USF Libraries
USF Digital Collections

An evaluation of conduit conceptualizations and model performance

MISSING IMAGE

Material Information

Title:
An evaluation of conduit conceptualizations and model performance
Physical Description:
Book
Language:
English
Creator:
Hill, Melissa Estelle
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Dual-permeability
Conduit-matrix fluid exchange
Non-Darcian flow
Karst hydrogeology
Model parameters
Dissertations, Academic -- Geology -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: The karst research community has known that traditional numerical groundwater flow codes ignore the non-Darcian, dual-permeability components of flow that can occur in karst aquifers. In this study, the potential limitations of using such tools are quantified by evaluating the relative performances of 3 groundwater flow models at a test-site near Weeki Wachee, Florida, in the dual-permeability Upper Floridan aquifer. MODFLOW-2005 and MODFLOW-2005 Conduit Flow Process (CFP), a Darcian/non-Darcian, dual-permeability groundwater flow code recently developed by the U.S. Geological Survey, are used in this study. A monitoring program consisting of discharge measurements and high frequency data from 2 springs and monitoring wells penetrating the matrix and conduit networks of a karst aquifer was initiated to characterize the test-site and constrain new parameters introduced with MODFLOW-2005 CFP.The monitoring program spanned conditions prior to, during, and following convective and tropical storm activity, and a drought. Analytical estimates for Reynolds numbers, ranging from 10⁵ to 10⁶, suggest that turbulent flow occurs in portions of the underlying conduit network. The direction and magnitude of fluid exchange observed between the matrix and conduit network indicate the conduit network underlying the test-site drains the matrix. Head differences and observed responses in monitoring wells penetrating the matrix and conduit network indicate that the hydraulic conductivities between the 2 networks do not significantly differ from each other. A conceptual model for the spatial distribution of preferential flow pathways using multiple data types, including shallow recession limbs observed in discharge hydrographs indicate a slow responding aquifer with a high storage capacity, and a poorly integrated conduit drainage network with little to no point recharge.Model performances were evaluated by comparing observed hydrographs for discharge and monitoring wells penetrating the matrix and conduit network following convective and tropical storm events, and drought conditions, to simulated values from transient simulations. Model statistics for 32 target wells and sensitivity analysis were included in the evaluation. The dual-permeability model using the MODFLOW-2005 CFP Mode 1 displayed the highest performance with improved matches ranging from 12 to 40% between simulated and observed discharges relative to the laminar and laminar/turbulent equivalent-continuum models.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2008.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Melissa Estelle Hill.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 230 pages.
General Note:
Includes vita.

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:
aleph - 002000477
oclc - 318990617
usfldc doi - E14-SFE0002497
usfldc handle - e14.2497
System ID:
SFS0026814:00001


This item is only available as the following downloads:


Full Text

PAGE 1

An Evaluation of Conduit Conceptua lizations and Model Performance by Melissa Estelle Hill A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy Department of Geology College of Arts and Sciences University of South Florida Major Professor: Ma rk Stewart, Ph.D. Robert Brinkmann, Ph.D. Eric Oches, Ph.D. John Mylroie, Ph.D. Ronald Green, Ph.D. Date of Approval: April 8, 2008 Keywords: dual-permeability, conduit-matrix fluid exchange, non-Darcian flow, karst hydrogeology, model parameters Copyright 2008, Melissa Estelle Hill

PAGE 2

Para mis padres, mi hermano, y Mikko

PAGE 3

ACKNOWLEDGEMENTS The author graciously thanks the generous contributions from several individuals that supported this research by providing code, cave survey data, services, equipment, or software often without compensation. Notably, W. Barclay Shoemaker from the U.S. G eological Survey is acknowledged for providing the MODFLOW-2005 Conduit Flow Process. Jeff Petersen and Brett Hemphill from Karst Underwater Research Inc. provided cave survey data and Mike and Sandra Poucher assisted with the radiolocation effort. Brian Pease is also acknowledged for providing equipment and helpful advice that assisted with the radiolocation effort and James O. Rumbaugh, III is acknowledged for providing software. The author thanks A ngel Martin, Mark Barcelo, and Dave DeWitt from the Southwest Florida Wa ter Management District and Lee Florea from the U.S. Geological Survey for providing editorial review on several professional papers produced from this research. Margaret Gilchrist, Dave, DeWitt, Jason LaRoche, Mike Kelley, and Kevin Stover from the Southwest Florida Water Management District are acknowledged for assisting with the collection of field data and estimating water use quantities. The author thanks Mark Stewart, Robert Brinkmann, Eric Oches, John Mylroie, and Ronald Green for fruitful discussions and technical review of this manuscript. The author graciously thanks her family for thei r continuous encouragement throughout this project. The field data collected during this multiyear research project was partly funded by the Southwest Florida Wate r Management District. As such, processed field data and interpretations made by the author that were released to the public prior to the completion of this written dissertation may be published elsewhere without proper attribution to the author of this manuscript.

PAGE 4

i TABLE OF CONTENTS LIST OF FIGURES vi LIST OF TABLES x ABSTRACT xi CHAPTER 1: INTRODUCTION 1 1.0. Literature Review 3 1.0.1. Traditional Methods for Interpreting Conduit Locations or Preferential Flow Pathways 3 1.0.1.1. Borehole Observations 3 1.0.1.2. Fracture Traces 4 1.0.1.3. Tracer Tests 5 1.0.1.4. Troughs/Mounds in the Potentiometric Surface 6 1.0.1.5. Spring Hydrographs 7 1.0.1.6. Cave Surveys 7 1.0.1.7. Geophysical methods 8 1.0.1.8. Assessment of Karst Features 9 1.0.1.9. Aquifer Performance Tests 10 1.0.1.10. Statistical Methods 10 1.0.1.10.1. Fractals 11 1.0.1.10.2. Semivariogram Cloud Analysis 11 1.0.1.10.3. Poisson and Bayesian Statistics 12 1.0.1.10.4. Reinforced Random Walk 12 1.0.2. Numerical Groundwater Flow Model Types 13 1.0.2.1. Equivalent-Continuum Models 13

PAGE 5

ii 1.0.2.2. Discrete-Fracture Network Models 13 1.0.2.3. Dual-Conductivity Models 14 1.0.2.4. Dual-Cont inua Models 14 1.0.3. Previous Applications of Numerical Models in Karst Aquifer Settings 15 1.0.3.1. Equivalent-Continuum Models 15 1.0.3.2. Discrete-Fracture Models 16 1.0.3.3. Dual-c onductivity/permeability /porosity models 17 1.0.3.4. Dual-continua Models 19 1.0.3.5. Hybrid In tegrated Models 20 1.1. Project Objectives 23 1.2. Overview 23 CHAPTER 2: APPROACH AND METHODOLOGIES 25 2.0. Introduction 25 2.1. Task 2 Site selection 25 2.2. Task 3 Determine if Non-Darcian Flow Occurs in the Conduit Network and Estimate Reynolds Numbers 26 2.3. Task 4 Develop a Concept ual Model for the Conduit Network 28 2.4. Task 5 Characterization of the Hydraulic Response and the Direction and Magnit ude of Fluid Exchange Between the Matrix and C onduit Networks 31 2.5. Task 6 Development of t he 3 Numerical Groundwater Flow Models 38 2.6. Task 7 Sensitivity Analysis 41 2.7. Task 8 Evaluation of t he Model Performances 42 CHAPTER 3: CONCEPTUAL MODEL OF STUDY AREA 43 3.0. Background 43 3.1. Hydrogeologic Setting of Study Area 46

PAGE 6

iii 3.1.1. Physiographic Regions 47 3.1.2. Soils 48 3.1.3. Land Use 49 3.1.4. Water Use 49 3.1.5. Karst 50 3.1.6. Water Quality 53 3.1.7. Stratigraphy 54 3.1.7.1. Undifferentiated Sands 54 3.1.7.2. Hawthorn Group 54 3.1.7.3. Suwannee Limestone 55 3.1.7.4. Ocala Limestone 55 3.1.7.5. Avon Park Formation 55 3.1.8. Hydrogeologic Framework 56 3.1.9. Previous Estimates of Transmissivity and Storativity Ba sed on Aquifer Performance Tests and Flow Net Analysis 60 3.1.10. Tracer Tests 62 3.1.11. Recharge 64 3.2. Does Non-Darcian Flow O ccur in the Underlying Conduit Network? 66 3.3. Conduit Conceptualizations 68 3.3.1. Fracture Traces 68 3.3.2. Troughs in Aquifer Water Levels 69 3.3.3. Borehole Porosity Descriptions 72 3.3.4. Cave Survey Data 74 3.4. Characterization of Aquifer Response and Fluid Exchange Between the Matrix and Conduit Networks 80 3.4.1. Monitoring Wells 81 3.4.2. Well Hydrographs 83 3.4.3. Spring Hydrograph 89

PAGE 7

iv 3.5. Discussion and Proposed Conceptual Model 92 CHAPTER 4: NUMERICAL ANALYSI S 103 4.0. Introduction 103 4.1. Model Calibration 103 4.2. Grid 106 4.3. Boundary Conditions 107 4.4. Estimates of Gr oundwater Withdrawals 108 4.5. Net Recharge 115 4.6. Discharge 116 4.7. Pool Stages 118 4.8. Spring Conductance Coeffi cients 119 4.9. Water Budget 124 4.10. Range of Hydraulic Conductivity Values 125 4.11. Equivalent-Continuum Model ( Laminar Flow) 125 4.12. Equivalent-Continuum Model (Laminar/Turbulent Flow) 126 4.13. Dual-Conductivity Model (Laminar /Turbulent Flow) 128 4.14. Sensitivity Analysis 131 4.14.1. Boundary Conditions 131 4.14.2. Model Parameters 133 4.15. Results of Sensitivity A nalysis 140 4.16. Model Performance Evaluati on Results 142 4.16.1. Spring Discharges 143 4.16.2. Wells 146 4.16.3. Simulated Matrix and Conduit Water Levels and Head Differences 148 4.16.4. Model Statistics 149 4.17. Discussion 150 CHAPTER 5: CONCLUDING REMARK S 156 5.0. Project Conclusions 156 5.1. Limitations/Recommended Paths Forward 159

PAGE 8

v 5.2. Contribution of This Res earch 160 REFERENCES CITED 162 APPENDICES 181 Appendix A 182 Appendix B 186 Appendix C 190 Appendix D 191 Appendix E 198 Appendix F 215 ABOUT THE AUTHOR End Page

PAGE 9

vi LIST OF FIGURES Figure 2-1. Location maps of te st-site with location for springs and monitoring wells 32 Figure 2-2. Transect locations used to measure discharge at Twin Ds Spring 37 Figure 3-1. Location and site map of study area 44 Figure 3-2. Previous interpretation of spring basin boundaries (from Jones et al., 1997) 45 Figure 3-3. Delineation of study area 46 Figure 3-4. Physiographic regions in the vicinity of study area 48 Figure 3-5. Soils in the study area 49 Figure 3-6. Water table surface map for the study area using May 2004 water levels 50 Figure 3-7. Location of karst f eatures in the study area 52 Figure 3-8. Interpretation of buri ed limestone surface at test-site based on kriging of 241 lithologic logs 53 Figure 3-9. Thickness of Upper Floridan aquifer (top) 59 Figure 3-10. Location of esti mated transmissivities in m2/d (top) and estimated storativitie s (dimensionless) 61 Figure 3-11. Location of Crescent Lake relative to Weeki Wachee Spring 63 Figure 3-12. Location of capped, recharge, and discharge zones with nested well sites 65 Figure 3-13. Plots of specific disch arge versus hydraulic gradient for Weeki Wachee Spring (top) and Twin Ds Spring 67 Figure 3-14. Inferred fracture traces from Jones et al. (1997) 69

PAGE 10

vii Figure 3-15. Potentiometric su rface maps for the Upper Floridan aquifer 71 Figure 3-16. Location of wells with reviewed lithologic logs (top) and location of wells intercepting a cavity 0.3 m in height (bottom) 73 Figure 3-17. Ratio of large (> 1.5 m in height) and small (< 1.5 m in height) cavi ties intercepted on the Brooksville Ridge (top) and North Gulf Coastal Lowlands and Coastal Swamps 74 Figure 3-18. Modes of elevation for intercepted cavities from borehole porosity descriptions and surveyed underwater caves 77 Figure 3-19. Projected conduit net works for Twin Ds and Weeki Wachee Springs on land surface with radiolocation verification sites 79 Figure 3-20. Barometric pressure recorder captures the decrease in pressure as Tropical Storms Frances and Jeanne pass over the study area 81 Figure 3-21. Albino crayfish in WW-F well 82 Figure 3-22. Shaded area represents the aerial extent of OneRain, Inc.15 min rainfall used with the well and spring hydrographs 83 Figure 3-23. Water-levels in matrix Weeki Wachee Deep (WW Deep) and WWSpg-ECK wells and the conduit (WW-F) well versus rainfall 84 Figure 3-24. Hydrographs fo r WWSpg-Eck (matrix) well and WW-F (conduit) well during events A through D 86 Figure 3-25. Head difference bet ween the WWSpg-ECK (matrix) well and the WW-F (conduit) well during storm events 87

PAGE 11

viii Figure 3-26. Distribution and rain fall quantities associated with events A through D 88 Figure 3-27. Hydrographs of hour ly water levels for the ROMP matrix wells versus rainfall 89 Figure 3-28. Hydrograph of monthly discharge at Twin D's Spring versus monthly rainfall 90 Figure 3-29. Rating curve for estimating discharge at Twin D's Spring based on average monthly discharge measurements and water levels at WW-F 91 Figure 3-30. Water-levels in the WW-F index well with Twin D's and Weeki Wachee Springs pool stages 92 Figure 3-31. Proposed conceptual model for preferential flow pathways in the Upper Floridan aquifer 100 Figure 4-1. Location of targets (monitoring wells in the Upper Floridan aquifer) used to calibrate groundwater flow models 104 Figure 4-2. Difference in daily water levels from September 30 through October 31, 2005 105 Figure 4-3. Lateral model boundari es 108 Figure 4-4. Non-domestic (top) and domestic (bottom) reported and estimated groundwater withdrawal quantities 113 Figure 4-5. Distribution of drains in model domain 117 Figure 4-6. Estimated spring conductance coefficients for Twin D's Spring (top) and Weeki Wachee Spring 123 Figure 4-7. Hydraulic conductivity values (ECM K) used in the laminar and laminar/tu rbulent equivalentcontinuum models (top) and in the dualconductivity (DCM K) models ( bottom) 127

PAGE 12

ix Figure 4-8. Location of model cells that contain conduit nodes in the dual-conductivity m odels 130 Figure 4-9. Net flux (m3/d) across steady-state, dual-conductivity model boundaries 133 Figure 4-10. Plots showing the e ffect of varying the storage coefficient and specific yield values in the mantle and Upper Floridan aquifer 138 Figure 4-11. Plots showing the e ffect of reducing conduit wall conductance and conduit diameter in the dualconductivity model 139 Figure 4-12. Location of turbulent flow in the combined steadystate/transient, laminar /turbulent equivalentcontinuum model for stress peri od 5 141 Figure 4-13. Plots of observed and simulated discharges for the laminar equivalent-continuum (ECM MODFLOW-2005), the dual-conductivity (DCM MODFLOW-2005 CFP Mode 1, the laminar equivalent-continuum model using the hydraulic conductivity array (K array) 145 Figure 4-14. Plots of observed and simulated water levels in the Weeki Wachee Deep (matrix well), WW-F (conduit well), and ROMP 98 (matrix well) 147 Figure 4-15. Plot of observed water levels for the Weeki Wachee Deep matrix well (WW Deep), WWSpgECK matrix well, and the WW-F c onduit well 149

PAGE 13

x LIST OF TABLES Table 1-1. Summary of previous investigations using numerical models in karst aquifer settings 22 Table 3-1. Hydrogeologic framework (modified from Miller, 1986) 57 Table 4-1. Reported and estimated 2002 water use for Hernando and Pasco Counties 109 Table 4-2. Groundwater withdraw als from the Upper Floridan aquifer for each stress period 114 Table 4-3. Estimated evapotranspiration rates (modified from HydroGeoLogic, Inc., 2007) 115 Table 4-4. Observed discharges for Twin D's and Weeki Wachee Springs 118 Table 4-5. Observed pool stages for Twin D's (TD) and Weeki Wachee (WW) Springs 119 Table 4-6. Estimated spring conductanc e coefficients 122 Table 4-7. Estimated spring conductance coefficients and drain conductances used in groundwater flow models 124 Table 4-8. Effect of varyi ng model boundary conditions on simulated discharge 132 Table 4-9. Effect of varying net recharge, hydraulic conductivity, well flow rate, general-head conductance, and general-head values 136 Table 4-10. Comparison of model statistics for the 32 target wells among the 3 groundwater fl ow models 150

PAGE 14

xi An Evaluation of Conduit Conceptua lizations and Model Performance Melissa Estelle Hill ABSTRACT The karst research community has known that traditional numerical groundwater flow codes ignore the non-Da rcian, dual-permeability components of flow that can occur in karst aquifers. In this study, the pot ential limitations of using such tools are quantified by eval uating the relative performances of 3 groundwater flow models at a test-site near Weeki Wachee, Florida, in the dualpermeability Upper Floridan aquifer. MODFLOW-2005 and MODFLOW-2005 Conduit Flow Process (CFP), a Da rcian/non-Darcian, dual-permeability groundwater flow code recently developed by the U.S. Geological Survey, are used in this study. A monitoring program consisting of discharge measurements and high frequency data from 2 springs and monitori ng wells penetrating the matrix and conduit networks of a karst aquifer was initia ted to characterize the test-site and constrain new parameters introdu ced with MODFLOW-2005 CFP. The monitoring program spanned conditions prio r to, during, and following convective and tropical storm activity, and a drought. Analytical estimates for Reynolds numbers, ranging from 105 to 106, suggest that turbulent flow occurs in portions of the underlying conduit network. The dire ction and magnitude of fluid exchange observed between the matrix and conduit network indicate the conduit network underlying the test-site drains the matrix. Head differences and observed responses in monitoring wells penetr ating the matrix and conduit network indicate that the hydraulic conduc tivities between the 2 networks do not

PAGE 15

xii significantly differ from each other. A c onceptual model for the spatial distribution of preferential flow pathways using mu ltiple data types, including shallow recession limbs observed in discharge hydr ographs indicate a slow responding aquifer with a high storage capacity, and a poorly integrated conduit drainage network with little to no point recharge. Model performances were evaluated by comparing observed hydrographs for discharge and monitoring wells penetrating the matrix and conduit network following convective and tropical storm events, and drought conditions, to simulated values from transient simulati ons. Model statistics for 32 target wells and sensitivity analysis were included in the evaluation. The dual-permeability model using the MODFLOW-2005 CFP Mode 1 displayed the highest performance with improved matches ranging from 12 to 40% between simulated and observed discharges relative to the laminar and laminar/turbulent equivalentcontinuum models.

PAGE 16

1 CHAPTER 1: INTRODUCTION Karst aquifers differ from other po rous media aquifers because of the secondary porosity that develops in them as a result of di ssolution (Palmer, 1999a). They consist of triple porosity that includes: 1) intergranular, 2) fracture, and 3) conduit or cave porosities (P almer 1999a; Worthington et al., 2000a; Martin and Screaton, 2001; White, 2002). Intergranular porosity comprises the matrix network, whereas the conduit porosity comprises the conduit network (Palmer, 1999a; Worthington et al., 2000a; Martin and Screaton, 2001). Fracture porosity can be lumped with either the matrix or conduit network depending on aperture widths (White, 1988; Worthington et al. 2000a; Martin and Screaton, 2001). Dual-permeability arises from bot h the matrix and conduit networks (White, 1999). Groundwater flow in dual-perm eability karst aquifers often exhibits both Darcian (laminar) and non-Darcian (t urbulent) flow, with Darcian flow generally dominating in the matrix network and non-Darcian flow occurring in the conduit network (Martin and Screaton, 2001) Complicating the understanding of flow is head-dependent fluid exchange bet ween the matrix and conduit networks, which is site specific and can vary spat ially and temporally in the same aquifer depending on hydrogeologic c onditions and conductance between the matrix and conduit networks (Martin and Screaton, 2001; Bauer et al., 2003; Martin et al., 2006). Moreover, the dynamic hydraulic response observed in dual-permeability karst aquifers can differ from that of equivalent-continuum aquifers (Hess and White, 1988; Dreiss, 1989). The scientific karst community has been aware that trad itional numerical groundwater flow codes ignore the dualpermeability or non-Darcian component of flow that can occur in karst aquifers (Q uinlan et al., 1995; Mohrlok et al., 1997;

PAGE 17

2 Mohrlok and Sauter, 1997; Sasowsky, 2000; Wilson, 2002; Smith et al., 2005). Delineation of springhead, or well head protection zones, and statutory requirements for establishing minimum fl ows, impose a need for the development and verification of numerical models capab le of simulating the dual-permeability and dynamic hydraulic response observed in karst aquifers. The new challenge following the recent development of a public domain, Darcian/non-Darcian groundwater flow simulator (Shoemaker et al., 2008a), is the application of this new tool in ar eas where the spatial distribution and hydraulic properties of conduits may not be well known. Theref ore, the primary objective of this project is to eval uate the performance of three different groundwater flow models (an equivalent-continuum with laminar flow, an equivalent-continuum with both lami nar and turbulent flow, and a dualconductivity with both laminar and turbul ent flow and fluid exchange between the matrix and conduit networks) using three diff erent conceptualizations of conduits. To achieve the project s primary objective, eight tasks were assigned: 1) a literature review of the tr aditional methods used to inte rpret conduit locations or preferential flow pathways and the pr evious applications of numerical groundwater flow models in karst aquifer settings, 2) site se lection, 3) to determine if non-Darcian flow occurred in the conduit network underlying the test site, 4) development of a conceptual model for the conduit network, 5) characterization of the dynamic hydraulic response and the direction and magnitude of fluid exchange between t he matrix and conduit networks, 6) development of three num erical groundwater flow models with 3 different conceptualizations of the conduit network, 7) sensitivity analysis, and 8) evaluation of the model performances.

PAGE 18

3 1.0. LITERATURE REVIEW A literature review of traditional methods for interpreting conduit locations or preferential flow pathways and previous numerical modeling approaches applied to karst aquifer settings are discussed in the following sections. 1.0.1. Traditional Methods fo r Interpreting Conduit Locations or Preferential Flow Pathways Researchers have used a variety of methods to develop conduit conceptualizations or preferential flow pathways in dual-permeability aquifers. Traditional methods include: borehole observa tions, fracture traces, dye tracing, interpreting troughs or mounds in the potent iometric surface, spring hydrographs, cave maps, geophysics, assessments of karst features, aquifer performance tests, and statistical (i.e. fractals, semivariogram cloud analysis, and Bayesian) methods. 1.0.1.1. Borehole observations Borehole observations involving the use of geophysical logs, flowmeters, and video logging have been used to characterize secondary permeability in carbonate aquifers. Cunningham et al. (2004) coupled heat pulse flowmeter, geophysical, and video logging to identify permeable intervals in the upper portion of the Biscayne aquifer in southeas t Florida. Borehole observations were correlated with stratigraphic horizons. Si milar methods have been followed by others to estimate the occurrence of cavi ties intercepted by well borings. Wilson (2002) reviewed groups of up to 400 we ll logs for Polk, Orange, Marion, and Madison Counties in Florida. He estimates that the pr obably of intercepting a cavity 0.3 m in height ranges from 25% to as high as 50% per 30 m of drilling.

