Hydrogeological controls on the occurrence and movement of groundwater discharged at Magic Springs in the Spring Branch Creek drainage basin, Spring Branch, Texas

Citation
Hydrogeological controls on the occurrence and movement of groundwater discharged at Magic Springs in the Spring Branch Creek drainage basin, Spring Branch, Texas

Material Information

Title:
Hydrogeological controls on the occurrence and movement of groundwater discharged at Magic Springs in the Spring Branch Creek drainage basin, Spring Branch, Texas
Creator:
Childre, Mark Tilman
Publisher:
University of Texas
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Magic Springs (Texas, United States) ( 29.910506, -98.440617 )
Resource Management ( local )
Genre:
Thesis / Dissertation
serial ( sobekcm )
Location:
United States
Coordinates:
29.910506 x -98.440617

Notes

General Note:
"The hydrogeologic controls, flow velocities and paths, groundwater delineation, and physical characteristics in a joint controlled dendritic conduit-spring system have been characterized. The known conduit branches from C My Shovel (CM) Cave with 4475 meters (m) of measurable passages and tributaries. Surface entrance to CM Cave is located 1360 m upstream from discharge at Magic Springs. Four storm events were measured characterizing the dynamics. The maximum discharge of these four events was 1.2 m3/s (41 ft3/s) with 0.08 m3/s (3 ft3/s) baseflow conditions at Magic Springs. The characteristic shape and response of discharge are well defined with a rise time between 5.5 and 6.5 hours (hr). The half flow period time (t0.5) ranges between 12.9 and 15.7 hr, depending on peak discharge. The rise time and t0.5 occur in less than one day and the conduit volume exceeds 0.5 x 106 m3. The conduit-spring system drains within 3.7 to 7.5 days after the storm event. The thermal effects are event driven, maintaining 85% of the temperature change over 1300 m. The spring discharge has total dissolved solids around 350 mg/L and is chemically stable. The field component of this study include a karst density survey, four dye traces, and continuous monitoring of specific conductance, pressure, temperature, water-level stage height, and discharge at Magic Springs and in the conduit below CM Cave. The general karst density survey identifies caves and dolines within given area. There is a sinking stream that transfers flow from Spring Branch Creek into the conduit system and two focused regions in a karst plain having densities of 20 and 44 karst features/0.16 km2. Hydrographs and chemographs show patterns interpreted as pulses of dilute water recharging through exposed caves, sinkholes, and sinking streams. These pulses have minimal reaction with the rock or matrix during recharge, which is superimposed on baseflow from the joint controlled dendritic conduit-spring system in this karst terrane. The groundwater drainage basin has been defined. The dye tracing results identified groundwater piracy across surface water divides and helped define the groundwater drainage basin. Groundwater velocities were measured between 1800 m/d and 3000 m/d under baseflow conditions. The discharge at Magic Springs under these four storm events showed velocities between 8,700 and 15,120 m/d. An autosampler and charcoal packets were both employed during dye tracing. Both detected fluorescence from all four injection sites. The measured velocities ranged between 1865 up to 2929 m/d under baseflow conditions. All dye trace tests were conducted under baseflow. Under baseflow conditions, dye was only traced to the Magic Springs locations from the eleven charcoal monitoring locations.
Restriction:
Open Access - Permission by Author(s)
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:
K26-01538 ( USFLDC DOI )
k26.1538 ( USFLDC Handle )
11867 ( karstportal - original NodeID )

USFLDC Membership

Aggregations:
Added automatically
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item has the following downloads:


Full Text
Description
"The hydrogeologic controls, flow velocities and paths,
groundwater delineation, and physical characteristics in a
joint controlled dendritic conduit-spring system have been
characterized. The known conduit branches from C My Shovel
(CM) Cave with 4475 meters (m) of measurable passages and
tributaries. Surface entrance to CM Cave is located 1360 m
upstream from discharge at Magic Springs. Four storm events
were measured characterizing the dynamics. The maximum
discharge of these four events was 1.2 m3/s (41 ft3/s) with
0.08 m3/s (3 ft3/s) baseflow conditions at Magic Springs. The
characteristic shape and response of discharge are well
defined with a rise time between 5.5 and 6.5 hours (hr). The
half flow period time (t0.5) ranges between 12.9 and 15.7 hr,
depending on peak discharge. The rise time and t0.5 occur in
less than one day and the conduit volume exceeds 0.5 x 106
m3. The conduit-spring system drains within 3.7 to 7.5 days
after the storm event. The thermal effects are event driven,
maintaining 85% of the temperature change over 1300 m. The
spring discharge has total dissolved solids around 350 mg/L
and is chemically stable. The field component of this study
include a karst density survey, four dye traces, and
continuous monitoring of specific conductance, pressure,
temperature, water-level stage height, and discharge at Magic
Springs and in the conduit below CM Cave. The general karst
density survey identifies caves and dolines within given
area. There is a sinking stream that transfers flow from
Spring Branch Creek into the conduit system and two focused
regions in a karst plain having densities of 20 and 44 karst
features/0.16 km2. Hydrographs and chemographs show patterns
interpreted as pulses of dilute water recharging through
exposed caves, sinkholes, and sinking streams. These pulses
have minimal reaction with the rock or matrix during
recharge, which is superimposed on baseflow from the joint
controlled dendritic conduit-spring system in this karst
terrane. The groundwater drainage basin has been defined. The
dye tracing results identified groundwater piracy across
surface water divides and helped define the groundwater
drainage basin. Groundwater velocities were measured between
1800 m/d and 3000 m/d under baseflow conditions. The
discharge at Magic Springs under these four storm events
showed velocities between 8,700 and 15,120 m/d. An
autosampler and charcoal packets were both employed during
dye tracing. Both detected fluorescence from all four
injection sites. The measured velocities ranged between 1865
up to 2929 m/d under baseflow conditions. All dye trace tests
were conducted under baseflow. Under baseflow conditions, dye
was only traced to the Magic Springs locations from the
eleven charcoal monitoring locations.



PAGE 2

DEDICATION My work is dedic at ed to those who have supported me throughout my endeavor None of my success could be realized without guidance, help a n d support given by my higher power, Joh n Harp, my dear mother Ruth and Sunset Ridge Church of Christ Thank you for providi ng me with constant inspiration, motivation, a nd hope. Nothing is impossible!

PAGE 3

HYDROGEOLOGIC CONTROLS ON THE OCCURRENCE AND MOVEMENT OF GROUNDWATER DISCHARGED AT MAGIC SPRINGS IN THE SPRING BRANCH CREEK DRAINAGE BASIN : SPRING BRANCH, TEXAS by MARK TILMAN CHILDRE, B S THESIS Presented to the Graduate Faculty of The University of Texas at San Antonio In P artial Fulfillment Of the Requirements For the Degree of MASTER OF SCIENCE IN GEOLOGY THE UNIVERSITY OF TEXAS AT SAN ANTONIO College of Sciences Department of Geological Sciences May 2013

PAGE 4

iv ACKNOWLEDGEMENTS A special appreciation must be offered to my committee. Funding was provided by the Department of Geological Sciences, Center for Water Research, and South Texas Geological Society Jones Amesbury Research Grant. Edwards Aquifer Authority provided the dye s the lab for analysis of the traced water samples and support from Anastacio Moncada. Special thanks to Geary Schindel and Steven Johnson for their initial guidance as I proposed the test, planning, and methods. Great support and guidance was given by Richard Slattery (USGS) for surface waters and stream flow measurement s The cavers of south central Texas were not just desired but needed. Tom Rogers and Travis Klaassen provided a water tank generator and muscle Work was provided by Nathan Summa r at the springs I would not have been able to remove the instrumentation and cabling from CM Cave without Zach Schudrowitz and Jill Orr. Special t hanks to Tom Rogers, Wade McDaniel, Joe Schaertl, and Lydia Hernandez for their cave work. This work requi red an accurate map of the cave and conduit system Ben Hutchins Terry Holsinger and numerous others dedicated to cave mapping provided the work Terry Holsinger provided the needed introductions to the property owners. Kurt Menking provided the latest detailed aerial mapping and the instrument storage container. Without the support, assistance, and access granted by the property owners, this study would not have been possible. I was granted access to the springs by Wi ll McAllister IV. The entrance to CM Cave and unending support was given by Joe Eisenhauer. Thank you deeply, Sp ecial thanks must be offered to Wally who owns t he Henderson Ranch ; location of the si nking stream. May 2013

PAGE 5

v HYDROGEOLOGIC CONTROLS ON THE OCCURRENCE AND MOVEMENT OF GROUNDWATER DISCHARGED AT MAGIC SPRINGS IN THE SPRING BRANCH CREEK DRAINAGE BASIN: SPRING BRANCH, TEXAS Mark Childre B S The University of Texas at San Antonio, 20 1 3 Sup ervising Professor: Alan R. Dutton Ph.D. The hydrogeologic controls, flow velocities and paths, groundwater delineation, and physical characteristics in a joint controlled dendritic conduit spring s ystem have been characterized The known conduit branches from C My Shovel (CM) Cave with 4475 meters (m) of measurable passage s and tributaries. S urface entrance to CM Cave is located 1360 m upstream from discharge at Magic Springs. Four storm events were measured characterizing the dynamics The max imum discharge of these four events was 1. 2 m 3 / s (41 ft 3 /s) with 0. 08 m 3 /s (3 ft 3 /s) base flow conditions at Magic Springs. The characteristic shape and response of discharge are well defined with a rise time between 5.5 and 6.5 hours (hr) The half flow period time (t 0.5 ) ranges between 1 2.9 and 15 .7 h r depending on peak discharge The rise time and t 0.5 occur in less than one day and the conduit volume exceed s 0.5 x 10 6 m 3 The conduit spring system drains within 3.7 to 7 .5 d ays after the storm event The thermal effects are event driven maintain ing 85% of the temperature change over 1300 m. The spring discharge has total dissolved solids around 350 mg/L and is chemically stable The field component of this study include a karst density survey, four dye traces, and continuous monitoring of specific conductance, pressure, temperature, water level stage height,

PAGE 6

vi and discharge at Magic Springs and in the conduit below CM Cave The ge neral karst density survey identifies caves and dolines within given area. There is a sinking stream that transfers flow from Spring Branch Creek into the conduit system and tw o focused regions in a karst plain hav ing densities of 20 and 44 karst features /0.1 6 km 2 Hydrographs and chemographs show patterns interpreted as pulses of dilute water recharging through exposed caves, sinkholes, and sinking streams. These pulses have minimal reaction with the rock or matrix during recharge which is super im pose d on baseflow from the joint controlled dendritic conduit spring system in this karst terrane. The groundwater drainage basin has been defined. The d ye tracing results identified groundwater piracy across surface water divides and helped define the groundwater drainage basin. G roundwater velocities were m easured between 1800 m/d and 3000 m/d under baseflow conditions The discharge at Magic Springs under these four storm events showed veloc ities between 8,700 and 15,120 m/d. An a utosampler and charcoal packet s were both employed during dye tracing Both detected fluorescence from all four injection sites. The measured velocities ranged between 1865 up to 29 29 m/d under baseflow conditions All dye trace tests were conducted under base flow. Under baseflow conditions, d ye was only traced to the Magic Springs locations from the eleven charcoal monitoring locations.

