China Gongyu Li and Wanfang Zhou ................................................................. 147-156 The Sandstone Karst of Pine County, Minnesota Beverley Lynn Shade, E. Calvin Alexander, Jr. and Scott C. Alexander ..................................................................................... 157-166

PAGE 8

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 V A Proposed Hypogenic Origin of Iron Ore Deposits in Southeast Minnesota Karst E. Calvin Alexander, Jr. and Betty J. Wheeler ......................................... 167-176 Down the Rabbit Hole: Identifying Physical Causes of Sinkhole Formation in the UK Tamsin Green .............................................................................................. 177-188 Edwards and Trinity Aquifers, Hays and Travis Counties, Central Texas Brian B. Hunt, Brian Smith, Alan Andrews, Douglas Wierman, Alex S. Broun, and Marcus O. Gary ........................................................................................... 189-200 Preexisting Hypogenic Conduits E. Calvin Alexander, Jr., Scott C. Alexander, Kelton D.L. Barr, Andrew J. Luhmann, and Cale T. Anger ................................................. 201-210 GIS Databases and Mapping of Karst Regions Creation of a Map of Paleozoic Bedrock Springsheds in Southeast Minnesota Jeffrey A. Green and E. Calvin Alexander, Jr. ......................................... 211-222 Media, Sinkholes and the UK National Karst Database Vanessa J. Banks, Helen J. Reeves, Emma K. Ward, Emma R. Raycraft, Hannah V. Gow, David J. R. Morgan and Donald G. Cameron .......... 223-230 Shallow Depressions in the Florida Coastal Plain: Karst and Pseudokarst Thomas L. Dobecki ..................................................................................... 231-240 Sinkhole Vulnerability Mapping: Results from a Pilot Study in North Central Florida Clint Kromhout and Alan E. Baker ................................................ 241-254 A Semi-Automated Tool for Reducing the Creation of False Closed Depressions from a Filled LiDAR-Derived Digital Elevation Model John Wall, Daniel H. Doctor and Silvia Terziotti ....................................... 255-262 History and Future of the Minnesota Karst Feature Database Robert G. Tipping, Mathew Rantala, E. Calvin Alexander, Jr., Yongli Gao and Jeffrey A. Green .......................................................................................... 263-270 Legacy Data in the Minnesota Spring Inventory Gregory Brick ............................................................................................... 271-276

PAGE 9

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE VI Development of Cavity Probability Map For Abu Dhabi Municipality Using GIS and Decision Tree Modeling Yongli Gao, Raghav Ramanathan, Bulent Hatigoplu, M. Melih Demirkan, Mazen Elias Adib, Juan J. Gutierrez, Hesham El Ganainy and Daniel Barton Jr. .......................................................................................... 277-288 Evaluation of Cavity Distribution Using Point-Pattern Analysis Raghav Ramanathan, Yongli Gao, M. Melih Demirkan, Bulent Hatipoglu, Mazen Elias Adib, Michael Rosenmeier, Juan J. Gutierrez, and Hesham El Ganainy .................................................................................... 289-298 A Method of Mapping Sinkhole Susceptibility Using a Geographic Information System: A Case Study for Interstates in the Karst Counties of Virginia Alexandra L. Todd and Lindsay Ivey-Burden ........................................... 299-306 Finding Springs in the File Cabinet Mason Johnson and Ashley Ignatius ........................................................ 307-310 New Methodologies and Approaches for Mapping Forested Karst Landscapes, Vancouver Island, British Columbia, Canada (abstract) ............................................ 311-312 Contamination of Karst Aquifers Evaluation of Veterinary Pharmaceuticals and Iodine for Use as a Groundwater Tracer in Hydrologic Investigation of Contamination Related to Dairy Cattle Operations Larry Boot Pierce and Honglin Shi .......................................................... 313-318 Virginia Yingling ........................................................................................... 319-326 Tracking of Karst Contamination Using Alternative Monitoring Strategies: Hidden River Cave Kentucky Caren Raedts and Christopher Smart ...................................................... 327-336 Spatiotemporal Response of CVOC Contamination and Remedial Actions in Eogenetic Karst Aquifers Ingrid Y. Padilla, Vilda L. Rivera and Celys Irizarry ................................... 337-346 Determination of the Relationship of Nitrate to Discharge and Flow Systems in North Florida Springs Sam B. Upchurch ........................................................................................ 347-354

PAGE 10

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 VII Geophysical Exploration of Karst The Million Dollar Question: Which Geophysical Methods Locate Caves Best Over the Edwards Aquifer? A Potpourri of Case Studies from San Antonio and Austin, Texas, USA Mustafa Saribudak ..................................................................................... 355-364 Rollalong Resistivity Surveys Reveal Karstic Paleotopography Developed on Near-Surface Gypsum Bedrock Lewis Land and Lasha Asanidze ............................................................... 365-370 Integration and Delivery of Interferometric Synthetic Aperture Radar (InSAR) Data Into Stormwater Planning Within Karst Terranes Brian Bruckno, Andrea Vaccari, Edward Hoppe, Scott T. Acton and Elizabeth Campbell .................................................................................... 371-380 Detection of Voids in Karst Terrain with Full Waveform Tomography Khiem T. Tran, Michael McVay and Trung Dung Nguyen ...................... 381-386 Characterization of Karst Terrain Using Geophysical Methods Based on Sinkhole Analysis: A Case Study of the Anina Karstic Region (Banat Mountains, Romania) Laurentiu Artugyan, Adrian C. Ardelean and Petru Urdea ................... 387-398 Investigation of a Sinkhole in Ogle County, Northwestern Illinois, Using Near-Surface Geophysical Techniques Philip J. Carpenter and Lauren M. Schroeder ......................................... 399-406 Study on Monitoring and Early Warning of Karst Collapse Based on BOTDR Technique Zhende Guan, X. Z. Jiang, Y. B. Wu, Z. Y. Pang ....................................... 407-414 Pre-Event and Post-Formation Ground Movement Associated with the Bayou Corne Sinkhole Cathleen E. Jones and Ronald G. Blom .................................................. 415-422 The Application of Passive Seismic Techniques to the Detection of Buried Hollows Michael G. Raines, Vanessa J. Banks, Jonathan E. Chambers, Philip E. Collins, Peter F. Jones, Dave J. Morgan, James B. Riding and Katherine Royse .......................................................................................... 423-430 Using Electrical Resistivity Imaging to Characterize Karst Hazards in Southeastern Minnesota Agricultural Settings (abstract) Toby Dogwiler and Blake Lea ................................................................... 431-432

PAGE 11

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE VIII Karst Management, Regulation, and Education The Cost of Karst Subsidence and Sinkhole Collapse in the United States Compared with Other Natural Hazards David J. Weary ............................................................................................ 433-446 Hazard of Sinkhole Flooding to a Cave Hominin Site and its Control Countermeasures in a Tower Karst Area, South China Fang Guo, Guanghui Jiang, Kwong Fai Andrew Lo, Qingjia Tang, Yongli Guo and Shaohua Liu ......................................................................................... 447-454 Case Studies of Animal Feedlots on Karst in Olmsted County, Minnesota Martin Larsen ............................................................................................... 455-464 Evaporite Geo-Hazard in the Sauris Area (Friuli Venezia Giulia Region NE Italy) Chiara Calligaris, Stefano Devoto, Luca Zini and Franco Cucchi ........ 465-470 Building Codes to Minimize Cover-Collapses in Sinkhole-Prone Areas George Veni, Connie Campbell Brashear and Andrew Glasbrenner .................................................................................. 471-476 Cars and Karst: Investigating the National Corvette Museum Sinkhole Jason S. Polk, Leslie A. North, Ric Federico, Brian Ham, Dan Nedvidek, Kegan McClanahan, Pat Kambesis, Michael J. Marasa and Hayward Baker ........................................................................................... 477-482 Sparse Data; A Case in Kermanshah Province, Iran Kamal Taheri, Milad Taheri and Fathollah Mohsenipour ........................ 483-492 Modeling of Karst Systems Numerical Simulation of Karst Soil Cave Evolution Long Jia, Yan Meng, Zhen-de Guan and Li-peng Liu ............................ 493-500 Experimental and Numerical Investigation of Cover-Collapse Sinkhole Development and Collapse in Central Florida Xiaohu Tao, Ming Ye, Dangliang Wang, Roger Pacheco Castro, Xiaoming Wang and Jian Zhao ................................................................................. 501-506 Accounting for Anomalous Hydraulic Responses During Constant-Rate Pumping Tests in the Prairie Du Chien-Jordan Aquifer System Towards a More Accurate Assessment of Leakage Justin L. Blum ................................................................................................ 507-520