PAGE 19

4 A major drawback with utilizing borehole observations to identify secondary permeability features is t hat the method relies heavily on the researchers interpretation, which often reduces three dimensional features, such as fractures, to a two dim ensional interpretation. Additi onally, the probability of intercepting a conduit can be low (from 0.4 to 3%) in Paleozoic limestone (Worthington et al., 2000b). Moreover, Safko and Hickey (1992) point out that caution must be exercised when interpre ting secondary permeability features in well boreholes. They reviewed borehole data consisting of caliper, flowmeter, temperature and video logs at four site s in east Florida and conclude that some of the apparent secondary permeability features are in fact arti facts of the boring process (i.e. drilling induced collapse). This is particularly important for poorly indurated rocks, such as those that co mpose the Upper Floridan aquifer. 1.0.1.2. Fracture traces Fracture traces refer to natural linear trends in soil tonal patterns, vegetation, or sinkholes that reflect high permeability intervals. They are typically defined by length or size using predev elopment aerial phot ographs and can be verified using various geophysical methods. In terpretations of fracture traces to locate vertical fracture intervals in insoluble rocks, or subsurface conduits in soluble rocks is well established (Lattman and Parizek, 1964; Jones et al., 1997). In karst terrains, sinkhole lineaments can provide useful information because they are surface manifestations of subsurface cavities and reflect the upgradient terminus of conduits (Worth ington, 1999). Jones et al. (1997) interpreted fracture traces using closed depressions greater than 0.08 km2. Moore (1981) used several geophysical tec hniques to verify fracture traces interpreted from elongated sinkholes and soil tonal patterns at the Cross Bar Wellfield in Pasco County, Florida. The relationship between water quality and fracture trace length at t he Cross Bar Wellfield was ex amined by Williams (1985) to determine the vertical hydraulic c onnection. Williams (1985) suggests that

PAGE 20

5 fracture traces with surface expressi ons shorter than 2 km do not produce discernable geochemical signatures, suggesti ng that the features support diffuse recharge, or do not have a substantia l hydraulic connection to the underlying aquifer. Others have also noted limitations a ssociated with the use of fracture traces in karst terrains. For example, Wilson (2002) advises that while fracture traces may reveal the location of fracture or joint controlled conduits, they fail to recognize many anastomotic conduits, or those parallel to bedding, which often develop in soluble rocks with low dips. Mo reover, visually connecting sinkholes is also highly dependent on the researchers interpretation, or inherent bias. For example, Armstrong et al. (2003) note that the ability to discern closed topographic contours diminishes at lower elevations. 1.0.1.3. Tracer tests Tracer testing is a powerful tool that can be used to delineate spring basins, quantify travel times, and identif y hydrogeologic connections (Quinlan et al., 1995). A tracer test typically involves introducing a tracer, or multiple tracers, into the groundwater flow network at a loca tion of interest (e.g. a well or sinkhole) and collecting samples from discharge point s or other areas of interest. The tracer can be any measurable parameter. Chemical tracers include: dyes, isotopes, gases, and ions. Biological tracers include pollen, yeasts, viruses, and microbial spores (Davis et al., 1985). Na tural tracers, such as temperature or specific conductance, can be used to qualitatively describe hydraulic connection between sites (Martin and Dean, 1999). Accidental tracers, introduced during contaminant spills (Schindel, 2003) can also provide useful information. The goal in most dye tracer tests is to inject a tracer, or mu ltiple tracers, at a concentration below the level of visi ble detection, but above the level of detection for the instrumentation. Wort hington and Smart (2003) used data from

PAGE 21

6 203 quantitative tracer tests to develop tw o regression equations for estimating the concentration levels below visual detection. Tracer tests can be qualitative or quantitative depending on the objectives of the project and can prov ide additional useful information. For example, estimates for the volume of conduits in a spring basin can be estimated if the drainage area, average transit time in t he conduit network, and the discharge are known (Wilson 2002). Drawbacks of tracer testing include the potential for: i) dilution, which results in lost or nonrecoverable tracers, and ii) saturation, which involves the injection of dye quantities at or above the level of visible detec tion. It may also be impractical to perform tracer tests near major pumping centers, where municipal wells cannot go offline for extended periods if needed. Finally, quantitative tracer tests may confirm turbulent flow, but they provide little information about the location or geometry of the conduits betw een the injection and recovery sites. 1.0.1.4. Troughs/mounds in the potentiometric surface Troughs or mounds in the potentiometr ic surface or aquifer water levels are additional methods used to identify t he location of conduits or preferential flow pathways. Troughs form during drought conditions because the conduits serve as drains for groundwater stored in the matrix. Conversely, mounds may appear in the potentiometric surface above subsurface conduits during water level highs and, depending on the head cond itions, groundwater may actually drain from the conduits into the su rrounding matrix (White, 1999; Worthington 1999; Screaton et al., 2004). Accurately del ineating the potentiometric surface, however, requires a dense data network.

PAGE 22

7 1.0.1.5. Spring hydrographs Spring hydrographs provide an opportunity to evaluate the response or flashiness of the aquifer following high re charge events (White, 1988; Florea and Vacher, 2006; Florea and Vacher; 2007). Sma ll spring basins that are dominated by conduits typically exhibit steep, shor t recession limbs following high recharge events. Conduit fed springs with large bas ins that are dominated by the matrix exhibit shallow, longer recession li mbs (White, 1988; Florea and Vacher, 2006; Florea and Vacher, 2007). 1.0.1.6. Cave surveys Cave surveys provide the most useful information regarding the geometry and speleogenesis of a conduit network (Palmer, 1991). Surveys involve selecting stations throughout the cave network. Distances between stations and the height and width of passages are meas ured at each station. Surveys can be performed in air-filled (Florea, 2006a; Florea, 2006b) or water-filled cave networks (Karst Underwater Research, Inc., 2008a). Recently, sonar mapping has been utilized to survey portions of the underwater caves in the Woodville Karst Plain (White, 2002). The accuracy of cave survey maps can be highly dependent on the experience and skill of the survey team. Therefore, cave survey data should be verified using geoph ysical methods, such as radiolocation, when highly accurate surveys are required. Wilson (2002) used cave surveys to pr ovide a first approximation of cave density and geometries in north Flori da. Based on 15 cave surveys and 339 height to width ratios he estimated that many conduits in Florida are three times wider than they are high, suggesting that many conduits are elliptically shaped, possibly forming along bedding planes, prev ious water table elevations, which would exaggerate horizontality regardless of rock dip, or forming mixing zones. He also estimates that cave density in north Florida was no less than 1, 638

PAGE 23

8 km/km3. Assuming a nonpreferential, ubitiqu itous distribution results in an average distance of 100 m both vertically and horiz ontally between cave passages, although he correctly pointed out that it is not likely that karstification is ubiquitous. Using data from 223 surveyed caves, he also estimated that the median cave length in north Florida is 38 m. Finally, he obser ved that of 139 surveyed cave passages a bimodal distribution with a primary mode of N 72 W and a secondary mode of N 9 E exists. Florea (2006a) and Brinkmann and Reeder (1995) observed a similar bim odal distribution fo llowing a NE-SW and NW-SE orientation for air-filled caves in west-central Florida. Wilson (2002) further states that the pr imary mode may not parallel th e potentiometric gradient due to the low frictional resistance in c onduits. He further indicates that the database he used for his calculations consisted of surveys performed in waterfilled conduit networks (Wilson, 2002). This distinction can be important because air-filled cave geometries used as analogs for water-filled conduit geometries may be misleading due to the fact that the speleogenesis and resulting geometries of air-filled and water-filled caves may differ significantly (White, 2002). For example, caves formed in the vadose zone in low gradient strata typically are oriented parallel to the direct ion of dip, whereas those formed in the phreatic zone in low gradient strata, are typically oriented parallel to strike (Palmer, 1999a). This becomes particularly important in an area, such as Florida, that has experienced multiple transgr essive and regressive sequences. According to Florea et al. (2007) cavescale porosity in the Upper Floridan aquifer has been controlled primarily by Quaternary sea level fluctuations. Explorational bias is a limitation wi th cave survey data. That is, the surveyed cave passages do not represent the complete conduit network. 1.0.1.7. Geoph ysical methods Microgravity and resistivity geophysi cal methods can be used to detect shallow subsurface cavities and fracture s. Both methods are commonly coupled

PAGE 24

9 and include drilling or core sampling (Wood and Stewart, 1985; Crawford et al., 1999). Microgravity involves detecting c hanges in the density of subsurface media. Dense subsurface media produce ab ove normal variations in gravity and less dense media or voids produce lows. Micr ogravity is affected by factors such as distortion of the signal by epikarst, elevation, and latitude that requires the establishment of a base station to moni tor tidal and instrumental drift (Wood and Stewart, 1985; Crawford et al., 1999). Resistivity is a measure of a medium s resistance to conduct an electrical current. Resistivity is affected by saturation, porosity, and the bulk resistivity of the media and is typically performed using a variety of electrode arrangements (Parasnis, 1986). A limitation with resist ivity is that a c onsistent geophysical signature does not exist for airfilled or water-filled conduits. Pease (1997) has extensively tested and enhanced radioloc ation, which is a geophysical method that involves placin g a transmitter in the conduit network and a mobile receiver at land surface. The land surface location above a conduit passage is determined by identifying the null that occurs in the electromagnetic field directly above the transmitter. Radi olocation can be used in both air-filled and water-filled cave networks. Resistivity and radiolocation are limi ted in that their accuracy generally diminishes with increasing tar get depth. Their efficacy is also affected by the bulk conductivity of the media and cultural noise that may be present at a site (McNeill, 1990; Pease, 1997). 1.0.1.8. Assessment of karst features Veni (1999) proposes a geomorphologically based approach for assessing karst features in areas where tracer tests or geophysical methods are not available or feasible. The process involves identifying karst feat ures in the field within an area of interest, followed by ex cavation and exploration of the feature.

PAGE 25

10 A determination for the permeability of features and degree of hydraulic connection with the underlying aquifer is made with the purpose of identifying environmentally sensitive features or areas. 1.0.1.9. Aquifer performance tests Aquifer performance tests are typically performed to measure transmissivity, storativity and leakance. In terpretation of results from aquifer performance tests can be misleading howev er, as the limitations and conditions for the methods used to analyze them can be violated when applied to a karst aquifer (Sasowsky, 2000). Information perta ining to anisotropy can be obtained when a sufficient observation network is es tablished. For example, elongation of the cone of depression reflects the dire ction of maximum hydraulic conductivity and can be used to interpret conduit ori entations (Palmer, 1999b; Worthington and Ford, 1997). Aquifer performance tests can also provide useful information regarding hydraulic connection between matrix and frac tures. For example, results of an aquifer performance test conducted in the vicinity of ground-truthed fractures revealed that the degree of hydraulic connection between the matrix and fractures was highly variable (GeoTrans, 1988a). 1.0.1.10. Statistical methods Statistical methods can be practical for estimating permeable features when spatial data sets are difficult to ac quire as a result of conduit depths, or because of liability issues that are invo lved with performing cave surveys or dyetracer testing. However, the success of statistical methods is highly dependent on the available data, the appropriateness of the statis tical method selected for a particular region, and the problem that is being resolved.

PAGE 26

11 1.0.1.10.1. Fractals Fractals are geometric objects that ar e statistically self-similar regardless of scale. Curl (1999, 1986) observed that the distribution of cave lengths and geometries in a region exhibit a fractal, or self-similar nature. He utilized the Erlang stochastic process to describe the probability that a main conduit has n number of tributaries randomly lost and formed over time due to infilling, collapse, or dissolutional processes. Assuming that the rate of entrances opened is proportional to cave length and that the rate of entrances lost is proportional to the number of entrances, he speculated t hat cave distribution in a region, including entranceless caves, could be es timated using a Poisson distribution conditional on length. Comparison of data from 11 separate regions indicated that the Poisson conditional model perform ed better relative to the alternative models (Curl, 1999). However, a limitation with using a Poisson conditional model is that it provides no information on the connectivity of caves in a region. Mace et al. (2005) used fractal sca ling to estimate secondary porosity for portions of the Edwards aquifer in southcentral Texas. Joint apertures were measured along vertical and horizontal transects at 8 outcrop locations. Photograph exposures, photomosaics, and image analysis were utilized in addition to fractal statistics. Average es timates of secondary porosity using the Riemann zeta function were twice as high as estimates using direct summation (Mace et al., 2005). 1.0.1.10.2. Semivariogram cloud analysis Spatial data analysis involves iden tifying patterns in the geographical distribution of data. The semivariogram cloud method is a type of spatial analysis. The method was used by Kurtzmann et al. (2005) to correlate point head data with oriented 1 km scale fractures. Kurtzmann et al. (2005) demonstrate, using laboratory scale exper iments, that the hydraulic gradient,

PAGE 27

12 which is the primary cont rol on groundwater flow, over whelmed head variations due to regional joint and fracture pa tterns, whereas detrended data or head residuals generally provide a more reliable indication of fracture trends. The method was applied to a site-scale probl em involving groundwater contamination below an industrial park. Lineaments were mapped on aerial photographs and were verified or rejected based on fi eld mapping observations Results of the analysis were supported by slug, pumping, and tracer tests. The strength of the semivariogram method is that it prov ides information about the location of permeable features, but is limited in that it requires a dense monitoring well network. It also requires that a discernable difference in head exists between the matrix and conduit networks. 1.0.1.10.3. Poisson and Bayesian statistics Langevin (2003) used Poisson probabilit y density functions to randomly describe unknown parameters, such as fracture orientations, lengths, and transmissivities. The probability densit y functions were matched with field observations to keep them r epresentative of the site. Bayesian statistics can also be useful for model calibrations. Jiang et al. (2004) developed a code that uses Bayesi an inversion to calibrate MODFLOW models. The code was tested using an existi ng model of the Edwards aquifer in south-central Texas. Calibration of t he Edwards aquifer groundwater flow model using the Bayes code produced better fits to measured heads than previous calibration efforts using upsca ling and cokriging techniques. 1.0.1.10.4. Reinfo rced random walk Reinforced random walk methods are useful for investigations that integrate the physical and chemical pr ocesses associated with speleogenesis. Jaquet et al. (2004) used a reinforced rando m walk procedure to simulate conduit

PAGE 28

13 enlargement along fracture-intervals. Part icles had the possibility of following the same trajectories of previously dissolv ed particles, thereby retaining memory of prior trajectories and promoting prefer ential development of passages. 1.0.2. Numerical Groundwater Flow Model Types Numerical groundwater flow models are a tool for managing water resources and evaluating contaminate transport. Examples previously used in karst aquifer settings include: equivalent -continuum, discrete-fracture network, dual-conductivity, and dual-contin ua groundwater flow models. 1.0.2.1. Equivalent-continuum models Equivalent-continuum, numerical groundwater flow models have traditionally been used to simulate groundw ater flow in karst aquifers. They manage the matrix and conduit networ ks as one continuum. Equivalentcontinuum models are single continuum models since they represent the bulk transmissivites of the matrix and conduit networks. Conduits are traditionally incorporated by assigning high transmissivi ty values to cells at suspected conduit locations. Equivalent-continuum models t hat incorporate highly transmissive intervals (i.e. Scanlon et al., 2003; Smith et al., 2005) are sometimes referred to as fracture-zone continuum models (Langev in, 2003). Conduit properties can be incorporated into equivalent-continuum m odels using deterministic and stochastic methods. 1.0.2.2. Discrete-frac ture network models Discrete-fracture network models simula te flow only through the fracture network and hence are single continuum models. Matrix permeability is not represented in discrete-fractu re models, thereby restrict ing flow to the fracture

PAGE 29

14 network (Diodato, 1994). Conduit lengths, orientations, aperture, and hydraulic properties can be described explicitly, stochastically, or some combination thereof. 1.0.2.3. Dual-conductivity models Dual-conductivity models, which are also known as double permeability/double-porosity, or triple pe rmeability/triple porosity models simulate flow through the matrix and conduit net works separately and hence are multiple continua models (Diodato, 1994). They are a subset of full dual-continua models in that they only interact or permit fl uid exchange between the matrix and conduit networks at discrete locations withi n a common domain. Fluid exchange is expressed by a linear exchange term and the rate of fluid exchange typically can be adjusted by the user (Mohrlok and Sauter, 1997; Painter et al., 2007; Shoemaker et al., 2008a). The direction of fluid exchange (matrix to conduit or conduit to matrix) is typically headdependent. Until recently, most dualconductivity models utilized finite element grids which are more robust at handling irregularly shaped objects such as fractures or conduits (Mohrlok and Sauter, 1997; GeoTrans, Inc., 1988a). C onduit properties can be described deterministically, stochastically, or some combination thereof. 1.0.2.4. Dual-continua models Dual-continua models simulate flow through the matrix and conduit networks separately. They also permit fluid exchange between the matrix and conduit networks. The fundamental diffe rence between dual-conductivity and full dual-continua models is that in dual-c ontinua models, the matrix and conduit network overlap and interact at every node throughout the model domain, whereas in dual-conductivity models the ma trix and conduit netwo rks interact at discrete locations (Diodato, 1994). Given that the matrix and conduit networks

PAGE 30

15 overlap, conduit locations do not need to be described deterministically or stochastically in full dual-conti nua models (Mohrlok and Sauter, 1997). 1.0.3. Previous Applications of Nume rical Models in Karst Aquifer Settings Previous modeling efforts using equi valent-continuum, discrete-fracture network, dual-conductivity, and dual-continua models to capture the dynamic hydraulic response in karst aquifers are described in the following sections. 1.0.3.1. Equivalent-continuum models Equivalent-continuum models ar e often used for water management purposes. The Southwest Florida Wa ter Management District (SWFWMD) currently uses a variety of equiva lent-continuum models with governing equations that simulate laminar flow, but do not include turbulent flow, or fluid exchange between the matrix and conduit networks. Scanlon et al. (2003) evaluated the performance of lumped and distributed param eter equivalentcontinuum models. Both models were tested using discharge and head data at Barton Springs, in south-central Texa s. The lumped parameter model grouped parameters such as hydraulic conducti vity and elevation and specific yield and elevation. In the distri buted parameter model parameters such as hydraulic conductivity were spatially distributed. Zones of hydraulic conductivity were calibrated by trial and error and simulat ed heads were compared with observed heads in monitoring wells. Both models adequately simulated spring discharge, but the distributed parameter model permi tted a more comprehensive evaluation of aquifer head data and the effects of pumping on spring discharge. According to Scanlon et al. (2003) if average groundwat er fluxes are the primary objective, distributed parameter equiva lent-continuum models are generally considered adequate for karst aquifers except in ca ses involving contaminant transport, wellhead, or springhead protection zones, where dual-conductivity models are

PAGE 31

16 generally considered more appropriate (Q uinlan et al., 1995; Scanlon et al., 2003). In a more recent paper by Smith et al. (2005), significant differences between simulated and observed head values were noted in the Barton Springs equivalent-continuum groundwater flow model specifically in areas influenced by known conduits. Langevin (2003) used a fracture-zone c ontinuum model to estimate travel times from a potential we tland augmentation site to a municipal wellfield. Fracture-zone networks were characteri zed using probability density functions described from field observations. Uncertai nties in fracture-zone properties were quantified using Monte Ca rlo analysis. MODFLOW and MODPATH codes were used for the simulations and flow was assumed to be laminar. Shoemaker et al. (2008b) us ed a nine layer, steady-state, laminar/turbulent, equivalent-continuum model of the Biscayne aquifer in southeast Florida to evaluate the effe cts of turbulent groundwater flow on hydraulic heads and parameter sensitivit ies in preferential groundwater flow layers using the MODFLOW-2005 Conduit Flow Process (CFP) Mode 2 developed by Shoemaker et al., (2008a). They conclude that, 1) turbulent groundwater flow was laterally extensive in preferential groundwater flow layers within the Biscayne aquifer where high, hor izontal hydraulic conductivities exist and, 2) turbulent groundwater flow c ould increase, or decrease simulated hydraulic heads depending on the net volume into or out of a model cell, and 3) the sensitivity of composit e-scaled horizontal hydraulic conductivities decreased by as much as 70% when turbulent groundwat er flow was shut-off (Shoemaker et al., 2008b). 1.0.3.2. Discrete-fracture models Kraemer (1990) conducted a series of numerical experiments using randomly-generated fracture conceptualizat ions. The models consisted of twodimensional, laminar flow, steady-state conditions. The matrix was considered

PAGE 32

17 impermeable in the discrete-fracture m odels. The models were developed as a first approximation to understanding local-scale and regi onal-scale discrete flow in fractures. Results of the numerical experiments suggested that local-scale discrete flow was maintained at the regiona l-scale (fractal networks) for many of the realizations. Moreover a regional-scale fracture superimposed on local-scale fractures with transmissivities three or ders of magnitude less than the regionalscale fracture resulted in over 98% of the flow being transmitted through the regional-scale fracture. Two orders of magnitude difference in transmissivities between the regional-scale fracture superimposed on local-scale fractures resulted in 68% of the flow being transmi tted through the regional-scale fracture. 1.0.3.3. Dual-conductivity/pe rmeability/porosity models Painter et al. (2004) used a sub -regional, finite-difference, dualconductivity model using MODFLOW-D ual Continua Model (DCM) 1.0 to simulate discharge from Barton Springs in south-central Texas. Conduit locations interpreted from dye-tracer tests and tr oughs in the potentiometric surface were explicitly incorporated into the groundwater flow model. A comparison of simulated spring discharge using the dual-conductivity model which explicitly incorporated c onduits and an equivalent-continuum model that incorporated conduits across several cells (fracture-zone continuum model) was performed. Simulations using the fracture-zone continuum model yielded delayed responses in discharge rece ssions during stressed conditions or droughts and overestimated discharge durin g high flow conditions. Moreover, water levels in the simulations were mo re sensitive to recharge than observed responses. The dual-conductivity model appeared to perform more satisfactorily, however limitations in the code were noted so enhancements were made to MODFLOW-DCM 1.0 (Painter et al., 2004). The revised code, MODFLOW-DCM 2.0, a variant of MODFLOW-2000, was used to recalibrate the steady-state Barton Springs dual-conductivity model and to develop a twelve week transient

PAGE 33

18 simulation for the Santa Fe Sink/Rise loca ted in north-central Florida (Painter et al., 2007). In a separate study, GeoTrans, Inc. (1988a) used a steady-state, fracturezone continuum model (an equiva lent continuum model) as a tool to help identify the extent and orientation of regional fr actures, or fault zones, to simulate groundwater flow in Eocene limestones and dolostones underlying the area surrounding Rainbow Springs in north-centra l Florida. Inferred fracture traces and fault zones were incorporated into t he groundwater flow model. The actual extent, orientation and hydraulic properties of the fractures and fault zones were not known. Therefore, geophysical surveys and aquifer performance tests were performed at a fracture trace on the test-s ite. Conditions at the test-site were assumed to be representative of regional-scale conditions. Fracture zones at the test-site were verified using horizontal elec trical profiles, a DC resistivity method. Results of the aquifer performance test, which consisted of 3 observation wells penetrating the fracture, or solutional channel zone, and 1 observation well penetrating the matrix, revealed that at l east 1 of the 3 observation wells in the fracture zone exhibited a relatively lower degree of hydraulic connection with the matrix. It was speculated that fracture mineralization had reduced permeability at the observation well (GeoTrans, Inc., 1988a). Difficulty in simulating the predevel opment potentiometric surface was encountered when using the inferred, r egionally extensive fracture/fault zone delineations in the steady-state fracture-zone continuum model. An acceptable match between predevelopment and simulated water levels however, was obtained when using local-scale fractures with widths of 3 m extending from the top to the bottom of the Upper Floridan aqu ifer. This conceptual model of fracture traces was then incorporated into a transient, finite-element double-porosity model with laminar flow that incl uded fluid exchange between the matrix and conduit networks. The FRACFLOW code was used to simulate flow in the double-porosity model (GeoTrans, Inc., 1988a).

PAGE 34

19 One of the salient outcomes of the GeoTrans, Inc. (1988a) modeling effort is that it demonstrated the superiority of the double-porosity model over previous equivalent-continuum models. A prior attempt, using an equivalent-continuum model at the test-site, simulated a cessation of spring flow when recharge was shut-off, however the springs have disc harged continuously for the period of record, which includes droughts (GeoTrans, Inc., 1988b). The GeoTrans, Inc. (1988a) double-porosity model simulated continuous discharge from the springs when recharge was shut-off. However, the differences between observed and simulated discharges for the previ ous equivalent-continuum model and the double porosity model developed by GeoTrans Inc. (1988b) were not quantified. Hazlet et al., (2004) discuss devel oping a dual-permeability model to simulate discharge from Oligocene limest one. The study area is focused on the Woodville Karst Plain/Wakulla Springs basin in northwest Florida and involves an extensive characterization of the condui t network. Conduit surveys using sonar mapping (White, 2002), groundwater velocities estimated from quantitative dyetrace tests, and Reynolds number estima tes have been compiled (Kincaid et al., 2004). 1.0.3.4. Dual-continua models Mohrlok and Sauter (1997) com pared the performance of a dualpermeability and dual-continua model for groundwat er flow in a series of Jurassic limestones in Swabian Alb, Germany. T he Darcy-Weisbach equation was used to calculate flow through the conduits in the dual-permeabilit y model and Darcys equation was used to calculate flow in the matrix. Exchange between the matrix and conduit networks was governed by a head-dependent exchange term. The dual-permeability model consisted of a fi nite element grid. Conduit locations inferred from fracture traces observ ed in aerial and satellite images were explicitly incorporated into the mo del. The ROCKFLOW code was used to simulate flow in the dual-permeability model. The dual-continua model involved

PAGE 35

20 implementation of a MODF LOW extension that permi tted exchange between the overlapping continua. In the dual-continua model, flow through the conduits and matrix was calculated using Darcys equatio n. Given that the matrix and conduit network were connected at every node, conduit locations were not needed. Both the dual-permeability and dual-continua models adequately reproduced observed hydraulic heads and discharge (Mohrlok and Sauter, 1997). 1.0.3.5. Hybrid integrated models Recently, hybrid integrated numerical groundwater flow models that include the physical-chemical processes governing secondary permeability have been developed (Jaquet et al., 2004; B auer et al., 2003). These models are primarily focused on describing the evolut ion of karst features and transport through actively developing karst aquifers. Jaquet et al. (2004) used a stochasti c reinforced random walk procedure to model conduit geometries in the Barrois Karstic Limestones at the Meuse/Haute Marne Underground Resear ch Laboratory in France. Conduit locations were incorporated from field obs ervations of regional fault orientations and travel times were verified with tracer tests. The random walk procedure was initiated at regional fault zone locati ons and operated until the physical-chemical processes (speleogenesis) reached equilib rium. Given that conduit formation was enforced in a preferred manner, the simulated conduit network is highly dependent on the initial characterization of preferential pathways. Conduits were explicitly incorporated into the groundwater flow model (following the fracturezone continuum methodology). Darcian flow was assumed. Bauer et al. (2003) performed numeri cal experiments using the Conduit Aquifer Void Evolution (CAVE) numerical code. The simulations included turbulent flow through a single condui t surrounded by matrix. Fluid exchange between the matrix and conduit was permitted and integrated reaction kinetics. The primary objective of t he modeling effort was to ev aluate the effect of fluid

PAGE 36

21 exchange on conduit enlargement. Results indica te that dissolutional growth of the fracture increases many factors, c ontingent on the original diameter of the fracture, by permitting fl uid exchange between the ma trix and conduit. The numerical experiments also demonstrate th at the rate of dissolution enlargement is highly sensitive to the direction of fluid exchange. Transfer of fluids from the conduit into the matrix promotes deeper penetration of aggressive waters, which enhances dissolution, whereas fluids transfe rred from the matrix to the conduit are saturated with respect to calcium, t hereby diminishing conduit growth (Bauer et al., 2003). Table 1-1 summarizes specifications of previous modeling efforts.