PAGE 7

vii TABLE OF CONTENTS Acknowledg e ments ................................ ................................ ................................ ........................ i v Abstract ................................ ................................ ................................ ................................ ............ v List of Table s ................................ ................................ ................................ ................................ ix List of Figures ................................ ................................ ................................ ................................ .. x Cha pter One: Introduction ................................ ................................ ................................ .............. 1 Chapter Two : Hydrog eologic Setting ................................ ................................ ............................. 3 Geology ................................ ................................ ................................ .......................... 3 Cave Systems ................................ ................................ ................................ ................. 4 Study Area and Research Site ................................ ................................ ........................ 8 Chapter Three : Methods ................................ ................................ ................................ ............... 1 0 Karst Density Survey ................................ ................................ ................................ ... 1 0 Rainfall ................................ ................................ ................................ ......................... 11 Magic Springs Discharge ................................ ................................ ............................. 11 Specific Conductance, Pressure, and Temperature ................................ ...................... 12 Dye Tracing ................................ ................................ ................................ ................. 12 Chemical Analysis ................................ ................................ ................................ ....... 17 Chapter F our : Results ................................ ................................ ................................ ................... 19 Karst Density ................................ ................................ ................................ ............... 19 Hydrogeologic Data ................................ ................................ ................................ ..... 2 0 Therm al Effects ................................ ................................ ................................ ............ 24 Chemis try ................................ ................................ ................................ ..................... 24 Dye Tracing Results ................................ ................................ ................................ ..... 26

PAGE 8

viii Chapter Five : Discussion ................................ ................................ ................................ .............. 29 Hydrodynamic Response Two Pulse Recharge Event ................................ ................. 29 Timing and Velocities ................................ ................................ ................................ .. 31 Thermal Tracing and System Response ................................ ................................ ....... 32 Shape and Characteri stics of Response Using Discharge ................................ ............ 34 Discussion on Dye Tr acing and Groundwater Drainage ................................ ............. 3 7 C hemistry ................................ ................................ ................................ ..................... 38 A quifer Volume and Mass B alance ................................ ................................ ............. 39 A Case of Antecedent Moisture Content ................................ ................................ ..... 42 Chapter Six : Conclusions ................................ ................................ ................................ .............. 45 Chapter Seven : Recommendations ................................ ................................ ............................... 48 ................................ ................................ ................................ ............................... 51 Appendix A: Cave Symbols ................................ ................................ ........................ 51 Appendix B: Discharge Wading Rod Measurement Tables ................................ ....... 5 7 Appendix C: Tables ................................ ................................ ................................ .... 66 Appendix D: Figures ................................ ................................ ................................ ... 71 References ................................ ................................ ................................ ................................ .... 101 Vita

PAGE 9

ix LIST OF TABLES Table 1 Chemical characteristics of dyes ................................ ................................ ........ 66 Table 2 The characteristics of discharge during peak storm conditions .......................... 66 Table 3 Rainfall totals for Magic Springs and Honey Creek ................................ ........... 67 Table 4 Dye tracing results from the Magic Springs CM Cave conduit network ............ 68 Table 5 Titration results ................................ ................................ ................................ ... 68 Table 6 Karst thermal patterns ................................ ................................ ......................... 68 Table 7 Chemical Analyses for Spring Branch Quadrangle ................................ ............ 69 Table 8 Total response in discharge flow following storm event ................................ .... 70 Table 9 Conduit spring characteristic parameters ................................ ........................... 70

PAGE 10

x LIST OF FIGURES Figure 1 Stratigraphic Nomenclature and Geologic Sections Gulf Coastal Plain ............. 71 Figure 2 Paleogeographic Map of the Early Cretaceous Period ................................ ....... 72 Figure 3 a Paleogeography of the Comanche Shelf Geologic after deposition Glen Rose .. 72 Figure 3b Geologic cross section of study area ................................ ................................ ... 73 Figure 4 Spring Branch Creek drainage basin and research site ................................ ....... 74 Figure 5 Correlation between water (stage) height and discharge ................................ .... 75 Figure 6 Monitoring C My Shovel Cave injection site #1 ................................ ................ 76 Figure 7 No La Vi Cave survey of injection site # 2 ................................ .......................... 77 Figure 8 Cave Crack injection site #3 ................................ ................................ ............... 77 Figure 9 injection site #4 accessed with 16 meter rappel ..................... 78 Figure 10 Dye injecti on locations ................................ ................................ ....................... 79 Figure 11 Morphology of Magic Springs CM Cave conduit system ................................ .. 80 Figure 12 a Intensity response using Fityk ................................ ................................ ............ 81 Figure 12b Concentration correlated with Fityk area ................................ ............................ 81 Figure 13 Four storm events recorded in year 2012 ................................ ............................ 82 Figure 14 a Hydrograph response for February 18 event ................................ ...................... 83 Figure 14b Hydrograph response for March 9 event ................................ ............................ 83 Figure 14c Hydrograph response for March 20 event ................................ .......................... 84 Figure 14d Hydrograph response for May 10 event ................................ .............................. 84 Figure 15 Cross sectional profile of CM Cave at discharge ................................ ................ 85 Figur e 16 Nomenclature for discharge response ................................ ................................ 85 Figure 1 7 Response in SpC exhibits a two pulse behavior ................................ ................. 86

PAGE 11

xi Figure 1 8 Two pulse system with sinking stream and high density karst areas .................. 87 Figure 19 Magic Springs CM Cave conduit network schematic ................................ ......... 88 Figure 20 Tem perature response across 1304 meters of passage for 3 storm events .......... 89 Figure 21 Seven karst t ypes ................................ ................................ ................................ 90 Figure 22 Dye tracing results Magic Springs CM Cave conduit network system .............. 91 Figure 23 Comparison of dye tracing results between CM Cave and No La Vi Cave ....... 92 Figure 24 Concentration from dye injected in No La Vi Cave ................................ ........... 92 Figure 25 A comparison of concentration between two caves ................................ ............ 92 Figure 26 Hydrograph demonstrates two pulse discharge ................................ .................. 93 Figure 27 Time difference during recharge event with two pulse response ....................... 93 Figure 28 Discharge response and partitioning is based on maximum discharge ............... 94 Figure 29 Ratio of recharge sinking stream divided by maximum discharge ..................... 95 Figure 30 Time delay between recharge and discharge and conduit velocity at MS .......... 95 Figure 31 Groundw ater drainage defined for the conduit network ................................ ..... 96 Figure 32 Control volume for thermal calculations ................................ ............................ 97 Figure 33 Recession hydrographs show conduit geometry drives discharge response ....... 97 Figure 34 Hydrodynamics charact erized 1) rising rate and 2) flow peri od time (t 0.5 ) .... 98 Figure 35 Geochemical analysis Piper diagram ................................ ................................ .. 98 Figure 36 Dynamics from storm events in May 2012 ................................ ......................... 99 Figure 3 7 Dynamics from storm events in March 2012 ................................ ...................... 99 Figure 3 8 Evapotranspiration at Honey Creek ................................ ................................ .. 100

PAGE 12

1 CHAPTER ONE: INTRODUCTION The Spring Branch Creek surface drainage basin was chosen as a study location based on history and previous work by George Veni (1994) Veni co mpleted his dissertation on the hydrogeology of the lower member of the Glen Rose formation (Lower Glen Rose) as a stratigraphic setting Veni (1994) completed his research south of the Guadalupe River and t his investigation i s conducted north of the river in the same formational setting. Similar com position and texture occur in both areas. Karst landscapes in limestone terrane are the result of bedrock dissolution where recharged water dissolves calcite and dolomite enlarging fractures and joints, forming sinks and caves. As w ater moves through carbonate rock, a conduit flow aquifer system can have a residence ti me as short as hours or days and result in short, flashy storm pulses at the springs. Surface water features (e.g. springs, streams, and rivers) interact complexly through karst during e xchange with the groundwater system. Locally at the research site, runoff is intercepted by karst features and sinking streams which drain through the conduit system to Magic Springs. Karst system processes include aquifer recharge by sinking streams, fr actures, and conduit connections between surface water and groundwater Where are the recharge points located? T he recharge features connected through what path ? W hat does the dynamic response show a flashy or more matrix response ? Can thermal signals help define conduit dynamics ? To answer these questions, an understanding of surface and groundwater interactions in karst settings is needed as the surface water interacts with structure and texture, replenishing the groundwater in diverse environments t hrough intricate processes. I nnovations and techniques were used with instrumentation, dyes, and sampling equipment Terry Holsinger (personal communication) has maintained and expanded a karst

PAGE 13

2 work Water quality instrumentation and monitoring provided adequate data to assess the topics of discussion. Exploration of C My Shovel Cave (CM Cave) began with unearthing the entrance drop in 1989 covered by Texas Caver ( Rambo 1990). James Loftin was g iven the privilege of naming while surveying, he opened into the conduit 2 8 meters (m) below the land surface. This conduit is a joint controlled dendritic system that is composed of heterogeneous and anisotropic systems, wh ose characteristic parameters may be used to understand the hydraulics, thermodynamics, and interconnectivity. Water quality sondes, pressure transducers, and autosamplers were used during tests from January 29, 2 012 until this investigation was complete on August 22, 2012. These data characterize the interaction between the conduit system and the discharge at the springs. Hydrodynamics, thermal forcing, and dye tracing techniques were used to characterize th e conduit system Luhmann et al. (2011) and Covington et al. (2011) provided an excellent insight for therm al forcing and Raeisi et al. (2007) was used for hydrodynamics S pecial attention paid to contributions by Arthur Palmer, Derek Ford, and William Wh ite. Johnson et al. (2010) used as a reference when p rinciples of dye tracing provided the flowpath, interconnectivity, and velocities Ford & Williams (2007) gave guidance for understanding the characteristic shape and duration of the dynamic response Banta & Slattery (2011) provided an understanding of the local evapotranspiration. Parallel work on geochemistry in Honey Creek may be reviewed in Musgrove et al. (2010).

PAGE 14

3 CHAPTER TWO: HYDROGEOLOGIC SETTING Geology This study site is stratigraphically in the lower member of Glen Rose Limestone (Lower Glen Rose) atop the contact with the Hensel Member of the Pearsall Formation (Figure 1 ). The Hensel Member acts as a local hydraulically confining unit in the Spring Branch, Texas area The bedding dips approximately 0.5 degree to the southeast at the contact between the Lower Glen Rose and Hensel Member. There were three major transgressive regressive cycles of sedimentation that make up the Trinity Group separated by disconformities such as the contact at the study site between the Hensel Member and Lower Glen Rose. The sea transgressed westward with subsidence and eustatic sea level changes covering an eroded pre Cretaceous surface leaving islands of Pr oterozoic metamorphic and igneous rock (Barker & Ardis, 1996). The Pearsall Formation is composed of three members; Pine Island Shale Cow Creek Limestone, and Hensel m ember s The Pine Island ( Hammett ) Shale Member is highly burrowed and is made up of m ixed clay, silt, and calcareous mud 130 feet thick. This stratigraphic unit interfinger vertically with the overlying Cow Creek Limestone (Barker & Ardis, 1996). The Cow Creek Limestone appear s to have been deposited offshore under gradually shoaling co nditions because of a regressive sea in a high energy period giving beach conditions south of the Llano Uplift as the San Marcos Arch formed ( Figure 3a ) The lower part of the Cow Creek Limestone is composed of fine to coarse grained calcareous sandstone or calcarenite, while the middle part of the Cow Creek Limestone is composed of silty calcarenite, and the upper part is composed of coarse grained fossiliferous calcarenite with poorly sorted quartz grains and chert pebbles (Barker & Ardis, 1996)

PAGE 15

4 The He nsel Member crops out north of the study area in Gillespie County. The Hensel Member is composed of poorly cemented clay, quartz, and calcareous sand, chert, and dolomite gravel as much as 200 feet thick. The gravel beds occur at the base of this stratig raphic unit. The shallow marine deposits of the Bexar Shale Member of the Pearsall Formation are the downdip equivalent of the Hensel Member (Barker & Ardis, 1996). T he intercontinental sea reaches for connection with the Ancestral Gulf of Mexico (Figure 2) Uplift and shallow seas prevailed. The Glen Rose Limestone is composed of sandy fossiliferous limestone and dolostone that are characterized by beds of calcareous marl, clay stone and shale and include thin layers of gypsum and anhydrite throughout t he eastern Edwards Plateau D eposit ion began 112.5 Ma before present in a shallow marine to intertidal environment, grading north to the terrestrial Hensel Member. Figure 3 shows the f inal deposition on the Comanche Shelf and the San Marcos Platform 106.5 Ma [(Hammond, 19 8 4), (Barker & Ardis, 1996) ( Mancini & Puckett, 2005 ] The local t opo graphy of the study site is characterized as a gently rolling landscape that is dissected by s teep and narrow drainages with karstificat ion along fractured rock. Recharge occurs in the Spring Branch Creek drainage basin into the conduit system above the Hensel contact with discharge at Magic Springs and other possible locations. The Glen Rose Limestone has regions with a pronounced cha nge of competent limestone and incompetent marl, giving a stair step appearance this terrain provides excellent drainage through karst features which discharge at Magic Springs. Cave Systems George Veni (personal interview) discussed the history of caves n orth of the Guadalupe River. Veni shared that cavers have been visiting and exploring various caves in the Magic

PAGE 16

5 Joe Eisenhauer owned C My Shovel (CM) C ave and sink plain welcom ing caver s as he reveal ed his sinkhole paradis e. Beginning in 1990, James Loftin led a series of trips digging sinkholes and opening caves. Terry Holsinger was aware of the caves and started a karst density survey by gathering people together or locating sites using GPS. The survey waned in the lat e 90s and 2000s but has now been resurrected and used as a base for the current karst density survey. Andy Gluesenkamp established a resurvey of CM Cave working with the 15 th International Congress of Speleology in 2009. Andy was joined by Ben Hutchins an d Ben Tobin who were the managers of CM Cave. Ben Hutchins and his survey team have completed but have yet to release the final detailed cave map to Texas Speleological Survey Cave systems generally consist of a matrix of intersecting passageways that f orm distinctive patterns (Palmer, 1991). The joint controlled dendritic cave system within the Spring Branch Creek drainage basin provides interconnect ion horizontally with vertical recharge The cave openings that have vertically penetrated the marl layer provide the conduit with direct recharge within this cave system. Observations from within the conduit show cave pattern changes. The water generally goes down dip in joints the passages have very low ceilings in some places and a more vadose format in a higher gradient environment. The ceiling height varies in this section but generally is ve ry low In the middle of these low passages, joint ing is confirmed above. Once the cave turns abruptly in a southwest direction and runs along strike the form is more phreatic. There is a plug above the rapids and the final sump which limits discharge. The gradient would be much less in this latter section running along strike.