PAGE 12

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 IX Numerical Simulation of Spring Hydrograph Recession Curves for Evaluating Behavior of the East Yorkshire Chalk Aquifer Nozad Hasan Azeez, Landis Jared West and Simon H. Bottrell ............. 521-530 Study on the Critical Velocity of Groundwater to Form Subsidence Sinkholes in Karst Area Fuwei Jiang, Mingtang Lei and Dai Jian-ling .......................................... 531-536 Evaluation of First Order Error Induced by Conservative-Tracer Temperature Approximation for Mixing in Karstic Flow Philippe Machetel and David A. Yuen .................................................... 537-548 Engineering and Geotechnical Investigations in Karst Concepts for Geotechnical Investigation in Karst Joseph A. Fischer and Joseph J. Fischer .................................................. 549-558 Sinkhole Physical Models to Simulate and Investigate Sinkhole Collapses Mohamed Alrowaimi, Hae-Bum Yun and Manoj Chopra ..................... 559-568 Monitoring the Threat of Sinkhole Formation Under a Portion of US 18 in Cerro Gordo County, Iowa Using TDR Measurements Kevin M. OConnor and Matthew Trainum ............................................. 569-578 Predicting Compaction Grout Quantities in Sinkhole Remediation Edward D. Zisman ....................................................................................... 579-586 Mobility Grout to Mitigate Future Sinkhole Development in a 2,787.1 Square Meter (30,000 SF) Maintenance Facility ....................................................................................... 587-594 Successful Foundation Preparations in Karst Bedrock of the Masonry Section of Wolf Creek Dam David M. Robison ........................................................................................ 595-604 Hydrocompaction Considerations in Sinkhole Investigations Edward D. Zisman and Stephen West ...................................................... 605-611

PAGE 13

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE X ORGANIZING COMMITTEE Conference Co-Chairs George Veni, Ph.D., P.G., National Cave and Karst Research Institute (NCKRI), Carlsbad, New Mexico Kelton Barr, P.G., Braun Intertec Corpora tion, Minneapolis, Minnesota Jim LaMoreaux, Ph.D., PELA Geoenviron mental, Tuscaloosa, Alabama Program Co-Chairs Lynn B. Yuhr, P.G., Technos, Inc., Miami, Florida sociates, LLC, Tampa, Florida Proceedings Managing, Assistant, and Copy Editors Daniel H. Doctor, Ph.D., U.S. Geological Survey, Eastern Geology & Paleoclimate Science Center, Reston, Virginia Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and Na tional Cave and Karst Research Institute, Carlsbad, New Mexico J. Brad Stephenson, P.G., L.R.S., CB&I Federal Services, Knoxville, Tennessee Rebel Cummings-Sauls, Kansas State Uni versity Julie Fielding, University of Michigan Registration Sean Hunt, Department of Natural Re sources, Minneapolis, Minnesota Audrey Van Cleve, Hydrologist, Minnesota Pollution Control Agency (retired), Minne apolis, Minnesota Treasurer Jeanette Leete, Minnesota Department of Natural Resources, Minneapolis, Minne sota Session Chairs Upper Mississippi Valley Karst Aquifers Gregory Brick, Ph.D., Minnesota Depart ment of Natural Resources, St. Paul, Min nesota Karst Hydrology Samuel V. Panno, Illi nois State Geological Survey, Champaign, Illinois Karst Geology John Barry, Minnesota Dept. of Natural Resources, Rochester, Minnesota GIS Databases and Mapping of Karst Re gions Jason S. Polk, Western Kentucky University, Bowling Green, Kentucky Contamination of Karst AquifersMing Ye, Florida State University, Tallahassee, Florida Geophysical Exploration of Karst Mus tafa Saribudak, Ph.D., P.G., Environmental Geophysics Associates, Austin, Texas Karst Management, Regulation, and Edu cationDavid J. Weary, U.S. Geological Survey, Reston, Virginia Modeling of Karst SystemsMing Ye, Flor ida State University, Tallahassee, Florida Engineering and Geotechnical Aspects of Karst Lynn Yuhr, Technos, Inc., Miami, Florida Field Trips Co-chairs E. Calvin Alexander, Jr. Ph.D., Department of Earth Sciences, University of Minneso ta, Minneapolis, Minnesota Jeff Green, Minnesota Department of Natu ral Resources, Rochester, Minnesota Short Courses Joe Fischer, Ph.D., P.E., Geoscience Ser vices, Clinton, NJ Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and Na

PAGE 14

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 XI tional Cave and Karst Research Institute, Carlsbad, New Mexico E. Calvin Alexander, Jr., Ph.D., University of Minnesota Jeff Green, Minnesota Department of Natu ral Resources Lynn Yuhr, Technos, Inc., Miami Florida Invited Speakers Yongli Gao, Ph.D., University of TexasSan Antonio, San Antonio, TX Beck Scholarship Subcommittee E. Calvin Alexander, Jr. Ph.D., Department of Earth Sciences, University of Minneso ta, Minneapolis, Minnesota Ira D. Sasowsky, Ph.D., P.G., Geosciences, University of Akron, Akron, Ohio Beck Scholarship Silent Auction John Barry, Minnesota Department of Nat ural Resources, Rochester, Minnesota Hotel and Conference Facilities Audrey J. Van Cleve Hydrologist, Min nesota Pollution Control Agency (retired), Minneapolis, Minnesota Kelton Barr, P.G., Braun Intertec Corpora tion, Minneapolis, Minnesota Melinda Erickson, U.S. Geological Survey, Mounds View, Minnesota Circulars Samuel V. Panno, Illinois State Geological Sur vey, Champaign, Illinois Logo Samuel V. Panno, Illinois State Geological Survey, Champaign, Illinois Jeff Green, Minnesota Department of Natu ral Resources, Rochester, Minnesota Program with Abstracts Brian Smith, Ph.D., Barton Springs/Ed wards Aquifer Conservation District, Aus tin, Texas Brian Hunt, Barton Springs/Edwards Aqui fer Conservation District, Austin, Texas Justin Camp, Barton Springs/Edwards Aquifer conservation District, Austin, Tex as Website Gheorghe Ponta, P.G., Geological Survey of Alabama, Tuscaloosa, AL Public Relations Lanya Ross, President, Minnesota Ground Water Association, St. Paul, Minnesota George Veni, Ph.D., P.G., National Cave and Karst Research Institute (NCKRI), Carlsbad, New Mexico Professional Organizations Liaisons Kelton Barr, Braun Intertec Corporation, Minneapolis, Minnesota Wanfang Zhou, ERT, Inc., Knoxville, Ten nessee Members at Large Scott Alexander, University of Minnesota and Minnesota Ground Water Association, Minneapolis, Minnesota Philip Carpenter, Northern Illinois Univer sity, DeKalb, Illinois Ralph Ewers, Ewers Water Consultants, Inc., Richmond, Kentucky Bashir Memon, PELA GeoEnvironmental, Tuscaloosa, Alabama Deana Sneyd, Golder Associates, Inc., At lanta, Georgia

PAGE 15

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE XII FOREWORD Welcome to the Fourteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Im pacts of Karst in lovely Rochester, Minnesota! This venue is the farthest north that the Sinkhole Conferenceas it is better knownhas met since its inception in 1984. The setting will provide conference participants with a unique opportunity to view karst phenomena within the glaciated and driftless terrains of southeastern Minnesota, and a sig within the upper Mississippi Valley region. and hosting the Sinkhole Conference, and the 2013 Sinkhole Conference that was held at NCKRI headquarters in This year, NCKRI is pleased to partner with the Minnesota Ground Water Association (MGWA) for hosting the ing a common forum for scientists, engineers, planners, educators, attorneys, and other persons concerned with through the excellent research and information presented at the conference and published within these Proceedings. I am exceptionally pleased that the papers and abstracts within this volume aptly represent the current state of the science, as well as cover notable recent occurrences of sinkholes and other karst phenomena. At the time of this writ this particular sinkhole opened up in the town of Seffner, in the very same place where Jeffrey Bush tragically lost his life on the 28th of February, 2013 when the ground beneath the room where he lay caved in and swallowed him. within the public eye. Once again, all sinkholes had become newsworthy, no matter how small, or whether they oc curred in actual karst areas. A number of truly spectacular sinkhole collapses followed the 2013 tragedy in Seffner, Florida, including the event at the National Corvette Museum in Bowling Green, Kentucky on the 12th of February, 2014 that destroyed a number of vintage automobiles (see the paper by Polk et al., p. 479-484), as well as the continued saga of salt dome collapse threatening an entire community within Bayou Corne, Louisiana (see the paper by Jones and Blom, p.415422). Sinkholes have also generated news outside of the United States in recent years. For example, periods of unusually heavy rainfall in the winter of 2013-2014 triggered numerous sinkholes in the United Kingdom, sparking a similar that the information contained within these Proceedings will serve as a reference for many future studies. Daniel Doctor U.S. Geological Survey Reston, Virginia

PAGE 16

14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 XIII EDITED BY: Daniel H. Doctor U.S. Geological Survey 12201 Sunrise Valley Dr. MS 926A, Reston, VA 20192 USA Phone: 703-648-6027 E-mail: dhdoctor@usgs.gov Lewis Land New Mexico Bureau of Geology and Mineral Resources and the National Cave and Karst Research Institute 400-1 Cascades Ave. Carlsbad, NM 88220 USA Phone: 575-887-5508 E-mail: lland@nckri.org J. Brad Stephenson CB&I Federal Services 312 Directors Drive Knoxville, TN 37923 Phone: 865-694-7336 E-mail: brad.stephenson@cbifederalservices.com