PAGE 37

MODEL AUTHOR CODE CONDUIT CONCEPTUALIZATIONS FLOW REGIME Scanlon et al. (2003) MODFLOW transient, evaluated performance of lumped & distributed parameter approaches for determining effects of regional-scale pumping on spring discharge laminar Fracture-zone Continuum Langevin, (2003) MODFLOW & PATH3D transient, fractures stochastically generated using probability density functions laminar EquivalentContinuum Conduit Flow Process Shoemaker et al. (2008b) MODFLOW2005 CFP steady-state, testing effects of turbulence on heads & parameter sensitivities laminar & turbulent Discrete-Fracture Kraemer, (1990) FRACNET steady-state, synthetic models statistically generated fractures laminar Painter et al. (2004; 2007) MODFLOWDCM transient, conduits inferred from troughs potentiometric surface, tracer testing, survey cave data laminar & turbulent GeoTrans, Inc. (1988a; 1988b) FRACFLOW transient, fractures from photolinears & steady-state fracture-zone continuum model laminar Dual-Conductivity/ Permeability/Porosity Mohrlok & Sauter, (1997) ROCKFLOW Transient conduits from fracture traces laminar & turbulent Dual-Continua Mohrlok & Sauter, (1997) MODFLOW EXT transient, conduits connected with matrix at each node laminar Fracture-zone Continuum Jaquet et al. (2004) GARST transient, reinforced random walk at fault zone locations laminar Hybrid Integrated Dualpermeability Bauer et al. (2003) CAVE transient, synthetic models, single conduit, speleogenesis laminar & turbulent Table 1-1. Summary of previous investigations using numerical models in karst aquifer settings. 22

PAGE 38

23 1.1. PROJECT OBJECTIVES The role that the conceptualization of conduits play in determining the ability of a groundwater flow model to capture the dynamic hydraulic response in a dual-permeability karst aquifer will be investigated. Three conduit conceptualizations will be incorporated into 3 groundwater flow models: a laminar, equivalent-continuum model with bulk hydraulic conductivity values for the matrix and conduit networks, a laminar/turbulent eq uivalent-continuum model with bulk hydraulic conductivity values for the matrix and conduit networks, and a laminar/turbulent dual-conductivity model with water-filled conduits explicitly incorporated, as well as fluid exchange between the matrix and conduit networks. Performance of the 3 groundwater flow models will be evaluated by comparing, 1) observed and simulated water levels for 32 observation wells, 2) simulated and observed water levels from moni toring wells penetrating the matrix and conduit networks, and 3) observed discharge hydrographs will be compared to simulated discharges from transient simulations. A satisfactory outcome of this resear ch effort is to quantify the differences in model performance. The performance of the 3 models, which have 3 different conceptualizations of conduits, may prov ide insight into the application of Darcian/non-Darcian groundwater flow simula tors in karst aquifers where conduit locations, elevations, orientations, or hydraulic properties of the conduits are poorly understood. 1.2. OVERVIEW The remainder of this dissertation is divided into chapters that detail the tasks of the research project. Chapter 2 discusses the approach and methodologies and Chapter 3 describes the proposed conceptual model for the study area. Chapter 4 discusses devel opment of the num erical models, sensitivity analysis, and performance of the groundwater flow models. The

PAGE 39

24 project conclusions, limitations, recommended paths forward, and the contribution of this research are presented in Chapter 5.

PAGE 40

25 CHAPTER 2: APPROACH AND METHODOLOGIES 2.0. INTRODUCTION To achieve the projects primary objective, eight ta sks were assigned: 1) a literature review of the traditional methods used for understanding conduit locations or preferential flow pathways and the previous applications of numerical groundwater flow models in karst aquifer settings, 2) site se lection, 3) to determine if non-Darcian flow occurred in the conduit network underlying the testsite and if so to estimate Reynolds numbers, 4) develop a conceptual model for the conduit network, 5) characterization of the hydraulic response and the direction and magnitude of fluid ex change between the matrix and conduit network, 6) development of the 3 numer ical groundwater flow models, 7) sensitivity analysis, and 8) evaluation of the model performances. The results of the literature search (Task 1) were di scussed in Chapter 1. The approach and methodologies of Tasks 2-8 are presented below. 2.1. TASK 2 SITE SELECTION Task 2 involved defining a se t of criteria that ultimately directed site selection. Site selection was guided by matching a karst aquifer with the following criteria: i) fresh water is discharged from the spring(s), ii) the spring(s) is/are not tidally influenced, and iii) information perta ining to the spatial distribution of the conduit network is available. Twin Ds Spring, sometimes spell ed Twin Dees, (Jones et al., 1997; Champion and Starks, 2001) or called Li ttle Spring, (Scott et al., 2004) and Weeki Wachee Spring near the Gulf of Me xico in west-central Florida were

PAGE 41

26 selected as the test sites for this study (Figure 2-1). The springs were selected because they match all three criteria used to guide site selection. 2.2. TASK 3 DETERMINE IF NONDARCIAN FLOW OCCURS IN THE CONDUIT NETWORK AND ESTIMATE REYNOLDS NUMBERS Task 3 involved determining if non-Da rcian flow occurs in the conduit network underlying the study area and if so to estimate the Reynolds numbers. The Reynolds number is the ratio of inertial to viscous forces and is an indicator of Darcian or non-Darcian flow condition s (Freeze and Cherry, 1979). This task was important because it ju stifies the use of a Darc ian/non-Darcian groundwater flow simulator and because estimates of the Reynolds numbers are needed for the laminar/turbulent numerical groundwater flow models. In a previous study in the vicinity of Weeki Wachee Spring, Yobbi (1989) assumed that flow was laminar. He stated that, The Upper Fl oridan aquifer is characterized by an overall high transmissivity caused by solution of limestone and dolomite. Transmissivities are highest in areas i mmediately surrounding large springs and decrease away from the springs . and may exceed 1,000,000 m2/d . Hickey (1984) [however] was able to confirm that flow in the aquifer is Darcian, (Yobbi, 1989). The Hickey (1984) study involved aquifer performance tests that were performed in southeast Pinellas County, Florida. The tests were not performed in close proximity to any major di scharge points and the hydrogeologic characteristics of the aquifer in Pinella s County, where the aquifer performance tests were performed, differs from the hydrogeologic char acteristics of the aquifer in the vicinity of Weeki Wachee Spring. Estimates of Reynolds numbers and groundwater velocities should be estimated through quantitative dye tracing. In this study a dye-trace test was originally proposed, but was not funded. Informatio n pertaining to conduit geometries in the conduit network near t he springs, however, was available, as

PAGE 42

27 the water-filled caves have been partially surveyed by cave divers. Therefore, analytical estimates of the Reynolds num bers were calculated using water levels in a nearby well with an open interval t hat breaches the roof of an underwater cave (WW-F well shown in Figure 2-1), pool stages, measured groundwater temperatures in the conduit network, and t he cave survey data provided by Karst Underwater Research, Inc. (KUR). For the analytical estimates of the Reynolds numbers, values for density ( = 997 kg/m3) and viscosity at 24 C ( = 9 x 10-4 kg/m-s; Chemical Rubber Company Press Inc., 1995) were select ed based on temperatur e measurements obtained in the conduit monitoring well an d springs. The hydraulic radius ( R ) of a circular pipe was assumed for these esti mates. The hydraulic gradient between the monitoring well penetrating the conduit network and the springs was calculated using differences in water le vel elevations between the conduit well and the pool stages divided by the distances between the conduit well and the springs. Reynolds numbers were calculated using (White, 1988): R Nr (eq. 2-1) where: Reynolds Number (dimensionless), Nrwater density at 24 C (M/L3), specific discharge (L/T), AQ / hydraulic radius for a conduit with a circular geometry (L), r R 2 dynamic viscosity at 24 C (M/L-T). Cave divers indicate that Twin Ds Spring vent is roughly a vertical, circular conduit with a diameter of approximately 0.9 m (Karst Underwater Research, Inc. 2008c). The vent at Weeki Wachee Spring is a vertical fracture that reduces to a 0.9 x 2 m restricti on (Karst Underwater Research, Inc., 2008b). Vents at both Twin Ds and Weeki Wachee Springs open into larger chambers

PAGE 43

28 with varying complex geometri es (Karst Underwater Research, Inc. 2008b; Karst Underwater Research, Inc., 2008c). For thes e analytical estimates, diameters of 0.9 and 5 m were selected for the conduit geometry (hydraulic radius). This was considered adequate as the governing equ ations in the dual-conductivity, Darcian/non-Darcian groundwater flow si mulator assumes a cylindrical pipe (Shoemaker et al., 2008a). Discharge measurements at Weeki Wachee Spring has varied from discrete measurements near the vent to composite measurements performed downstream of the confluence for Twin Ds Spring run and the Weeki Wachee Spring run (Yobbi, 2004). A rating curve for estimating discharge at Weeki Wachee Spring was developed by the U. S. Geological Survey (USGS) using measured discharge values and measured water levels in the Weeki Wachee Deep well (Figure 2-1; Knochenmus and Yobbi, 2001; Yobbi, 2004). The USGS rating curve used to estimate discharge at Weeki Wachee Spring includes composite measurements of discharge quantities from both Weeki Wachee and Twin Ds Springs (Knochenmus and Yobbi, 2001). Therefore, measured discharges at Twin Ds Spring collected in this study were subtracted from the estimates of discharge for Weeki Wac hee Spring, which are a composite of discharge from both Weeki Wachee and Twin Ds Springs, to obtain a more representative estimate of discharge for Weeki Wachee Spring. Discharge values were then divided by the conduit cross-se ctional areas to estimate specific discharges (AQ / ). 2.3. TASK 4 DEVELOP A CONCEPT UAL MODEL FOR THE CONDUIT NETWORK The next step involved developin g a conceptual model for conduit locations or preferential flow pathways. This information was needed because understanding the location of conduits or preferential flow pathways has direct application in the groundwater flow models. Task 4 was accomplished by

PAGE 44

29 compiling various types of data traditiona lly used to infer preferential flow pathways into a geographical information system (GIS) database. Data types used to infer preferential flow pathways include: a physical inventory of surface karst features, fracture tr aces inferred from the alig nment of closed topographic depressions, inferred troughs from water levels within the aquifer, elevation modes of conduits inferred from borehole porosity descriptions, caliper logs, and surveyed underwater caves. Moreover, r adiolocation was used, when possible, to verify cave survey data. Hydrogeolog ic data consisting of aquifer thickness and well/spring hydrographs were also used to develop the conceptual model for preferential flow pathways. An inventory of known karst features was compiled following a modified version of that proposed by (Veni, 1999) to assess karst features. The inventory of karst features included: springs, underwater cave entrances, and sinkholes. Sinkholes determined by Trommer (1987) to have a substantial hydraulic connection with the underlying aquifer are included in the inventory. Not all sinkholes were included in the database, primarily because some features can be filled-in with sediment (Hill and DeWitt, 2004) and may not be as well connected with the underlying aquifer. Inferred fracture traces from a prev ious study performed by Jones et al. (1997) based on the alignment of clos ed topographic depressions with minimum areas of 0.08 km2 and a fracture trace verified with a geophysical method (vertical electrical soundings) in a previous study conducted by Wood and Stewart (1985) were compiled in the GIS database. Troughs in the potentiometric surface were inferred using contours of aquifer water levels at the start of a drought in May 2001 and May 2006. These time frames were selected because the lowest pool stages recorded at Weeki Wachee Spring for the period of record which dates back to 1929, were observed in June 2002 and July 2007. Therefore, troughs may be more identifiable in aquifer levels near the st art of the droughts th at produced the record low pool stages at Weeki Wachee Spring.

PAGE 45

30 Borehole geophysical data included calip er, resistivity, induction, and video logs. Although less reli able than the methods listed above, well completion reports were also examined to identify potential permeable intervals. Intervals described as conduit, cavity, loss of circulation, or no recovery with a minimum height of 0.3 m were interpreted as cavity intervals. In this study, cavity heights rather than cavity diameters ar e used because caliper logs are not available for each borehole and lithologic decriptions are based on observations noted in the vertical dimension. The use of 0.3 m cavity intervals in this study does not imply that smaller features ar e unimportant. Indeed, much smaller karst features with apertures (> 1 cm) can exhibit turbulent flow (Worthington, et al. 2000b; White, 1988). Larger ka rst features (underwater caves) are included in the dual-conductivity model and smaller kars t features (vugs) are included in the laminar/turbulent equivalent-continuum model. Underwater cave survey data was provided by Karst Underwater Research, Inc. (2008a). The accuracy of cave survey data collected near Twin Ds Spring was verified using radioloc ation, which utilizes Faradays Law. A voltage was applied to a conductor (coil) placed in the conduit network which produced an electromagnetic field (trans mitter). An observer on the surface equipped with a receiver that uses near field very low frequency technology can identify the location of the transmitter below the surface by locating the position where the null occurs at the receiver. A detailed explanation of radiolocation is provided in Pease (1997). Radiolocat ion has not been performed at Weeki Wachee Spring. Changes in aquifer thickness interpre ted by Florida Ge ological Survey (2008) were compared to the location of inferred troughs in the potentiometric surface. Well/spring hydrographs were al so collected during the monitoring phase of this project to evaluate the matrix and conduit network responses to high recharge events.

PAGE 46

31 2.4. TASK 5 CHARACTERIZATION OF THE HYDRAULIC RESPONSE AND THE DIRECTION AND MAGNITUDE OF FLUID EXCHANGE BETWEEN THE MATRIX AND CONDUIT NETWORK A high frequency (15 minute) monitoring program was initiated to characterize the hydraulic response of the aquifer and the direction and magnitude of fluid exchange between the ma trix and conduit networks. Six wells (WW-F, WW-3, Weeki Wachee Deep, WWSpg-ECK, WHC#6, and WHC#7) and two vents (Twin Ds Spring and Weeki Wachee Spring) were instrumented with probes for the high frequency monitoring pr ogram. The WW-F well is also known as the Weeki Wachee F well (USGS site no. 283043082344101) and the WWSpg-ECK well is also known as the We eki Wachee Springs well (USGS site no. 283104082341801). The USGS site num ber for the Weeki Wachee Deep well is 283201082315601. Locations of t he monitoring sites are provided in Figure 2-1. Moreover, 4 additional wells (ROMP Centralia, ROMP 105, ROMP 98, and ROMP 97) that are part of the long-term Regional Observation Monitoring Program (ROMP) maintained by the SW FWMD are included in this study. The ROMP wells are located fart her from the springs relati ve to the 6 wells (WW-F, WW-3, Weeki Wachee Deep, WWSpg-ECK, WHC#6, and WHC#7) previously discussed (see Figure 2-1). The ROMP we lls are equipped with recorders that log hourly aquifer water levels.

PAGE 47

32 Figure 2-1. Location maps of test-sit e with location for springs and monitoring wells. The high frequency (15 minute) monito ring program lasted approximately 6 months and covered conditions prior to, during, and following the wet season of 2004. The 2004 wet season involved monito ring during normal convective storm

PAGE 48

33 activity and 2 tropical storm events (J ordan, 1984). Less frequent monitoring (primarily hourly aquifer water leve l measurements and monthly discharge measurements) continued through May 2006, which included drought conditions. The high frequency monitoring phase invo lved collecting: 1) water levels, temperature, and specific c onductance at 15 minute intervals in the matrix and conduit networks and 2) performing weekly to bimonthly discharge measurements at Twin Ds Spring. Hourly barometric pressure data for the study area were recorded at a National Oceanographic and Atmospheric Association (NOAA) station 12 km from Weeki Wachee (N ational Oceanographic and Atmospheric Association, 2004). Fift een minute rainfall was compiled by OneRain, Inc. The rainfall data combine Doppler radar estimates of rainfall distribution with rainfall quantities reco rded at local rain gauges. The lower frequency monitoring phase involved daily wa ter level measurements in a smaller number of monitoring wells and monthly discharge measurements at Twin Ds Spring. The number of wells monitored duri ng the 6 month period varied at a given time because 2 of the monitoring wells were constructed for public supply (WHC # 6 & 7, Figure 2-1), therefore monitoring ceas ed when the production wells went online. Additionally, one of the probes malfunctioned for an extended period of time (WW-3, Figure 2-1) during the monitoring phase. The sixth monitoring well (Weeki Wachee Deep) is maintained by the USGS (Figure 2-1). Three of the probes were initially se t to monitor param eters at 30 minute intervals, but were changed to 15 minute in tervals prior to the start of the 2004 wet season. Additionally, two of the ROMP well reco rders temporarily went offline. Geophysical logs were run on monitoring wells that had not been previously logged (Appendix A) Geophysical logs for m onitoring well WW-3 can be found in Hill and DeWitt (2004). Geophysica l or lithologic logs for the ROMP wells are available from the SWFWMD. Four of the high frequency monitoring wells monitored the matrix network withi n an 8 km radius of the springs and one

PAGE 49

34 of the monitoring wells (WW-F) breaches the roof of an underwater cave with a height of 8 m. The presence of the underwater cave was verified with video logging. Based on the geophy sical or lithologic logs the ROMP wells do not breach the roof of an underwater cave or feature comparable to that encountered at the WW-F well. Theref ore, in this study the open intervals for the ROMP monitoring wells are considered to be representative of the matrix. Once the caliper geophysical logging was complete, In-Situ Trolls (model numbers 4000, 8000, 9000, and a MiniTroll) equipped with pressure transducers and thermometers were deployed in the open intervals of the monitoring wells. The 9000 Trolls were also equipped with c onductivity meters. All instruments were verified to be operating within calib ration limits prior to deployment. The Trolls were deployed with vented cables, with the exception of the 9000 Trolls which were non-vented. Barometric pr essure was subtracted from the nonvented probes that recorded absolute pressure. Twin Ds and Weeki Wachee Springs were also equipped with instrumentation. Probes placed in the spring vents monitored temperature and conductivity at 15 minute intervals. Addi tionally, a gauge was installed to monitor spring pool elevations at Twin Ds Spri ng and is currently maintained by the SWFWMD. Spring pool elevations at W eeki Wachee Spring are maintained by the USGS, site id 02310500. The suite of probes deployed in the springs consisted of YSI sondes model numbers 600 XLM. The conduit well was also equipped with a YSI instrument because the In-Situ Troll dep loyed in that well was not equipped with a conductivity meter. The YSI sondes m onitored temperature, conductivity, and pressure. All sondes were verified to be operating within calibration limits prior to deployment. At Twin Ds Spring, the s hallower spring vent, the sonde was deployed at a depth of 5 m below pool stage in the spring vent while the sonde in the deeper Weeki Wachee Spring was deployed in the spring vent at a depth of 15 m below pool stage. The target depths were selected to ensure that the probes were recording samples repres entative of the spring discharge. On

PAGE 50

35 August 17, 2004, the sondes deployed in bot h spring vents had to be replaced as the original sondes were required el sewhere. At Twin Ds Spring, the replacement sonde was deployed to a depth of 5 m, while the sonde in Weeki Wachee Spring, which was rated for a lowe r pressure than the original sonde, was deployed to a depth of 3 m. The replacement sondes were verified to be operating within calibra tion limits prior to deploymen t. However, the conductivity meter on the replacement probe used for Twin Ds Spring began to decrease significantly recording the minimum specific conductance (151 S/cm) recorded for the duration of the m onitoring program. The meters, which actually measure conductivity, normalize the conductivity va lue to 25 C to provide specific conductance. There was no identifiable s ource for the decrease in specific conductance, in fact rainfall prior to and during the minimum were significantly less compared to rainfall from tropical storms that arrived later during the monitoring program. Near the conclusi on of the monitoring program, the conductivity meter in the replacement probe in Weeki Wachee Spring also appeared to drift for no apparent reason. The observed drifts in specific conductance were not corroborated by wa ter level changes, or temperature changes. Thus, the specific conductance data collected during the monitoring program are considered suspect and we re therefore, not included in the interpretations. Instrument specifications varied from .1 to 0.25 C for temperature (YSI, Inc., 2007; In-Situ, Inc., 1995; In-Situ, In c., 2000; In-Situ, In c., 2003a; In-Situ, Inc., 2003b). Verification of water level measurements using a hand tape at the wells indicate that recorded water levels are, on average, accurate to within 0.03 m. Discharge measurements were originally performed at Twin Ds Spring on a weekly basis, but were reduced to bi-monthly, and eventually monthly measurements as significant changes in discharge were not observed. Discharge measurements were performed 2 weeks be fore and 3 days after the passage of Tropical Storm Frances and 2 days before and 12 days after the passage of

PAGE 51

36 Tropical Storm Jeanne. Monthly dischar ge measurements were performed for a year and resumed 8 months later to capture discharges under low flow conditions. Measurements were temporar ily discontinued after a year because previous measurements had been performe d at similar pool stages. A rating curve was developed in this study to esti mate discharge at Twin Ds Spring using measured discharge values and water levels in the WW-F well (Figure 2-1). Discharge measurements were performed at Twin Ds Spring by setting up two transects immediately downstream (< 15 m) from the head spring at Twin Ds, Figure 2-2.