PAGE 17

6 When diffuse water passes through a porous cap it enters fractures in the underlying soluble beds and enlarges more or less uniformly. These fractures may have enlarge d when the contact between the Hensel and Lower Glen Rose was below the ancestral Guadalupe River. Eventually the permeability of the soluble r ock exceeds that of the cap rock. W ater that does enter the soluble rock is highly aggressive and gradually widens the fractures into vertical passages T his establishes the basic framework of a joint controlled dendritic cave system. T he Guadalupe Rive r breaches the permeable cap rock, the water table drops, with only vadose water passing through the cave system (Palmer, 2007) This explanation might address the vadose and phreatic passages within the CM Cave system. The conduit system is hydraulically interrelated and spread horizontally along formation al contact following subsurface jointing. Each major recharge through a sinkhole or cave generally contributes to a solution conduit. A secondary network overprint may be interpreted from the high density karst areas (Palmer, 2007) Caves fed by sinking streams are exposed to severe floods during heavy rainfall. CM Cave is a floodwater cave with anastomotic network, or spongework pattern based on bedding plane partings. Mazes are superimposed on this dendritic cave pattern. Floodwater effects are strongest wh ere blockage occurs, similar to the plug before the rapids (Palmer, 2007). This plug below CM Cave entrance has resulted in up to a 3 meter rise during storm events (communication Loftin) flooding most or all conduit passages above sump through the rapids Some piracies in his studies result ed in the development of conduit groundwater divides in the higher elevation passages. Much of the previous paleogeology has been studied and documented.

PAGE 18

7 Veni (1994) defined the geomorphology of Lower Glen Rose cav es where they form preferentially in purer limestones, highly permeable honeycomb limestones, along joints and seldom along faults. He also investigated the physical hydrogeology Veni (1994) calculated the mean aquifer storativity, permeability, transmi ssivity, the volume of the conduit, and diffuse flow storage for karst features within the Guadalupe River drainage basin. Veni (1994) studied two caves, Honey Creek Cave (HCC) and Cave Without a Name (CWN), to determine the dynamics and geometry. The ri se time for HCC clusters around 10 hours, with a range of 4 hours up to 22 hours. Veni (1994) determined that the recession coefficients ( ) demonstrated that the baseflow of HCC is supported by high volumes of water in storage. Calculation shows that = 0.00272 The half flow period time (t 0.5 ) is 254.8 days, not unusual for a representative long, low slope recessions from established aquifers with high storage capacities. Both HCC and CWN responded rapidly in less than one hour from precipitation time and returned to basef low conditions within 3 to 7 d ays Conduits fed by c aves sinking streams, sinks, and other karst features on rolling terrain are exposed to intense flo oding during periods of high rainfall. The largest conduit flows tend to occur when all portions of the watershed are contributing storm flow simultaneously to the outlet, creating subterranean phreatic flow especially near the upstream and downstream sum ps relative to CM Cave entrance. In this study, bare Lower Glen Rose lithology in the Edwards Plateau provides the rapid infiltration in the high density karst areas and provides bank flooding sinking stream conditions. The rapid runoff into dolines cave s and sinking streams cause water to rise to storm levels (> 0.7 m 3 /s or 25 ft 3 /s) in less than one day Karst features were added to the survey map GPS data

PAGE 19

8 provided by Holsinger (personal communication). An example would be Figure 22 new karst feature s around CM Cave entrance. Study Area a nd Research Site The highest elevation in the Spring Branch Creek watershed is the western hilltop of Twin Sisters (Figure 4 ) in southwest Blanco County at 522 m (1,779 ft) above mean sea level (msl). T he watershed drains to Magic Springs at 340 m (1115 ft) msl and finally discharges in the Guadalupe River (Figure 4 ) at an elevation of 296 m (970 ft) msl The w ells listed in t he Texas Water Development Board (TWDB) database draw from the lower member of the Glen Rose Limestone. TWDB also maintains records on chemical analyses for specific wells This study area in Spring Branch Creek drainage basin is bounded by the Guadalupe River to the south, Twin Sisters hills to the north, and highlands to the east and the west (Figure 4) Spring Branch Creek incised the Lower Glen Rose The Cow Creek Member and Hensel Member mark the base of the Middle Trinity Aquifer although locally the Hensel is an aquitard. G roundwater in the Lower Glen Rose drains to a conduit system. Spring Branch Creek has discharge near Magic Springs following a 14.4 kilometer (8.95 mile) long limestone bedded waterway that bisects the drainage basin. Spring Branch Creek watershed is composed of eight tributary watersheds coveri ng the drainage basin, dissecting the upper and lower members of the Glen Rose; the Hensel, and Cow Creek members Spring Branch Creek begins its flow in the upper member of the Glen Rose Limestone in Blanco County crossing Kendall County before discharge from Magic Springs in Comal County. The primary entrance to the conduit system is CM Cave 1360 m upstream from the discharge at Magic Springs. The surveyed distance of the primary conduit in this system is 3755 m (2. 33 miles) in length plus an additional 720 m of tributaries to the first upstream sump.

PAGE 20

9 Divers are continuing their exploration beyond the first sump. There are several tributaries that provide recharge to this conduit and which require additional investigation. The Balcones escarpment resulted from normal faulting which occurred during the early Miocene Epoch T he morphology of the local hilly terrain developed as rivers and streams incised after formation of the escarpment. The research site is located inland of the escarpment at the northwest edge of the ramped, enechelon fault zone at the southeast edge of the Hill Country P lateau N o faults have been identified within the karst survey area Deep Pit and Magic Springs in th e cave mapping Both chemical and physical conditions resulted in dissolution and erosion. E ros ion and the environmental climate in the Magic Springs location resulted in thin stony soil s. The primary land usage is ranching with some farming although residential development is becoming much more common.

PAGE 21

10 CHAPTER THREE : METHODS The primary objective of this research wa s to defin e the discharge response and the groundwater drainage basin at Magic Springs The methodology used for this investigation covers both hydraulic and thermal effects as well as dye tracing processes. S pecific conductance, water temperature, water pressure, and barometric pressure were monitored with data loggers F low velocities, hydrodynamics thermal forcing, and direction through dye tracing were defined T he research plan include s : a) karst survey, b) monitoring various hydrochemical aspects of flow in the conduit and springs, c) thermal forcing and d ) dye tracing results W ater samples were collected and chemically analyzed Magic Springs was instrumented with a water quality sonde, pressure transducer, and an autosampler CM Cave was instrumented with a water qu ality sonde. S ites in Spring Bra nch Creek, Cypress Creek, and Curry Creek were monitored for dye to delineate the groundwater drainage basin Karst Density Survey A karst density survey was performed in order to identify high ly permeable terrain and areas susceptible to contamination of the conduit Terry Holsinger (personal communication) provided preliminary surveys A few of the cave and sink locations identified by Holsinger were verified for confidence resulting in a more complete survey around CM Cave. The survey was completed after a 200 x 200 meter grid system was established. Each sink or cave location was recorded using a global positioning system (GPS) and transferred to a geographic information system (GIS). This study is responsible for identifying the GPS location or naming 43 caves and 49 sinks on the Eisenhauer Ranch.

PAGE 22

11 Rainfall Rainfall was recorded at two locations. Will McAllister IV maintains a daily record of rain readings at the springs throughout th e research period (Table 3). R ainfall data w as also recorded using five minute intervals at station USGS 08167347 Honey Creek Site 1C near Spring Branch, TX. T he Honey Creek site is 4.3 kilometers (km) to the southwest of Magic Springs A comparison bet ween these two sources for rainfall data for these four storm events is shown in Table 3 The analysis required real time dynamic data and the USGS station data because of the five minute recording interval. Magic Springs Discharge A rating curve McBirney between the depth of the water and discharge (Figure 5) and was constructed and applied for calculation of discharge from the measured pressure (Rantz, 1982 ) The depth of the water may be determined from the pressure transducer. The correlation shown Figure 5 may be used when calculating discharge. The discharge from the Magic Springs conduit system was monitored using InSitu TROLL 9000 water quality sonde serial number 33494 using a 300 psig pressure transducer at th e CM Cave instrumentation site. A second InSitu TROLL 9000 water quality sonde serial number 33444 has the same pressure transducer and was installed at Magic Springs disch arge. An InSitu Rugged TROLL 100 pressure transducer with operational range of 30 ft was used at the springs. The 30 ft probe was installed at the spring discharge on May 9, 2012 prior to the May storm event. The 30 ft transducer was used until August 2 2 This transducer was installed again o n October 13, 2012 hoping to record another storm and was in service until

PAGE 23

12 December 8, 2012. The accuracy of this transducer is over an order of magnitude greater than the 300 psig transducer but the latter was common in all storm events and was used in the analyses. The data from this transducer is available for future research. Specific Conductance, Pressure, and Temperature Two types of water quality monitoring devices were used : a) Mul ti Parameter TROLL 9000 and b) YSI 556 MPS (Multi Probe System) The TROLL 9000 has its own internal data logger and can be deployed remotely measur ing specific conduct ance water temperature, water pressure, and barometric pressure. The YSI 556 MPS ha s an external data logger, which is not weatherproof and must be connected real time to the water quality sonde. The YSI simultaneously measures conductivity, temperature, and barometric pressure T he maintenance, calibration, and measurement procedures followed those recommended by ( Radke et al. 2006 ) The TROLL 9000 and the YS I 556 MPS were calibrated using the same water source on March 4, 2012. A calibration was conducted after completion of test revealed a difference of 76 S/cm for SpC. The difference was ramped from zero in March to the June 30 calibration difference. The difference between SpC in the cave entrance and SpC at the springs shown on the hydrograph are accurate on March 4 and June 30 2012. Dye Tracing Dye tracing is an effective way to determine apparent velocities (point to point), fluid travel time, and flow directions in karst aquifers (Johnson et al., 2010) T he charcoal receptors and a Teledyne Isco 6712 full size portable auto sampler were used to s ample for dyes at the monitoring points A Teledyne Isco 6712 full size portable auto sampler includes a pump, collection hardware, and software programs for operation. The Isco 6712 controller was set to time paced sampling, for each water sample collec ted. Table 4 describes each dye injection,