PAGE 17

NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE XIV which increased each year. The TVA began constructing the most expensive cutoff wall ever built, drilling 750 18-inch diameter holes along the dams centerline and 163 feet, extending 25 to 103 feet below the river bed. In April 1963 the TVA announced it was abandoning Hales Bar Dam, due to increasing leakage. Biography Dr. J. David Rogers holds the Karl F. Hasselmann Chair in Geological Engineering at the Missouri University of Science & Technology in Rolla, Missouri. He is pres ently representing the geological and geotechnical en gineering professions on the National Academies panel that has been charged with examining Levees and the National Flood Insurance Program: Improving policies and practices, being funded by FEMA. Dr. Rogers has served as principal investigator for research funded by the NSF, U.S. Geological Survey, National Geospatial Intelligence Agency, Federal Highway Administration, Department of Defense, and several state departments of transportation. He has served on numerous panels, including the Mississippi Delta Science & Engineering Special Team, the Coastal Louisiana Recovery Panel, the NSF Independent Levee Investigation Team and USGS Investigation Teams evaluating the impacts of Hurricanes Katrina and Rita, the NSF team evaluating Resilient and Sustainable Infrastructure Networks team Dr. Rogers received his B.S. degree in geology from California Polytechnic University at Pomona, his M.S. degree in civil engineering from the University of Cali fornia, Berkeley, and his Ph.D. in geological and geo technical engineering at the University of California, Berkeley. He served on the Berkeley faculty in civil en gineering for seven years prior to accepting his current position in 2001. Hales Bar Dam was built on the Tennessee River 33 miles downstream of Chattanooga by a private com pany to generate power in 1905-1913. The dam site was selected because it was the narrowest reach in the downstream end of the Walden Ridge Gorge. The site is underlain by Mississippian Bangor Limestone on the Three different contracts failed to complete the dam 1913 diamond drill core holes were used to explore the site and a series of reinforced concrete caissons 40x45 ft on upstream side and 30x32 ft on the downstream side were installed. Excessive leakage soon appeared near the eastern abutment, and gradually increased. Sound ings were made in 1914 to ascertain the areas of gross leakage. Shortly thereafter, rags were placed over suc tion holes on the river bed and concrete pumped over these. Once a leak was stemmed, leakage would resume the leaks by inserting hay bales, old mattresses, chicken wire, and even corsets! In 1919 the owners began drill ing grout holes from the inspection gallery within the dam and pumping hot asphalt into the voids. This was phalt grout into the dam foundation, using 6,266 lineal feet of boreholes with average hole depth of 92 ft. By 1922 the problem appeared solved, but leakage gradual ly resumed between 1922-1929, rising to the same level as had been observed previously. In 1930-1931 a new program of exploration was un dertaken, using dyes and oils to identify conduits under the dam. Leakage was found to vary between 100 and 1200 cubic feet per second (cfs). When the dam was ac quired by the Tennessee Valley Authority (TVA) in 1939 age. Dye tests revealed that the leakage varied between They also noted seepage boils forming in the gravel bars, INVITED SPEAKER HALES BAR AND THE PITFALLS OF CONSTRUCTING DAMS ON KARST J. David Rogers Missouri University of Science and Technology Department of Geological Sciences and Engineering 157 McNutt Hall, 1400 Bishop Ave. Rolla, Missouri, 65409

PAGE 19

1 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 systematically from the upstream springs to the downstream springs. The nitrate concentrations have been increasing at four springs from 1985 to 2014 and at two surface sampling points from 2001 to 2014. The nitrate concentration of another surface sampling point increased from 2001 to 2006, decreased from 2006 to 2012, and increased from 2012 to 2014. Snowmelt and rainfall runoff was sampled on 2 March 2012 and showed no detectable nitrate in the runoff from a watershed with no row-crop agriculture, with row-crop agriculture. All of these trends illustrate the dominance of agricultural sources of nitrate in Trout Brook. Introduction Karst Complex surface and groundwater interactions are dominated by karst processes in southeastern Minnesota. Karst features often include caves, sinkholes, springs, stream sieves, and sinking streams. A springshed is the subsurface and surface areas that provide the discharge to a spring. A stream sieve describes a losing reach of a surface stream where specific water sinking points, stream sinks, are not evident. Karst features result from water containing carbonic acid which dissolves the carbonate in soluble bedrock. Water quality is a concern because karst features allow rapid groundwater velocities and short residence times. Karst springs provide the source water to premier trout streams in southeastern Minnesota. Trout Brook in the Miesville Ravine Park Reserve in southeast Dakota County and south of Miesville, Minnesota is one of these trout streams. Location, Geology, and Topography Trout Brook (Figure2) is located in the southeastern part of Douglas Township (T113N, R17W) of Dakota County, south of Miesville, Minnesota in the Miesville Abstract Trout Brook in the Miesville Ravine County Park of Dakota County Minnesota is the trout stream with the highest nitrate concentration in the karst region of southeastern Minnesota. Water quality data from 1985 and 1995 (Spong, 1995) and from 2001, 2002, 2006, 2010, and 2014, collected by the Dakota County Soil and Water Conservation District (Dakota SWCD, 2014) document an increasing level of nitrate in Trout Brook. A karst hydrogeologic investigation was designed to measure nitrate levels at sampling points along the stream and to increase our understanding of the source and movement of nitrates throughout the length of Trout Brook. Eighteen springs and seeps have been located in the Main Branch and tributaries of Trout Brook. A previously unreported flowing section and stream sieve, Weber Sieve, were found above what had been thought to be the head of perennial flow in the East Branch of Trout Brook. Two new sinkholes developed after the 14-15 June 2012 flood in a field northeast of the East Branch of Trout Brook. This investigation included springs, synoptic stream flow measurements, and a dye trace of a sinking stream in the Trout Brook drainage. baseflow of Trout Brook was from discrete springs. However, synoptic baseflow and nitrate measurements show that only 30-40 percent of the total flow in Trout Brook is from discrete springs, and the rest appears to be from distributed groundwater discharge directly into the stream. Both the discrete springs and the distributed recharge occur along reaches of Trout Brook that drain of the regionally important Shakopee aquifer. Dye traces have confirmed flow-paths from Weber Sieve to LeDuc and Bridgestone Springs and have begun to define springsheds for these head water springs. Nitrate concentrations and chloride/bromide ratios decreased KARST HYDROGEOLOGIC INVESTIGATION OF TROUT BROOK Joel T. Groten U.S. Geological Survey (University of Minnesota Graduate School Project), 2280 Woodale Drive, Mounds View, MN, 55112, USA, jgroten@usgs.gov E. Calvin Alexander Jr. University of Minnesota, 450 McNamara Alumni Ctr. 200 Oak St. SE, Minneapolis, MN, 55455, USA, alexa001@umn.edu

PAGE 20

2 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The Shakopee Formation unconformably overlies the Oneota Dolomite. The Shakopee and Oneota Dolomite collectively form the Prairie du Chien aquifer, which is one of the most heavily used aquifers in Dakota County. An unconformity between these two formations represents a 10 million year subaerial erosion episode that left enhanced porosity and permeability in the top of the Oneota. During and after the deposition of the Shakopee, karst solution processes expanded the HTZ up into the lower Shakopee. Runkel et al., (2003) and Tipping et al., (2006) report that this mid-Prairie du Chien HTZ is one of the primary features of the hydrostratigraphy of southeastern Minnesota. The source water for Trout along bedding plane fractures, and through solutionally enlarged porosity and permeability and anastomosing karst conduits. The Trout Brook surface watershed is largely an intensively cultivated gently rolling upland. The main land use type in the Trout Brook watershed is row-crop agriculture, and many farmers irrigate their crops due to the sandy soil. Livestock feedlots in the watershed are also potentially significant sources of nitrate. Very little surface water flows on the upland except during spring snowmelt and after the largest and most intense precipitation events. Most routine precipitation not consumed by evapotranspiration rapidly infiltrates to groundwater. South of Meisville the surface drainage abruptly incises steep-sided valleys to form the West and East Branches of Trout Brook. The East and West the Miesville Ravine Reserve Dakota County Park. The lower reaches of both branches of Trout Brook and the Main Branch widen downstream. All branches of Trout Brook meander across steep-sided, flat-bottomed valleys. Historical Spong studied Trout Brook, Dakota County, in 1985 (Beaver, LeDuc, Fox, and Swede Springs) in 1985 and two springs (LeDuc and Swede Springs) in 1995. The two sampling events in 1985 and 1995 were collected during baseflow periods (with no significant stormflow) in years with normal precipitation (Ron C. Spong, written communication, 2012). Flow rates were measured at Trout Brooks springs and streams in 1985 Ravine County Park. The area is underlain by a thin cover of flood plain alluvium, colluvium and Illinoian glacial outwash, loess and till (Hobbs et al., 1990). The glacial sediments rest unconformably on the lower Ordovician Shakopee Formation of the Prairie du Chien Group (Mossler, 1990). The Shakopee Formation is a mixture of limestone and dolomite with the New Richmond Sandstone near the bottom of the formation. Figure 1. Baseflow nitrate-nitrogen concentration as a function of the percentages of row crop agriculture in the karst region of southeaster Minnesota. Figure 2. Trout Brooks streams, springs, and nitrate levels.