PAGE 52

37 Figure 2-2. Transect locations used to measure discharge at Twin Ds Spring. Arrows point to the spring vent. Top left photograph shows an overflow channel that is active during high and average flow conditions, but is inactive during low flow conditions. Bottom right photograph taken during the cessation of flow shows the location for the transect on the main spring run channel that is active during high, average, and low flow conditions. A calibrated measuring rod was used to measure depth at 0.15 m to 0.30 m increments (depending on the transect lengt h). Point velocities were measured at target depths of 60% below the water-air interface. This was done in an effort to avoid measuring reduced velocities produced by drag along the bottom, or higher velocities resulting from wind near the water-air interface. Total depths were typically below 0.9 m, but occasionally two measurements (depths of 80% Twin Ds Spring cessation of flow Twin Ds Spring cessation of flow Twin Ds Spring cessation of flow Twin Ds Spring high flow Twin Ds Spring high flow

PAGE 53

38 and 20% below the water-air interface) were averaged. Velocities were measured using a Marsh McBirney model 2000 portable flowmeter. The Marsh McBirney probe was zeroed to establish a ba seline prior to use by submerging it in a closed reservoir with no flow (Marsh McBirney, Inc., 1990). The depth(s) at each meas urement point along the transect was multiplied by the point velocity and summed for each transect. Reported discharges are the sum of total transects. Discharge measur ements were periodically verified by performing spot checks that consisted of duplicate measurements performed by USGS personnel using different instrum entation. Differences of 0.03 m3/s or less, were observed during spot checks. 2.5. Task 6 DEVELOPMENT OF THE 3 NUMERICAL GROUNDWATER FLOW MODELS Task 6 focused on development of the numerical groundwater flow models. Three types of groundwater flow mo dels were used in this study: 1) an equivalent-continuum with laminar flow, 2) an equivalent-continuum with both laminar and turbulent flow, and 3) a dual-conductivity model with both laminar and turbulent flow, as well as fluid exchange between the matrix and conduit networks. Detailed differenc es between the 3 types of groundwater flow models were discussed in Chapter 1. MODFLOW-DCM 2.0, a proprieta ry Darcian/non-Darcian, dualconductivity groundwater flow code develop ed by Southwest Research Institute (Painter et al., 2007) was in itially used in this project. However, MODFLOW-DCM 2.0 cannot simulate flow in multilayer aqui fers or tiered conduit networks in single layer aquifers (Painter et al., 2007). Si gnificant numerical instability during steady-state simulations was encountered using MODFLOW-DCM 2.0 in this study when the conduit network underly ing Twin Ds Spring was explicitly incorporated into the dual-conductivity model using cave survey data provided by Karst Underwater Research, Inc. (2008a) Conduit cells are represented using

PAGE 54

39 the minimum cell width in MODFLOW-DCM 2.0, which is a variant of MODFLOW-2000 (Painter et al., 2007). The steady-state simulation using MODLOW-DCM 2.0 converged when a less detailed representation of the conduit network (spread across multiple cells) underlying Twin Ds Spring was applied. Moreover, numerical instability was encountered in the dual-conductivity model during the transient simulation using MODFLOW-DCM 2.0, even when a less detailed representation of the conduit network (spread across multiple cells) underlying Twin Ds Spring was utilized. The transient dual-conductivity model using MODFLOW-DCM 2.0 faile d to reach normal completion and inconsistently failed to converge during various stress per iods in the transient simulation. The source for the instability was not resolv ed, but the comparable transient, laminar equivalent-continuum model usin g MODFLOW-2005 (Harbaugh, 2005) consistently achieved normal comple tions and did not demonstrate any numerical instability. All numerical m odeling efforts with MO DFLOW-DCM 2.0 in this study ceased. Instead, a public domain, Darcian/non-Darcian dualpermeability groundwater flow code (MODFLOW-2005 CFP) developed by the USGS (Shoemaker et al., 2008a) was used for laminar/turbulent simulations in this study. MODFLOW-2005 (Harbaugh, 2005) was used for the laminar equivalent-continuum models. The numerical instabilit y issues encountered with MODFLOW-DCM 2.0 were not encount ered with MODFLOW-2005 CFP. MODFLOW-2005 CFP version 1.2.01, compiled on February 12, 2008 (Shoemaker et al., 2008a) can operate in 1 of 3 modes. A brief description of the salient points of the code is provided below, but readers are referred to the MODFLOW-2005 CFP manual (Shoemaker et al., 2008a) for specific details on its use. MODFLOW-2005 CFP Mode 1 is used for dual-conductivity models and couples the groundwater flow equation to a discrete network of cylindrical pipes (Shoemaker et al., 2008a). Laminar flow in the cylindrical pipes is governed by the Hagen-Poiseuille equat ion and turbulent flow is governed by the DarcyWeisbach equation (Shoemaker et al., 2008a). The cylindrical pipe network can

PAGE 55

40 be fully or partly water-filled. Volumetric fluid exchange between the matrix and conduit networks is assumed laminar and is computed using (from Shoemaker et al., 2008a): ) (matrix conduit exchangehh Q (eq. 2-2) where: conduit conductance (L2T-1), head in conduit (L), head in surrounding matrix (L). MODFLOW-2005 CFP Mode 1 has two options for simulating conduit conductance (). The first option permits the user to assign a conduit wall conductance for each node in the cylindrical pipe network (Shoemaker et al., 2008a). The second option permits MODFLO W-2005 CFP Mode 1 to internally calculate the conduit wall conductance using the pipe geometry. This second option requires the user to insert the conduit wall permeability for each node (Shoemaker et al., 2008a). In this study, t he first option that pe rmits the user to assign the conduit wall conductance was utilized. Groundwater flow in the matrix using MODFLOW-2005 CFP Mode 1 is governed by the standard groundwater fl ow equation used in MODFLOW-2005. Additional parameters in cluded with Mode 1 are gr oundwater temperature, conduit locations, elevations, diameter, tortuosity, roughness, upper and lower Reynolds numbers, and direct recharge to the conduit nodes if applicable, such as for a test-site with a sinking stream (Shoemaker et al., 2008a). MODFLOW-2005 CFP Mode 2 was used for the laminar/turbulent equivalent-continuum models. Both laminar and turbulent flow can occur in Mode 2, but fluid exchange betw een the matrix and conduit networks does not occur. The conduit network is not represented by a cylindrical pipe network, but rather conduithmatrixh

PAGE 56

41 as vugs embedded in the matrix. Users assign a mean void diameter and upper and lower Reynolds numbers (Shoemaker et al., 2008a). Moreover, turbulent flow in Mode 2 is not governed by t he Darcy-Weisbach equati on. Mode 2 utilizes a turbulent hydraulic conductivity computed as a power function of the Reynolds number to simulate horizontal flow in preferential flow layers representing laterally extensive, well-integrated, c onduit networks consisting of vuggy porosity (Shoemaker et al., 2008a). Once the head difference between adjacent cells exceeds the critical gradi ent determined from the Rey nolds number, turbulent flow is invoked. Hydraulic conductivities for the appropriate cells are decreased producing a nonlinear relationship betw een specific discharge and hydraulic gradients (Shoemaker et al., 2008a). A third option, Mode 3, is a combination of both Modes 1 and 2 (Shoemaker et al ., 2008a). Only Modes 1 and 2 are evaluated in this study. Three pairs of combined steady-state/t ransient groundwater flow models were developed for the study area. Hydraulic conductivity, conductance between the matrix and conduit networks, and the lower and upper Reynolds numbers were the only parameters permitted to va ry between the 3 di fferent groundwater flow models. This was done to keep t he evaluation of model performances comparable. A detailed discussion of m odel development is presented in Chapter 4. 2.6. TASK 7 SENSITIVITY ANALYSIS A sensitivity analysis on boundary conditions and several model parameters was performed using the st eady-state and steady-state/transient dual-conductivity model. Net recharge, hydr aulic conductivity, well flow rate, and general head conductance were varied by adjusting calibrated values with multipliers of 0.01 and 100 following a modifi ed procedure of that used in Martin and Whiteman (1990). Additionally, general head values were evaluated by adjusting general head values to negativ e 1.52 m and positive 1.52 m from

PAGE 57

42 calibrated values. Sensitivity of net re charge, hydraulic conductivity, well flow rate, general head conductance, and general head values were evaluated in terms of the residual mean, residual st andard deviation, and residual sum of squares for 32 target wells. Sensitivity of drain conductances and pool stages for Twin Ds and Weeki Wachee Springs were evaluated in terms of their effect on simulated discharge. Sensitivity of drain conductances was per formed by varying calibrated values by multipliers of 0.01 and 100. The sensitivity of pool stage was evaluated by adjusting the observed pool stages by negative 0.30 m and posit ive 0.30 m from observed values. 2.7. TASK 8 EVALUATION OF THE MODEL PERFORMANCES Task 8 involved quantifying differences in the performance among the 3 combined steady-state/transient groundwater flow models. Differences in the model performances were evaluated by comparing: 1) observed discharge hydrographs following convective and tropical storm events, as well as drought conditions, to simulated dischar ges using MODFLOW-2005 and MODFLOW2005 CFP Modes 1 and 2, 2) observed and si mulated water levels for 32 target wells, 3) simulated water levels and head differences between the matrix and conduit network to observed values from monitoring wells penet rating the matrix and conduit network, and 4) model statisti cs in terms of the residual mean, residual standard deviation, and residual sum of squares for 32 target wells.

PAGE 58

43 CHAPTER 3: CONCEPTUAL MODEL OF STUDY AREA 3.0. BACKGROUND Twin Ds and Weeki Wachee Springs are two of several discharge points that comprise the Weeki Wachee Sp ring Group. The Weeki Wachee Spring Group differs from other first magn itude spring groups in the SWFWMD. Specifically, it consists of fewer vents and portions of explorable conduit networks. Weeki Wachee Spring, Twin D's Spring, Unnamed Spring No. 3, Salt, Mud, and Jenkins Springs are several of the vents that comprise the spring group. Unlike Twin D's and Weeki Wachee Springs, Unnamed Spring No. 3, Salt, Mud, and Jenkins, discharge brackish wa ter and are tidally influenced (Champion and Starks, 2001). The tidally influenced v ents are located a few miles west of Weeki Wachee and Twin Ds Springs, Figure 3-1. Since density effects are not addressed in this study, focus is di rected on Twin D's and Weeki Wachee Springs.

PAGE 59

44 Figure 3-1. Location and si te map of study area. Prior to the initiation of the sp ring characterization and monitoring program, only sporadic discrete measurements of discharge had been performed at Twin D's Spring (Champion and Star ks, 2001) and no rating curve had been developed to estimate discharge at Twin Ds Spring, prior to this study. The spring basins have not been rigor ously delineated with dye-trace testing and historically have been delinea ted using the potentiometric surface (Figure 3-2, Jones et al., 1997). Therefore, ther e is uncertainty associated with the size of the spring basins. W eeki Wachee is a first magnitude ( 3 m3/s) spring (Scott et al., 2004; Meinzer, 1927) and Twin Ds, based on average flow conditions from June 2004 through Ma y 2006, is a third magnitude ( 0.3 m3/s)

PAGE 60

45 spring (Meinzer, 1927). During the monito ring period, discharge at Twin Ds Spring occasionally exceeded 0.3 m3/s during high recharge events. Figure 3-2. Previous interpretation of sp ring basin boundaries (from Jones et al., 1997). Based on the previous interpreta tion, groundwater flow in the Weeki Wachee Spring basin generally flows in a northwest direction. Average discharge for Twin Ds and Weeki Wachee Springs during the monitoring period was 0.2 m3/s and 5 m3/s, respectively. The spring pool at Twin D's Spring during the study period was typically 0.76 m higher than the spring pool at Weeki Wachee Spring. During t he period of record for Weeki Wachee Spring, which dates back to the late 1920s, a cessation of flow has not been observed, however during the course of this project a cessation of flow was documented at Twin D's Spring. The study area encompasses approximately 1,479 km2. It includes the central and western portions of Hernando County and t he northwestern and

PAGE 61

46 central northern portions of Pasco County. The Gulf of Mexico bounds the western portion of the st udy area (Figure 3-3). Figure 3-3. Delineation of study area. 3.1. HYDROGEOLOGIC SETTING OF STUDY AREA The next few sections discuss the hydrogeologic setting for the study area. This includes a discussion of t he physiographic regions, soils, land use, water use, karst, water qua lity, stratigraphy, and the hydrogeologic framework for the study area. Additional background info rmation about the test-site, such as previous estimates of transmissivities, tracer tests, and delineation of the capped, recharge, and discharge zones are discussed.

PAGE 62

47 3.1.1. Physiographic Regions The study area includes the Coastal Swamps, North Gulf Coastal Lowlands, and the southern half of the Brooksville Ridge physiographic regions (shown in Figure 3-4) and is bounded by the offshore Drowned Karst physiographic region to the west (White, 1970). The Coastal Swamps physiographic region includes the areas consisting of continuous swamps coastward. The swamps occur adjacent to sand-starved coastal areas where transgression occurred directly over expos ed limestone (White, 1970). The North Gulf Coastal Lowlands consists of mari ne terraces interpreted as paleoshorelines that reflect higher sea levels of Pa mlico (9 m NGVD) and Wicomico (+30 m NGVD) ages (White, 1970). The presence of the terraces has been interpreted as an indication that the site has transit ioned from a sand-rich coast to a sandstarved coast (White, 1970). The southern half of the Brooksville Ridge, which is roughly 97 km in length, varies from 16 to 24 km in wid th (White, 1970). The highest elevations (from 53 to 61 m NGVD) follow a linear northwest-southeast trend and occur on the western flank of the ridge. The Brooksv ille Ridge is capped, for the most part, with Miocene age siliceous sands and clays, which reduced denudation rates relative to the surrounding uncapped, lowlan d regions (i.e. the North Gulf Coastal Lowlands and the Western Valley to the east of the Brooksville Ridge; White 1970). A breach in the Miocene age sands and clays is present on the southern half of the Brooksville Ridge in the nort heastern portion of the study area. Airfilled caves are found on t he breached portion of the Brooksville Ridge (Florea, 2006a, Florea 2006b).

PAGE 63

48 Figure 3-4. Physiographic regions in the vicinity of study area. 3.1.2. Soils The soils in the study area primarily consist of well drained, low water holding capacity soils (Figure 3-5). Howe ver, poorly drained, high water holding capacity soils occur on the Brooksville Rid ge and in the Coastal Swamps (Hyde et al., 1977).

PAGE 64

49 Figure 3-5. Soils in the study area (m odified from Hyde et al., 1977). Brown and light green colors are inte rmediate transitional zones. 3.1.3. Land Use Land use in the study area ranges from upland forests and agriculture on the eastern side of the study area to urban and wetland areas on the western section. Additionally, limestone quarries lie in the north central portion of the study area (Jones et al., 1997). 3.1.4. Water Use Water uses in the study area inclu de agricultural, industrial, mining, domestic self-supply, public supply, and recreational uses. Water use estimates

PAGE 65

in the study area have traditionally consisted of reported and estimated quantities (Southwest Florida Water Management District, 2004). Total water use estimates are discussed in Chapter 4. 3.1.5. Karst The study area is characterized by inte rnal drainage, a water table that is upwardly concave (Figure 3-6), springs water-filled caves, and sinkholes. The study area has been overprinted with multiple episodes of karstification during the Cenozoic Era in response to sea level fluctuations, (Florea, et al., 2007) that has resulted in tiered passages and co mplex cave geometries. Wilson (2002) and other cave divers (Karst Underwater Research, Inc., 2008b) describe the wet, water-filled caves as consisting of predominately horizontal ellipticallyshaped passages that are generally subparallel to depositional layering with relatively fewer passages along vertically elongated fractures. Figure 3-6. Water table surface map for the study area using May 2004 water levels based on kriging the location and water levels of 67 wells. Arrow indicates approximate location of Weeki Wachee Spri ng. Inset is the location for the 67 wells used to generate the surface map for this study. 50

PAGE 66

51 Air-filled caves immediately north of the northern boundary for the study area have been described as forming at intersecting fractures (Florea, 2006a; Brinkmann and Reeder, 1994) and laterally ext ensive cavities resulting in a plussign shape (Florea, 2006a). Some of the ca ves are believed to have formed at the water table (Florea, et al., 2007). It is also conceivable that some of the large oval, or circular chambers may have originated at former mixing zones (Reeder and Brinkmann, 1998) as their descriptions resemble the flank margin caves discussed in Mylroie and Carew (2000). Today, dry, air-filled caves lie on the topographically high Brooksville Ridge and wet, water-filled caves lie in the topographically low, North Gulf Coastal Lowlands, Coastal Swamps, and offshore Drowned Karst physiographic regions. Figure 3-7 is an inv entory of known karst features in the study area. Karst features include: springs, water-filled caves, and sinkholes such as Peck, Blue, Hernasco, and Bear (sinkholes show n in Figure 3-7) that are fairly well connected to the underlying karst aquifer (Trommer, 1987). Many of the springs in the study area are proposed to be fo rmer sinkholes (recharge points) that reversed into focused discharge points in response to sea level rise (Upchurch and Randazzo, 1997).

PAGE 67

52 Figure 3-7. Location of karst f eatures in the study area. Soluble carbonates lie below a thin m antle that varies from less than a meter to 61 m in thickne ss based on review of 241 lithologic logs. Sinkholes, or vertical conduits, are numerous in the st udy area. Figure 3-8 is a surface map of the underlying soluble carbonates.

PAGE 68

53 Figure 3-8. Interpretation of buried limest one surface at test-site based on kriging of 241 lithologic logs. Inset is location of wells used to generate the surface map for this study. 3.1.6. Water Quality Geochemically, discharge from Twin Ds and Weeki Wachee Springs is very similar (Jones et al., 1997). The arithmetic mean of Ca/Mg ratios, which is an indication of the type of rock groundwat er is flowing through (White, 1988), for water samples collected from Weeki Wach ee Spring from July 1994 through July 2002 and Twin D's Spring from July 1994 through January 1995 is 9.5, which suggests that flow is primarily thr ough limestone rather than dolostone. The coefficient of variation of total bica rbonate, or hardness, for water samples collected from Weeki Wachee Spring is 5%, which suggests the spring is diffuse (White, 1988). Although Weeki Wachee Spring is conduit fed, it is possible to retain diffuse flow characte ristics when the spring basins ar e large, as is likely the

PAGE 69

54 case for Weeki Wachee Spring, or when ther e is a substantial contribution of flow from the matrix (White, 1988). 3.1.7. Stratigraphy The study area consists of, from top to bottom, undifferentiated sands, the Hawthorn Group, the Suwan nee Limestone, Ocala Limestone, and Avon Park Formation (see Table 3-1; Miller, 1986). 3.1.7.1. Undifferentiated sands Scott (1997) termed the Post-Miocene st rata described in Miller (1986) as undifferentiated sands. The undifferentiated sands are reworked terrace deposits that consist of quartz sand (Scott, 1997) with interbedded residual clays as evidenced by review of lithologic logs. Th icknesses of the undifferentiated sands vary over the study area from less than a meter to 61 m in th ickness, in karst features, based on driller repor ts reviewed in this study. 3.1.7.2. Hawthorn Group In the southern and eastern portions of the study area, along the southern half of the Brooksville Ridge, the Hawt horn Group lies below the undifferentiated sands (Miller, 1986). The Hawthorn Group c onsists of interbedded clay, sand, and silt. Phosphate deposits are also pres ent and vary in concentration. The Hawthorn Group is absent in the central and western portions of the study area (Miller, 1986). Thickness of the Hawthorn Group varies from 0 to 23 m and it occurs at elevations ranging from 23 to 30 m NGVD (Florida Geological Survey, 2008). Recently, the Hawthorn Formation has been designated as the Hawthorn Group, rather than the Ha wthorn Formation (Scott, 1997).

PAGE 70

55 3.1.7.3. Suwannee Limestone The Suwannee Limestone is a weathered, fossiliferous marine carbonate (Miller, 1986). Moldic porosity coupled wi th extensive weathering makes the Suwannee Limestone highly permeable in the study area. Drillers often note a loss of circulation as they approach/br each the top of the Suwannee Limestone. The Suwannee Limestone has been removed by erosion from the northwestern section of the study area. Thickness of the Suwannee Limestone varies from 0 to 46 m in the study area and occurs at el evations ranging from 0 to 23 m NGVD (Florida Geological Survey, 2008). 3.1.7.4. Ocala Limestone The Ocala Limestone has been described as consisting of the Inglis, Williston, and Crystal River Formations. Mill er (1986) concludes that the Inglis, Williston, and Crystal River Formations are not distinguishable and are lithologically similar. He de scribes the Ocala Limestone as a, white, generally soft, somewhat friable, por ous coquina . loosely bound by a matrix of micritic limestone, (Miller, 1986). The Ocala Limest one also contains many cavities and conduits and therefore is the most permeable strata in the study area. Thickness of the Ocala Limestone varies from 30 to 61 m in the study area and occurs at elevations of 0 to -46 m NGVD (F lorida Geological Survey, 2008). 3.1.7.5. Avon Park Formation The Avon Park Formation is descri bed as a chalky, microfossiliferous, carbonate possessing localized, intergranular evaporites. The Avon Park Formation varies from a limestone to a dolostone (Miller, 1986). Intergranular evaporites consist of gypsum and anhydrite. The top of the Avon Park Formation occurs at elevations of -30 to -91 m NG VD (Florida Geological Survey, 2008).

PAGE 71

56 3.1.8. Hydrogeol ogic Framework The term aquifer in this report refe rs to a porous medium that is perennially or ephemerally saturated with water and excludes fine-grained, low hydraulic conductivity media. Using this def inition, two aquifers exist in the study area: the surficial and Upper Floridan aquifers. Semi-confining and low permeable strata occur in the Hawthorn Group and middle confining unit II. The surficial aquifer consists of the saturated portions of the undifferentiated sands. It is discontinuous or ephemeral throughout the central and western portions of the study area, where sands have a low water holding capacity. A distinct perennial surficia l aquifer is present however, on the Brooksville Ridge where the Hawthorn Group is present as evidenced by hydrographs for nested well sites (Appendix B). The Floridan aquifer system is subdivided into the Upper Floridan aquifer and Lower Floridan aquifer (Miller, 1986). Th e Upper Floridan aquifer in the study area consists of the Suwannee Limestone, Ocala Limestone, and the Avon Park Formations (Miller, 1986) and has a thick ness ranging from 188 to 300 m (Florida Geological Survey, 2008). The base of t he Upper Floridan aquifer is demarcated by the middle confining unit II (MCU II) des cribed in Miller (1986) which refers to the low permeable horizon in the Avon Park Formation where continuous intergranular evaporites exist. The top of the middle conf ining unit II occurs at elevations of -293 m to -274 m NGVD in the study area (Florida Geological Survey, 2008). The Lower Floridan aquifer is absent in the study area. The hydrogeologic framework for the st udy area is summarized in Table 3-1.

PAGE 72

57 Table 3-1. Hydrogeologic framew ork (modified from Miller, 1986). Thickness of the Upper Flori dan aquifer based on the delineation interpreted by the Florida Geological Survey (2008), used in the groundwater flow models discussed in Chapter 4, is provided in Figure 3-9. The Florida Geological Survey (2008) inte rpolates the top of the Upper Floridan aquifer using borehole data from wells that penetrate ve rtically consistent carbonates lacking an overlying thick mantle of clastics. Using this criteria smoothes the surface irregularities displayed in Figure 3-8. Maximum thickness occurs in the southeastern section of the study area with minimum thickness occurring in the northwestern and northeastern portions of the study area. The surficial aquifer was set equal to the top of land surface elevation using a compilation of USGS digital elevation models (Labins, 2003) minus the top of the Upper Floridan aquifer. The Hawthorn Group was assigned a thickness of 0.3 m and was placed Epoch Stratigraphic Unit Description Aquifer/Confining Unit Model Layer PlioceneRecent Undifferentiated Sands quartz sand & residual clays surficial aquifer 1 Miocene Hawthorn Group clays, silts, sands, & phosphates semi-confining unit mantle 2 Oligocene Suwannee Limestone weathered, fossiliferous limestone Ocala Limestone friable coquina in a matrix of micritic limestone microfossiliferous carbonate Upper Floridan aquifer 3 Eocene Avon Park Formation dolomitic-limestone with intergranular evaporites MCU II lower horizontal no-flow boundary

PAGE 73

58 between the surficial aquifer and Upper Flor idan aquifer. In the west and central portions of the study ar ea, where the Hawthorn Group is absent, the layer represents residual, weathered clays for the undifferentiated sands (Goddard, 1998). Elevations for the top of the Upper Floridan aquifer were decreased by 0.3 m where the Florida Geological Survey (2008) delineations ex ceed the top of land surface in the digital elevation model. Total thickness for the surficial aquifer, Hawthorn Group, and Upper Floridan aquif er is provided in Figure 3-9.

PAGE 74

59 Figure 3-9. Thickness of Upper Flori dan aquifer (top), modified from (Florida Geological Survey, 2008) and thickness of the surficial aquifer, Hawthorn Group, and the Upper Floridan aquifer with land surface superimposed (bottom). Land surface elevation modified from Labins ( 2003). Arrows point to the approximate location of Weeki Wachee and Twin D's Springs.

PAGE 75

60 3.1.9. Previous Estimates of Transmi ssivity and Storativity Based on Aquifer Performance Tests and Flow Net Analysis Transmissivites estimated from aqui fer performance tests (PW-4, PW-5, N12, shown on Figure 3-10) or flow net analysis (Weeki Wachee Springs, shown on Figure 3-10) that assume isotropic, homogeneous, laminar flow conditions can be highly variable (i.e. 11,958 to 147,460 m2/day; Hydro-Environmental Associates, Inc., 2003; Gilboy and Moore, 1982a; Sinclair, 197 8) in the study area. This variability is due to secondar y permeability, the quality or conditions during an aquifer performance test, and t he analytical solution methods used to interpret aquifer performance tests, which usually assume laminar flow, that may or may not be valid at a particular site depending on the scale of karst features relative to the scale of an aquifer perfo rmance test. Because of the assumptions and observations noted during the aquifer per formance tests or flow net analysis, low weight was applied to them in this study as accurate indicators of transmissivity. Locations for additional aquifer performance tests not shown on Figure 3-10 are availa ble from the SWFWMD. Storativity of the Upper Flori dan aquifer estimated from aquifer performance tests in the study area Figure 3-10, varies from 0.002 to 0.0007 (Hydro-Environmental Associates, Inc., 2003; Gilboy and Moore, 1982a). Low weight was applied to t he storativity estimates obtained from the aquifer performance tests for the reasons previous ly discussed. Storage values of 0.05 for the Upper Floridan aquifer and 0.15 for the Hawthorn Group and surficial aquifer were used in the transient st ress periods for the groundwater flow models. These are consistent with values used in a previous dual-porosity model of the Upper Floridan aquifer north of the study area (i.e. GeoT rans, Inc., 1988a).

PAGE 76

61 Figure 3-10. Location of esti mated transmissivities in m2/d (top) and estimated storativities (dimensionless) shown at bottom. Values compiled from HydroEnvironmental Associates, Inc., 2003; Gilboy and Moore, 1982a; Sinclair, 1978.

PAGE 77

62 3.1.10. Tracer tests No known successful quantitative dy e-trace tests have been performed near Weeki Wachee Springs. There is anec dotal evidence from an accidental tracer test that suggests that a high degree of connection may exist among south-southeast trending conduits and Weeki Wachee Spring. In March of 1976, the Deltona Corporation wa s excavating Century Lake. Water from Century Lake was being pumped to Crescent Lake which is connected to Century Lake via a canal. On or about March 19, 1976, water levels in Crescent Lake began to drop even though water from Cent ury Lake was being pumped in to Crescent Lake. It was suspected that a sinkhole in the lake bottom may have opened due to the loading. Crescent Lake is located approx imately 3 km south-southeast of Weeki Wachee Spring, Figure 3-11. On March 21, 1976, water quality at Weeki Wachee Spring began to visibly degrade. Maximu m degradation of water quality occurred on March 23, 1976. This is the only documented even of cloudiness, or high turbidity reported for Weeki Wachee Spring during the period of record that dates back to the late 1920s (Court Complain t, 1976). The location of the suspected sinkhole beneath Crescent Lake is not known.