PAGE 24

13 timing, concentration, and velocities. Dye concentration was measured using a Perkin Elmer LS50B fluorescence spectrometer. Three non toxic, fluorescent organic dyes were used in testing for this study: uranine (fluoroscein), eosin, and sulfo rhodamine B (SRB). These dyes were chosen for the experiments because of their non toxicity, cost effectiveness, and ease of detection. All dyes used are fluorescent an d used as colorants in medicine, foods, cosmetics, and industrial applications. Table 1 lists the names, molecular weights, Chemical Abstracts Service (CAS) registry number and excitation wavelengths of the dyes used in this series of tracer tests. The dye v olume for each injection site was calculated using equation 1 ( Worthington and Smart 2003) 1 w here m is the mass of dye injected in grams Q is the output discharge in m 3 /s c is the peak recovery dye concentration in g/m 3 a nd D = distance in meters between injection and recovery points Dye masses and injection locations Four c aves were chosen for dye injection points since these features are direct paths to the groundwater system. Wells were not used as injection points for this study because they may not be as integrated into the regional groundwater flow system as caves or other dynamic karst features. The procedure of dye injection consisted of prewetting injection points with water volumes appropriate for each site injecting the dye, and then flushing the dye with additional water to carry it into the aquifer where possible. Prewetting is thought to reduce adsorption of the dye on rock and soil through the vadose zone to the water table

PAGE 25

14 The four dye injection site s were identified and distributed across the study site and are shown on Figure 10. T wo locations were outside the Spring Branch Creek surface drainage basin. The first location was injected with 10.85 grams of uranine dye directly into the conduit 1859 m upstream from the spring discharge (Figure 6 ) B oth an apparent and actual velocity with flow direction will be determined The entrance to C My Shovel Cave ( Figure 6 a) marks the first dense karst area which provides recharge to the conduit and aqui fer 27.2 m below. This cave has an initial vertical drop of 6.7 m followed by an off cambered 9.6 meter vertical traverse followed by another vertical drop of 10.9 m. At the base is a horizontal passage that leads to the conduit flow. A communications l ink ( Figure 6 b) allows direct monitoring and download capability at the surface ( Figure 6 c). The second injection site was located in a siphon at the base of No La Vi Cave (Figure 7 ) 40 m downdip from the conduit in the Cypress Creek surface drainage basin This flowpath was estimated to be 105 m along subsurface jointing from the injection site to the visual point of entrance up dip and upstream at survey marker MM5 The mass of the dye injected was 168 grams of eosin and was flushed with 1500 liters of water. No La Vi Cave is located approximately 40 m downdip from the passage and 66 m from the entrance to C My Shovel Cave. The pit has a difficult approach at the entrance, progressing into a vertical section, with a drain in the marl layer over 15 m be low the surface. Cave Crack was chosen as the third dye injection site 198 m downdip from the passage in the Cypress Creek drainage and 222 m from the CM Cave entrance (Figure 8 ). The mass of the dye injected was 146 grams of sulfo rhodamine B and was fl ushed with 30 00 liters of water transported by Joe Eisenhauer.

PAGE 26

15 (Figure 9) was chosen as the fourth dye injection point where the drain is 465 m from the closest mapped passage (October 2012) for C My Shovel Cave. It is suspect Cave and C My Shovel Cave. The dyes were locked up and possibly adsorped following injection because of the noted change in the morphology of the clay marl layer near the drain for this feature. The mass of the dye injected was 254 grams of ura nine and was flushed with 1500 liters of water. The flushing was discontinued during a rain storm. The relative location of all dye injection sites may be seen in Figure 1 0 Cave Crack lies within the Cypress Creek surface drainage basin but drains part ly into Magic Spring in the Spring Branch Creek surface drainage basin. All of the dye injection caves follow a specific pattern in their formation morphology. A ll local caves develop vertically from stoping in the Rust member and are confined by the ma rl matrix above the Honey Creek member. The estimated elevation of the marl matrix is located at the Honey Creek and Rust members of the Glen Rose Formation at 357 m msl (Clark et al., 2013) or approximately 17 m (55 ft) above the spring discharge overflow Once penetration of the confining bed has been achieved the water takes a 17 meter vertical drop (Figure 1 1 ) into the conduit system. are still working on the marl in much the same fashion that C M Cave did years ago The marl layer is the contact between the base member and the entrance member of the Lower Glen Rose. Once the marl has been penetrated, the recharge is percolates to the base into the cond uit system in less than one day A sketch show ing recharge usin g a sinking stream and autogenic capture (Gunn, 1986) is drawn in Figure 11 with similar characteristics to the local terrane around Magic Spring s Morphologically, joint controlled dendritic caves consist of passages that join each other as

PAGE 27

16 tributaries. The sketch (b) represents the environment between No La Vi Cave and CM Cave. These caves and other karst features form from drains through sinkholes generated from aggressive waters going from paleokarst to the current karst terrane. The mar l matrix marks contact between Honey Creek and Rust members with different joint orientation Monitoring locations There were four dye injection points, one active monitoring site, and 11 passive monitoring sites chosen during this experiment. The autos ampling system was the active at the Magic Springs monitoring site. Passively, the charcoal receptors were distributed in the research area and in the adjacent surface water sheds. A total of 11 sites were selected and are shown in Figure 4 During moveme nt of tracers through monitored sites, fluorescent dyes were adsorbed and accumulated onto activated carbon samplers. The carbon dye receptors were deployed in flowing water of springs, streams, and caves as well as in intermittent streams. Sample collecti on and analysis An ISCO 6712 programmable automatic water sampler (autosampler) was deployed at the Magic Springs discharge. The autosampler was programmed to collect 24 samples at one two or four hour intervals. Water sampling was started before dye injection to collect samples to test for possible background fluorescence. At the end of each automatic cycle, each bottle was decanted into a 13 mm screw top glass vial and marked with identification prior to analysis using the Perkin Elmer LS50B fluores cence spectrometer. Three standards were prepared for the uranine, eosin, and sulfo rhodamine B dyes that were used in the tracer tests. The solutions for these dyes were prepared based on the dye mass and then diluted with deionized water with filtering through a 0.2 micron filter for three concentrations for each.

PAGE 28

17 The Perkin Elmer LS50B measure s fluorescence in intensity units, which is directly proportional to the concentration of dye. Dye peaks were separated from background fluorescence of water by fitting the curves to the Pearson VII statistical function using Fityk software (Wojdyr, 2010) The difference between sample and background fluorescence is the net intensity. Net intensity measurements were converted to a concentration using the Figure 1 2 calibration curve In Figure 12, t he intensity response is graphed using Fityk with water response as background. The concentration may be determined from the area produced by Fityk. Breakthrough curves were prepared from laboratory results. The initial travel time, duration, and peak c oncentrations were calculated. The breakthrough curves, measure the dye response from the cave injection sites. Chemical Analysis Three water samples were collected and analyzed for major cations, anions, and stable isotopes. The work was done by Texa s State University in San Marcos using the Schwartz Isotope Lab. The instrument used is a Mod el 908 0008, Los Gatos Research Liquid Water Stable Isotope Analyzer. This instrument uses Off Axis Integrated Cavity Output Spectroscopy (OA ICOS) laser technolo gy to simultaneously measure Hydrogen and Oxygen stable isotopic ratios from liquid water samples. If the solution is bicarbonate, the equivalence point will always be at a pH of 4.5. The titrations of water samples are to insure that these solutions are bicarbonates and there are no other alkaline substances that may be affecting the equivalence point. The analysis performed with the data collected in a titration experiment with HCO 3 using the equation: 2

PAGE 29

18 w here Alk is the HCO 3 alkalinity of the sample B is volume of acid titrant added from the initial pH to the bicarbonate eq uivalence point in milliliters, C a is concentration of acid titrant in milliequivalents ( meq ) per milliliter CF is correction factor, V s is vol ume of the sample in milliliters This should provide the bicarbonate alkalinity with an equivalence point of pH=4.5.

PAGE 30

19 CHAPTER FOUR : RESULTS Karst Density As a starting point, Terry Holsinger provided a karst density survey within the Spring Branch Creek drainage basin above Spring Branch Road. Magic Spring s Eisenhauer Ranch and other karst features like are included in his study This study expanded the survey includ ing a complete inventory of all featu res on the Joe Eisenhauer Ranch and the Henderson ranch. Two high density karst areas and a sinking stream have been defined as focused recharge Overall, the combined karst density survey area (Figure 18) encompassed 3. 52 km 2 revealing 146 sinks, pits, and caves. An investigation satisfactorily verified some of the previous survey points within the Twin Oaks ranch area giving coordinate system agreement T he study identified two high density karst regions and a sinking stream In Figure 1 8 there were 4 4 karst features in a 0.1 6 km 2 (400 m x 400 m) network which provide d the first pulse ( pulse I ) recharge near the entrance to CM Cave. There is a second area with 20 features per 0.1 6 km 2 of high density karst located 3 700 to 3900 m past the first sump al ong the conduit but on the surface providing pulse recharge. The sinking stream takes water that has pooled ( by observation) and develops a bank over flow condition recharg ing the groundwater tributaries of the spring conduit system All of the water discharged in Spring Branch Creek during a storm event enters Cool Creek Cave until the storm event breaches the top of the pool increas ing the surface water drainage with the surface tributary. Combine the sinking stream with the second dense area defin ing pulse II recharge Rapid recharge characteristics have been identified down dip 0.5 o from CM Cave entrance. Dye has been detected from Cave Crack 200 m downdip during baseflow conditions using a charcoal packet located at CA18 (see note in center, Figure 22). Based on the flowpaths

PAGE 31

20 and also visual observation of dye at MM5/AA1 confirms Cave Crack and No La Vi, as well as other surrounding karst features, are part of the network overlay. Both No La Vi and Cave Crack support the seco ndary dendritic or network overprint hypothesis The dye injected in No La Vi Cave (Table 4) visibly appeared within 4 hours at the MM5/AA1 survey marker (see note below legend, Figure 22) The dye discharged from No La Vi Cave moved up stream (relative t o cave flow) and up dip being detected using the autosampler w ithin 18 hours at Magic Springs under baseflow conditions Table 4 shows th e dye tracing results from analysis of autosampler bottles and from the charcoal packet s Charcoal packets are much more sensitive and show positive but not concentrated enough for the autosampler bottle Low levels of dyes at Cave Crack and at after s torm event s occurred on July 9 through 11 No significan t change s in discharge were visible using the 300 psi transducer. Hydrogeologic Data The response time of karst aquifers depends, at a minimum, on three factors: (a ) the contribution of allogenic r echarge and i n ternal runoff (b) the carrying capacity and in t ernal structure of the conduit sy stem and (c) the area of the ground water bas i n (Ford & Williams, 2007) Table 2 shows the overall shape is defined by the maximum discharge rising time, and half flow time. The recession coefficient [ (hr 1 ) ] defines the slope on the logarithm of discharge T he response shape may be determined with the maximum discharge, half flow period time (t 0.5 ), rise time, recession time, and the rising rate. The rising time is invariant and th e t 0.5 is dependent on maximum discharge

PAGE 32

21 M ore than one recession constant may be defined for these four storm events (Figure 33). T his study will only be concerned with the initial t 0.5 even though frequently there will be two or more straight log linear line segments on the discharge Figure 16 is a schematic of a storm hydrograph showing the nomenclature for the discharge at the spring s In this example, t he discharge pulse recorded on the hydrograph i s initiated by a storm event which defines zero in the case A lag time could not be recorded because of the remote location of the rain gage 4.3 km to the southwest The arrival of the storm flow initiates a rapid rise in the spring d ischarge that forms the rising limb and gives the rising time Spring flow reaches a maximum discharge, Q max at the crest, and then discharge begins to decrease i n the recession 1 i mb. The time from recharge through the rising limb to Q max is the peak lag time. The schematic hydrograph of a storm event was chosen as an example of nomenclature i n Figure 16 T he peak lag time is a combination of the initial lag and the rise time. A recession coefficient ( ) may be determined and is characteristic of that c onduit spring system. T he half flow period time (t 0.5 ) may be calculated with maximum discharge and baseflow The t 0.5 for this system is independent of baseflow and total flow (Ford & Williams, 2007) From this data, a recession coefficient may be calc ulated. These storm events produced the maximum discharge shown in Table 2 for each rain event defined in Table 3. The total rainfall recorded at Magic Springs may be compared against Honey Creek located 4.3 km to the southwest. The total rain event and the maximum are highlighted. Table 3 provides a comparison between the overall maximum discharge, total rain at Magic Springs, and the instantaneous rate from Honey Creek.