PAGE 21

3 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 versus the nitrate concentrations in streams at baseflow. Trout Brook was selected for study because it has the highest nitrate concentration at baseflow of monitored streams in southeastern Minnesota. Trout Brook is the uppermost data point, below 18 mg/L, on the above graph. Monitoring An initial survey of Trout Brook occurred in February 2011, and a second round of sampling occurred in July 2011. A systematic water sampling campaign began in October 2011 and ended in October 2012 Eighteen discrete springs documented along Trout Brook are shown in Figure 2. All of the springs emerge where the steep valley walls meet the flat valley floor. All but two of the springs, Beaver and Swede, emerge where Trout Brook has meandered up against the base of the valley walls. Beaver and Swede are buffered from Trout Brook by beaver dam induced wetlands. Geochemistry The field and analytical methods used in this work are described in Alexander and Alexander (2011). The nitrate concentrations at Fox, LeDuc, Beaver, and Swede Spring have been increasing with time (Figure 3). The nitrate concentration of Fox Spring is increasing at the greatest rate of 0.42 ppm/year while the concentrations at Beaver, Swede, and LeDuc Springs are increasing at rates of 0.25, 0.18, and 0.11 ppm/year, respectively. These rates were calculated over a 27 year span. The rate of increase for Fox Spring is almost twice as high as the next highest rate, that of Beaver Spring. These increases are likely due to changes in farming practices over time and the intensity of farming on the contributing springsheds. Figure 2 shows the concentrations of these four springs compared to the other springs in Trout Brook as colorcoded dots. The springs discharging into the West Branch of Upper Trout Brook have the highest nitrate levels. The springs discharging into the East Branch of Upper Trout Brook have low to moderate nitrate levels. The three yellow dots on the Main Branch indicate moderate nitrate levels and the springs further downstream have the lowest nitrate levels. The nitrate concentrations in spring resurgence appear to be controlled by location. This relationship is likely (Spong, 1995). These data are important because they document water quality of the springs 27 and 17 years time trends. The Dakota County SWCD measured baseflow and obtained grab samples during storm events at Trout Brook during 2001, 2002, 2006, and 2010. Flow water quality parameters and were reported to the State of Minnesota. Automated stage monitoring was also in place, but the data is suspect due to the flashy nature of this stream (Dakota Co. SWCD, 2010). Row-Crop Agriculture Versus Nitrate Reactive nitrogen is an environmental concern. In the hydrosphere it can lead to eutrophication, toxic algae blooms, and hypoxia. In the atmosphere it contributes to acid rain, deposition of nitrogen in forests leading to nitrogen saturation which can alter the soil, and global warming. Reactive nitrogen can also be harmful to humans due to air pollution and contamination of drinking water. The U.S. Environmental Protection Agency (EPA) reports that greater than 10 parts per million (ppm) of nitrate nitrogen in drinking water can have adverse health effects (U.S. EPA, 1990). (In the rest of this work, the word nitrate is synonymous with nitrate-nitrogen.) The Hastings Area Nitrate Study (HANS, 2003) was carried out in Dakota County, MN near the study area. That study highlights the nitrate contamination problem in groundwater and considered three possible sources: row-crop agriculture, feedlots, and septic systems. Groundwater samples collected in 2000 showed nitrate levels varied among the aquifers. The Shakopee aquifer had the highest concentration of nitrate at 15 ppm, the Quaternary aquifer was next at 8.7 ppm, and the Jordan aquifer had the lowest at 1.85 ppm. Figure 1 from Watkins (2011) shows the percent of rowcrop agriculture plotted against nitrate plus nitrite in streams at baseflow in the karst region of southeastern Minnesota. Groundwater discharge supports the baseflow of streams and rivers in this karst region, with discrete springs typically providing a substantial portion. A linear relationship is present with R value of 0.70. This indicates a strong relationship between the percentages of row-crop agriculture in southeast Minnesota watersheds

PAGE 22

4 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The chloride/bromide ratios of Trout Brooks springs and streams on Figure 5 are averaged values from samples collected from 2011 to 2012. The figure also shows the discrete chloride/bromide ratios of the samples collected in the Main Branch of Trout Brook during the 28 October 2011 synoptic stream flow measurement campaign. The West Branch Springs have the highest ratios indicating the greatest anthropogenic impact. The East Branch Springs have the second highest chloride/bromide ratios. The Main Branch Springs have the lowest ratios. Figure 5 shows that the chloride/bromide ratios are a function of distance downstream. The chloride/bromide ratios at springs seem to be decreasing towards the southeast. This reduction is occurring from the headwater springs of the West Branch towards Swede Spring. The 28 October 2011 chloride/bromide ratios show that there was not a significant decrease from the upstream to the downstream end of the Main Branch. The total of all the averaged ratios of the springs are shown as the purple X. This total averaged value is very similar to the averaged ratios of all samples from TB3 (denoted by the dark green circle), as well as the discrete chloride/bromide ratio of the sample from TB3 on 28 October 2011. This is evidence that the stream channel has similar chloride/bromide ratios as the mixed water in the stream channel. The lower chloride/bromide ratios in the Main Branch springs do not significantly lower the chloride/bromide ratios of the mixed water. determined by the springshed of each spring. The springs further downstream probably involve longer flowpaths draining deeper parts of the aquifers. The contributing springsheds probably vary substantially by different types and percentages of land use. Figure 4 shows the nitrate concentrations at baseflow in the East, West, and Main Branches of Trout Brook between 2001-2012. Nitrate levels in the West Branch [as represented by the results of sample TB2] increased at the greatest rate, 0.35 ppm/year. From 2001-2006 nitrate in the East Branch [represented by the sample nitrate concentrations decreased from 2006-2012. The scope of this study did not include further analysis of this phenomenon. The Main Branch [as represented by sample TB3] was collected the furthest downstream and had a nitrate increase of 0.09 ppm/year. The TB3 sample near the end of the Main Branch is a collection of all the water mixing with surface water and groundwater in the streamshed, derived from baseflow and runoff events. At baseflow that water is almost entirely a variable mixture of discrete spring flow and distributed groundwater discharge, as there is no significant surface runoff to Trout Brook during baseflow. Chloride/bromide ratios are useful in groundwater studies because chloride and bromide are conservative anions that travel with the groundwater and can be used to identify the source water that is recharging the aquifer. Chloride/bromide ratios provide indications of anthropogenic impacts on waters and are explained further by Anger and Alexander (2010). Figure 4. Nitrate baseflow concentration is a function of time at the Main Branch of Trout Brook and its tributairies. Figure 3. Nitrate concentration as a function of time at four springs.

PAGE 23

5 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Synoptic Flow and Nitrate Assessment A 28-29 October, 2011 synoptic study, of the flow and nitrate levels of Trout Brook at baseflow conditions was conducted. The data were used to understand the interaction between flow and nitrate and to determine or from distributed discharge into the stream channels. Flow measurements were taken at 32 locations in the Main Branch, tributaries, and springs of Trout Brook to estimate the water contribution from the springs to the overall flow of this trout stream. Measurements were taken upstream and downstream of where the identified springs discharge into East, West, and Main Branches of Trout Brook. The approximate spring flow was calculated by subtracting the flow of the stream measured upstream of the spring, from the flow measured downstream of the spring. If a spring had formed a channel with enough water, then direct measurements in the spring run were also taken to calculate the flow. All direct flow measurements of the springs were averaged with the flow measurements calculated from the upstream/downstream subtraction. Unfortunately, the flow of some of the springs (with no measureable separate channel) were within the uncertainty of the up and downstream stream flow measurements. In those remaining cases, limits on flow in those spring was visually estimated. Flow was also measured at particular sites in the stream to distribute the measurements, in order to interpret reaches that may be gaining or losing flow. Every location where a flow measurement was taken, a water sample was also obtained. Water samples were also retrieved at spring orifices. This was done to understand the mixing concentrations of nitrate and chloride/bromide ratios. The initial hypothesis that the source of baseflow in the Main Branch of Trout Brook and its tributaries would be primarily from discrete springs is falsified. Data from the two-day synoptic flow measurements show that Approximately 30-40 percent of the total flow at the sampling point TB3, close to the outlet of Trout Brook, is from spring water. The remaining flow is apparently Swede Spring has an average chloride/bromide ratio of 348. This is the lowest ratio detected in the springs and is closest to the ratio of rainwater, which varies from 200-250 in Minnesota. The West Branch Springs have the highest ratios, ranging from 630-680. These ratios are comparable to those Runoff Versus Land Use Surface water runoff was sampled on 2 March 2012 from the recession of 1.67 inches of rain that occurred on 29 February 2012. Two samples were obtained from two different ravines. Each ravine has its own small watershed, which were delineated using the watershed tool in ArcGIS: ArcMap 10.1. Land use polygons were created using aerial photographs. In 2011-2012, the smaller watershed (darker green on Figure 2, next to fluorescent green) did not contain any row-crop agriculture while the larger watershed (fluorescent green on Figure 2) contained 40 percent row-crop agriculture. The nitrate concentration was below the detection limit in the samples collected from the watershed with no row-crop agriculture. The samples from the watershed with row-crop agriculture had a nitrate concentration of 10.6 ppm. The 10.6 ppm nitrate in its runoff event from the watershed with row-crop agriculture is consistent with the results in Figure 1. The rapid groundwater velocities and short residence times in karst does not allow enough time for significant nitrate reduction. The nitrate concentrations at springs may be a key indicator of the percent row crop agriculture of their springsheds due to the rapid, direct water flow in karst. Figure 5. Chloride/bromide ratios are a function of distance from the confluence of East and West Branches of Upper Trout Brook.