PAGE 78

63 Figure 3-11. Location of Crescent Lake relative to Weeki Wachee Spring. Jones et al. (1997) report an attempted dye-trace test at an underwater cave (Diepolder) located approximately 4 km east of Weeki Wachee Spring. Twenty-three kilograms of optical bright ener were released into the underwater cave, but the dye was lost and was never recovered (Jones et al., 1997). However, water samples from nearby monitoring wells were not tested, only Weeki Wachee and Twin Ds Springs were sampled for the optical brightener. Additional limitations with the experiment al design of the dye-trace test can be found in Jones et al. (1997).

PAGE 79

64 3.1.11. Recharge The area from Weeki Wachee Spring westward to the Gulf of Mexico is a discharge zone for the Upper Floridan aquifer. The region east of Weeki Wachee Spring is a recharge zone with internal dr ainage. Diffuse recharge occurs in the undifferentiated sands and point recharge oc curs at a few sinkholes (Trommer, 1987; shown in Figure 3-7). The Upper Florid an aquifer varies from unconfined to semi-confined conditions in the study area due to the presence or absence of the Hawthorn Group, semi-confining residual cl ays in the undifferentiated sands, or variation in vertical hydraulic conducti vities for the Upper Floridan aquifer. Lithologic, hydrologic, and physiographic boundaries were used to delineate the capped, recharge, and discharge zones (Figure 3-12). Hydrographs for nested well sites monitoring the surficial and U pper Floridan aquifer are provided in Appendix B. Note the degr ee of hydraulic separati on between the surfical and Upper Floridan aquifer wells located in the capped zone, relative to the coincidence of surficial and Upper Flori dan aquifer levels for the recharge and discharge zones.

PAGE 80

65 Figure 3-12. Location of capped, recharge, and discharge zones with nested well sites. Note the breached portion of the capped zone. Rainfall data for the study area are currently collected using two sources: rainfall gages and Doppler Radar. OneRain, Inc. combines Doppler Radar, which is a representation of rainfall distributi on, and calibrates rainfall quantities using rain gages distributed throughout the st udy area. Hoblit and Curtis (2005) demonstrate that calibrated Doppler Radar provides a more accurate spatial distribution of rainfall versus Thiessen pol ygons or kriging techniques, therefore data sets compiled and processed by OneR ain, Inc. are used in this study for rainfall estimates. Previous sensitivity analysis on regional groundwater flow models demonstrate that recharge is a sensitiv e model parameter in the Weeki Wachee area, (Yobbi, 2000). Average annual rainfall is 137 cm over the test-site (Knochenmus and Yobbi, 2001). Evapotranspiration losses estimated fo r the southern and central portions of the SWFWMD using the HSPF model (Geurink et al., 2000) that were later

PAGE 81

66 compiled for the SWFWMD by HydroGeoLogic, Inc. (2007) yield an annual evapotranspiration rate of 91 cm (HydroGeoLogic, Inc., 2007). This evapotranspiration rate is comparable to ranges (69 to 107 m) estimated by Knochenmus and Yobbi (2001). The Hydr oGeoLogic, Inc. (2007) estimate of evapotranspiration losses, coupled with the OneRain, Inc. recharge estimates, yields average annual net recharges over th e study area of 66, 51, and 23 cm for 2004, 2005, and 2006, respectively. 3.2. DOES NON-DARCIAN FLOW O CCUR IN THE UNDERLYING CONDUIT NETWORK? Anecdotal evidence from cave divers indi cate that turbulent flow occurs in portions of the conduit networks under lying Weeki Wachee and Twin Ds Springs. Divers describe portions of the conduit network underling Twin Ds Spring where constrictions occur as having raging flow (Karst Underwater Research, Inc., 2008c). Discharge velocities through the vent at Weeki Wachee Spring under average flow conditions are so high that divers are not able to access the conduit network except when flows are abnormally low (Karst Underwater Research, Inc. 2008b). Darcian or laminar conditions require that the relation between discharge and hydraulic gradient be: i) li near and ii) that the intercept occur at (0,0). That is, in the absence of a hydraulic gradient there should be no discharge. A plot of specific discharge versus hydraulic gradient between the WW-F well and Weeki Wachee Spring illustrates an intercept that deviates from (0,0) under low gradient conditions for conduit diameters of 0. 9 m, suggesting that non-Darcian, or turbulent flow may occur, see Figure 3-13. A similar plot for Twin Ds Spring demonstrates a change in slope (nonlinearity) for 0.9 m diameter conduits under low gradient conditions. Due to continui ty, larger conduit diameters of 5 m illustrate linearity under low gradient conditions, suggesting Darcian, or laminar flow may occur through larger passages. Reynolds numbers greater than 2,300

PAGE 82

67 indicate non-laminar flow in smooth walled pipes (White, 1988). In a rough walled conduit with many asperities, the Reynolds number for turbulent flow theoretically would be lower than 2,300 (White, 1988). In th is study, estimates of the Reynolds numbers for Weeki Wachee Spring and Twin Ds Spring are 106 and 105, respectively. The Reynolds numbers estimated in this study are a first approximation that when coupled with the anecdotal evidence from cave divers indicates that non-Darcian flow, and therefore the use of a Darcian/non-Darcian groundwater flow simulator may be applicable for the test-site. Data used to estimate the Reynolds numbers are provided in Appendix C. Figure 3-13. Plots of specific dischar ge versus hydraulic gradient for Weeki Wachee Spring (top) and Twin Ds Spring (bottom).

PAGE 83

68 3.3. CONDUIT CONCEPTUALIZATIONS A conceptualization of conduit locations and preferential flow pathways in the study area was developed by applying several of the traditional methods discussed in Chapter 1. In this study conduit locations and preferential flow pathways are interpreted where multiple types of data coincide. The term conduit in this study implies turbulent or lami nar groundwater flow as defined in Field (2002). Data used to interpret preferent ial flow pathways include: a physical inventory of karst features (shown in Figur e 3-7), fracture trac es inferred from the alignment of closed topographi c depressions, troughs in water levels within the Upper Floridan aquifer, elevation modes for conduits interpreted from borehole porosity descriptions, caliper logs, and su rvey data from underwater caves, and hydrogeologic data consisting of changes in aquifer thickness, and well/spring hydrographs. 3.3.1. Fracture Traces Fracture traces were interpreted over the study area in a previous study by Jones et al., (1997). In that study, the authors delineated fracture traces by identifying closed topographic depre ssions with minimum areas of 0.08 km2 Figure 3-14. The fracture traces interpre ted in the Jones et al. (1997) study indicate that a fairly well-integrat ed, conduit network underlies the Brooksville Ridge and that a regional-scale conduit network underlies the study area. A single fracture trace verified using vertical electrical soundings (Wood and Stewart, 1985) is included in Figure 3-14.

PAGE 84

69 Figure 3-14. Inferred fracture traces modified from J ones et al. (1997). Verified fracture trace modified fr om Wood and Stewart (1985). 3.3.2. Troughs in Aquifer Water Levels Minimum pool stages for Weeki Wachee Spring, which date back to the late 1920s, occurred during June 2002 and July 2007. Troughs in water levels were interpreted from published USGS c ontours of water levels in the Upper Floridan aquifer during May 2001 (Duerr, 2001) and May 2006 (Ortiz, 2007) near the start of the droughts t hat contributed to the minimum spring pool elevations. The 3 m contour interval us ed for the interpreted potentio metric surface maps is considered adequate for the purpose of th is study. The Weeki Wachee Trough in the vicinity of Weeki Wachee and Twin Ds Springs is clearly visible in the May 2001 potentiometric surface m ap. The Weeki Wachee Trough is also visible in the May 2006 potentiometric surface map and a second trough near the western boundary of the Brooksville Ridge is also distinguishable. The two troughs are merged in the May 2001 potentiometric surface map forming a single large trough, Figure 3-15. Overall, the location of the troughs coincide fairly well with

PAGE 85

70 decreases in Upper Floridan aquifer thi ckness as interpreted by the Florida Geological Survey (2008) see Figure 315. The Weeki Wachee Trough is also visible during water level highs in September 2004, but is less pronounced. There are uncertainties associated with the aquifer water level contours. Average monthly pool stages at Twin Ds and Weeki Wachee Springs are 3.70 m and 3.01 m, respectively for September 2004, however the 3 m contour shown in Figure 3-15c indicate that pool stages are less than 3 m. Additionally, the average monthly pool stage at Twin Ds Sp ring, even during a cessation of flow in May 2006, was 3.13 m, yet the contours shown in Figure 3-15b indicate it was less than 3 m. The proposed conceptual model of conduit locations and preferential flow pathways is developed using multiple types of data because of these and additional uncertainties.

PAGE 86

a b c d 71 Figure 3-15. Potentiometric surface maps for the Upper Floridan aquifer (UFA) near start of droughts for a) May 2001 (Duerr, 2001) and b) May 2006 (Ortiz, 2007). Potentiometric surface map for September 2004 (Blanchard and Seidenfeld., 2005) at high water level condi tions is provided for comparison (c). Troughs are less distinguish able during water level highs in September 2004, however the Weeki Wachee Trough is still visible. Contour interval is 3 m. Datum is NGVD. Location of Weeki Wachee Trough relative to aqui fer thickness (d) based on delineations interpreted by t he Florida Geological Survey (2008).

PAGE 87

72 3.3.3. Borehole Porosity Descriptions Statistical information pertaining to cavity heights and elevations were compiled by reviewing 320 lithologic and/or geophysical logs (Figure 3-16) from the Regional Observation Monitoring Program and well driller reports for the study area. Review of the lithologic logs indicate that 18% of the 320 logs in the study area intercept a cavity greater than or equal to 0.3 m in height (Figure 316). The probability of intercepting a cavity greater than or equal to 0.3 m is 10%. This probability compares favorably to an estimate based on the inventory of karst features (6%) shown in Figure 3-7 and is substantially higher than the low probabilities encountered (from 0.4 to 3%) in a Paleozoic limestone aquifer in Kentucky (Worthington, et al. 2000b). This estimate is also lower than the probabilities (from 25-50%) reported in Wilson (2002) for the Floridan aquifer.

PAGE 88

73 Figure 3-16. Location of wells with review ed lithologic logs (top) and location of wells intercepting a cavity 0.3 m in height (bottom). The well density (0.22 well/km2) in the western portion of the study area (North Gulf Coastal Lowlands) is nearly double the well densit y on the Brooksville Ridge (0.13 well/km2) because of the relatively higher degree of development in the southwestern portion of the st udy area. Ratios of large ( 1.5 m in height) to small (< 1.5 m in height) cavities interc epted for the North Gulf Coastal Lowlands and Coastal Swamps relative to the Brooksville Ridge, based on the borehole porosity descriptions, indicate a slightly higher degree of connection in the

PAGE 89

74 conduit network relative to the Brooksville Ridge (Figure 3-17). This result is corroborated by the inventory of karst features that indicate a relatively higher density of karst features in the western por tion of the study area, see Figure 3-7. Figure 3-17. Ratio of large ( 1.5 m in height) and small (< 1.5 m in height) cavities intercepted on the Brooksvill e Ridge (top) and North Gulf Coastal Lowlands and Coastal Swamps (bottom). 3.3.4. Cave Survey Data Exploration and mapping of the conduit networks underlying Twin Ds and Weeki Wachee Springs have been advanced by Karst Underwater Research, Inc. (2008b; 2008c). Discharge at Weeki Wa chee Spring occurs via a single first magnitude vent along a vertical fractu re with an elbow of approximately 45

PAGE 90

75 occurring at a depth of approximately 42 m (Wetterhall, 1965) or 56 m as reported in Jones et al. (1997) or 44 m as report ed in Karst Underwater Research, Inc. (2008b). The elbow conti nues to a depth of approximately 62 m (Jones et al., 1997) or 50 m (Karst Under water Research, Inc. 2008b) where it opens into a large room. Approximatel y 2 km of the Weeki Wachee conduit network have been surveyed by Karst Underwater Research, Inc. (2008b) which extended the original distance of explor ation performed by Sheck Exley in the 1980s (Karst Underwater Research, Inc. (2008b). Explorat ion of the conduit network at Weeki Wachee Spring has hist orically been hindered due to the large velocities and highly turbulent discharge at the spring vent, which under average flow conditions is too high for divers to penetrate. Karst Underwater Research, Inc. (2008b) was able to explore more of the conduit network due to relatively lower discharge velocities associated with the 2006-2007 drought. The full extent of the conduit network has not been expl ored. Based on the limited available data, it appears that much of the explored conduit network at Weeki Wachee Spring consists of large, horizontal elli ptically-shaped conduits sub-parallel to depositional layers, with the exce ption of the vent which follows a vertical fracture (Karst Underwater Research, Inc., 2008b). The maximum relative frequency for t he elevation of surveyed passages for the conduit network underlying W eeki Wachee Spring was 28% at -79 m NGVD followed by 26% at -73 m NGVD, (Figure 318). Comparison with the borehole data in the North Gulf Coasta l Lowlands and Coastal Swamps (Figure 3-18) displays shared conduit elevations at -73 and -79 m NGVD, but at a lower relative frequency (< 4%). Twin Ds Spring is fed by a local c onduit network with rooms that exceed 30 m in diameter (Champi on and Starks, 2001). Entrance to the conduit network is through a vertical circular conduit 0.9 m in diameter (Karst Underwater Research, Inc, 2008c). Two vents exist at Twin Ds, but one of them is currently occluded with debris, therefore all discharge, when flowing, occurs from a single vent.

PAGE 91

76 Approximately 1 km of the conduit net work underlying Twin Ds Spring has been mapped, based on cave survey data provided by Karst Underwater Research, Inc. (2008a), but the total ext ent of the conduit network has not yet been explored. Cross-sections of surveyed passages at Twin Ds suggests that the geometry of the netwo rk has been influenced by both fracture sets and bedding (Karst Underwater Research, Inc. 2008a). Narrow, vertically elongated passages and horizontal elliptically-shaped passages sub-parallel to depositional layering have been surveyed (Karst Underwater Research, Inc., 2008b; Champion and Starks, 2001). According to Palmer (1991) conduits forming along bedding planes and closely spaced joints within specific beds would yield branchwork and anastomotic passages. Wilson (2002) described many of the water-filled caves in Florida as anastomotic. The maximum relative frequency for t he elevation of surveyed passages based on the cave survey data for the conduit network underlying Twin Ds Spring was 18% and occurred at an elevat ion of -37 m NGVD (Figure 3-18). Comparison with the borehole data in the North Gulf Coastal Lowlands and Coastal Swamps displays some shared conduit elevations, but at a lower relative frequency (< 4%).

PAGE 92

Figure 3-18. Modes of elevat ion for intercepted cavities from borehole porosity descriptions and surveyed underwater caves. Bin interval equals 6 m. Elevations reported for Weeki Wachee Spring exclude those from the vertical fracture entrance (fissure). Additionally, survey data noted as "bogus" by cave divers in the Weeki Wachee Spring survey data are omitted. Cave survey data for Weeki Wachee and Twin D's Springs provided by Karst Underwater Research, Inc. (2008a). The principle axes of orientation for the surveyed portion of Weeki Wachee Spring (not including the vertical fracture at the entrance or the WW-F segment) based on cumulative passage length using azimuth bin intervals of 10 degrees is 210-220 and 130-140. The prin ciple axes of orientation for the surveyed portion of Twin Ds is 200-210 and 160-170. To date, an underwater cave connec tion between the conduit networks underlying Twin Ds and Weeki Wachee Springs has not been identified by cave divers (Karst Underwater Research, Inc. 2008b). Nor has an underwater cave 77

PAGE 93

78 connection been made between the WW-F conduit well and Twin Ds or Weeki Wachee Springs (Karst Underwater Research, Inc. 2008b). Radiolocation was performed to verify the cave survey data for the conduit network underlying Twin D's Spring. Radiolocation can be used to verify the extent of the entire surveyed conduit network, but for the purposes of this study it was adequate to verify the survey data at only two locations, since the conduits will not be represented as accurately in the numerical groundw ater flow models. Figure 3-19 displays the projected c onduit network on land surface and the two verification sites. The projection of the conduit network on land surface satisfactorily matches the target underwater passages. Radiolocation has not been performed at Weeki Wachee Spring.

PAGE 94

79 Figure 3-19. Projected condui t networks for Twin Ds and Weeki Wachee Springs on land surface with radiolocation verificati on sites for Twin Ds Spring. Only the major passages for Weeki Wachee Spring conduit network are shown. Survey data for the conduit network underlying Weeki Wachee Spring is provisional. Twin Ds Conduit Network Weeki Wachee Conduit Network

PAGE 95

80 3.4. CHARACTERIZATION OF AQUI FER RESPONSE AND FLUID EXCHANGE BETWEEN THE MATRIX AND CONDUIT NETWORKS Aquifer response is affected by several factors that include, but is not limited to, the quantity and duration of rainfall, the aerial extent of rainfall in the spring basin, vertical infiltration rate s, the extensiveness and interconnection among the conduit network, and antecedent conditions. Frequent, short duration rainfall events occur over the study area during the wet season and are related to convective storm acti vity (Jordan, 1984). Three major hydrologic events occurred during the characterization program which recorded data prior to, during, and after the 2004 wet season. One event involved Hurricane Frances, which made landfall as a Category 2 hurricane over the Labor Day weekend in September of 2004 on the eastern peninsula of Florida (U.S. Geological Survey, 2007). The storm, which was roughly the size of Texas, produced copius amounts of rainfall over large areas. The storm passed over the study area as a tropical storm. The second event involved Hurricane Jeanne which made landfall as a Category 3 hurricane on the eastern peninsul a of Florida on September 25, 2004 (National Geographic, 2007). Hurricane Jeanne was smaller in size relative to Hurricane Francis, but produced significant rainfall. The storm passed over the study area as a tropical storm. Figure 3-20, a plot of hourly barometric pressure recorded at the NOAA stat ion, (roughly 12 km from Weeki Wachee Spring) documents the decrease in barometric pr essure associated with passage of the tropical storms. The third event involved a drought. T he drought began in the spring of 2006 late in the monitoring program after the high frequency (15 minute) monitoring probes had been removed. Di scharge measurement s were performed during the drought and captured a cessation of flow at Twin D's Spring.

PAGE 96

81 Figure 3-20. Barometric pressure recorder captures the decrease in pressure as Tropical Storms Frances and Jeanne pass over the study area. Gaps represent missing data. Characterizing aquifer response and the direction and magnitude of fluid exchange between the matrix and conduit networks focused on monitoring water levels, and temperature for wells in the ma trix and conduit networks. Additionally discharge measurements at Twin Ds Sp ring were performed to determine if water level responses in upgradient wells open to the conduit network and matrix match the spring hydrograph. 3.4.1. Monitoring Wells Several observations are apparent from the geophysica l logs performed on the monitoring wells shown in Appendi x A (see Figure 2-1 for well locations): 1) the WW-F well breaches the roof of an underwater cave, 2) the WW-F well is in a horizon with a slightly higher clay content relative to the WWSpg-ECK well,

PAGE 97

82 and 3) the spontaneous potential, resistance, and resistivity logs reveal that there is no upwelling of sulfate-rich water in the WWSpg-ECK well and that the open interval is well above the MCU II. Upwellin g of sulfate-rich water was observed in the WW-3 well (Hill and DeWitt, 2004). Re sistance, resistivity, and spontaneous potential logs are not avail able for WW-F, but specific conductance data indicate that upwelling of sulfate-rich water does not occur and that the open interval is well above the MCU II. A video log performed on the conduit well (WW-F) when the high frequency probe was removed captured an al bino crayfish on video (Figure 321). The WW-F well is approximately 386 m from Twin Ds Spring and 797 m from Weeki Wachee Spring. Figure 3-21. Albino crayfish in WW-F well. Breakdown visible in background when driller intercepted the roof of an underwater cave (photograph taken by Kevin Stover from the SWFWMD with the author on De cember 17, 2004).

PAGE 98

83 3.4.2. Well Hydrographs Comparisons of 15 minute rainfall upgr adient of the monitoring wells and springs (Figure 3-22) versus 15 minute wa ter levels in the matrix wells and conduit well illustrate aquifer response to recharge during the 2004 wet season (Figure 3-23). Typically rainfall over the entire spring basins is used for comparisons in well/spring hydrographs to evaluate aquifer response, but since the spring basins have not been delineated by dye-trace testing, a more conservative comparison of rainfall over a smaller area in close proximity to the monitoring wells and springs was selected to avoid including rainfall outside of the spring basins. Fifteen minute OneRain, Inc. data were compiled over the entire study area to verify that using rainfa ll over a relatively smaller area (Figure 3-22) were not affecting the interpretation of data. No significant differences were observed using rainfall compiled over the ent ire study area relative to the area in Figure 3-22, therefore the interpretation using rainfall compiled for the smaller area highlighted in Figure 3-22 is considered valid. Figure 3-22. Shaded area repres ents the aerial extent of OneRain, Inc. 15 minute rainfall used with the well and spring hydrographs.

PAGE 99

84 Figure 3-23. Water levels in the ma trix Weeki Wachee Deep (WW Deep) and WWSpg-ECK wells and the conduit (WW-F) well versus rainfall. Rainfall quantities represent total 15 minute rainfall for the shaded area (132 km2) in Figure 3-22. Events A and C ar e from convection storm activity and B and D are from tropical storm activi ty. The high frequency (15 minute) monitoring wells with the most complete records are illustrated in Figure 3-23. Wells that went on-line (WHC # 6 and 7) or have incomplete data due to malfunctions (WW-3) were not included in Figure 3-23. Data recorded fo r the wells with incomplete data sets corroborate the results observed for the wells shown in Figure 3-23. Inset illustrates water levels in the WWSpgECK matrix well and the WW-F conduit well that are 920 m apart.

PAGE 100

85 Hydrographs of water levels in the matrix and conduit network illustrate the response to two local-scale, short duration, convective storm events, A and C, and two regional-scale, relatively longer duration, tropical storm events, B and D (Figure 3-23). Locally intense, short term events associated with convective storm activity generally did not produce a significant response in the matrix or conduit networks relative to the regional-sca le longer duration events associated with tropical storm activity in September 2004. Minor responses of less than 0.05 m are observed following the local-scale, s hort duration storm events (A and C). Conversely, a closer look at the matrix (WWSpg-ECK) well and conduit (WW-F) well during September 2004, show a relatively larger increase of 0.5 m at the WWSpg-ECK well following Tropical Storm Frances. Water levels continued to increase following Tropical Storm Frances (event B, Figure 3-23 inset) and a relatively shorter duration, lower volume event (C), before plateauing roughly a week after the passage of Tropica l Storm Frances, with a total increase in water levels of 0.7 m at the WWSpgECK well. An increase in water levels slightly above 0.1 m was observed at the WWSpg-ECK following passage of Tropical Storm Jeanne (event D). Close inspection of water levels dur ing each storm event at the WWSpgEck (matrix) well and the WW-F (conduit) we ll, which are approximately 920 m apart, suggests that the conduit network appr oaches equilibrium with the matrix network during events A and B (Figure 324). Plots of water level, or head difference in the matrix and conduit networks indicate that the water levels in the conduit network generally do not exceed those in the matrix network during and after convective or tropical storm events (Figure 3-25). Conduit network heads exceeded those in the matrix only briefl y during event A, producing a flux of groundwater from the conduits into the ma trix network. During the brief reversal, water level differences between the ma trix and conduit network were minimal and well within measurement error (i.e < 0.03 m, see Chapter 2). The minor head differences observed between the matr ix and conduit network during high

PAGE 101

86 recharge events indicate that the condui t network feeding the springs is not connected to point sources of recharge. Figure 3-24. Hydrographs for the WWSpg-E ck (matrix) well and WW-F (conduit) well during events A through D. Bar width s represent 15 minute rainfall intervals for a 132 km2 area (shaded area in Figure 3-22) Events represent continuous, uninterrupted rainfall upgradient of the m onitoring wells and springs, even during event D, with minimal quantities occurri ng near 16:00 on September 26, 2004.

PAGE 102

87 Figure 3-25. Head differences between the WWSpg-ECK (matrix) well and the WW-F (conduit) well during storm events A (convective storm activity) and B (tropical storm activity). Negative values reflect higher heads in the conduit network. In addition to the high frequency (15 minut e) monitoring well hydrographs, the distribution and total quantities of rainfall during each event were evaluated for the 132 km2 area upgradient of the monito ring wells and springs (shaded area in Figure 3-22). Figure 3-26 is the quantity of rainfall over the 132 km2. Rainfall quantities in Figures 3-23 through 3-24 ar e the 15 minute summation of rainfall over the total shaded area (132 km2) displayed in Figure 3-22. The largest quantities of rainfall occur during event B, followed by events D, A, and C (Figure 3-26). The maximum rainfall quantities a ssociated with the two local-scale events (A & C) were approximately 29% of the maximum quantities associated with event B (Tropical Storm Frances). Esti mates of evapotranspiration rates for September are slightly less than in July (HydroGeoLogic, Inc., 2007). Net recharge values are therefore, higher for ev ent B relative to events A, C, and D. Considering the maximum rainfall in a 15 minute interval, event A is the largest followed by events C, B, and D (Figure 3-23).