PAGE 33

22 Table 2 shows the results from analysis of timing and duration for each storm event The time required in all 4 events for pulse I and pulse II may be determined from the CM Cave specific conductance ( SpC ) response The timing by comparison between the response at CM Cave and the springs was define d and velocities calculated. Both event timing are dependent on the discharge The May storm event gives an excellent example concerning t he flow dynamics of a focused two pulse recharge system (Figure 17) The timing differences estimated storm velocities, and the pulse II delay have been marked in Figure 14d. The timing is dependent on the discharge during the period being timed The measurement of time may be made for any similar distinct points from the springs and CM Cave SpC response The initial pulse appears to have the proper timing based on the response shape demonstrated by dye but then repeats with a second pulse. The recharge event for the May storm has time duration of 21 hours before the phase II response. This gives evidence for a baseflow recharge quick flow recharge baseflow event The hydrogeologic control expressed by pulse I and pulse II recharge and is dependent on average discharge On February 18 an event was recorded following a storm three weeks earlier which ended a year long drought. The February 2012 event (Figure 14a) began just after midnight with a strong downpour shown in Table 3 T he first recharge event occurs from baseflow conditions (Figure 1 4 a ) T he SpC response then returns towards baseflow followed by another drop as pulse II dilutes SpC further. Within 17 days the second recorded event occurred on well saturated soil following the February event. The early March event was in response to three significant storm s over 1.8 d ays (Figure 14b) Both the entrance to CM Cave and the spring discharge were instrumented for

PAGE 34

23 dynamics and timing The defini tion of the recession response characteristics and may be calculated The rainfall totals and rates may be reviewed in Table 3. Timing between the re sponse at CM Cave and Magic Springs may be analyzed since both instrumentation sites were recorded. The timing delays and the two pulse response may be reviewed in Figure 1 4 a d. The third recorded event occurred just after midnight on March 20 ; 11 days later The total rainfall and rainfall rates may be reviewed in Table 3 Two points were used to evaluate the differences between the two instrumentation sites. Figure 1 4 c identifies the timing and the pulse response. A dry period of 51 days transpired before the May storm event occurred resulting in a logarithmic decline in discharge Evapotranspiration was dominating the water budget so that a major storm event occurred 6 days earlier with no significant change in discharge. This rain sequence delivered substantial rain (Table 3) on May 6, satisfying the antecedent moisture content in the soil so that the next response to a storm surge would be positive On May 11, the storms began delivering 1.56 inches with immediate recharge response. The instrumentation sites are separated by 13 04 m of measured conduit and provide data to under stand the flow concepts These concepts are essential in developing a hypothesis concerning the recharge arriv al in two pulses (Figure 1 7 ) The example shown in Figure 1 7 is the May event and identifies the average discharge as a variable for the period between the pulses so that the flow dynamics may be understood Thermal effects Temperature data was recorded at the springs and within the conduit. The control volume is the interior of the cave passage ( Figure 3 2 ) Convective heat transfer is more dominant as turbulence increases. The data recorded for evaluation of this temperature exchange

PAGE 35

24 as a secondary tracer may be reviewed in Figure 20 These data were recorded at CM Cave and Magic Springs from which the t iming and shape of the response may be determined. Equilibrium exists before each storm event so the temperatures are set to the same level. When a conduit is full of water, convection in the water and conduction in the rock are the primary processes S ince convection and conduction act in series, if one of the two processes is significantly slower at exchanging heat than the other, then the slow process will be rate limiting, and exchange rates can be approximated by only considering the slow (limiting) process. The evaluation showed that conduction was an order of magnitude higher process. These hydrographs (Figure 20) show the thermal data recorded real time throughout the study time. The temperature response at the springs will be 85% to 87% of the t emperature response 1304 m up stream. This lack of radial heat flux may be compared to very high lateral heat flux from velocity during a storm event and classifies this as a Pattern I or thermally ineffective system with event scale variability (Luhmann et al., 2011) The temperature response mimics the SpC response and the timing of temperature transfer is dependent on maximum discharge Chemistry Water samples were collected during three visits to the springs on August 22, September 15, and December 8. Two bottles were captured, sealed, and measured the cation complex and anion complex. The third bottle was designated to be used during alkalinity titrations. The data sets measured represent the acidimetry titration readings of water samples to det ermine alkalinity ( Radtke et al 200 6 ). All isotope data are reported relative to Vienna Standard Mean Ocean Water ( VSMOW ) Cations, anions, pH, and alkalinity were measured and reported in Table 7. Alkalinity was

PAGE 36

25 measured using 1.6 N H 2 SO 4 titration to a pH inflection point. The titration results have been analyzed with the results shown in Table 5. W ater is electrically neutral and the charge balance error should be zero. To determine the accuracy of water analyses, charge balances were calculated by means of the following equation: 3 where all variabl es are in meq/L and the charge balance ( CB ) error is in p ercent. Charge balance errors were between 2.9% and 7.0% for the samples T he closest well had a charge balance error of 3.4%; all well within limits. In 1966, TWDB well 6805502 constituents were identified and is the well that is located 4,140 m north o f the springs and is the closest Neutralization is the process whereby an acid and base react with one another to form a salt and water. A titration experiment is used as an application to determine neutralization of a sample with unknown purity by measu ring the amount of an acid added until the equivalence point is attained at neutrality. Once enough of the acid has been added to neutralize the base, it is possible to determine how much base exists in the solution. For typical water samples the activity is generally from HCO 3 and the amount of acid required to achieve the pH equivalence point of 4.5 establishes the alkalinity of the solution. Table 5 shows the pH, equivalence point, and alkalinity for these three samples. Two samples were collected fr om the spring. The data for the first is during baseflow conditions. The second was above base flow with a slight increase in water height

PAGE 37

26 Dye Tracing Results Results include hydrographs for each of the four traces measur ing groundwater velocities, travel time, flow direction, and establish the degree of hydraulic connectivity between different creek drainage basins. Tests were conducted for characterization of the role of karst terrane. There are seven simple karst t ypes which may significantly influence tracer tests between the point of inflow (IN) and the point of outflow (OUT) in a karst system. Discharge into the conduit is q, discharge out of the conduit is Q, tracer mass injected into the conduit is m i and tracer m ass recovered is T T These t ypes are defined by Fields & Nash (1997) and may be interconnected with any of the others as conduit systems are established (Figure 21) Magic Springs are actually three separate and distinct springs within 100 m of the overf low. This is a combination of Type I V a recharge followed by a Type IIIb discharge from Figure 21 (Field & Nash, 1997) The tracing results showed conditions that were measured with actual velocities between 1800 and 3000 m/d under base flow conditions while storm velocities ranged from 8400 m/d to 15,120 m/d Pirating is implied based on positive tracer tests for t he locations of the injection site 2 (No La Vi) and 3 (Cave Crack) and will impact the definition of the groundwater drainage basin. Figu re 22 identifies all four injection sites and the methodology and velocities where appropriate. Both apparent velocity (V a ) and measured velocity (V m ) were calculated (Table 4) based on the data and understanding the flow concept through the s ystem. The sinuosity for CM Cave injection 1 has been calculated as 2.03. The sinuous behavior of other injection sites may be calculated by V m /V a given in Table 4 Two traces showed that groundwater which should drain

PAGE 38

27 in Cypress Creek crosses the surface water div ide, discharging at Magic Springs. Two dyes showed up in water samples and two dyes were recorded using the charcoal packets; all positive. During analysis, the text file was created from the original intensity plots generated by the Perkin Elmer LS50B. The process is executed for each intensity plot. Once convergence is achieved, the data for the sample, date, center, height, and area should be copied and loaded in the concentration data file folder. The dye regression standards given in Figure 23 may be used to convert the intensity to concentration using the area generated by Fityk program. The concentration response may then be graphed and compared. Once all the samples from one run have been analyzed, a plot showing concentration as a functio n of time may be made (Figure 24 ). The concentration curves for the CM Cave and the No La Vi Cave injection may be reviewed in Figure 25 in values and Figure 26 as a comparison The response in CM Cave will give the quickest response time and highest per cent recovered from field investigations. The response given by No La Vi Cave illustrates the spongework terrane near the entrance to CM Cave. Each dye s regression curve may be used to derive concentration as a function of area from Fityk. Dye solutions were prepared on the basis of mass and diluted with deionized water filtered through a 0.2 micron filter to produce dye concentrations in the range that was expected in the water samples. The area is based on the curve fit provided by Fityk for the data, minus the background. This area will be used in the conversion to concentration. The concentration for uranine may be determined by the equation in Fi gure 23 : 4 from the regression curves shown in (B). Once the concentrations have been determined, calculations based on discharge and t may be made and the total mass of the dye detected

PAGE 39

28 summed up (Figure 23 ). A percentage recovered is calculated. The tan line illustrates the calculations and the area represented on the graph. Knowing the discharge (Q) and the concentration (C) the total mass (M R ) of the dye recovered may be calculated by summing the t across th e concentration curve using: 5 The mass (M R ) retrieved may be compared against the total mass (M T ) of the dye initially injected using: 6 The percentage recovered (% R ) provides an initial look at the mass balance q uestion. Once the percentage recovered is calculated, the concentration curve may be compared against the secondary tracer of SpC during a storm event.

PAGE 40

29 CHAPTER FIVE : DISCUSSION The three main issues that characterize the multiple focused recharge are to : 1) use t he shapes and durations illustrated in the hydrograph s of focused recharge events and how these events affect the system, 2) gain insight how temperature variab i l ity characterize the exchange of heat flux radially and laterally and 3) gain in sight on concentration shapes, flowpaths and velocities from tracer test s Hydrodynamic Response Two Pulse Recharge Event The characteristic response curves show bimodal behavior as the response transitions from baseflow to diffuse and recharge flow through both pulse I and pulse II Dye t racing from injection sites 1 and 2 indicate a unimodal response. The hydrographs for SpC, discharge and rainfall for all four storm events may be compared in Figure 14 a through 14d This bimodal behavior indicates two or more separate but distinct recharge paths beyond the entrance to CM Cave This behavior al response supports t he hypothesis that r echarge enters the groundwater system at the two or more focused locations (e.g. high density karst areas, s inking stream s other karst features ) The shape of the SpC response at CM Cave and Magic Springs provides evidence that the concentration from the dye injections is modified from an external source ( labeled ulse II in Figure 1 8 ) during a storm event The shape of the SpC response to the storm event reflects multiple focused recharge The second pulse is from a separate source upstream from the CM Cave entrance above njection L ocation Figure 22 Figure 19 is a dimensioned schem atic of this conduit system with the discharge at Magic Springs Actually, there are three springs within 100 meters

PAGE 41

30 The first high density karst area surrounds the entrance to CM Cave and the second high density area is located 2451 m upstream with the dominant sinking stream is located up an infeeder Combined, the sinking stream and second high density karst area constitute the pulse II recharge. This is a schematic of the joint controlled dendritic pattern cave system; sur veyed and the distances are listed next to each trunk. The network maze pattern may be overprinted in the high density karst areas. The initial sump and the final sump below the rapids are shown. It has been confirmed by Loftin that f looding occurs behi nd the final sump and phreatic flow conditions exist during storm events This means that w ater rose about two and one half to three m eters based on information given by Loftin The time required is illustrated in a hydrograph with transgression of the bimodal response for these two pulses Beginning with baseflow conditions, before the rain event dilute recharge enters from the storm event from the SpC response The dilute recharge continues until the matrix component returns The second high density karst recharge then reaches the instrumentation near the second peak in CM SpC providing another pulse of dilute recharge flow. Th e karstic terrane sinking str eam topology, and the conduit characteristics are the primary controls on the rapid bimodal response Figure 28 d efines the differences between pulse s I and II based on SpC history using the second SpC peak time location as the maximum flow for pulse II The time period for the conduit rising limb are characteristic of this conduit spring system. The spring system response also is characteristic based on t 0.5 and maximum discharge Any hydraulics that happens within this rise time period would be the same and should discharge With the rise time defined, t he location for maximum