PAGE 24

6 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Flow measurements were also taken on 31 March 2012. This was done to see how much flow was gained from the confluence of the East and West Branch to the stream section between Beaver and Hill on the 28-29 October 2011 study. It was found that approximately 4.5 cfs of water was gained along this portion of the Main Branch. The only discrete spring in this area is Beaver Spring, which was discharging approximately 0.7 cfs of water into the Main Branch. Approximately 3.8 cfs of water was from distributed discharge into the stream channel. discharge into the stream channel. However, the flow inputs have changed in Trout Brook. In 1985 the East Branch was contributing more flow to Trout Brook than the West Branch (Spong, 1995) and in 2011 the West Branch was contributing more flow than the East Branch. This illustrates the changes that occur with flow regimes over time. These changes could be from climate, anthropogenic activities such as irrigation, changes in land use, or from changes in the steam channel itself by Weber Run Dye Traces Weber Run is located in the E of the SE of sec 22 of Douglas Township on private land on the northwest corner of the County Park. Weber Run is shown on Figure 2. Surface flow in Weber Run is fed by several source springs and then, under all but high flow conditions, sinks completely in a stream sieve (indicated by the from distributed groundwater discharge into the stream channel because no surface runoff was observed during the synoptic measurements. This result is different from the pattern seen in many southeastern Minnesota trout streams. The water of Trout Brook is from the mid-Prairie du flowing and gaining reach of Trout Brook is entirely in that HTZ stratigraphic interval. Runkel et al. (2003) and Tipping et al. (2006) have shown that water flows in the HTZ through a broad range of solution enlarged conduits, of porosity. The larger conduits and flow features are connected to the surface via smaller, more distributed discharge features. The channels of Trout Brook were eroded deeper into the bedrock during the Pleistocene, under conditions of glacial low base levels. Those deeper channels were back-filled with glacial sediments at the end of the last glacial cycle, forming the relatively flat bottom of the Trout Brook Valley. Much of the distributed discharge is in reaches where Trout Brook flows across this sediment backfill. The spring with the greatest contribution to flow in Trout Brook is Beaver Spring. Beaver Spring discharges into a beaver pond and flows through a swamp before it discharges into the Main Branch at approximately 0.82 cfs. This spring provides only 6.5 percent of the total flow of Trout Brook at the site TB3. The approximated flow from the remaining measured springs can be seen in Table 1 (in Appendix I). Figure 6 displays nitrate concentration and measured flow versus distance from the confluence of the East and West Branches of Trout Brook. Flow is displayed on the left vertical axis. Nitrate concentration is on the right vertical along the streams from the confluence of the East and West Branches to form the Main Branch of Trout Brook. The nitrate concentration decreases only slightly downstream, from ~ 16 ppm below the confluence to ~ 13 ppm at TB3. In contrast, flow in the Main Branch increases by about a factor of 3 over the same reach, from about 4 cfs to over 12 cfs. This relationship can be seen in Figure 6. Distributed groundwater inflow dominates the Main Branch of Trout Brook and the nitrate content of that water is apparently in the 13 to 15 ppm range. Figure 6. Flow and nitrate concentrations are a function of distance from the confluence of the East and West Branches of Trout Brook.

PAGE 25

7 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 for three points in the surface streams. The study period 2011-2012 was abnormally dry but significant floods occurred on 6 May 2012 and 14 -15 June 2012. Our stream and spring flow measurements indicate that 30-40 percent of the water in Trout Brook is from identified discrete springs. As surface runoff contributes to Trout Brook steam flow only during and immediately after significant precipitation events, distributed groundwater discharge into the stream channel, therefore makes up 60-70 percent of the baseflow. Nitrate concentrations ranged from about 9 ppm in springs near the downstream end of Trout Brook to over 25 ppm at the upstream springs in the West Branch of Trout Brook. Chloride/ bromide ratios decreased systematically from the upstream springs to the downstream springs. The nitrate concentrations in four of the springs increased at rates ranging from 0.42 to 0.11 ppm/year from 1985 to 2012. The nitrate concentrations at one upstream site, and the downstream surface water sampling points, have increased at similar rates from 2001 to 2012. The nitrate concentration at another upstream surface water sampling point increased from 2001 to 2006 but decreased from 2006 to 2012. Runoff from a 29 February 2012 winter rain event was sampled on 2 March in two small sub-watersheds along the East Branch of Trout Brook. Runoff from a sub-watershed containing only forest and CRP land contained no detectable nitrate. Runoff from a sub-watershed that was 40 percent row-crop land contained 10.6 ppm nitrate. Dye tracing documented karst aquifer flow-paths from the Weber Run stream sieve to LeDuc and Bridgestone Springs with flow velocities in the range of 15 to 27 meters per day. The results of this study indicate that row-crop agriculture in the surface and subsurface drainage basins of Trout Brook is the primary cause of the waters elevated concentrations of nitrate. This conclusion is supported by the MPCAs correlation between the percentages of row-crop agriculture and the nitrate concentrations in run-off and stream samples. blue-dashed line). The source springs (also shown on Figure 2) are a series of small springs and seeps on the southwest stream bank. On 28 December 2011 Rhodamine WT was introduced in Weber Run. A second dye trace was introduced in Weber Run on 12 April 2012 using eosine dye. Four positive detections of Rhodamine WT and three positive detections of eosine at LeDuc Spring and one detection of Rhodamine WT confirm that at least a portion of the water of LeDuc and Bridgestone Springs comes from the Weber Sieve area of Weber Run. The 29-day travel time between Weber Sieve and LeDuc Spring measured in the first dye trace corresponds to a groundwater flow velocity of ~27 meters/day in the southsoutheast direction. The first dye trace also estimated groundwater flow from Weber Sieve to Bridgestone Springs to be 15-20 meters/day. The 79-day travel time between Weber Sieve and LeDuc Spring corresponds to a ~10 meters/day groundwater velocity during the second dye trace. As the 28 December 2011 and 12 April 2012 traces took place under low flow conditions in an exceptionally dry winter, spring, summer, and fall, the flow velocities under normal, higher flow conditions may be faster. Conclusion Trout Brook is a trout stream in Dakota Countys Miesville Ravine Park Reserve. An MPCA survey found Trout Brook had the highest baseflow nitrate concentrations measured in southeastern the karst hydrogeology and water quality in Trout Brooks water to gain information on the source and movement of nitrates through the landscape. This investigation located springs, stream sinks, sinkholes, and other karst phenomena in the Trout Brook watershed. Synoptic surveys of the stream and spring flows. Periodic water samples were chloride/bromide ratio time trends. Two dye traces were conducted initiating springshed mapping for the springs. This study combined existing, historical data from 1985 and 1995 with our 2011-2012 results to quantify nitrate time trends for four springs. Data from the 2001, 2002, 2006, and 2010 Dakota SWCD surveys, combined with our 2011-2012 results, permitted documentation of shorter-term time trends