PAGE 103

88 Figure 3-26. Distribution and rainfall quant ities associated with events A through D. Event A lasted 4 hours and 45 minutes, event B, 36 hours and 45 minutes, event C, 5 and half hours, and event D, 19 hours. Event s represent continuous, uninterrupted rainfall upgradient of t he monitoring wells and springs. Hydrographs of hourly water levels in the matrix ROMP wells (Figure 3-27) that are located farther from the spri ngs than the high frequency (15 minute) monitoring wells (well locations shown in Figure 2-1) display a similar response to that observed in the high frequency matrix wells (WWSpg-ECK and Weeki Wachee Deep) and the WW-F conduit well shown in Figure 3-23.

PAGE 104

89 Figure 3-27. Hydrographs of hourly water levels for the ROMP matrix wells versus rainfall during convective storm activity (A and C) and tropical storm activity (B and D). 3.4.3. Spring Hydrograph The ratio of Qmax/Qbaseflow, a measure of the flashi ness of a spring, (White, 1988; Florea and Vacher 2006; Florea and Vacher, 2007) was evaluated at Twin Ds Spring prior to and following Tropical Storms Frances and Jeanne. The calculated ratios of discharges measured prior to and following both tropical storms were 2. Similar ratios were obt ained for Weeki Wachee Spring, indicating that both Twin Ds and Weeki Wac hee Springs are slow-responding. The hydrograph for Twin Ds Spring based on monthly average discharge and rainfall for the duration of the monitoring pr ogram is given in Figure 3-28.

PAGE 105

90 Figure 3-28. Hydrograph of monthly discharge at Twin D's Spring versus monthly rainfall. Gap represents missing data. Hydrograph for Weeki Wachee Spring was not included as Weeki Wachee Deep (WW D eep) is used to estimate discharge for Weeki Wachee Spring (see Chapter 2) A hydrograph of WW Deep is shown in Figure 3-23. Monthly average rainfall is for a 132 km2 area (shaded area in Figure 3-22). A rating curve developed from a linear regression model (shown on Figure 3-29) was developed in this study to esti mate discharge for Twin D’s Spring. The rating curve uses measured discharges at Twin D’s Spring and water levels in the WW-F conduit well. The locations for the index well (WWF) and the spring vents are provided in Figur e 3-19. Discharge and water levels were monitored over a range of conditions varying from high flow to cessation of flow at Twin D’s Spring, which provided an opportunity to identify limitati ons with the linear regression model developed in this study to estimate discharge at Twin D's Spring. Figure 3-30 is a plot of water leve ls in the index well (WW-F) and the pool stages for Weeki Wachee and Twin D's Springs. Water levels in the WW-F well were observed to drop slightly below t he pool stage at Twin D's Spring. Well and spring pool elevations where verified to be accurate when resurveyed. The

PAGE 106

91 differences, if not due to measurement e rrors, may suggest t hat a shift in the basin boundary for Twin D's Spring occurs at low discharge levels. In either case, it appears Twin D’s Spring may be an overflow for Weeki Wachee Spring. Twin D’s likely flows during high water level cond itions because it c aptures a portion of groundwater that would otherwise discharge at Weeki Wachee Spring. Dye-trace testing is needed to resolve this issue and for interpreting defensible spring basin boundaries. The rating curve developed for estimating discharge at Twin D’s Spring, which uses water levels in WWF as the explanatory variable, may be less reliable at low flow conditions, w hen pool stage at Twin D’s Spring exceed water levels in the WW-F index well. At this time, there are no other monitoring wells in closer proximity to Twin D's Sp ring that are better candidates for an index well. The rating curve developed to estimate discharge for Twin D’s Spring in this study does not include the data when the wa ter levels in the WW-F index well drop below the pool stage at Twin D's Spring. Figure 3-29. Rating curve for estimati ng discharge at Twin D's Spring based on average monthly discharge measurements and water levels at WW-F. Rating curve was developed using only thos e data where water levels at WW-F exceeded pool stage values at Twin D's Spring. Twin D's Q = 0.56x 1.81 R2 = 0.96 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 3.003.203.403.603.804.004.204.40Water Levels WW-F (m NGVD)Q (m3/s)n = 12

PAGE 107

92 Figure 3-30. Water levels in the WW-F index well with Twin D's and Weeki Wachee Springs pool stages. 3.5. DISCUSSION AND PR OPOSED CONCEPTUAL MODEL Although a first approximation, Reynol ds numbers estimated in this study using hydraulic gradients calculat ed from observed values, measured groundwater temperatures, and cave survey information are comparable to Reynolds numbers (105) obtained using quantitative tracer breakthrough curves for underwater caves, with average diameter s ranging from 10 to 80 m, in the Woodville Karst Plain of north Florida (K incaid et al., 2004; Hazlett et al., 2004). They are also comparable to estimates (105-106) reported in Martin (2003) for the Santa Fe Sink/Rise. Admittedly, t here are some uncertainties in how representative the estimated Reynolds numbers for Weeki Wachee and Twin D’s Springs are to actual numbers usi ng the approach applied in this study. Comparable Reynolds number s are obtained using larg er diameter (5 m) conduits, yet plots of specific dischar ge versus hydraulic gradient indicate 2.50 2.75 3.00 3.25 3.50 3.75 4.00 4.25 4.50 May-04Aug-04Nov-04Feb-05Jun-05Sep-05Dec-05Apr-06Time (month)Elevation (m NGVD) Twin D Pool Elevation WW-F Weeki Wachee Pool Elevation

PAGE 108

93 laminar flow through larger diameter c onduits (Figure 3-13). However, plots of specific discharge versus hydraulic gradi ent for relatively smaller diameter conduits (0.9 m shown in Figure 3-13) coupled with anecdotal evidence of “raging” flow (Karst Underwater Research Inc., 2008c) described by cave divers that is too difficult to navigate (K arst Underwater Research, Inc., 2008b) however, indicates that non-Darcian, turbul ent flow occurs in portions of the underlying conduit network near Weeki Wachee and Twin D's Springs, particularly where constrictions occur. As conduits widen and constrict, and mounds of breakdown (Karst Underw ater Research, Inc., 2008b) are encountered flow likely varies from la minar to turbulent and may also vary temporally depending on hydrologic condi tions. Estimated values for the Reynolds numbers have direct applicati on in Darcian/non-Darcian groundwater flow simulators (Shoemaker et al., 2008a). Uncertainty with the Reynolds number estimates performed in this study, in terms of whether or not la minar or turbulent flow has a significant effect on simulated spring flows, which is important for the objectives of this project, will be in vestigated in the next chapter. Not all of the traditional methods used for understanding conduit locations, or preferential flow pathways discussed in Chapter 1, are applied in this study. Methods that were not applied include se mivariogram cloud analysis, dye-tracer tests, and fractal analysis. An aquifer performance test was not performed as part of this study and most previous aqui fer performance tests lacked a sufficient number of observation wells to determi ne the direction of maximum hydraulic conductivity, with the exception of one test at the Cross Bar Wellfield where an evaluation of aquifer anisotropy reveal ed a NW-SE direction of principal transmissivity (Gilboy and Moore, 1982a; Gilboy and Moore, 1982b). This is consistent with the orientat ion of a ground-truthed fracture trace at the Cross Bar Wellfield (Wood and Stewart, 1985) and passage orientations based on cumulative lengths identified in air-filled caves (Florea, 2006a; Brinkmann and Reeder, 1995).

PAGE 109

94 Fracture traces inferred from t he alignment of closed topographic depressions were the least corroborated by other data types. Moreover, there are questions regarding accuracy and resolution of fracture traces that are corroborated by other data. For example, the fracture traces in the vicinity of the Weeki Wachee Trough do not intersect Weeki Wachee or Twin D’s Springs (see Figure 3-14). These findings support W ilson’s (2002) argument that fracture traces that are not verified using geophysical methods, or that are not corroborated by additional data may be le ss effective tools for interpreting conduit locations, or preferential flow pathways in the Upper Floridan aquifer. The location of troughs in aquifer water levels (Figure 3-15) coincide fairly well with decreases in Upper Floridan aqu ifer thickness as interpreted by the Florida Geological Survey (2008). Uncert ainties with the resolution of the water levels contours were previously discussed. Specifically, the location of the 3 m contour is not always supported by pool stages at Twin D’s and Weeki Wachee Springs. In general, troughs in aquifer water levels at the test-site are very broad subtle features due to the low hydraulic gr adients, high permeability of the matrix, and the absence of large water level diffe rences (< 0.03 m) between the matrix and conduit networks as s hown in Figure 3-25. The probability of intercepting a cavity based on borehole porosity descriptions in this study (10%) is hi gher than the probability for a Paleozoic limestone aquifer (from 0.4 to 3%; Worthington et al., 2000b), but are conservative relative to estimates obtained by Wilson (2002) based on borehole porosity descriptions for the Floridan aquife r in several other counties throughout Florida. He estimates that the probability of intercepting a cavity 0.3 m in height ranges from 25 to 50% (Wilson, 2002). Elevation modes for surveyed underwa ter cave passages near discharge points occur with relative frequencies ranging from 18% to 28% at -37, -73, and 79 m NGVD (see Figure 3-18) Although there are some shared modes with the borehole data, the relative frequencies for elevation modes based on borehole porosity descriptions are substantially lower (less than 4%).

PAGE 110

95 Ratios of large ( 1.5 m) and small (< 1.5 m) cavity heights from the borehole porosity descriptions compare favorably with an inventory of known karst features in the study area, shown in Figure 3-7. That is, the density of karst features is higher in th e North Gulf Coastal Lowlands and Coastal Swamps indicating a slightly higher degree of connection among the conduit network in the western portion of the st udy area relative to the so utheastern corner of the study area where maximum thi ckness occurs in the Uppe r Floridan aquifer. This interpretation differs from t hat in Armstrong et al. (2003) In the Armstrong et al. (2003) study, the presence of depressions or closed topographic contours were correlated with transmissivity values used in a calibrated groundwater flow model. The study area encompassed Hernando, Pasco, and portions of Hillsborough County in Flori da. Results of the Armstr ong et al. (2003) study indicate a relatively higher doline density on the Brooksville Ridge relative to the North Gulf Coastal Lowlands and Coasta l Swamps, indicating that relatively higher hydraulic conductivities underlie the Brooksville Ridge. However, they note that the ability to di scern closed topographic contours diminishes at lower elevations (Armstrong et al., 2003). Similarly, in the Jones et al. (1997) study, the density of inferred fracture traces indicates that a well-integrat ed conduit network underlies the Brooksville Ridge and that a regional-scale conduit net work underlies the study area. It is likely, that the relatively higher densit y of inferred fracture traces on the Brooksville Ridge is related to the diffi culty with discerning topographic features in topographically low areas noted in the Armstrong et al. (2003) study. In this study, a different conceptualiz ation of preferential flow pathways is proposed. Relatively higher hydraulic cond uctivities underlie the western portion of the study area based on the coin cidence of multiple data types. Possible sources for differences between the results obtained in this study and probability estimates for intercepting a cavity by Wilson (2002) and the density of karst features interpreted in the Armstrong et al. (2003) study, include but are not limited to, 1) the criteria used to delin eate conduits in borehole

PAGE 111

96 porosity descriptions, 2) subjectivity of the reviewer, 3) the reliance of the evaluators on a single data source rather than combining multiple data types, and 4) limitations associ ated with data sources, uncertainties in model parameters, and assumptions used to estima te the probabilities and/or delineate the occurrence of karst features. Well and spring hydrographs exhibit shallow, longer recession limbs rather than the steep, shorter recession limbs that are typical of conduit hydrographs in flashy karst aquifers. Hydrographs for the matrix wells and conduit well and Weeki Wachee and Twin D’s Springs exhibit a slow response to convective and tropical storm activity. Hydrographs of hourly water data for the matrix ROMP wells, which are located farther from the springs also displayed a similar response to convective and tropical storm activity (see Figure 3-27). This indicates that the similarity in response to convective and tropical storm activity between the WWSpg-ECK matrix well and th e WW-F conduit well, shown in the inset for Figure 3-23, is not a result of t he proximity of the wells to the springs as the ROMP matrix wells locat ed farther from the springs exhibit a similar response to recharge events A,B, C, and D shown in Figures 3-23 through 3-24 and 3-27. Moreover, wells and springs in the Upper Floridan aquifer nort h and northeast of the study area that do not interact direct ly with surface water sources exhibited a similar response to convective and tropi cal storm activity (Florea and Vacher, 2007). The shallow, long recession limbs following high recharge events observed in well and spring hydrographs is corroborated by water quality and temperature data. The coefficient of vari ation of total bica rbonate, or hardness, for water samples collected from Weeki Wachee Spring is 5%, which indicates the springs are diffuse flow systems (W hite, 1988). Additionally, observations noted in the hydrographs are corroborated by temperature data. That is, shifts in temperature during events A and B, when wa ter levels in the matrix and conduit network approached equilibrium, did not exc eed instrument precision 0.1 C in the monitoring wells or 0.25 C in the springs, as discussed in Chapter 2,

PAGE 112

97 suggesting that most of the flow is coming from the matrix network. In a flashy conduit network, larger temperature fluc tuations would be observed (Martin and Dean, 1999). Most previous studies on fluid exchange between the matrix and conduit networks in the Upper Floridan aquife r have been performed north of the study area at the Santa Fe Sink/Rise netw ork in Alachua County, where shallow conduits interact directly with surface water sources (Martin and Screaton, 2001; Martin and Dean, 2001; Martin et al., 2006, and Screaton et al ., 2004). The Santa Fe Sink/Rise conduit network, reveals a mo re dynamic bidirectionality of fluid exchange that varies tempor ally based on hydraulic conditions (Martin et al., 2006). Additionally, larger temperature r anges are observed in the relatively shallower Santa Fe Sink/Rise network (Martin and Dean, 1999). Conversely, in the vicinity of Weeki Wachee Spring, heads in the matrix network generally exceed those in the conduit network even dur ing tropical storm ac tivity, as shown in Figures 3-23 through 3-25. T he unidirectionality of flux observed in this study is related to the high conduit wall conduc tance between the matrix and conduit networks, the elevation of the conduits, which are relati vely deep with respect to the elevation of the water table, and the absence of direct recharge from a surficial point source. The probes monitoring the conduit network (WW-F well, Twin D’s and Weeki Wachee Springs) also recorded a rela tively constant temperature of 23.6 C during the high frequency monitoring period. This corroborates the direction of flux observed at the condui t well (WW-F) and suggests that the conduit network in the vicinity of Weeki Wachee Spring pr imarily receives water from the matrix. Moreover, the response observed in the hydrographs for both the conduit and matrix networks, which closely mi mic each other, implies that a high conductance, or permeability exists betw een the two networks (see Figures 3-23 through 3-25). Water level differences in the matrix and conduit networks do not differ significantly from each other ev en during high recharge events (< 0.03 m, shown in Figure 3-25). Moreov er, the matrix ROMP wells located farther from the

PAGE 113

98 springs exhibit a similar response to the high frequency monitoring wells located closer to the springs. A corollary is that the hydraulic conductivity does not differ significantly between the matrix and condui t networks, which is prevalent in eogenetic karst aquifers (Vacher and Mylr oie, 2002). In telogenetic karst aquifers, where significant contrasts in hy draulic conductivity between the matrix and conduit network exists, hydrographs for the conduit network would be expected to differ significantly from the observed response in the matrix network (Florea and Vacher, 2006). Results from the multiple types of data collected in this study (i.e. inventory of karst features, modes of conduit elevations, and well/spring hydrographs) imply that the Upper Flor idan aquifer does not contain a regionalscale, laterally extensive conduit network across the study area, but rather consists of smaller scale, karst feat ures embedded in the matrix network. Moreover, Weeki Wachee Spring has not ceased to flow during the period of record (1929 to present) and although the spring basin may be large, a regionalscale, well-integrated conduit network would result in a cessation of flow. This does not imply that smaller-scale karst f eatures are not import ant. Turbulent flow can occur in karst features with aperture s greater than 1 cm (Worthington et al., 2000b; White, 1988) which is import ant for contaminant transport. Based on the coincidence of multiple data types using: an inventory of karst features, inferred fracture traces, troughs in aquifer water levels, changes in aquifer thickness, conduit elevation modes interpreted from borehole porosity descriptions and cave survey data, well and spring hydrographs following convective and tropical storm activity, the magnitude and di rection of fluid exchange between the matrix and conduit networks, water quality data, and temperature data, a concept ual model for the location s of preferential flow pathways is proposed (Figure 3-31). Pref erential flow pathways occur in the western portion of the study area and underlie the breached portion of the Brooksville Ridge in the northeast co rner of the study area. This conceptualization of preferential flow pathways for the study area is used to

PAGE 114

99 guide the approximate locations of relative ly higher hydraulic conductivities for the groundwater flow models discussed in the next chapter. Specifically, relatively higher bulk hydraulic conductivities will be assigned across broad areas where multiple types of data coinci de in the laminar and laminar/turbulent equivalent-continuum models. Smaller karst features representing vuggy porosity will be included in the laminar/turbul ent, equivalent-continuum model using MODFLOW-2005 CFP Mode 2. U nderwater cave survey data (shown in Figure 3-19) will be explicitly incorporated in to the dual-conductivity model. Moreover, an inferred connection between Weeki Wa chee Spring and the WW-F conduit well, and Century Lake based on the anecdotal accidental tracer event is used to extrapolate the location of conduits ex tending beyond the surveyed passages. Fluid exchange between the matrix and conduit networks will be active in the dual-conductivity model. The conduits underlying Twin D’s and Weeki Wachee Springs are perennially water-filled, even when Twin D’s Spring ceases to flow. Conduit elevations in the dual-conductivi ty model were loosely constrained using modes of elevation from the cave survey data (Figure 3-18). Since the caves are perennially water-filled this will be a less sensitive parameter for the test-site relative to one where the conduit netwo rk is partly water-filled. Inferred preferential flow pathways in Figure 3-31 based on the coincidence of multiple data types (inventory of kars t features, troughs, fracture traces) will be treated as the bulk hydraulic conductivities for bot h the matrix and conduit networks. These features will not be explicitly incorpor ated into the dual-conductivity models and there will be no fluid exchange between t he matrix and conduit networks at these locations. Instead relatively higher bulk hydraulic conductivities will be used for these less defined, broad pref erential flow pathways located farther from the springs.

PAGE 115

100 Figure 3-31. Proposed conceptual model for pref erential flow pathways in the Upper Floridan aquifer near Weeki Wac hee, Florida. Relatively higher permeabilities occur in the vicinity of troughs and generally ar e higher in the western portions of the study area where the density of karst features is higher relative to the southeastern portions of t he area. Relatively higher permeabilities also underlie the breached portion of the Brooksville Ridge located in the northeast corner of the study area. Inferred fractu re traces from the Jones et al. (1997) study (Figure 3-14) t hat are not corroborated by additional data types are omitted from the proposed conceptual mode l of preferential flow pathways. The verified fracture trace from Wood and Stewart (1985) in included in the conceptual model of prefer ential flow pathways. May 2006 Upper Floridan aquifer (UFA) contours from (Ortiz, 2007). Contour interval is 3 m and the datum is NGVD. The conceptual model of a poorly in tegrated conduit network proposed in this study is consistent with previous conceptual models proposed for westcentral Florida (GeoTrans, Inc., 1988a; Florea, 2006a; Florea and Vacher, 2007). In the GeoTrans, Inc. (1988a) study, discu ssed in Chapter 1, a calibrated steadystate model could not be achieved usi ng a well integrated conduit network based

PAGE 116

101 on inferred fracture traces or faults in t he vicinity of Rainbow and Silver Springs. Florea (2006a) and Florea and Vacher ( 2007) conclude that caves in westcentral Florida are discontinuous. A conceptual model for a poorly integrated conduit network does not violate continuity. Previous work per formed by Kraemer (1990) discussed in Chapter 1, and empirical data collected in this study shown in Figures 3-23, 3-24, 3-25, and 3-13 provide evidence supporti ng that a poorly integrated conduit network is hydraulically possible and need no t violate continuity. Kraemer (1990) performed a series of numeric al exercises using a discr ete-fracture model that consisted of overlaying local-scale and r egional-scale fractures. He concludes that a regional-scale fracture super imposed on local-scale fractures with transmissivities three orders of magnit ude less than the regional-scale fracture resulted in over 98% of the flow bei ng transmitted through the regional-scale fracture. Two orders of magnitude diffe rence in transmissivities between the regional-scale fracture superimposed on loca l-scale fractures resulted in 68% of the flow being transmitted through the dom inant regional frac ture (Kraemer, 1990). Figures 3-23 through 3-25 provide em pirical data indicating that the hydraulic conductivity for the matrix and conduit networks do not differ significantly from each other. Moreover, the similarity of water level responses in the high frequency monitoring wells located close to the springs and the ROMP wells located farther from the springs indica te that the hydraulic conductivities in the matrix and conduit netwo rks do not significantly differ from each other, suggesting that although the primary effect of the matrix is storage, the matrix itself plays a role in transport. Fi gure 3-13 shows that specific discharge decreases for larger diameter condui ts using the observed low hydraulic gradients and may be laminar. Laminar flow likely occurs throughout most of the study area, with turbulent flow occurri ng in the conduit network, particularly where constrictions are encountered and near areas of focused discharge. Worthington et al. (2000a) estimates that the proportion of flow in conduits can be 94% or higher. Indeed, this occurs in some karst aquifers, however based on

PAGE 117

102 the numerical studies performed by Kr aemer (1990), coupled with the observed responses for matrix wells, a conduit well and springs in this project, and others (Florea 2006a; Florea and Vacher, 2007) it appears that the proportion of flow in the conduit network relative to the matrix in the Upper Floridan aquifer may be lower than that of some previously studied karst aquifers located elsewhere in North America and on other continents. Vacher and Mylroie (2002) demonstrated that the matrix permeability of the Upper Floridan aquifer is higher relative to matrix permeabilities for telogenetic karst aquifers. For example, average matrix hydraulic conductivities (10-6 m/s) measured from core samples in the Ocala Limestone (Florea 2006b; Florea and Vacher, 2007) are high relative to mean matrix values for comparable volumes measured for other dual-pe rmeability karst aquifers (i.e. 10-8 m/s for the San Antonio segment of t he Edwards aquifer in south-central Texas, Mace and Havorka 2000; Halihan et al. 2000, and 10-11 m/s in the Ste. Genevieve Formation in central Kentucky, Worthingt on et al. 2000b). It is the absence of significant contrasts between the matrix and conduit network hydraulic conductivities that permits a poorly integrated conduit network embedded in the matrix to be hydraulically possible without violating continuity.

PAGE 118

103 CHAPTER 4: NUMERICAL ANALYSIS 4.0. INTRODUCTION One of the first questions to ans wer before developing any numerical model is, what is the purpose of the model? The primary purpose for developing the numerical models in this study was to evaluate the performance of 3 groundwater flow models with 3 differ ent conceptualizations of conduits. The equivalent-continuum model using MODFLOW-2005 (Harbaugh, 2005) simulates water levels and discharges using the bulk hy draulic conductivities for the matrix and conduit networks. The dual-conductivity model using MODFLOW-2005 CFP Mode 1 (Shoemaker et al., 2008a) simulates groundwater flow through a discrete conduit network consisting of large underwater caves embedded in the matrix. The laminar/turbulent equi valent-continuum model using MODFLOW-2005 CFP Mode 2 (Shoemaker et al., 2008a) simulate s groundwater flow through a matrix consisting of vuggy porosity. 4.1. MODEL CALIBRATION Fully calibrated, site-scale models were not necessary to satisfy the purpose for developing the numerical models. Models were calibrated to observed water levels and discharges in the Upper Floridan aquifer only. The surficial aquifer (top layer), an interm ediate layer represent ative of residual weathered clays in the recharge and disc harge zones of the model domain, and the Hawthorn Group in the capped zone (shown in Figure 3-12) were combined to represent the mantle (Undifferentiated Sands and Ha wthorn Group, see Table 3-1) overlying the buried karst terrain. No wells were used to calibrate water

PAGE 119

104 levels in the mantle. This is sufficient as the conduit networks are located in the Upper Floridan aquifer and because it is t he primary source of groundwater at the test-site. The mantle was included bec ause in the transient simulations it provided additional storage and attenuated the simulated response of the Upper Floridan aquifer to net recharge events. Thirty-two target wells distributed throughout the model domain within the Upper Floridan aquifer were used for calib ration (Figure 4-1). Spring discharge was calibrated relative to two springs (Twin Ds and Weeki Wachee Spring) as discharge measurements or rating curves for estimating discharge from these springs are available (see Chapter 3). Locations for Weeki Wachee and Twin Ds Springs are provided in Figure 2-1. Figure 4-1. Location of targets (monitori ng wells in the Upper Floridan aquifer) used to calibrate groundwater flow model s. Cross Bar WF represents the Cross Bar Wellfield.