PAGE 42

31 flow and maximum diffusivity are known and will the same for all four events. Atkinson (1977) stated that the a nalysis of spring hydrographs have his torically revealed that diffusivity may be estimated such that 50% of the spring discharge is recharge and 50% is diffuse flow H y drograph analysi s h a s a long history i n ka r st hydrology going b a c k t o t h e pion e eri ng work in the 1960s. This approximation w as used while evaluating the data in Figure 28 Figure 29 estimates the ratio between the second high density karst area and the sinking stream based on maximum discharge T he May 10 storm event was chosen because there was no subsequent rain, it had the lowest maximum discharge and the discharge shape offered a n additional distinction between the two pulses. Once the May 10 event hydraulics was understood the pulse events may be based on the SpC response shape. Once this evaluation is appl ied to all storm events a r atio curve may be derived (Figure 29 ) Therefore the sinking stream increases its ratio with increasing maximum discharge and eventually dominates the flow pattern. Continuing research could be designed to specifically identify the threshold points and flow patterns associated with the sinking stream Timing and Velocities The geometry, hydraulic conductivity, and changes in head across the conduit provide timing differences between the pulse I recharge and the discharge. Thi s time delay noted in Figure 1 4d and is dependent on discharge In Figure 30 the time delay may be correlated with a high degree of determination (R 2 ) discharge An understanding of the dynamics will allow calculation of the average flow velocity. The conduit velocities for these storms ranged between 350 m/hr to 630 m/hr (8400 m/d to 15,120 m/d) Data w ere recorded and monitored at both the pulse I recharge ar ea and the discharge at Magic Springs where t he timing difference correlate s with average discharge There are 13 04 m

PAGE 43

32 between the monitoring locations through a conduit with one sump. The time delays measured may be determined, similar to the measurement shown in Figure 1 4d Once the input has been defined, the time delay may be plotted against total discharge with a degree of determination greater than 0.93 (Figure 30 ). Therefore, the time difference between the recharge and the discharg e is dependent on average discharge Thermal Tracing and System Response Temperature variations at karst springs and cave streams can provide information indicating aquifer geometry recharge and system response These patterns can be divided into two types: those produced by discharge with ineffective heat exchange (e. g. conduits where water rapidly flows through the aquifer before it can equilibrate to the rock temperature ) ; and those produced by flow paths with effective heat exchange in small fractures and pore space. Luhmann et al. (2011) describes four variants of heat transfer (Table 6) T here may be other patterns, particularly in other climates but Pattern 1 refl ect s local observations in south central Texas. Th e four storm event variant is an event scale with temperature fluctuation that occurs over the 3.7 and 7.5 day storm event Luhmann et al. (2011) identifies this system as a n ineffective heat transfer or Pattern 1 (Table 6) system with a Stanton number less than one T he cross sectional profile of discharge at Magic Springs is shown in Figure 15 Magic Springs exhibits Pattern 1 c haracteristics where water enters quickly as recharge at a different temp erature than the aquifer The hydrographs shown in Figure 20 show that 85% to 87% of the temperature change is transmitted over 1. 3 k m T his system has an ineffective heat exchange with event scale variability from storm events. T he shape and the response of the temperature parameters show rapid transient times through the conduit system similar to SpC, as discussed in the Timing and Velocity section Diagnostics of the hydrodynamic data confirms this claim.

PAGE 44

33 Covington (2011) provides research specifically on both full and partial pipe flow thermal characteristics with original research by Gnielinski (1976). St anton number (St) and Nusselt number (Nu) may be calculated (Covington et al., 2011) so that the heat flux and conve ctive influences may be known. As Nu increases the flow in the boundary layer becomes much more convective and turbulent O bservation of turbulence visible in the discharge flow confirms that Nu > 1 The Stanton number (St) may be used to define the heat flux during heat transfer and also indicates the flow exposure time. St is a dimensionless number which measures the ratio of heat flux transferred radially into the walls to the heat flux into the fl uid, or laterally during flow from velocity. Essentially, St represents the ratio of heat flux into the conduit wall to the heat flux along the conduit It is used to characterize heat transfer in convection flows. 7 where V is th e velocity of the fluid, c p is the specific heat of water at 20 o C or 4.177 kJ/kgK by Wark (1977), and the convective heat transfer coefficient (h conv) is between 20 and 100 W/m 2 K When St is large (St >>>1) then the exchanged heat is effective allowing water flowing through the conduit to equilibrate with the rock temperature. Th is is not the case in the Magic Springs conduit system based on the storm data The analysis indicates a transfer of 85 to 87% of temperature change (Figure 20) with an ineffective heat exchange system T herefore St < 1 and the exchanged heat is ineffective with event scale variability which is supported by the hydrographs as the thermal signature move 1.3 km through the conduit to the spring. Based on hydrograph data, most of the energy is transferred downstream, not to the conduit walls. Karst conduit networks are often ineffective enough to allow transport of non equilibrium water deep into or all of the way through the system. Temperature signals play an important role

PAGE 45

34 in defining internal aquifer structures. During storm events the hydrograph gives evidence of a thermally inefficient pattern based on the temperature response through the system. This system is seasonally variable perennial spring that responds rapidly to any significant storm event based et al. (201 1 ) publication. Therefore, the conduit system may be classified as thermally ineffective with event scale variability. There are values that define the shape and response and are important defining parameters. A method for estimating the conduit hydraulic diameter (D H ) of a pore or conduit having a noncircular opening may be made (Figure 32) An ellipse provided the best fit within the opening, and the length of the short axis (a) is measured. The value for the variable, a, in the definition of a characteristic length during the March 20 event at Magic Springs overflow is 1.18 m which is the minimum diameter or depth of the water in the conduit at the storm half flow period. For this research: 1.66 m 8 using a cross section of the conduit two m upstream from the discharge over the weir Figure 15 provides the concept of the discharge from the conduit. The conduit area and the perimeter may be calculated using the conduit cross section and c alculations of characteristic length ( L ) using Figure 32 are based on an example from a storm event on March 20 : 9 where A is the area and P is the perimeter in Figure 15 S hape and Characteristics of Response Using Discharge The shape of the discharge response curve will identify the dynamics and timing of the discharge flow. Ford & Williams (2007) observed that widespread recharge from a precipitation

PAGE 46

35 event over a karst basin results in discharge at the springs exhibiting : 1) a lag time before response occurs 2) 3 ) recession as spring discha rge returns towards its pre sto on either limb although best seen on the recession. When the hydrograph is at its peak, storage in the karst system is at its maximum, and after a long pe riod of recession storage is at its minimum. The rate of withdrawal of water from storage, from the springs or from pumping ; is indicated by the slope of the subsequent recession curve ( ). The discharge of a spring is a function of the volume of water h eld in storage and that the half flow period (t 0.5 ) is the time for the maximum flow to half when returning to baseflow. If e = then the flow recession has the relationship: 10 where is the recession constant A t 0.5 is defined as the time required for the maximum discharge. S ubstituti on into equation 1 0 gives: 1 1 and 1 2 The parameter t 0.5 has the following properties: 1) it is independent of Q 0 and Q t and of the time elapsed between them, 2) it is sensitive to change and can take value s in the range zero to infinity, 3) it can be calculated using equation 1 2 and is simply related to 4) it is a direct measure of the rate of recession and therefore ca n be used as a means of characterizing exponential baseflow recession (Ford & Williams, 2007).

PAGE 47

36 There is a linear relation between hydraulic head and flow rate (commonly found in karst at baseflow conditions), and the curve can be expressed as a straight l ine with slope if plotted as a semi logarithmic graph. It can be represented in logarithmic form from which may be determined from : 13 shown in Table 2 Each of the storm response characteristics are given in Table 2 defining the complete storm event. For each storm event in Table 2, a rising response may be calculated for the rising limb, and the recession flow response at Magic Springs defined by the ha lf flow period time (t 0.5 ), rising time, recession curve, and baseflow (Figure 3 3 a ) The rising rate defines the initial dynamic response to storm event and is invariant with values around 6 hours with no strong correlation to discharge while the t 0.5 ranges between 1 2.9 to 15 .7 hours The storm events are superimposed and plotted for comparison Figure 3 3 b The sum of the rise time and half flow time is less than one day. Figure 3 3 shows the results of the hydrodynamics and the relationship with maximum discharge Figure 34a schematically shows the rising time, t 0.5 and a. All of the storm flow responses have been normalized and overlain in Figure 3 3 b A miniTROLL SN 20177 was used for February 18. The other events were recorded using the internal transducer in the TROLL 9000 SN 33494 Figure 35 summarizes the characterization of the rise time particula r to this spring conduit system, and the dynamics during half flow period time. The half flow period time vary with the maximum flow. T he recession curve defines head loss and hydrauli c conductivity with depletion back to baseflow within 3.8 to 7 .5 days (Figure 14) . Therefore, the analysis shows the rapid response to storm events which gives

PAGE 48

37 credence to the previous claim that the M agic Springs and the CM Cave conduit is a Pattern 1 thermal system (Luhmann et al., 2011). Discussion on Dye Tracing and Groundwater Drainage FDA approved non toxic dyes were used and a better understand the flow velocities and the flow paths was obtained. The flow paths further define the groundwater drainage and the groundwater drainage and the surface drainage are not the same. The surface drainage from Cypress Creek basin is being partially or completely pirated by the groundwater drainage which is discharged at Magic Springs. This experiment requires redefinition of the groundwater drainage although further dye tracing during both base f low Another note is that there are no springs in the Cypress Creek surface drainage. There is one spring further downstream from Magic Springs but access has bee n problematic because of property ownership and development and is in the Spri ng Branch Creek drainage basin. Out of the 4 tracers, two crossed surface water drainage basins. The No La Vi Cave which is closest, is part of the pulse I recharge path as is C ave Crack which is around 200 m downdip Both caves are well within the Cypress Creek drainage. Therefore, both caves have given tracer data and gives evidence that Cypress Creek is pirated and that the groundwater drainage system does not solely rely on Spring Branch Creek. Neighboring Cypress Creek has no notable springs and flow is strictly runoff. The tracing results give evidence that this conduit cave system in the Comal and Kendall county area is joint controlled dendritic pattern with a network m aze overprint bearing approximately 45 o and 315 o azimuth in the subsurface All of these karst surface features bear approximately 300 o

PAGE 49

38 Two of the dye detections were made using an autosampler and two of the dyes only showed on the charcoal packets. The dyes detected with the autosampler show interconnection, apparent velocity, and % recovered at the Magic Springs overflow. The charcoal packets are more sensitive but only tell if the dye was present. Modification of the methodology that supplies recharg e must be made based on injection sites and stream paths. There were four dye injection points of which two were in the Cypress Creek surface drainage basin requiring adjustment of the groundwater drainage. The sinking Spring Branch Creek significantly i mpacts the surface drainage directly to the conduit system and adds to the modification shown in Figure 3 1 Therefore, the groundwater basin grew to the east within the pirated Cypress Creek. The groundwater drainage was greatly increased on the west by expansion because Spring Branch Creek is a sinking stream that directs flow into the CM Cave conduit system. In Figure 31, f our cave locations were chosen as dye injection sites with monitoring at the spring discharge. Two locations were in the Spring Bra nch drainage and two locations in the Cypress Creek surface drainage. This discontinuity required to refine the groundwater drainage basin discharging at Magic Springs. The Spring Branch surface drainage basin includes part but not all of the groundwater drainage basin. C hemistry A comparison can be made by reviewing the data from the closest well to the research site. The TWDB Well number 6805502 was 4,140 m north of Magic Springs and the results are from 1966. A Piper diagram (Figure 3 5 ) was constructed for the 1 6 wells located within the Spring Branch Creek, Cypress Creek, or Curry Creek within the Spring Branch Quadrangle T he

PAGE 50

39 three water samples were retrieved from Magic Springs. The Piper diagram confirms the close relationship betw een the closest well and the spring sample analyses. There were three water samples collected at the Magic Springs. MS 120822 and MS121208 were taken during baseflow conditions while MS 120915 while receiving diffuse flow. The water level rose by 6 cm wh ich may be compared against the water levels shown in Figure 15. For this sample, a six cm rise at the discharge during a storm event. These analyses in Table 7 provide calculated total dissolved solids and a charge balance and then compared against the 26 wells drawing from the Lower Glen Rose. Then the data may be compared against the closest well which i s 4140 m north of the spring. The three samples taken during the research period are not very different regardless of the source The three samples taken from Magic Springs have been chemically analyzed and are plotted in the Figure 3 5 Piper diagram. Aquifer Volume and Mass Balance The study site and drainage is illustrated in Figure 3 1 This figure shows the conduit (red), the recharge points and flow lines (blue), monitoring sites, and the dye injection locations. These flow lines identify three sources of recharge all discharging at Magic Springs and possibly the spring mile downs tream. The sinking stream and the second high density karst area are approximately the same distance above CM Cave entrance; both contribute to pulse II recharge. This analysis requires knowledge concerning flow volumes, areas, and concentration results. The results from the tracing tests, the volumes, and areas have been calculated. Tracing studies used in the determination of subsurface flow conditions in karst terranes are greatly influenced by various subsurface flow patterns in combination located be tween the inflow and outflow points of the aquifer. Seven types of karst patterns are known to exist (Figure 21 ).