PAGE 26

8 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Minnesota Climatology Working Group. 2012. Flooding Rains: June 14, 2012. 02 November 2012. fl ood14_150612.htm Mossler, JH. 1990. Bedrock Geology, Plate 4 in (Hobbs, Howard C. and Nancy Balaban eds.) Geologic Atlas, Dakota County, Minnesota, County Atlas Series Atlas C-6, Minn. Geol. Survey, St. Paul, MN. Runkel, AC, Tipping, RG, Alexander, EC, Jr., Green, JA, Mossler, JH, Alexander, SC. 2003. southeastern Minnesota. Report of InvestigationsMinnesota Geological Survey 61 (105 pp., 2pls.). Spong, RC. 1995. Optical Brighteners: Indicators of Sewage Contamination of Groundwater, LCMR Report. Tipping, RG, Runkel, AC, Alexander, EC Jr., Alexander, SC. 2006. Hydrostratigraphic porosity in part of the Cambrian Sandstone aquifer system of the cratonic interior of North America: Improving predictability of hydrogeologic properties. Sedimentary Geology 184: 305-330. U.S. EPA. 1990. Fact sheet: National pesticide survey nitrate: Washington, D.C., U.S. Government Printing Office, 4 p. U.S. EPA. 2011. Reactive Nitrogen in the United States: An Analysis of Inputs, Flows, Consequences, and Management Options. A Report of the Science Advisory Board. Watkins, J. 2011. The Relationship of Nitrate-Nitrogen Concentrations in Trout Streams to Row Cropland Use in Karstland Watershed of Southeast Minnesota, 2011 GSA Annual Meeting, Poster Session No. 108. References Alexander, SC, Alexander, EC, Jr. 2011. Field and Laboratory Methods, Hydrogeochemistry Laboratory, Department of Geology and Geophysics, University of Minnesota, 19 pp. Alexander, SC. 2005. Spectral Deconvolution and Quantification of Natural Organic Material and Fluorescent Tracer Dyes, in ( Barry F. Beck, ed.) Sinkholes and the Engineering and Environmental Impacts of Karst Proceedings of the Tenth Multidisciplinary Conference, Geotechnical Special Publication 144, ASCE, Reston, VA, p. 441-448. Anger, CT, Alexander, EC, Jr. 2010. Monitoring Hydrogeologic Response to Aggregate Mining Data Assessment and Groundwater Evaluation, Prepared for the University of Minnesota, March 26, 2010. Bierman, P, Rosen C, Venterea, R, Lamb, J. 2011. Survey Minnesota Department of Agriculture, 26 pp. Bomberg, CW, Greenwaldt, BA, Alexander, EC, Jr. 2011. Sources of High Nitrate Levels in Trout Brook (poster), 2011 Univ. of Minn. Earth Sciences Dept., NSF REU Summer Intern Presentations. Dakota County SWCD. 2010. North Cannon Water Quality Monitoring Report. 29 November 2012. DNR/MPCA Cooperative Stream Gauging. 2012. Little Cannon River nr Cannon Falls, CR24 (39016001). 03 November 2012. www.dnr.state. mn.us/waters/csg/site_report.html?mode=get_ site_report&site=3901600 HANS. 2003. Hastings Area Nitrate Study. Dakota County Environmental Management, 70 pp. Hobbs, Howard C., Saul Aronow and Carrie J. Patterson. (1990). Surficial Geology, Plate 3 in (Hobbs, Howard C. and Nancy Balaban eds.) Geologic Atlas, Dakota County, Minnesota, County Atlas Series Atlas C-6, Minn. Geol. Survey, St. Paul, MN. Luhmann, AJ, Covington, MD, Peters, AJ, Alexander, SC, Anger, CT, Green, JA, Runkel, AC, Alexander, EC, Jr. 2011. Classification of thermal patterns at karst springs and cave streams. Ground Water 49 (3): 324-335. Miesville Ravine Park Reserve Dakota County. Walk-Through of the 2005 Miesville Ravine Park Reserve Master Plan.

PAGE 27

9 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 ratios of road salt, 1,000 to 10,000. Road salt is applied throughout the winter. The temperature of the spring is variable and higher than its pre-settlement temperature. Nicollet (1841) recorded the temperature of Coldwater Spring multiple times in summer of 1836 as 46F (7.8 C) and multiple times in winter of 1837 as 45.5F (7.5 C). More recently the temperature of Coldwater Spring fluctuates smoothly between 10.7 and 13.1 C. The higher temperature of the springs discharge also indicates an anthropogenic source of heat within the spring-shed or spring recharge area. The spring water is coldest in May and June and warmest in October and November. The temperature of the water in Wetland A fluctuates from 6.4 to 13.8 C in a pattern that is opposite of that in Coldwater Spring. Coldwater Spring also contained a significant, increasing nitrate-nitrogen component which ranged from 2.5 to 5.2 ppm with dips at the same times as the chloride pulses. Wetland As nitrate-nitrogen level varied between 0.2 to almost 6 ppm with large pulses at the same time as Coldwater Springs dips. A 2014 study performed by the US Geological Survey came to the conclusion that increasing chloride levels in lakes and streams are likely driven by increasing road salt application, rising baseline concentrations, as well as an increase in snowfall in the Midwestern area of the US during the time of the study (Corsi 2014). The significant chloride, temperature and nitrate levels are likely to be driven by anthropogenic sources. HUMAN IMPACTS ON WATER QUALITY IN COLDWATER SPRING, MINNEAPOLIS, MINNESOTA Sophie M. Kasahara Civil, Environmental and Geoengineering Department, University of Minnesota, 122 Civil Engineering Building, 500 Pillsbury Dr. SE., Minneapolis, MN 55455, kasah007@umn.edu Scott C. Alexander Earth Sciences Department, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, alexa017@umn.edu E. Calvin Alexander Jr. Earth Sciences Department, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, alexa001@umn.edu Abstract Coldwater Spring in Minneapolis, Minnesota was the water supply for Fort Snelling from the 1840s to 1920. The spring site has been declared a sacred site by some monitored the water chemistry of Coldwater Spring to document human impacts on the springs water quality. Temperature, dissolved oxygen, conductivity, pH and anions were monitored weekly and cations and alkalinity Wetland A from 15 February 2013 through 18 January 2015. Coldwater Springs water flows through fractures in Platteville Limestone of Ordovician age. The basic chemistry of Coldwater Spring should be the calcium magnesium bicarbonate water typical of carbonate springs. However, on an equivalent basis, Coldwater Springs water currently contains almost as much sodium as calcium + magnesium and more chloride than bicarbonate. The chloride concentrations are about 100 times the levels from 1880. Maguire (1880) reported the chloride levels of Coldwater Spring were about 4.5 ppm. During the current study the chloride content in the spring increased from about 320 ppm from March 2013 to about 410 ppm in December 2014. In April, May, and June of 2013 and 2014, the chloride rose about 100 ppm in three month-long pulses. The chloride concentration of the water in Wetland A ranges from about 400 ppm to over 600 ppm with a pattern that is a mirror image of chloride component has a chloride to bromide ratio of 2,500 300, well within the range of chloride to bromide

PAGE 28

10 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Mississippi River National River and Recreation Area (NPS, 2012) and is open to the public. The Minnesota Department of Health test results of the spring water in 2005 revealed the presence of bacteriological contamination of total coliform indicating organisms, but did not detect E. coli. The NPS has posted signage at the spring indicating the water is not suitable for drinking. Figure 1 is an aerial view of Coldwater Spring area. trunk highway/light rail interchange is immediately west of the spring. The large white squares visible northwest of the spring are buildings of a VA Medical center. Runway 22-4 of the MSP International Airport extends into the southwest corner of Figure 1. St. Paul, Minnesota is across the Mississippi River to the eastnortheast. The groundwater in the Coldwater Spring region typically flows from west to east toward the Mississippi River, and recharges from precipitation and lakes such as Lake Nokomis and Mother Lake. The bedrock geology of this area is of 26 feet of Platteville Limestone and a variety of soils, which overlays 3 feet of Glenwood Shale (Howe 2014). Both these layers overlie a highly permeable layer of St. Peter Sandstone. The Platteville Limestone is often described as being two different layers, Upper and Lower Platteville, because of the differing amounts of shale content. In the Lower Platteville Plattevilles composition, while the Upper Platteville has a much lower shale contents. The Lower Platteville Limestone thus has a lower conductivity and acts as a Introduction Spring, Minneapolis, Minnesota, to evaluate the impact wetland (Wetland A). This report expands and adds a second year of data to Kasaharas (2014) report on the In the Twin Cities Metropolitan area (TCMA), roughly 350,000 tons of de-icing road salt is applied to the TCMA this salt dissolves in snowmelt that either flows overland to surface water bodies or infiltrates to recharge the water salt applied in the Twin Cities area stays in the regions watershed (Rastogi, 2010). The road salt used is about and is generically referred to as NaCl (Keseley, 2007). This road salt runoff can lead to high levels of salinity in freshwater areas due to the dissolution of NaCl in the water. This is harmful to freshwater aquatic life, regional mammals and birds, and plants native to Minnesota. Salinity Effects on Wildlife 30 days at salinity concentrations of 220-240 ppm, trout behavior is affected at levels as low as 250 ppm, and the overall diversity of aquatic species decreases as the salinity concentration rises (Keseley, 2007). According to the Minnesota Pollution Control Agencys (MPCA) 2010 draft report, 11 metro-area streams have levels of chloride concentration above 230 ppm (Homstad, 2010). Road salt particles may attract moose and deer to roadsides, where cars or trucks may strike them, and birds, such as sparrows, can die after eating only two salt particles. Plants as far as 200 feet away from the roadside ppm concentration can lead to damage to coniferous species such as the pine tree (Keseley, 2007). Because groundwater is the source for drinking wells, high levels of salinity can also affect humans on restricted-sodium diets (Rastogi, 2010). The Sample Sites 11 47.81 W. Coldwater Spring is an important Native American cultural site (NPS, 2012) and is an integral part of the history of Fort Snelling and Minneapolis. The spring site is now a part of the National Park Services Figure 1. An aerial view of the Coldwater Spring site.