PAGE 120

105 Three dynamic steady-state models were calibrated by trial and error to October 2005. During this time daily groundwater levels were reasonably stable and between the extremes of high water level conditions following the tropical storms and low water conditions during the drought. Daily water levels in 6 monitoring wells varied by 0.27 m or less between September 30 and October 31 of 2005, equivalent to 6.5% of the median water level change for the same wells from June 2004 through May 2006 (Figure 42). Locations for the monitoring wells are provided in Figure 4-1. Figure 4-2. Difference in daily water le vels from September 30 through October 31, 2005. Combined steady-state, transient groundwater flow models using 25 stress periods were developed. The firs t stress period was set to steady-state and the remaining 24 stress periods (June 2004 through May 2006) were transient. Each stress period consists of the number of days for each representative month. For ex ample, the stress period representing June consists of 30 days, whereas the stress period repr esenting October consists of 31 days. The time frame selected for the transient stress periods capture conditions prior

PAGE 121

106 to, during, and after the passage of two tropical storms in 2004 and a cessation of flow at Twin Ds Spring in May 2006. This provides a fairly rigorous evaluation of model performance for the various groundwater flow models. With the exception of hydraulic conductivity, c onduit wall conductance between the matrix and conduit networks, and t he higher and lower Reynolds numbers, hydraulic parameters are consistent across the three models. The average absolute difference between observed and simulated water levels, relative to the 32 target wells, for each of the 3 groundwater flow models for the entire 25 stress periods wa s 0.77 m or less. Maximum absolute differences between the simulated and obser ved values for 31 of the 32 targets for each of the 3 groundwater flow models were 2.79 m or less for the entire 25 stress periods of the combined steady-s tate/transient simulations. Maximum absolute differences between simula ted and observed values for the 32nd target (ROMP TR 17-3, which is located near the southwestern boundary of the model domain far away from Weeki Wachee and T win Ds Springs, shown in Figure 41) for all 25 stress periods in the 3 groundwater flow models was 3.33 m. The average simulated discharges for Weeki Wa chee Spring for all 25 stress periods in the 3 groundwater flow models was 77% or higher of observed average discharges for Weeki Wachee Spring and 45% or higher for Twin Ds Spring. The match between observed and simulated water levels and discharges are considered an adequate calibration for the purpose of this study. 4.2. Grid Three, 3-dimensional, 3-layer groundwater flow models were developed. The models consist of uniform grids with cell widths of 152 m with 285 columns and 236 rows. Although computationally sl ower, small cell dimensions were selected to provide a more accurate spatial resolution for drains and conduit networks in the groundwater flow models. A uniform grid was selected because it permitted a relatively better comparison of the conduit conceptualizations among

PAGE 122

107 the various groundwater flow models. Coor dinates for the southwest corner of the model grids in UTM (m) Zone 17 are 329427 Easting and 3133534 Northing. The model domain boundaries extend furt her westward of the study area boundaries described in Chapter 3. The western boundary for the groundwater flow models was shifted further west in t he Gulf of Mexico to minimize boundary effects. The model domain is also sli ghtly south of the study areas northern boundary and slightly east of the study areas eastern boundary shown in Figure 3-1. This was done to reduce computational run times since the western boundary was shifted westward, thereby in creasing the number of active model cells. The extent of the model domain is shown in Figure 4-3. 4.3. BOUNDARY CONDITIONS The boundaries of the model domain were delineated pr imarily along noflow, specified-head, and general-head boundaries. No-flow boundaries generally parallel flow lines. Model boundaries in t he Gulf of Mexico were simulated as specified-head or general-head boundaries. General-h ead boundaries are also designated where lateral flux occurs along the northern, eastern, and southern boundaries (Figure 4-3). General-head boundaries were updated for each monthly stress period in the transient si mulations by interpolating monthly head values for monitoring wells located near general-head boundaries.

PAGE 123

108 Figure 4-3. Lateral model boundaries. 4.4. ESTIMATES OF GROUNDWATER WITHDRAWALS The Upper Floridan aquifer is the pr imary source of water supplies in the study area. Published water-use estima tes for 2003 onward were not available from the SWFWMD at the time of model development. Therefore, a similar methodology to that used by the SWFWMD to estimate groundwater withdrawals was applied in this study using 2002 publ ished water use estimates (Southwest Florida Water Management District, 2004). Estimated water use for Hernando and Pasco Counties in 2002 are provided in Table 4-1. No-flow General-head Specified-head County Boundary

PAGE 124

109 Table 4-1. Reported and estimated 2002 water use for Hernando and Pasco Counties in million liters per day (MLD) and million gallons per day (MGD). Quantities include agricultural (AG), indus trial, mining, domestic self-supply, public supply, and recreational uses. Nongroundwater withdrawals account for less than 3 MLD of Hernando County totals and less than 8 MLD of Pasco County totals reported in Table 4-1 (Southwest Florida Water Management District, 2004). Table 4-1 shows that water use esti mates involve compiling both reported and estimated water use quantities that in clude agricultural, industrial, mining, domestic self-supply (private landowner wells), public supply, and recreational, uses. Florida Administrative Code, Chapter 40D-2 requ ires that average permitted withdrawals equal to or greater than 378,500 MLD, maximum permitted withdrawals of 3,785,000 MLD or more on any single day, or withdrawals from a 15 cm or larger diameter well to be r eported to the SWFWMD (Southwest Florida Water Management District, 2004). It is unc lear if the report ed quantities are read from calibrated flowmeters. This limitat ion is not addressed to criticize the reporting procedures, but rather to convey to the reader the uncertainty in the water use estimates. All nonrep orted quantities are estimated. In this study, reported water use quantities for 2004, 2005, and 2006 were compiled. Nonreported quantities were es timated for 2004, 2005, and 2006 by applying a multiplier to 2002 estimated quant ities. That is, multipliers based on COUNTYWATER USEAGINDUSTRIALMININGDOMESTIC SELF-SUPPLYPUBLIC SUPPLYRECREATIONALTOTAL (MLD)(MLD)(MLD) (MLD) (MLD) (MLD)(MLD) HERNANDO REPORTED23842 0 79 8 169 ESTIMATED80.20.3 4 1 8 22 Couny total 191 PASCO REPORTED8150 0 322 8 353 ESTIMATED5302 26 8 11 100 Couny total 453 (MGD)(MGD)(MGD) (MGD) (MGD) (MGD)(MGD) HERNANDO REPORTED0.41011 0 21 2 44 ESTIMATED20.040.07 1 0.3 2 5 Couny total 49 PASCO REPORTED240 0 85 2 93 ESTIMATED140.10.6 7 2 3 27 Couny total 120

PAGE 125

110 the ratio of permitted and es timated water use for 2002, t he most recent year for which published water use estimates were available at the time of model development, were applied to nonreported quantities. Domestic self-supply water use quantities reported for 2002 were applied to 2004, 2005, and 2006 as these quantities are assumed to have changed littl e to none during the specified years. Using this methodology, water use types and county totals for 2004, 2005, and 2006 were comparable, to slightly higher, to 2002 water use estimates. Mining water use estimates reported in Table 4-1 do not account for recirculated quantities. Therefore, mine water use permits were reviewed and consumptive uses were estimated based on information provided by the permittees. Consumptive use was estimated by summing total losses (i.e. truck washing, personal sanitary, product ent rainment, and evaporative losses from recirculation ponds) while excluding recirculated quantities for each mining water use permit in the study area. Additionally, the water use estimate s reported in Table 4-1 do not account for artificial recharge from irrigation or septic tank l eakage. It is assumed that some percentage of groundwater withdrawals may return to the Upper Floridan aquifer based on a previous study of nitr ate sources in groundwater and spring discharge (Jones et al., 1997). Application of residential fertilizers was listed as highly significant and recreational (golf courses) as significant sources of nitrates in the study area, indicating that some portion of public supply and recreational uses recharge the Upper Floridan aquifer. Moreover, estimates of recharge from reclaimed water sources published by the Department of Environmental Protection (Department of Environmental Protection, 2005) indicate that 40% of residential lawn watering, 10% of recreational irrigati on, and 25% of agricultural irrigation returns to t he Upper Floridan aquifer. In the absence of empirical data, 10% percent of repor ted public supply, agriculture, and recreational uses is assu med to return to the Upper Floridan aquifer for wells located in the recharge zone of the model domain (shown in Figure 3-12) with the exception of the Cross Bar Wellfield where public supply

PAGE 126

111 withdrawals are transported offsite. These quantities were included in the model by reducing reported quantities by 10%. No adjustment s were made to estimated quantities, other than domestic, self-supply quantities in the recharge zone. Moreover, no adjustments were made to re ported or estimated quantities located in the capped or discharge zones of t he model domain (shown in Figure 3-12). Adjustments were made to domestic se lf-supply quantities for wells in the recharge zone of the model domain to account for artificial recharge from septic tank leakage. No adjustments were m ade to domestic quantities in the discharge or capped zones of the model domain. Fo rty percent of dom estic, self-supply quantities were considered consumed and t he remaining 60% were assumed to return to the Upper Flori dan aquifer as septic tank leakage. The quantities were added to the model by reducing the dom estic withdrawal quantities in the recharge zone of the model (shown in Figure 3-12) by 60%. This was done because: i) the location, distribution, and type of septic tanks (mounded vs. nonmounded) are not known, ii) nitrate leve ls in groundwater and spring discharge substantiate that a percentage of septic t ank leakage in the study area recharges the Upper Floridan aquifer (Jones et al., 1997), iii) the percentage of septic tank leakage (60%) is consistent with estima tes based on per capita use (Ross, 2006) and iv) is consistent with estimates used in groundwater flow models north of the study area (Knowles et al., 2002). Admittedl y, there is uncert ainty associated with the estimates of septic tank leakage used in this study (60%). In the absence of empirical data these estimated quantities of septic tank leakage are only a guess. Because of the uncertainties asso ciated with quantities for artificial recharge, the distribution of septic tanks, and sensitivit y of net recharge, which is the most sensitive model parameter (discu ssed later in this chapter), artificial recharge estimates were accounted for by adjusting reported well flow rates and domestic self-supply estimates as appropriate, which is a less sensitive model parameter (third) relative to net rechar ge. Moreover, artifi cial recharge was accounted for only in the recharge zone of the model domain (shown in Figure 312) where water levels for monitoring wells in the surficial aquifer vary from dry to

PAGE 127

112 having water levels that coincide with t hose in the Upper Floridan aquifer (shown in Appendix B), indicating that the surficial aquifer and Upper Floridan aquifer are hydraulically well connected in the re charge zone. One can argue that the artificial recharge estimates should be accounted for as net recharge to the uppermost layer, rather than as a well flow rate in the Upper Floridan aquifer, but it was in the authors judgment, based on t he limitations with artificial recharge discussed above, coupled with the sensitivit y of net recharge, that it would be unwise to adjust the most sensitive model parameter using quantities that are admittedly a guess. This is an area that c an be revisited in the future, if and when empirically based quantitative data for artificial recharge becomes available. Reported and estimated withdrawal quantities for nondomestic and domestic wells in the steady-state model s are provided in Figure 4-4. Annual average water use estimates in the model domain for 2005 were 218 MLD (57 MGD). Annual average water-use esti mates for 2004 and 2006 are not provided since only 5 to 7 months of those calendar years were simulated in the transient simulations. However, estimated groundw ater withdrawals for each monthly stress period in the combined steady-sta te/transient models are provided in Table 4-2. Minimum monthly groundwat er withdrawals of 138 MLD (37 MGD) occurred in September 2004 when the two tropical storms passed over the study area and maximum monthly groundwater withdrawals of 324 MLD (86 MGD) occurred in May 2006 during the drought.

PAGE 128

113 Figure 4-4. Nondomestic (top) and dom estic (bottom) reported and estimated groundwater withdrawal quantitie s in the steady-state m odels. Values reported in million liters per day (MLD).

PAGE 129

114 Table 4-2. Groundwater withdr awals from the Upper Fl oridan aquifer for each stress period in the combined steady-state /transient groundwater flow models. Quantities represent withdr awals within the model domai n which includes parts of Hernando and Pasco Counties. MLD repr esents million liters per day. MGD represents million gallons per day. Stress Period Groundwater Withdrawals MLD Groundwater Withdrawals MGD Steady-state 218 57 June_04 286 75 July_04 187 49 Aug_04 181 48 Sept_04 138 37 Oct_04 185 49 Nov_04 190 50 Dec_04 180 48 Jan_05 193 51 Feb_05 217 57 Mar_05 210 55 Apr_05 260 69 May_05 293 77 June_05 223 59 July_05 195 52 Aug_05 176 46 Sept_05 225 59 Oct_05 218 57 Nov_05 227 60 Dec_05 175 46 Jan_06 230 61 Feb_06 208 55 Mar_06 266 70 Apr_06 311 82 May_06 324 86

PAGE 130

115 4.5. NET RECHARGE Net recharge was estimated by adjusting rainfall data to account for evapotranspiration losses as follows: Net Recharge = Rainfall Evapotranspiration eq. 4-1 As previously discussed in Chapter 3, recharge values obtained from OneRain Inc. and estimates of evapotraspiration compiled by HydroGeoLogic, Inc., (2007) were used in this study. Net recharge was applied to the uppermost layer (layer 1 shown in Table 3-1) in the groundwater flow models. Table 4-3 summarizes estimates of evapotranspiration losses. Table 4-3. Estimated evapotranspirati on rates (modified from HydroGeoLogic, Inc., 2007). Month Average Evapotranspiration (cm/month) January 4 February 5 March 7 April 8 May 9 June 11 July 12 August 12 September 10 October 7 November 4 December 3 Annual Average (cm) 92

PAGE 131

116 Appendix D provides the distribution and quant ities of net recharge for each stress period in the combined steady-state/transient simulations. 4.6. DISCHARGE Discharge occurs in the western portion of the model domain (Figure 4-5). Point discharge occurs at spring vent s and diffuse discharge occurs in the coastal swamps. Both point and diffuse di scharges are represented as drains in the groundwater flow models. Values for discharge, pool stages, and spring conductance coefficients from observ ed data are used, when available, and estimated for drains lacking observed dat a. Observed discharges for Twin D's and Weeki Wachee Spring are provided in Table 4-4. The rating curve for Twin D's Spring shown in Figure 3-29 was used to estimate discharge when discrete measurements were not performed. Discharge for Weeki Wachee Spring was estimated following the procedure discu ssed in Chapter 2. A rating curve previously developed by the USGS for Bobhill Spring (Knochenmus and Yobbi, 2001) which is part of the Aripeka Spri ngs Group (shown in Figure 4-5) was not used in the calibration process as focu s was directed to Twin D's and Weeki Wachee Springs since relatively more dat a are available for these two springs.

PAGE 132

117 Figure 4-5. Distribution of drains in model domain.

PAGE 133

118 Table 4-4. Observed discharges for Twin D's and Weeki Wachee Springs (field measurements are accurate to 0.03 m3/s, as discussed in Chapter 2). Date Twin D's Spring m3/s Weeki Wachee Spring m3/s Remarks* Steady-state 0.25 4.87 Estimated June 2004 0.10 4.57 Measured July 2004 0.13 4.54 Measured August 2004 0.22 4.72 Measured September 2004 0.52 5.94 Measured October 2004 0.60 6.47 Measured November 2004 0.51 6.27 Measured December 2004 0.45 5.90 Measured January 2005 0.39 5.51 Measured February 2005 0.27 5.30 Estimated March 2005 0.30 5.08 Measured April 2005 0.26 4.86 Measured May 2005 0.22 4.75 Measured June 2005 0.12 4.65 Measured July 2005 0.33 5.01 Estimated August 2005 0.33 5.14 Estimated September 2005 0.29 5.05 Estimated October 2005 0.25 4.87 Estimated November 2005 0.19 4.71 Estimated December 2005 0.15 4.55 Estimated January 2006 0.12 4.39 Estimated February 2006 0.03 4.33 Measured March 2006 0.03 4.24 Measured April 2006 0.01 4.09 Measured May 2006 0.00 3.89 Measured *Refers to measured or estimated dischar ge for Twin D's Spring. See Chapter 2 for a detailed discussion of discharge es timates for Weeki Wachee Spring. 4.7. POOL STAGES In the combined steady-state transi ent simulations, pool stages for Twin D's and Weeki Wachee Springs are updated for each stress period. Table 4-5 lists the observed pool stages for both springs.

PAGE 134

119 Table 4-5. Observed pool stages for T win D's (TD) and Weeki Wachee (WW) Springs Date Twin D's Spring m NGVD Weeki Wachee Spring m NGVD Remarks Steady-state 3.57 2.84 Measured June 2004 3.44 2.69 Measured July 2004 3.47 2.72 Measured August 2004 3.52 2.78 Measured September 2004 3.70 3.01 Measured October 2004 3.72 3.10 Measured/WW Provisional November 2004 3.69 3.06 Measured/WW Provisional December 2004 3.65 2.98 Measured/WW Provisional January 2005 3.62 2.89 Measured/WW Provisional February 2005 3.59 2.86 TD Interpolated/WW Provisional March 2005 3.57 2.83 Measured/WW Provisional April 2005 3.55 2.82 Measured/WW Provisional May 2005 3.54 2.80 Measured/WW Provisional June 2005 3.51 2.80 Measured/WW Provisional July 2005 3.62 2.91 Measured/WW Provisional August 2005 3.62 2.92 Measured/WW Provisional September 2005 3.60 2.88 Measured/WW Provisional October 2005 3.57 2.84 Measured/WW Provisional November 2005 3.51 2.80 Measured/WW Provisional December 2005 3.49 2.72 Measured/WW Provisional January 2006 3.41 2.71 Measured/WW Provisional February 2006 3.37 2.70 Measured/WW Provisional & Interpolated March 2006 3.33 2.69 Measured/WW Provisional April 2006 3.25 2.63 Measured/WW Provisional May 2006 3.13 2.56 Measured/WW Provisional 4.8. SPRING CONDUCTANCE COEFFICIENTS Huang (1994) demonstrated that the spring conductance coefficient, or drain conductance, is a sensitive model parameter. Therefore, spring conductance coefficients were estima ted following the procedure described in Huang (1994):

PAGE 135

120 )( ph Q C eq. 4-2 where: spring conductance coefficient (L2/T) discharge (L3/T) aquifer head (L) pool stage (L) Estimated spring conductance coe fficients are calculated using the head difference between the spring pool stage, the aquifer head, at the WW-F conduit well, and spring discharge (), eq. 4-2. Twin D's fewer estimates of the spring conductance coe fficient shown in Table 4-6 than Weeki Wachee Spring as the pool stage at Twin D's Spring dropped slightly below water levels at the WW-F well, suggesting measur ement errors, or a possible shift in the spring basin boundary under low flow conditions (as shown in Figure 3-30). Therefore, those values were not used to estimate the spring conductance coefficient. Individual estimates based on observed values are summarized in Table 4-6. Rating curves were developed using the data summarized in Table 4-6 to obtain a more representative spring conduc tance coefficient for Twin D's and Weeki Wachee Springs using head differ ences between the pool stages and the WW-F conduit well and observed spring disc harges (shown in Figure 4-6). Spring conductance coefficients were also estimated for Weeki Wachee Spring using head differences between the pool stage and the WWSpg-ECK matrix well and the Weeki Wachee Deep matrix well, respectively. Spring conductance coefficients of 4 x 105 m2/d (n=6) and 9 x 104 m2/d (n=16) were obtained using the WWSpg-ECK and Weeki Wachee Deep well s, respectively. Locations for the WWSpg-Eck and Weeki Wachee Deep wells are provided in Figure 2-1. The estimated coefficient values using the WW-F conduit well we re considered valid C Q h pph Spring hasQ

PAGE 136

121 approximations for the purposes of this study since limited data exist for the WWSpg-ECK well and because Weeki Wachee Deep is farthest from the springs relative to the WW-F and WWSpg-ECK wells. Spring conductance coefficients, used in the groundwater flow models are summarized in Table 4-7.

PAGE 137

122 Table 4-6. Estimated spring conductance coefficients. Twin D's Spring Q m3/s Head Difference WW-F & Twin D's Spring m Spring Coefficient Twin D's Spring m2/s Spring Coefficient Twin D's Spring m2/d Weeki Wachee Spring Q m3/s Head Difference WW-F & Weeki Wachee Spring m Spring Coefficient Weeki Wachee Spring m2/s Spring Coefficient Weeki Wachee Spring m2/d 0.10 0.01 10.00 864000 4.57 0.76 6.01 519264 0.14 0.02 7.00 604800 4.54 0.77 5.90 509760 0.22 0.06 3.67 316800 4.72 0.80 5.90 509760 0.52 0.49 1.06 91690 5.94 1.12 5.30 457920 0.60 0.56 1.07 92571 6.47 1.18 5.48 473472 0.51 0.46 1.11 95791 6.27 1.09 5.75 496800 0.45 0.32 1.41 121500 5.08 0.83 6.12 528768 0.39 0.19 2.05 177347 4.86 0.77 6.31 545184 0.30 0.09 3.33 288000 4.75 0.77 6.17 533088 0.26 0.06 4.33 374400 4.65 0.73 6.37 550368 0.22 0.05 4.40 380160 4.33 0.64 6.77 584928 0.12 0.01 12.00 1036800 4.24 0.63 6.73 581472 4.09 0.60 6.82 589248 3.89 0.56 6.95 600480 Min91690 Min457920 Max1036800 Max600480 Median302400 Median530928

PAGE 138

123 Figure 4-6. Estimated spring conductance coefficients for Twin D's Spring (top) and Weeki Wachee Spring (bottom) using head differences between respective pool stages and the WW-F conduit well, and observed discharges. Conductance = 7 x 104 m2/d R2 = 0.920.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.00.10.20.30.40.50.6Head Difference WW-F-Twin D's (m)Discharge (m3/s)n=12 Conductance = 3 x 105 m2/d R2 = 0.973.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 0.00.20.40.60.81.01.21.4Head Difference WW-F-Weeki Wachee (m)Discharge (m3/s) n=14

PAGE 139

124 Table 4-7. Estimated spring conduct ance coefficients and drain conductances used in groundwater flow models. Report ed observed value shown in Figure 4-6. Twin D's SpringWeeki Wachee Spring m2/d m2/d Value used in models 7 x 1046 x 105Observed value 7 x 1043 x 105 4.9. WATER BUDGET Previous attempts have been made to es timate discharge from the coastal swamps based on analytical methods (flow net analysis) that assume homogeneous, isotropic, laminar flow conditions and require accurate transmissivities (Knochenmus and Yobbi 2001). An analytical water budget was not developed in this study due to insuffici ent data, specifically: 1) coastal swamp and tidally influenced spring discharges are not well known, and 2) the assumptions associated with the flow net method, which are likely violated near discharge points base on the analytical estimates for Reynolds numbers 105 and 106 discussed in Chapter 3, coupled with descriptions of rigorous flow by cave divers (Karst Underwater Research, In c., 2008c). A water budget for the Weeki Wachee and Twin Ds spring basins wa s not performed because there are too many unknowns. As stated in Chapter 3, the full extent of the conduit networks underlying Weeki Wachee and Twin Ds Springs are not known. Additionally, the size of the spring basins has not been delineated with dyetrace testing. Moreover, quantitative dye-tr ace testing has not been perfo rmed that is useful for estimating groundwater velocities in t he media underlying the springs.

PAGE 140

125 4.10. RANGE OF HYDRAULIC CONDUCTIVITY VALUES The range of hydraulic conductiviti es for the matrix and conduit or preferential flow pathways (4 to 3,810 m/d) is consistent among the three models, but the distribution varies upgr adient of Weeki Wachee and Twin Ds Springs and in the western portion of the model domain. Matrix hydraulic conductivities in the groundwater flow models are 2 to 5 orders of magnitude higher than measurements for core si ze samples from the Ocala Limestone reported in Florea (2006b) and Florea and Vacher (2007). However their measurements, which represent the matr ix, are on a relatively much smaller scale. Kiraly, (1975) and Ha lihan et al. (2000) report th at permeabilities increase with scale in dualpermeability karst aquifers. Quantitative dye-trace tests performed in the vicinity of Sulphur Sp ring in west-central Florida and the Woodville Karst Plain in northwest Flor ida vary from 2,200-6,000 m/d (Wallace, 1993; Kincaid et al., 2004) and are comparabl e to conduit hydraulic conductivities in the groundwater flow model s developed for this study. 4.11. EQUIVALENT-CONTINUU M MODEL (LAMINAR FLOW) Highly permeable zones or preferential flow pathways were incorporated into the equivalent-continuum groundwater flow models across broad areas including multiple cells. In order to simu late the observed flux at the springs, the highly permeable zones were extended ou tward from the springs guided by the proposed conceptual model of preferential flow pathways discussed in Chapter 3 (shown in Figure 4-7). Preferential flow pathways extended the entire vertical thickness of the Upper Floridan aquifer. Re latively higher hydraulic conductivities were extended across broad areas in the vicinity of preferential flow pathways shown in Figure 3-31. Constraining relatively higher hydraulic conductivities to narrower zones along the preferential flow pathways shown in Figure 3-31 decreased the match between observed and simulated water levels for the 32

PAGE 141

126 target wells. MODFLOW-2005 (Harbaugh, 2005) was used for the laminar equivalent-continuum models. 4.12. EQUIVALENT-CONTINUUM MODE L (LAMINAR/TURBULENT FLOW) The dual-conductivity groundwater flow model was developed by incorporating a mean void diameter and lower and upper Reynolds numbers for the transition from laminar to turbulent and turbulent to laminar flow. A mean void diameter of 6 cm was arbitr arily selected to represent vuggy porosity in the Upper Floridan aquifer for the numerical simula tions. There are no active cylindrical pipe networks in the laminar/turbulent equi valent-continuum model. An upper Reynolds number greater than 10 (from Freeze and Cherry, 1979) was originally selected for the laminar, equivalent-contin uum model, however it was gradually decreased to 2 to invoke turbulence. It is acknowledged that a Reynolds number of 2 indicates laminar flow, however the transition from laminar to turbulent flow, which is affected by grid discretizati on (Shoemaker et al., 2008b) was needed to invoke turbulence. MODFLOW-2005 CFP Mode 2 (Shoemaker et al., 2008a) was used for the laminar/turbulent equivalent -continuum models. Preferential flow pathways extended the entire vertical thickness of the Upper Floridan aquifer. Relatively higher hydraulic conductiviti es were extended across broad areas in the vicinity of preferential flow pat hways shown if Figure 3-31. The same hydraulic conductivity array used in the laminar, equivalent-continuum model was used in the laminar/turbulent, equ ivalent-continuum simulations.