PAGE 51

40 In review, the mean tracer residence time T t may be estimated from: 14 by using the method of moments (Fields and Nash, 1997). A rough estimate of the maximum volume of the karst conduit can be assessed. Tracer mass recovery at a spring where discharge was measured during the tracer study allows for a rough estimate of the maximum conduit volume. This is achieved by: 15 If a single discharge value is used as a mean spring discharge then the karst conduit volume may be estimated by 16 where Q is mean spring discharge (L 3 T 1 ) and V is volume of individual karst conduits (L 3 ). If the floods are high, then more of the passage is phreatic so the data was taken from the March 20 storm event. The maximum discharge was 70 m 3 /min (41 ft 3 /s) and the half flow period was 15.2 hours. The volume has been estimated for the March 20 storm ev e nt by integrating the flow rate [ Q(t) ] for each time period. The values may be substituted in equation 16 : Using a passive tracer, the concentrations may be determined. For the March 20 event, 33.6% was recover ed at the Magic Springs overflow. There are losses which must be assessed since dye is adsorbed or absorbed. These experiments are not designed for detailed analysis although a maximum estimate of conduit volume during the March 20 event may be calculated : 17 No estimate has been made on loss of the tracing dyes.

PAGE 52

41 Since phreatic and vadose flow has been characterized the volume discharged was not at a limiting maximum In support there is a restricted sump 668 m upstream from the discharge below which the flow is vadose. Above the water has been verified multiple feet thick vertical coverage in CM Cave where phreatic flow conditions exist for the next 3100 m. The actual conduit volume will be higher. The calculations for the other two storm events may be reviewed in Table 8 where the volumes have been calculated for three storm events. The estimates for the calculation of total volume have been made with consideration that 33.6 % of the dye was recovered The output is a Type IIIb pattern for the CM Cave conduit at Magic Springs (Figure 21) since discharge occurs from three locations This will result in reduced tracer recoveries and unequal discharges because of dilution and di version to charcoal packets or unmonitored discharge points. The weir where all flow discharge characteristics were made is monitoring an overflow state with a small discharge spring within 10 m from the weir and similar discharge where the baseflow will persist as the aquifer drains. Since there are three known springs near the conduit discharge, this configuration may be similar to Type IIIb. Consider that there is an additional spring within mile from the Magic Springs that may be applicable to rec overy of dyes during the tracing experiment. A Type IVa may simulate the input, but Type IIIb will be used in this evaluation. Access problems would not allow samples to be taken at remote spring. Additional downgradient discharge locations must be consi dered concerning the quantity of dye recovered. The maximum amount dye recovered was 33.6% and occurred when the dyes were released directly into the conduit.

PAGE 53

42 The evapotranspiration (ET) value may be estimated from research by Banta and Slattery (2011). The ET value for the March 20 event is 34% of the averaged maximum of the ET response, or is 1.7 mm/d. This study used specific conductance as a passive tracer during storm events to estimate the values for baseflow and recharge. All mixing occurs in the recharge, not affecting the value or concentration for baseflow. Recharge includes the conduit flow and the diffuse flow. The values for the flow and the event time are given in Table 8 so that calculation might be made for the two March storm events From the data in Table 8 estimates may be made for flow discharged from the other three springs. A Case of Antecedent Moisture Content Veni (1994) discussed standard concepts where an aquifer's response to recharge changes according to antecedent rainfall. The effect of precipitation i s proportional to the volume and is inversely proportional to the time since it fell. The greater the antecedent rain, the greater will be the discharge following the subsequent rain. Studi es show that a s time increases between antecedent and current rain events the level of discharge will decrease logarithmically (Linsley et al., 1982). Around the karst, the soil conditions are thin. The response to storm events correlates with antecede nt moisture remaining after ET Across the high ground the coverage is consistently rocky with some soil. No i ncrease in discharge may occur i f either the antecedent or subsequent rain was insufficient. Also a decrease in discharge may be expected when t he time between the antecedent and subsequent rain was too great. In this latter case, most of the water will be absorbed by plants and soils, with little to no

PAGE 54

43 recharge for the local aquifer. Time and lack of rain both conspire to reduce the discharge a s summer approaches. Banta & Slattery (2011) evaluated the local evapotranspiration (ET) experienced in the Guadalupe River basin during their research on the hydrologic effects of ashe juniper ( Juniperus ashei ) removal as a brush management conservation practice. An assessment of the evaporation and transpiration (ET) was made at Honey Creek State Natural Area by Banta and Slattery (2011). The Honey Creek location is 4.3 km to the southwest of Magic Springs, located at the base of the Lower Glen Rose Me mber, and has similar karst features. The hydrograph that may be observed in Figure 3 6 illustrates two different results to storm events in May which is depend ent on the saturation of the soil The two storm event was recorded in May when 78% of the max imum change in ET had taken place (Figure 3 8 ). The time between storm events was 47 days. Figure 3 7 shows the response during March when only 22% above the minimum on the 9 th or 34% in ET on the 20 th The ET value may be estimated from research by Banta and Slattery (2011). The ET value for the March 20 event is 1.7 mm/d. The storm events recorded in March begin to show the effect of ET in the water balance. The primary understanding is the percent drop in SpC and the amount of rainfall for the March 9 and March 20 event. The last recorded storm event gives the best example of how dominant ET becomes late in the spring and through the summer. The rain incurred on May 6 was more intense with higher rain rates than in the previous storms and there was no response. After the ground was saturated throughout the next week, the next downpour gave immediate results. This delay is because of ET

PAGE 55

44 A comparison of the reaction from three storm events reveals the driving force. The rainfall totals for each storm event may be reviewed in Table 3 along with the energy imparted based on maximum discharge There was a reduction in SpC w hich is also noted in Figure 3 7 The discharge gives the energy of the storm The March events responded with 26.22 ft 3 /s on the 9 th and 41.17 ft 3 /s on the 20 th A higher energy storm (March 20) sees a lower change in SpC which would be opposite to what one would expect until ET is considered. On May 6, t here was a significant storm event wi th no response in SpC the last rai nstorm occurred 47 days ago and the logarithmic decline in antecedent moisture was dominant. For the case given on May 6, spring was in full bloom, the ground was dry as time passed, and t he antecedent moisture content was very low The response which mi ght have been discharged was either supplying the plants or evaporated in the May heat not recharging the aquifer. R ain occurred daily a fter May 6 continuously soaking the soil until the May 10 storm occurred On the 10 th a total of 1.22 inches of rain a ccumulated over 1.91 hours with t he maximum rate of 0.21 inches/5 minutes. Considering this rainfall event is equivalent to March 20, the maximum discharge on May 10 is surprisingly less in significance. Since the soil was saturated on the 10 th the resp onse in SpC was equivalent to both the storm event s o n March 9 and March 20 all because of antecedent moisture content from previous storms Therefore, the impact of ET increases from 1 .0 to 3.0 mm/d locally and impacts the influence of dilute water During May, 78% of the increase is observed and the ET is 2.5 5 mm/d giving credence to no response because of ET following a major storm event.

PAGE 56

45 CHAPTER S IX : CONCLUSIONS The conclusions reached for this research describe the velocities, flowpaths, and physical characteristics that provide a rapid response for Magic Springs conduit system. The physical controls, thermodynamics, tracing results, groundwater drainage delinea tion, mass balance, and hydrodynamic response are defined The Magic Springs conduit spring system has jointed passages which combin e into tributaries. Although the dendritic pattern is in jointed rock, an overprint with a dendritic or network maze occurs in the two high density karst areas. An anomaly was identified in the SpC hydrograph response. The response curve does not match the concentration curve indicating a path from another source for the recharge flow. Th e response correlates with a t wo pulse recharge event whose source is the 2 nd high density karst area and a sinking stream. Two high density karst areas and a sinking stream are terranes providing different conduit flow paths for the recharge event. Discharge provides coefficients of determination for the timing between the recharge and the springs, rising discharge and recession half flow time periods. Therefore, the time difference between the recharge and the discharge is dependent on average discharge The prediction of th e second pulse has been correlated to the shape of the SpC response allowing calculation of different conduit pathways for pulse II. Once the pathways are known, the volumes were calculated and the ratio b etween the pulses was calculated This ratio corr elate s best with the maximum discharge and shows the possibility of stag e height increasing above the sinking stream, dominating pulse II as discharge increases This conduit and spring system flow responds from immediate recharge through the first pulse and defines the thermodynamic reaction. Nu fell above the laminar and transitional

PAGE 57

46 thermal region near the base of the turbulent region. St is small (St <<<1) by orders of magnitude and the exchanged heat is ineffective and the thermal sign ature will move through the conduit to the spring Therefore, t he conduit system may be classified as thermally ineffective with event scale variability based on the requiremen ts given by Luhmann et al. (2011). This Magic Springs CM Cave conduit system has a rise time between 6 to 6.5 hours and has a half flow period time (t 0.5 ) between 1 2.9 and 15 .7 hours. The value for discharge to half period time ( t 0.5 ) is a direct measure of the rate of recession and therefore can be used as a means of characterizing exponential baseflow recession (Ford & Williams, 2007). The total time from storm event to t 0.5 is less than to one day N eighboring Cypress Creek has no notable sprin gs and flow is strictly runoff. Both No La Vi Cave and Cave Crack tracer data give s evidence that Cypress Creek is pirated and that the groundwater drainage system does not solely rel y on Spring Branch Creek. Therefore, the groundwater drainage basin mus t be redefined. The conduit system thermal forcing may be classified as thermally inefficient with 85% temperature retention over 1300 m following a storm event. T he dye tracing results identify surface creek piracy and requires redrawing the groundwat er drainage basin. Even during baseflow conditions (< 2 ft 3 /s or 3.4 m 3 /min) the measured groundwater velocities approached 3000 m/d and crossed surface water drainage basins. The spring discharge has total dissolved solids around 350 mg/L and is c hemically stable with an equivalence point for pH of 4.5. The discharge at Magic Springs responds rapidly to any significant storm event (below 15 ft 3 /s or 25 m 3 /min) fed by a minimum of three input flows for a recharge event. The timing,

PAGE 58

47 dynamics, and ch emistry require attention be paid to these local recharge points and the location for the inflow that influences the recharge using best management practices.