PAGE 29

11 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Data and Interpretation Although Coldwater Spring and Wetland A are less than 100 meters apart, their water chemistries and flow rates show differences. Both show significant seasonal fluctuations in many parameters which, in several cases, are mirror images of each other. Chloride Figure 2 displays the chloride concentrations in Coldwater Spring, Wetland A and the Big Culvert during the monitoring period. Historical Data During the 19 th Century Coldwater Spring was the water source of the U.S. Armys Fort Snelling in Minneapolis. Army Captain Maguire (1880) reported the chloride level in Coldwater Spring to be 0.26 grains per gallon which is equivalent to 4.5 ppm in modern units. [Early analyses of chloride were done by titration with a clear end-point and are quite reliable.] Coldwater Spring Figure 2 shows that the chloride content of Coldwater Springs water increased from 320 ppm in March 2013 to 410 ppm in January 2015. Superimposed on the increase of about 45 ppm/year were two pulses between roughly April and July of 2013 and 2014. The first pulse reached 432 ppm on 10 June 2013. The second pulse rose to 455 ppm on 20 May 2014. These chloride levels are about 100 times the levels from 1880. close to the springs throughout the late fall, winter and early spring. If the pulses are attributable to infiltration confining layer for the Upper Platteville Limestone. Water in the Upper Platteville Limestone has easy water flows through fractures in the Upper layers of Platteville Limestone of Ordovician age. Studies of this formation have shown that it is ineffective at filtering out many of the contaminants from the recharge areas, highways (NPS, 2012). The Wetland A sample site is a reconstructed wetland. The Wetland A previously had a building from the former Bureau of Mines, Twin Cities Research Center historic springhouse and reservoir were left intact as the restoration of the Coldwater Spring site was completed by the National Park (NPS, 2012). Wetland A is located The Big Culvert is a 6-foot diameter storm water outlet that drains the area southwest of the Coldwater Spring site. It flows perennially and was sampled twice Methodology Weekly visits were conducted to Coldwater Spring and Wetland A to record the temperature, to measure the conductivity, dissolved oxygen and pH levels of the spring using a Thermo Orion multi-meter, and to collect weekly anion samples. The Coldwater Spring samples were collected at the outfall pipe on the southwest corner of the springhouse ion chromatography [EPA 300.0]. Samples were collected monthly for alkalinity analyses by digital titration [ASTM 1067 02] and cation analyses by ICP/OES [EPA 200.7]. Water samples were collected from the two sites from the beginning of February 2013 to the middle of January 2015. The timing of this data collection was intended to study any impact to the springs salinity levels during snow melt, when deicing salt would presumably be washing into and through groundwater supplies. A temperature data logger was installed in Coldwater Spring. The data logger was inserted into a feed pipe along the north side of the springhouse about 3 meters from the outfall pipe. Figure 2. Chloride concentrations in Coldwater Spring, Wetland A and the Big Culvert.

PAGE 30

12 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE Coldwater Spring, Wetland A and Big Culvert The chloride to bromide ratios in all three sample sites scatter from about 2,000 to 4,500 well within the range of reported values for road salt. The ratios from the three samples of road salt taken at three different sites around the University of Minnesota campus are in the same range. These data suggest that road salt is the primary contributor to the high chloride levels in Coldwater Spring, Wetland A and the Big Culvert. Temperature Historical Temperature Data for Coldwater Spring In 1836 and 1837, a French explorer named Joseph Nicollet measured the temperature of Coldwater Spring in both the summer and the winter months (Nicollet, 1845). He described Coldwater Springs temperature as: Of the numerous springs that issue from the foot of thebluffs [adjoining Fort Snelling] there is one particularly deserving of notice. It is very abundant and perfectly shaded. It goes by the name of Bakers spring. Having taken its temperature three times a day during twenty days of the month of July, 1836, and then again during the following winter months, I never found it to vary more than 46F in July, and 45.5F in January. Converting the temperatures from Fahrenheit to degrees centigrade, the measured temperature average of Coldwater Spring was about 7.8 C in July and 7.5 C in the winter. This indicates that in the summer of 1836 and of road salt runoff, these data suggest it takes about five months for the winter road salt input to reach Coldwater Spring. The magnitude of each years pulse should depend on the severity of the winter which fluctuates significantly from year to year depending on the severity of each winter. Ignoring the chloride pulses, chloride levels in Coldwater Spring are very high and generally increased over the period of this study. Wetland A Wetland A chloride levels (Figure 2) were higher than the chloride levels at Coldwater Spring, and displayed an inverse relationship to Coldwater Spring. Wetland As chloride levels were low when Coldwater Springs chloride pulses were high and vice versa. The dips in Wetland A chloride concentrations were greater than the pulses in Coldwater Springs concentrations, but occurred at essentially the same time. Wetland As chloride levels ranged from 313 ppm to 632 ppm. This mirror image behavior of Coldwater Spring and Wetland As water quality is repeated in several of the other parameters discussed below. This behavior is cryptic and surprising. We are not aware of any other system that displays this mirror image behavior. Chloride is the most common anion in both Coldwater Spring and Wetland As water. Big Culvert Only one sample was collected from the Big Culvert. Big Culverts value is shown in Figure 2 and subsequent graphs. This result is similar to the data seen at Wetland A. Chloride to Bromide Mass Ratios Salts from different sources have different, characteristic chloride to bromide ratios (Davis et al., 1998). In addition to the sewage and rainfall Cl/ Br ratios, Panno et al. (2006) found Cl/Br ratios for softened water ranging between 175 and 1,122 and for agrichemical-affected water with tile drains, from 108 to 1974. The road salt ratio minimum was 1,164 and ranged up to 4,225. Measurements of the chloride to bromide ratios in Coldwater Spring, Wetland A and Big Culvert help to identify the source of the salt. Figure 3 shows the chloride to bromide mass ratios from the sample sites. It also shows the chloride to bromide mass ratios of three road salt samples taken from different sites on the University of Minnesota campus in winter 2014. Figure 3. Chloride to bromide mass ratios for Coldwater Spring, Wetland A, Big Culvert, and for two road salt samples taken from different sites on the University of Minnesota campus.

PAGE 31

13 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 snow is cleared from roads and parking areas allowing greater frost depths and much higher temperatures under summer sun. The lowest point in the temperature fluctuations occurs in early June. The highest temperatures occur in early November. In comparison the air temperatures (Figure 5) peak in late July or early August and the minimum temperatures occur in late January to early February. The temperature fluctuations in Coldwater Spring lag air temperature fluctuations by about four to five months. The data logger record (Figure 6) has relatively smooth, linear temperature profiles with minimal short term variation. The absence of hours to days temperature excursions indicate that the flow path to this sampling point is long enough to even out variations due to storm events but not long enough to average out seasonal effects. The data logger temperature record is surprisingly different from the thermometer readings record. The maximum and minimum temperatures of the data logger record are 0.3 to 1 C higher than those in hand measured temperature record using the thermometer. Discrepancies in the two temperature plots may be due to the location of the temperature samples taken. The data logger was located in one of the entrance pipes on the north side of the springhouse, and the thermometer samples were taken from the south side of the springhouse in the exit pipe, about ten feet from the data logger. Alternatively, the discrepancies may the winter of 1837, Coldwater Springs temperature was essentially constant. The temperature of shallow springs is typically the average annual air temperature of that region. The shallow ambient groundwater temperature in this part of Minnesota is about 8 C Temperature Records Figure 4 displays the weekly temperatures of Coldwater Spring and Wetland A. The temperature was measured using an ASTM calibrated glass thermometer. Figure 5 contrasts the weekly temperature records for Coldwater Spring and Wetland A with the average air temperatures on the days that samples were collected. The air temperature data, measured at the Minneapolis/ St. Paul Airport and which is two kilometers from Coldwater Spring, are from the National Weather Service website, as follows: www.crh.noaa.gov/mpx/ Climate/MSPClimate.php In Figure 6 the 15 minute interval data logger temperature record from Coldwater Spring is shown in blue. The gaps in the data logger plot are due to equipment problems. The weekly thermometer data from Coldwater Spring (from Figures 4 and 5) is shown in red. Temperature in Coldwater Spring The field thermometer temperature of Coldwater Spring (Figure 4) fluctuated seasonally between 10.7 C and 13.1 C in a sinusoidal pattern. This is a 3to 5degree increase from the 1836 and 1837 temperature levels. Since the temperature fluctuates over a 2.4 degree range, it is clear that Coldwater Spring is no longer a constant temperature spring. The increasing range of spring Figure 4. Weekly temperature fluctuations in Coldwater Spring and Wetland A. Figure 5. Comparison of air temperatures with the water temperatures in Coldwater Spring and Wetland A.

PAGE 32

14 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE The concentration of nitrate-nitrogen rose by about 2 ppm over two years, and also displayed a sinusoidal pattern. The annual nitrate minima occur almost simultaneously with the highest chlorides. Seasonal application of lawn Nitrate-Nitrogen Levels in Wetland A Wetland A had a wider nitrate-nitrogen level range than Coldwater Spring, varying between about 0.3 ppm and 6.0 ppm. Wetland As nitrate-nitrogen levels vary much more than do Coldwater Springs but are rising when Coldwaters levels are dropping and vice versa. Major Ion Chemistry Piper Diagrams are a 2D graphical technique for displaying (Piper, 1944). The lower left triangle is a ternary diagram Diagram where the total concentrations of the dissolved ions are shown by the vertical bars. Figure 8 illustrates that the waters from Coldwater Spring and the waters from but differ enough to be able to be able to be distinguished. Wetland As waters contain higher levels anthropogenic ions of sodium, chloride, and nitrate, than Coldwater Springs waters. Most springs from carbonate rocks in Minnesota fall closer to the bicarbonate corner of the anion triangle than do the waters Coldwater Spring and Wetland A. Solution of the Platteville Limestone produces calcium, magnesium and bicarbonate ions. Coldwater Springs be evidence for multiple, different water inputs to the Coldwater Springhouse. Such different sources (fed by different flow paths) would be the start of an explanation for the observed chemical and temperature variations. Temperature in Wetland A Wetland A thermometer temperatures ranged between 6.4 and 13.8 C with the temperature peaking at the end of August. Wetland A is also not feed by a constant temperature spring. The Wetland As water source temperature fluctuation is out of phase with the air temperature by about two months. Wetland A temperatures range from well above to below the average groundwater temperatures of roughly 8 C Wetland As temperatures can be explained as being driven by the air temperatures in a very shallow flow system. Nitrate-Nitrogen The 1974 Safe Drinking Water Act set a national maximum contaminant level (MCL) for nitrate of 10 ppm as an enforceable standard (Water, 2013). Although the nitrate levels in each sample site are below this drinking standard, maximum contaminant levels are purposely set as close as possible to the health goals. Nitrate-nitrogen is an indicator parameter of contamination, and is often accompanied with other pollutants such as pathogenic viruses, bacteria and synthetic organic compounds. Therefore, the elevated levels found in Coldwater Spring and Wetland A are significant and, in this urban setting, suggest contamination from human sources such as lawn Nitrate-Nitrogen in Coldwater Spring The nitrate-nitrogen levels in Coldwater Spring (Figure 7) varied between 2.6 ppm to 5.2 ppm during this study. Figure 6. Comparison of the data logger temperatures with the thermometer temperatures in Coldwater Spring. Figure 7. Nitrate-nitrogen levels in Coldwater Spring and Wetland A.