PAGE 59

48 CHAPTER SEVEN : RECOMMENDATIONS An understanding needs to be developed concerning t he northern section of the surveyed passage exhibits v adose joint controlled flow whereas p hreatic conditions exist 1300 m downstream to the springs. The juncture of the vadose and phreatic flow conditions occurs near survey marker MM 5/ AA 1 Vadose flow is drawn dow nward by gravity along joints down dip. When this vadose flow reaches the water table of the ancestral Guadalupe River, there is no tendency to continue down dip. Instead the flow found the fastest path to discharge, downdip or strike. The entrance to C M cave is located 92 m further downstream from survey marker, and then flow reaches the rapids 340 m toward the springs Observations have been made to the rapids 340 m beyond CM Cave. One hypothesis is that these flow conditions exist because this must have been a paleoenvironment with a water table at this juncture so that a transition between vadose and phreatic flow occurs. The second hypothesis is that this change in flow characteristics occurs within the Pulse I high density karst recharge area. This change in flow conditions must be explored as either hypotheses are viable This data would also provide information on understanding the paleo history of the flow patterns in this karst system. Investigate the phreatic and vadose passages within CM Cave. There must be some conclusion as to why these different flow conditions exist as CM Cave must ha ve formed under different hydraulic conditions. The upper section of the surveyed passage exhibits vadose joint controlled sections. Phreatic conditio ns exist 1300 m downstream to the springs from the juncture near survey marker AA1 and MM5 at UTM Zone 14N 554276 330933 (Figure 22) The juncture is located within the high density karst area with 44 features per 0.1 6 km 2 where increased flow from karst d ensity would have occurred during storm events. The vadose to

PAGE 60

49 phreatic zones indicate a change in flow patterns. Attention should be paid to Veni (1994) where discussion on the preferential paths and erosion associated with the Guadalupe River basin. E valuate the connectivity between the sinking stream at Cool Creek Cave to the infeeder 100 m away. Data needs to be collected using tracer results supporting the connectivity hypothesis A better understanding of the flow characteristics is needed Estab lish an activity to monitor or instrument the Cool Creek Cave and Magic Springs during a storm event so as to establish the actual contribution of this sinking stream input to the discharge shape. Dye tracing under baseflow conditions provided unusual resu lts but additional tracer testing is needed under storm events with the same 11 monitoring stations. Evidence has that the conduit possibly infiltrates other surface drainage systems during flood events which should be detected using these monitoring stat ions. Research should be designed to specifically identify the flow patterns of the recharge event between the sinking stream and the 2 nd dense area. Much of this research could be made by simple observation of the sinking stream at the Henderson Ranch. A recommendation is that the property owners establish the best management practices. Magic Springs has a unique geographic situation where the spring is in Comal County but the recharge sources are in Kendall County. First recommended would be a meeting regarding this water source. Additional conduit passage has been discovered in March 2013 of passages west of Cool Creek Cave. additional recharge sites and possibly addition al sinking streams. Research should be continued

PAGE 61

50 on this spring conduit system since the groundwater drainage basin will continue to change as new discoveries are uncovered

PAGE 62

51 APPENDIX A

PAGE 63

52 APPENDIX A

PAGE 64

53 APPENDIX A

PAGE 65

54 APPENDIX A

PAGE 66

55 APPENDIX A

PAGE 67

56 APPENDIX A

PAGE 68

57 APPENDIX B

PAGE 69

58 APPENDIX B

PAGE 70

59 APPENDIX B

PAGE 71

60 APPENDIX B

PAGE 72

61 APPENDIX B

PAGE 73

62 APPENDIX B

PAGE 74

63 APPENDIX B

PAGE 75

64 APPENDIX B

PAGE 76

65 APPENDIX B

PAGE 77

66 APPENDIX C

PAGE 78

67 APPENDIX C

PAGE 79

68 APPENDIX C

PAGE 80

69 APPENDIX C

PAGE 81

70 APPENDIX C

PAGE 82

71 APPENDIX D

PAGE 83

72 APPENDIX D

PAGE 84

73 APPENDIX D

PAGE 85

74 APPENDIX D

PAGE 86

75 APPENDIX D

PAGE 87

76 APPENDIX D

PAGE 88

77 APPENDIX D

PAGE 89

78 APPENDIX D

PAGE 90

79 APPENDIX D

PAGE 91

80 APPENDIX D

PAGE 92

81 APPENDIX D

PAGE 93

82 APPENDIX D

PAGE 94

83 APPENDIX D

PAGE 95

84 APPENDIX D

PAGE 96

85 APPENDIX D

PAGE 97

86 APPENDIX D

PAGE 98

87 APPENDIX D

PAGE 99

88 APPENDIX D

PAGE 100

89 APPENDIX D

PAGE 101

90 APPENDIX D

PAGE 102

91 APPENDIX D

PAGE 103

92 APPENDIX D

PAGE 104

93 APPENDIX D

PAGE 105

94 APPENDIX D

PAGE 106

95 APPENDIX D

PAGE 107

96 APPENDIX D

PAGE 108

97 APPENDIX D

PAGE 109

98 APPENDIX D

PAGE 110

99 APPENDIX D

PAGE 111

100 APPENDIX D

PAGE 112

101 REFERENCES Atkinson, T. C. (1977). Diffuse flow and conduit flow in limestone terrain in the Mendip Hills, Somerset (Great Britain). Journal of Hydrology, 35 (1 2), 93 110. Baker, E. T., Jr. (1995). Stratigraphic Nomenclature and Geologic Sections of the G ulf Coastal Plain of Texas. Open File Report U. S. Geological Survey, 94 161 34. Banta, J. R., & Slattery, R. N. (2011) Effects of brush management on the hydrologic budget and water quality in and adjacent to Honey Creek State Natural Area, Comal County, Texas, 2001 10. Scientific Investigations Report 2011 5226, 35. Barker, R. A., & Ardis, A. F. (1 996). Hydrogeologic framework of the Edwards Trinity aquifer system, west central Texas Washington : U.S. G.P.O.: Denver, CO. Birk, S., Liedl, R., & Sauter, M. (2004). Identification of localised recharge and conduit flow by combined analysis of hydraulic and physico chemical spring responses (Urenbrunnen, SW Germany). Journal of Hydrology, 286 (1 4), 179 193. doi: doi.org/10. 1016/j.jhydrol.2003.09.007 Blakey, R. C. (2005). Paleogeography and geologic evolution of North America: Early Cretaceous -115 Ma: http://jan.ucc.nau.edu/~rcb7/namK115.jpg Retrieved October 16, 2012 Clark, A. K., Blome, C. D., Morris, R. R., and Golab, J A., 2013. The Geologic Map and Hydrostratigraphy of the Guadalupe River State Park and Honey Creek Natural Area, Kendall and Comal Counties, Texas. U. S. Geological Survey, Scientific Investigations Map XXXX. [1 sheet]

PAGE 113

102 Clark, Amy R., Blome Charles D., & Faith Jason R.. 2009. Map Showing Geology And Hydrostratigraphy Of The Edwards Aquifer Catchment Area, Northern Bexar County, South Central Texas. U. S. Geological Survey Open File Report 2009 1008. 24, 1 she et Collins, E. W. (Cartographer). 2000. Geolo gic Map of the New Braunfels, Texas, 30 x 60 Minute Quadrangle Miscellaneous Map No. 39. U. S. Geological Survey, [1 sheet]. Covington, M. D., Luhmann, A. J., Gabrovsek, F., Saar, M. O., & Wicks, C. M. (2011). Mechanisms of heat exchange between water an d rock in karst conduits. [Article]. Water Resources Research, 47 doi: W10514 10.1029/2011wr010683 Field, Malcolm S. & Nash, Stephen G. 1997. Risk assessment methodology for karst aquifers: (1) Estimating karst conduit flow parameters. Environmental Moni toring and Assessment 47: 21. Ford, D., & Williams, P. W. (2007). Karst Hydrogeology and Geomorphology Chichester, England: John Wiley & Sons. Gnielinski, V. (1976). New Equations For Heat and Mass Transfer in Turbulent Pipe and Channel Flow. [Article]. International Chemical Engineering, 16 (2), 359 368. Gunn, J. (1986). A conceptual model for conduit flow dominated karst aquifers. IAHS AISH Publication, 161 587 596. Hammond, W. (1984). Hydrogeology of the Lower Glen Rose Aquifer, South Central Texas (Groundwater, Water Resources, Isotope Hydrology. The University of Texas at Austin Ph.D., The University of Texas at Austin, United States -Texas. Johnson, S., Schindel, G., & Veni, G. (2010). Tracing groundwater flowpaths in the Edwards Aquifer recharg e zone, Panther Springs Creek basin, northern Bexar County, Texas. Bulletin of the South Texas Geological Society, 51(3), 15 46.

PAGE 114

103 Linsley, Ray K, Jr; Kohler, Max A ; Paulhus, Joseph L H. ; 1982; Hydrology for engineers; McGraw Hill Book Co, New York, NY, United States (USA); 1 508. Luhmann, A. J., Covington, M. D., Peters, A. J., Alexander, S. C., Anger, C. T., Green, J. A., . Alexander, E. C. (2011). Classification of Thermal Patterns at Karst Springs and Cave Stream s. Ground Water, 49 (3), 324 335. doi: 10.1111/j.1745 6584.2010.00737.x Mancini, E. A., & Puckett, T. M. (2005). Jurassic and Cretaceous Transgressive Regressive (T R) cycles, Northern Gulf of Mexico, USA. Stratigraphy, 2 (1), 31 48. Musgrove, M., Stern, L. A., & Banner, J. L. (2010). Spring water geochemistry at Honey Creek State Natural Area, central Texas; implications for surface water and ground water interaction in a karst aquifer. Journal of Hydrology, 388 (1 2), 144 156. doi: 10.1016/j.jhydrol.2010.04 .036 Nance, H. S. (2004). Hydrochemical variability in the Edwards Trinity aquifer system, Edwards Plateau, Texas. Report Texas Department of Water Resources, 63 89. Palmer, A. N. (1991). Origin and morphology of limestone caves. Geological Society of America Bulletin 103 (1), 1 21. doi: 10.1130/0016 7606 Palmer, A. N. (2007). Cave Geology (2007 ed.). Dayton, OH: Cave Books. Pape, J. R., Jay L. Banner, Lawrence E. Mack, MaryLynn Musgrove, Amber Guilfoyle. (2010). Controls on oxygen isotope variability in precipitation and cave drip waters, central Texas, USA. Journal of Hydrology, 385 (1 4), 203 215. Phelps, R. M., C. Kerans, and R. G. Loucks. (2010). High resolution regional sequence stratigraphic framework of Aptian through Coniacian strata in the Coman che shelf. Central and South Texas: Gulf Coast Association of Geological Societies Transactions Gulf Coast Association of Geological Societies, 60 755 758.

PAGE 115

104 Radtke, D. B., Davis, J. V., & Wilde, F. D. (2006). Techniques for Water Resources Investigation s, Book 9; Specific Electrical Conductance. National Field Manual for the Collection of Water Quality Data; U.S. Geological Survey A6 (6.3), 22. Raeisi, E., Groves, C., & Meiman, J. (2007). Effects of partial and full pipe flow on hydrochemographs of Logsdon river, Mammoth Cave Kentucky USA. Journal of Hydrology, 337 (1 2), 1 10. doi: 10.1016/j.jhydrol.2006.11.015 Rantz, S. E. (1982). Measurement and Computation of Streamflow Volume 1. Measurement of Stage and Discharge. U.S. Geological Survey, Wate r Supply Paper 2175 284. Rambo, Bill. 1990. Eisenhauer Ranch Caves. Texas Caver 35 (5) October 1990, 96 98. Ryan, M., and J. Meiman. 1996. An examination of short term variations in water quality at a karst spring in Kentucky. Ground Water 34, no. 1: 2 3 30. Veni, G. (1994). Geomorphology, hydrogeology, geochemistry, and evolution of the karstic Lower Glen Rose aquifer, south central Texas. Ph.D., The Pennsylvania State University, United States -Pennsylvania. Wark, Kenneth. 1977. Thermodynamics Ne w York: McGraw Hill Book Company, 3 rd edition, 1 909. White, William B., 2002. Karst hydrology: recent development and open questions. Engineering Geology, v65 (2002), 85 105 Wojdyr, M. (2010). Fityk: A General purpose Peak Fitting Program. Journal of Applied Crystallography, 43 (5 1), 1126 1128. Worthington, S. R. H., & Smart, C. C. (2003). Empirical Determination of Tracer Mass For Sink To Spring Tests in Karst. Sinkholes and the Engineering and Environmental Impacts on Karst, Geotechnical Special Publ ication, 122

PAGE 116

VITA Mark Childre was born in San Antonio on March 7, 1959 and has worked in power generation, aircraft manufacturing, aquifer protection, and teaching professions He received his Bachelor of Science in mechanical engin eering from Texas A & M University while work ing as a student engineer at Texas Utilities Services. Following graduation, Mark accepted an offer from General Dynamics Fort Worth (Lockheed Martin Aeronautics) where he t ook a position in the F 16 Propulsion Analysis Group. He expanded his horizons by assisting in the design and development but leading the integration of propulsion controls o n a vectored thrust aircraft. He finished his work with Lockheed Martin while working on the development of the F 35 Lightning at Skunkworks in Palmdale, California. Mark became fascinated with geology while starting an automotive repair business in Colorado. Once return ing to San Antonio, he enrolled with The University of Texas at San Antonio studying geology with a specialty in karst hydrogeology. During his studies at the UTSA, he was granted a scholarship in the Minority Opportunities Research Experience Science Program where he studied the complexities of the Edwards Aquifer and mapped t he stratigraphy of the Kainer Formation The culmination of his work would be his thesis topic where he characterize d a spring conduit system, redefine d the groundwater drainage, and assess ed recharge to this groundwater conduit spring system as he receiv es his Master of Science degree


printinsert_linkshareget_appmore_horiz

Download Options

close
Choose Size
Choose file type
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.