PAGE 33

15 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 3. Coldwater Spring is no longer a constant temperature spring. Coldwater Springs temperature fluctuates smoothly by 2.4 C The temperatures are 3 to 5 C warmer than the temperatures of the spring in 1837. The temperature fluctuations are not in phase with seasonal air temperatures. Coldwater Springs waters are coolest in June and warmest in around Coldwater Spring has driven larger seasonal variation in soil temperature and higher temperatures overall. 4. Wetland A is fed by a shallow and more local ground water recharge area than Coldwater Spring. Wetland A has larger temperature fluctuations than Coldwater Spring and is only slightly out of phase with the seasonal air temperature. 5. Coldwater Springs nitrate-nitrogen levels rose from about 3 ppm in February 2013 to over 5 ppm in January 2015. The nitrate-nitrogen values in Coldwater Spring vary inversely with chloride. One potential source of nitrate-nitrogen is leaking sanitary sewage lines, but one would expect that input would be a more constant nitrate source year-round. A second source is larger pulses of which occur during the spring and summer in different seasons than the road salt applications. 6. Wetland As nitrate levels peak in the spring at 6 ppm but are generally lower than Coldwater Spring, around 1 ppm through most of the summer and winter. This may reflect the more application on the Coldwater Spring property and Future Work In addition to the potential for multiple water inputs to the Coldwater Springhouse, NPS staff and the Howe and drains of the Coldwater Spring area two old drain pipes to the east and one additional drain pipe to the south. All of these places where ground water returns to the surface need to be monitored and their chemical and physical properties documented. Dye tracing will be a significant tool to determine the possible recharge areas for the Coldwater Spring site in this complex urban environment. Alexander et al. (2001) liter. Both numbers are a measure of the large chloride and nitrate contamination of the spring and wetland water sources. The large concentrations of chloride in the waters from both sources are not balanced by the sodium concentrations. Cation exchange has occurred along the groundwater pathways in the time between the applications of the road salt, and when the dissolved road salt reaches the springs. Both Coldwater Spring and Wetland As waters have relatively constant elevated sulfate ion concentrations of about 80 to 90 ppm. These values are probably produced naturally from the oxidation of sulfides present in the Platteville Limestone. Summary and Conclusions 1. The temperature and water chemistry of the groundwater flowing from Coldwater Spring and Wetland A appear to have been significantly impacted by human activities. 2. The chloride concentrations in Coldwater Springs waters have increased 100-fold between 1880 and the present. The chloride contents in Wetland As waters are even higher than those in Coldwater Spring. The chloride/ bromide ratios in both springs suggest that road salt is likely the primary source of the chloride. The chloride contents of Coldwater Spring and Wetland A fluctuate in mirror image fashion: one is high when the other is low and vice versa. Figure 8. 3D Piper diagram of major cation and anion composition.

PAGE 34

16 NCKRI SYMPOSIUM 5 14TH SINKHOLE CONFERENCE of the dyes through the groundwater system at Coldwater Spring. Figure 9 shows the various input and monitoring points. We hope to have dye trace results to report by the 14 th Sinkhole Conference. Acknowledgments A special thanks goes to the National Park Service assistance and encouragement. Another special thanks to the University of Minnesotas Undergraduate Research documented connections to Coldwater Spring from two, nearby points, but did not monitor any of the other potential resurgences. A dye trace is currently being conducted at Coldwater Spring. Rhodamine and uranine/fluorescein dyes were poured into two separate rain gardens in the Veterans Hospital, located across Highway 55 from Coldwater Spring on 6 June 2015. Eight dye bugs were placed around the Coldwater Spring site to track the movement Figure 9. Map showing the relationships between the TH62/TH55 traffic interchange, the Minneapolis VA Medical Center and the Coldwater Spring Mississippi National River & Recreation Area. The two blue circles show Coldwater Spring (upper) and Wetland A Spring (lower). The orange and green arrows show the results of 2001 dye traces (Alexander et al., 2001). The two red triangles show the 2015 dye input points in the northern parking lot rain garden (upper) and southeastern parking lot rain garden (lower). The green triangles show the outfall pipes of drainage and storm water drains. The two black crosses are sampling locations under the foot bridges across the Coldwater and Wetland A spring runs. Charcoal detectors were placed and are being periodically changed at the eight locations shown by the blue circles, black crosses and green triangles.

PAGE 35

17 14TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 5 Maguire E. 1880. Press copies of office letters sent by Captain Edward Maguire, Chief Engineer, June 26, 1880. In: The Water. Department of Dakota., p. 2 MAR 4725. 2013. Minnesota Administrative Rules, Chapter 4725. Wells and borings. [Internet]. St. Paul (MN): State of Minnesota. Office of the Revisor of Statutes. Available from: https://www. revisor.mn.gov/rules/?id=4725 Mohseni O, Stefan HG. 2006. Study of environmental effects of de-icing salt on water quality in Minnesota. [Internet]. Minneapolis (MN): University of Minnesota Center for Transportation Studies. Available from: http://www.cts.umn.edu/ Nicollet JN. 1845. Report intended to illustrate a map of the hydrographical basin of the Upper Mississippi River. [Internet]. 1845. Document of the US House of Representatives. 28th Congress, 2nd Session, no. 52. Washington (DC): Blair and Rives. http://www.worldcat.org/ title/report-intended-to-illustrate-a-map-of-thehydrographical-basin-of-the-upper-mississippiriver/oclc/166586906?referer=di&ht=edition Mississippi National River and Recreation Area. [Internet]. [Place of publication unknown]: National Park Service. US Department of the Interior. Available from: http://www.nps.gov/miss/ parkmgmt/bomcurr.htm Panno S, Hackley K, Hwang H, Greenberg S, Krapac I, and Identification of Na-Cl Sources in Groundwater. Groundwater 44(2):176-187. PCCC. 2012. Preserve Camp Coldwater coalition. [Internet]. [Place of publication unknown]: Preserve Camp Coldwater Coalition. Available from: http://www.preservecampcoldwater.org/ index.html Piper, AM. 1944. A graphical procedure in the geochemical interpretation of water analyses. American Geophysical Union Transactions 25: 914-923. Rastogi N. [Internet] 2010. Does road salt harm the environment? [Place of publication unknown]: Slate.com. Available from: http://www.slate. com/articles/health_and_science/the_green_ lantern/2010/02/salting_the_earth.html Sander A, Novotny E, Mohseni O, Stefan H. 2007. Inventory of road salt use in the Minneapolis/St. Paul Metropolitan Area. Minneapolis (MN): St. Anthony Falls Laboratory. Available from: http:// purl.umn.edu/115332 and for the opportunity to present the preliminary results at the National Conference on Undergraduate Research in April 2014. SMK would also like to thank her father Hisanao for taking her sampling every week throughout the school year and the cold Minnesota winter, and to her mother Susan for driving her to and from the University of Minnesota campus during the summer. We thank Betty Wheeler for her editing of this paper. References Alexander EC Jr., Alexander SC, Barr K. 2001. Dye Tracing to Camp Coldwater Spring, Minneapolis, MN. Minn. Groundwater Assoc. Newsletter 20 (4): 4-6. River chloride trends in snow-affected urban watersheds: increasing concentrations outpace urban growth rate and are common among all seasons. Science of The Total Environment 508:488-497. www.sciencedirect.com/science/ article/pii/S0048969714017148 Davis SN, Whittemore DO, Fabryka-Martin J. 1998. Uses of Chloride/Bromide Ratios in Studies of Potable Water. Ground Water 36 (2): 338-350. Homstad M. 2010. Hold the salt. In: Minnesota Conservation Volunteer. January-February, 2010. St. Paul (MN): Minnesota Department of Natural Resources: 62-63. www.dnr.state.mn.us/ Spring. Minnesota Department of Transportation Right-of-Way. Kasahara SM. 2014. Chloride monitoring at