PAGE 142

127 Figure 4-7. Hydraulic conductivity values (ECM K) used in the laminar and laminar/turbulent equivalent-c ontinuum models (top) and in the dual-conductivity (DCM K) models (bottom). Multiple types of data used to develop the conceptual model of preferential flow pathways discu ssed in Chapter 3 (Figure 3-31) were used to guide the assignment of relatively higher hydraulic conductivity values over broad areas. Verified fracture trace from Wood and Stew art (1985). Inferred fracture traces from Jones et al. (1997) Potentiometric contours for the Upper Floridan aquifer (UFA) Ma y 2006 from (Oritz, 2007). Cont our interval is 3 m and datum is NGVD.

PAGE 143

128 4.13. DUAL-CONDUCTIVITY MODEL (LAMINAR/TURBULENT FLOW) For the dual-conductivity model, the same procedure used for assigning preferential flow pathways in the equi valent-continuum models was followed (Figure 4-7). However, surveyed underwater caves underlying Weeki Wachee and Twin Ds Springs were explicitly in corporated as cylindr ical pipes embedded in the matrix. Underwater cave survey data and an inferred connection with the WW-F well that breaches the roof of an underwater cave were used to designate conduit cells. To date, a humanly ent erable connection between the Weeki Wachee Spring and the WW-F well has not been made (Karst Underwater Research, Inc., 2008b). However, in this study a connection between the Weeki Wachee Spring conduit network and the WW-F well is inferred. The conduit networks in the vicinity of Weeki Wac hee and Twin Ds Springs are known to extend beyond the survey data (Figure 319). Therefore, the conduit networks were extrapolated approximat ely 2 km beyond the terminus of the cave survey data. The Weeki Wachee Spring conduit network is extrapolated to the approximate location of the accidental tr acer event near Crescent Lake (see Chapter 3). The extrapolat ed conduits represent t he unmapped portions of the conduit networks in the dual-conductivity models (shown in Figure 4-8). Inferred preferential flow pathways (other than the surveyed und erwater caves) were not explicitly incorporated into the dual-conductivity model as pipes, but rather were represented as the bulk hydraulic conducti vities for the matrix and conduits similar to that typically used in equiva lent-continuum models. These preferential flow pathways extended the entire vertical thickness of the Upper Floridan aquifer. MODFLOW-2005 CFP Mode 1 (Shoem aker et al., 2008a) was used for the laminar/turbulent dual-conductivity models. Conduit diameters were loosely constrained using underwater cave survey data provided by Karst Underwater Research, Inc. (2008a). Conduit diameters were slightly exaggerated because reducing the diameters by a factor of 0.01 essentially decoupled the matrix and conduit networks in the dual-conductivity models. Conduit elevations were also

PAGE 144

129 loosely constrained using the cave surv ey data provided by Karst Underwater Research, Inc. (2008a). The conduit netwo rks are perennially water-filled at the test-site, so elevation is not as sensitiv e of a parameter relative to a test-site simulating flow in a conduit network that is partly water-filled. Conduit wall conductances were constrained using the observed water levels and head differences between wells monitoring the ma trix and conduit network as shown in Figures 3-23 through 3-25. That is, condui t wall conductances were varied until an acceptable match between observed and simulated water levels, or head differences between matrix and conduit netwo rks in the vicinity of the WW-F conduit well were obtained. Direct rec harge was not applied to the conduit networks underlying Weeki Wachee and Twin Ds Springs as recharge to those networks is diffuse based on the observ ed response in well/spring hydrographs discussed in Chapter 3 and shown in Fi gures 3-23 through 3-25 and Figure 3-28.

PAGE 145

130 Figure 4-8. a) Location of model cells that contain conduit nodes in the dualconductivity models. Verified and provisio nal cave survey data provided by Karst Underwater Research, Inc. (2008a) was used to delineate conduit cells. The conduit cells were extrapolated approximat ely 2 km beyond the terminus of the cave survey data. The conduit network underlying Weeki Wachee Spring was extended to the approximate location of an accidental tracer event (discussed in Chapter 3). The location of the W eeki Wachee Trough using May 2006 Upper Floridan aquifer (UFA) water levels (Ortiz, 2007) and the location of inferred a b c Crescent Lake

PAGE 146

131 fracture traces (Jones et al., 1997) are shown for comparison. The circled waterfilled caves shown in the upper right cor ner represent the approximate location for the injection site for t he tracer test reported in J ones et al. (1997), b) conduit cells with cave survey data superimposed. Hydraulic conductivity values for model cells surrounding conduit cells are lower than hydraulic conductivity values assigned to conduit cells), c) the conduit tubes show connection among the conduit nodes. Note the inferred connecti on of the WW-F conduit well with the Weeki Wachee Spring conduit network. 4.14. SENSITIVITY ANALYSIS Sensitivity analysis was performed on boundary conditions and several model parameters. Sensitiv ity analysis provided useful information regarding the traditional parameters that affect model performance and insight into the sensitivity of new parameters introduced with MODFLOW-2005 CFP (Shoemaker et al., 2008a), which provides directi on into future data collection needs. Sensitivity analysis was only performed on the dual-conductivity steady-state and combined steady-state/transient simulations using MODFLOW-2005 CFP Mode 1 as this includes traditional MODF LOW parameters in addition to new parameters introduced wit h MODFLOW-2005 CFP. 4.14.1. BOUNDARY CONDITIONS Boundary conditions were evaluated for the steady-state, dual-conductivity model by changing specif ied-head boundaries to gener al-head boundaries and quantifying the effects of changing t he boundary conditions on spring flow at Weeki Wachee and Twin Ds Springs. The procedure was repeated by changing no-flow boundaries to general-head boundaries, and general-head boundaries to specified-head boundaries. C hanging the boundaries did not significantly affect simulated discharges at Weeki Wachee or Twin Ds Springs (Table 4-8). A previous model constructed in the vicinity of the study area by Blanford and Birdie (1993) has the southern, eas tern, and northern boundaries as no-flow

PAGE 147

132Weeki Wachee SpringTwin D's Spring (m3/s) (m3/s) Steady-state 4.92 0.27 General-head to specified-head 4.81 0.25 Specified-head to general-head 4.93 0.28 No-flow to specified-head 4.95 0.28 hydrologic barriers. They identify probl ems with simulating the potentiometric highs in the eastern portions of their model domain and speculate that it is related to the use of no-flow boundaries. Flux into and out of the study area occurs along portions of the northern, eastern, western, and southern boundar ies as head changes with time (see potentiometric surface maps Figure 3-15), therefore designation as general-head or specified-head boundaries are conceptua lly more defensible than no-flow boundaries. Net fluxes across the model domain are provided in Figure 4-9. Additional sensitivity analysis on general-head conductances and head values are presented later. Table 4-8. Effect of varying model boundar y conditions on simulated discharge at Weeki Wachee and Twin Ds Springs in the steady-state, dual-conductivity model.

PAGE 148

133 Figure 4-9. Net flux (m3/d) across steady-state, dual-conductivity model boundaries. 4.14.2. MODEL PARAMETERS Sensitivity analysis was performed using: 1) the steady-state, dualconductivity model (net recharge, hydrau lic conductivity, well flow rate, generalhead conductance, general-head value, spring conductance coefficients, pool stage), and 2) the combined steady-state/t ransient, dual-conductivity simulations (storage, conduit wall conduc tance between the matrix and conduit networks, conduit diameter, and the sensit ivity of turbulent flow on simulated discharge). Net recharge, hydraulic conductivity well flow rate, and general-head conductance for the steady-state, dual-conductivity model were varied by adjusting calibrated values with multip liers of 0.01 and 100, similar to the -6 x 10 5 6 x 10 No-flow General-head Specified-head County Boundary 4 -5 x 10 5 -5 x 10 4

PAGE 149

134 procedure used by Martin and Whiteman (1990). Additionally, general-head values were evaluated by adjusting general head values to negative 1.52 m and positive 1.52 m from calibrated values. Sensitivity of net recharge, hydraulic conductivity, well flow rate, general-head conductance, and general-head values were evaluated in terms of the residual mean, residual standard deviation, and residual sum of squares for the 32 target wells, (see Table 4-9). Sensitivity of spring conductance coefficients for Week i Wachee and Twin Ds Springs, pool stages (Table 4-9), storage values (shown in Figure 4-10), conduit wall conductance between the matrix and condu it networks, and c onduit diameters (shown in Figure 4-11) were evaluated in terms of their effect on simulated discharge at Weeki Wachee and Twin D s Springs. Sensitivity of spring conductance coefficients was performed by varying calibrated values by multipliers of 0.01 and 100. Pool stage was evaluated by adjusting the observed pool stages by negative 0.30 m and positiv e 0.30 m from observed values. Additionally, analysis was performed to quantify how much of the apparent change in the performance of the combi ned steady-state/transient simulations was due to actual changes in the distribut ion of hydraulic conductivities, which generally were decreased in the western portions of the model domain in the dual-conductivity models relative to t he equivalent-continuum models (see Figure 4-7). This was important because hydraulic conductivity is the second most sensitive parameter, as shown in Table 49. This could be accomplished in either one of two ways: 1) by inco rporating the hydraulic conductivities from the dualconductivity model into the laminar, equi valent-continuum model, or 2) by decoupling fluid exchange bet ween the matrix and conduit networks by setting the conduit wall conductance to zero. In this study, both options were applied to verify that similar results were obtained. Moreover, the effect of turbulent flow on simulated discharges at Twin Ds and Weeki Wachee Springs was evaluated for the combined steadystate/transient dual-conductivity m odel using MODFLOW-2005 CFP Mode 1, where laminar flow in the conduit network is simulated using the Hagen-

PAGE 150

135 Poiseuille equation and turbulent flow is simulated using the Darcy-Weisbach equation. This was accomplished by increasing the upper Reynolds number that is used to invoke turbulent flow. Additionally, the effect of varying the mean void diameter on simulated discharges in the laminar/turbulen t, equivalent-continuum model using MODFLOW-2005 CFP Mode 2 was investi gated. Lastly, 2 scenarios were performed to determine if results similar to those discussed in the GeoTrans, Inc. (1988a; 1988b) study could be replicated. The first scenario involved incorporating the inferred fracture traces plus the one verified fracture trace shown in Figure 3-14 into the laminar equivalent-continuum simulations. The incorporated fracture trac es extended the entire thick ness of the Upper Floridan aquifer and were assigned hydraulic conducti vity values of 3,810 m/d. Recharge was shut-off in the laminar, equivalent -continuum simulations for the second scenario.

PAGE 151

136 Table 4-9. Effect of varying net recharge, hydraulic conductivity, well flow rate, general-head conductance, and general-head values on the residual m ean, residual standard deviation, and the re sidual sum of squares for the 32 target wells for the steady-s tate dual-conductivity model. MultiplierResidual MeanResidual Standard DeviationResidual Sum of Squares mmm2NET RECHARGE 0.01 3.58 2.15 560.15 1 -0.39 0.98 35.86 100 -385.45 189.07 5898800.73 no recharge3.63 2.17 573.16 HYDRAULIC CONDUCTIVITY 0.01 -206.93 114.15 1783574.39 1 -0.39 0.98 35.86 100 8.67 6.53 3762.23 WELL FLOW RATE 0.01 -0.84 1.02 55.46 1 -0.39 0.98 35.86 100 62.10 26.25 145843.54 no pumping-0.82 1.01 54.71 GENERAL-HEAD CONDUCTANCE 0.01 -2 1 186 1 -0.39 0.98 35.86 100 -0.3 1 32 GENERAL-HEAD VALUE m Negative 1.520.4 1 47 Calibrated Value-0.39 0.98 35.86 Positive 1.52-1 1 76

PAGE 152

137 Table 4-9 cont. Effect of varying spri ng conductance coefficients and pool stages on simulated discharge for Weeki Wachee and Twin Ds Springs for the steadystate dual-conductivity model. MultiplierConductanceDischarge m2 / dm3/s DRAIN CONDUCTANCE Weeki Wachee Spring 0.01 6 x 1030.16 16 x 1 054.92 100 6 x 1076.92 Twin D's Spring 0.01 7 x 1020.00 POOL STAGE 17 x 1 040.27 100 7 x 1061.59 StageDischarge mm3/s Weeki Wachee SpringNegative 0.30 m2.545.52 Calibrated Value2.844.92 Positive 0.30 m3.154.31 Twin D's Spring Negative 0.30 m3.260.49 Calibrated Value3.570.27 Positive 0.30 m3.870.06

PAGE 153

138 Figure 4-10. Plots showing the effect of varying the storage coefficient and specific yield values in the mantle and the Upper Floridan aquifer (UFA) on simulated discharge for Weeki Wachee Spring (top) and Twin Ds Spring (bottom). Only the transi ent stress periods (2 through 25) are shown. Adjusted storage values are shown in the legend for each spring. 1 2 3 4 5 6 7 8 May-04Sep-04Jan-05May-05Sep-05Jan-06May-06Time (month)Q (m3s) Observed Mantle0.15_UFA0.05 Mantle0.15_ UFA0.1 Mantle0.25_UFA0.005 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 May-04Sep-04Jan-05May-05Sep-05Jan-06May-06Time (month)Q (m3s) Observed Mantle0.15_UFA0.05 Mantle0.15_UFA0.1 Mantle0.25_UFA0.005

PAGE 154

139 1 2 3 4 5 6 7 8 May-04Sep-04Jan-05May-05Sep-05Jan-06May-06Time (month)Q (m3s) Observed ECM MODFLOW-2005 Conduit Wall Conductance Reduced x 0.01 Conduit Diameter Reduced x 0.01 DCM MODFLOW-2005 CFP Mode 1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 May-04Sep-04Jan-05May-05Sep-05Jan-06May-06Time (days)Q (m3s) Observed ECM MODFLOW-2005 Conduit Wall Conductance reduced x 0.01 Conduit Diameter Reduced x 0.01 DCM MODFLOW-2005 CFP MODE 1 Figure 4-11. Plots showing the effect of reducing conduit wa ll conductance and conduit diameter in the dual-conductivity m odel by a factor of 0.01 on simulated discharge for Weeki Wachee Spring (top) and Twin Ds Spring (bottom). Only the transient stress periods (2 through 25) are shown.

PAGE 155

140 4.15. RESULTS OF SENSITIVITY ANALYSIS Sensitivity analysis indicates that net recharge is the most sensitive model parameter in terms of the residual m ean, residual standard deviation, and the residual sum of squares for the 32 target wells, followed by hydraulic conductivity and well flow rate (Table 4-9). Simulated discharge in the dual-conductivity model is sensitive to spring conductance coefficients and pool stage (Table 4-9). Observed pool stages fluctuated by 0.54 to 0.58 m for Weeki Wachee and Twin D's Springs, respectively duri ng the 24 months of study. Storage values ranging from 0.15 to 0.25 for the mantle and 0.05 to 0.003 for the Upper Floridan aquifer were ev aluated. Higher storage values for the mantle generally attenuate the response fo llowing the tropical storms (Figure 410). Adjustments to the conduit wall conductance between the matrix and conduit networks and conduit diameters produced a narrower range (Figure 411) in simulated discharges relative to adjustments made to the spring conductance coefficients observed in Table 4-9. In fact, reducing the parameters by a factor of 0.01 essentially decoupled the fluid exchange between the matrix and conduit networks as the simulated discharges match simulations in which the conduit and matrix netwo rks are decoupled. Simulated discharge at Weeki Wac hee and Twin Ds Springs using MODFLOW-2005 CFP Mode 1 varied by less than 1% if flow was laminar (Hagen-Poiseuille) or turbulent (DarcyWeisbach) in the conduit network. Decoupling the matrix and conduit netwo rks combined with using the hydraulic conductivity array originally used with the dual-conductivity simulations and its effect on simulated discharge at Weeki Wachee and Twin Ds Springs is discussed in section 4.16.1. Adjustments to reasonable mean void diameters and Reynolds numbers for the laminar/turbulent equivalent-continuum model using MODFLOW-2005 CFP Mode 2 invoked few occurrences of turbulent flow in the vicinity of Weeki

PAGE 156

141 Wachee and Twin Ds Springs, (see Figure 412). This may indicate that the grid spacing (152 m x 152 m) was too large, and therefore requiring a larger critical gradient to invoke turbulent flow, to effectively applying Mode 2 of MODFLOW2005 CFP as discussed in Shoemaker et al. (2008b). Figure 4-12. Location of turbulent flow in the combined steady-state/transient, laminar/turbulent equivalent-continuum model for stress period 5 (September 2004) using MODFLOW-2005 CFP Mode 2. Turbulent flow occurred just north of Weeki Wachee Spring. Values of 0 represent laminar flow to the right and front of the model cell, values of 1 r epresent turbulent flow to t he right and laminar flow to the front of the model cell, and values of 3 represent tu rbulent flow to the front and right of the model cell (Shoemaker et al., 2008a). Values of 2, which represent turbulent flow to the front and laminar flow to the right of the model cell did not occur during stress period 5. Incorporating the fracture traces shown in Figure 3-14 qualitatively decreased the match between observed and simulated water levels for the 32 0 1 3

PAGE 157

142 target wells in the combined steady-state/transient laminar, equivalent-continuum simulations. This is similar to the resu lts obtained in the GeoTrans, Inc. (1988a) study that indicated incorporating inferr ed fracture and fault traces into the laminar, equivalent-continuum model affected model calibration. Shutting-off recharge in the combi ned steady-state/transient laminar, equivalent-continuum simulations did not result in a simulated cessation of flow at Weeki Wachee Spring, however simulated discharges were, on average only 27% of observed discharges at Weeki Wachee Spring. Shutting-off recharge in the dual-conductivity model resulted in simulated discharges that were on average 38% of observed discharge at Week Wachee Spring. This is consistent with the observations noted in GeoTrans, Inc. (1988b). That is, using the dualconductivity model improves the match between observed and simulated discharges when recharge is shut-off. T he difference between this study and the previous laminar equivalent-continuum m odel discussed in the GeoTrans, Inc. (1988b) study however, is that simulated discharges at a first magnitude spring (Weeki Wachee Spring) did not show a cessation of flow when recharge was shut-off in the laminar, equi valent-continuum model. 4.16. Model Performance Evaluation Results The ability of laminar and laminar /turbulent equivalent-continuum and dual-conductivity groundwater flow model s to capture the dynamic hydraulic response during high recharge events and through drought conditions (June 2004 through May 2006) using 3 different conduit conceptualizations was evaluated. Model performance was evaluat ed by comparing the match between: i) observed and simulated discharge at T win D's and Weeki Wachee Springs, ii) observed and simulated water levels for 32 target wells, iii) observed and simulated head differences between the matrix and conduit network, and iv) model statistics in terms of the resi dual mean, residual standard deviation, and

PAGE 158

143 residual sum of squares for 32 target we lls in each of the 3 groundwater flow models. The well locations are provided in Figure 4-1. 4.16.1. Spring Discharges The transient stress periods (2 through 25) in the combined steadystate/transient, dual-conductivity model, on average, simulated 89% of observed discharges at Weeki Wachee Spring vers us 77% for the comparable laminar, equivalent-continuum simulations. Simulated discharges at Twin Ds Spring for the transient stress periods using the dual-conductivity model, on average, were 85% of observed values versus 45% fo r the laminar, equivalent-continuum model. During the last stress period (25), when the effects of a drought became apparent (i.e. a cessation of flow occu rred at Twin Ds Spring), the dualconductivity model simulated 86% of observed discharges at Weeki Wachee Spring versus 72% in the comparable laminar, equivalent -continuum model. Simulated discharges at Twin Ds during the same stress period were slightly higher, 0.1 m3/s in the dual-conductivity model versus 0.03 m3/s in the laminar equivalent-continuum model. The observed discharge at Twin Ds Spring during stress period 25 was 0 m3/s. Conversely, simulated discharges for Weeki Wachee Spring using the laminar/turbul ent, equivalent-continuum model showed a slight decrease in simulated spring flow s relative to the comparable laminar, equivalent-continuum model (F igure 4-13). Turbulent flow occurred in only a few cells near Weeki Wachee Spring (see Figur e 4-12). Simulated spring flows at Twin Ds Spring using the laminar/turbul ent, equivalent-continuum model showed little change relative to the comparable laminar equivalent-continuum model. Decreasing portions of the bulk hydraulic conductivity values upgradient of the springs, and in the northwest portion of the study area, while maintaining relatively higher hydraulic conductivity values in the conduit networks attributed to the improved performance in simulati ng discharges using the dual-conductivity model relative to the laminar equival ent-continuum model (K array shown in

PAGE 159

144 Figure 4-13). Eight perc ent of the improved perfo rmance at Weeki Wachee Springs in the dual-conductivity simula tions was due to decreasing the bulk hydraulic conductivities. Reducing bulk hydrau lic conductivities alone resulted in an average simulated discharge value abov e the average observed value in the equivalent-continuum model for Twin Ds Spring. Permitting fluid exchange between matrix and conduit networks in t he dual-conductivity model permitted a better match with observed values for both springs (Figure 4-13). Simulated discharges for each of the 3 groundwater flow models are provided in Appendix E.

PAGE 160

145 Figure 4-13. Plots of observed and si mulated discharges for the laminar equivalent-continuum (ECM MODFLOW2005), the dual-conductivity (DCM MODFLOW-2005 CFP Mode 1), the laminar equivalent-continuum model using the hydraulic conductivity array (K arra y) used for the dual-conductivity model, and the laminar/turbulent equivalent-cont inuum model (ECM MODFLOW-2005 CFP Mode 2). Only the trans ient stress periods (2 through 25) are shown.

PAGE 161

146 4.16.2. Wells The dual-conductivity model simulated s lightly higher water levels in the target wells relative to the laminar and laminar/turbulent equivalent-continuum models (Figure 4-14) due to decreases in bulk hydraulic conductivities for portions of the model domain (shown in Figure 4-7). However, the match between observed and simulated water leve ls for the 32 target wells for all 3 groundwater flow models is fairly well. In fact, the average absolute difference between observed and simulated water levels, relative to the 32 target wells, for each of the 3 groundwater flow models fo r the 25 stress periods is 0.77 m or less. Simulated water levels for each of the 3 groundwater flow models are provided in Appendix E. Hydrographs for all thirty-two target wells are shown in Appendix F.

PAGE 162

147 Figure 4-14. Plots of observed and simula ted water levels in the Weeki Wachee Deep (matrix well), WW-F (conduit well), and ROMP 98 (matrix well) for the laminar equivalent-continuum (ECM MODFLOW-2005), the dual-conductivity (DCM MODFLOW-2005 CFP Mode 1), the laminar equivalent-continuum model using the hydraulic conductivity array (K array) used for t he dual-conductivity model, and the laminar/turbulent (ECM MODFLOW-2005 CFP Mode 2) models. Only the transient stress periods (2 through 25) are shown.

PAGE 163

148 4.16.3. Simulated Matrix and Conduit Water Levels and Head Differences Observed matrix and conduit water levels in the vicinity of the WW-F well that breaches the roof of an underwater cave closely mimicked each other and shared a similar response to high rec harge events. Moreover, observed head differences between the WWSpg-ECK ma trix well and the WW-F conduit well, that are 920 m apart, were typically le ss than 0.03 m (shown in Figures 3-23 through 3-25). The observed water levels re flect a shorter time frame during the high frequency monitoring period relative to the time frame for the transient stress periods, but simulated water levels in the dual-conductivity model for both the matrix and conduit water levels in t he model cell that contains the WW-F well closely mimic each other and share a sim ilar response to the observed data shown in (Figure 4-15). Moreover, simu lated head differences for the matrix and conduit network in the vicinity of the WW-F well favorably match observed head differences between the WWSpg-ECK and WW-F wells.

PAGE 164

149 Figure 4-15. a) Plot of obs erved water levels for the Weeki Wachee Deep matrix well (WW Deep), the WWSpg-ECK matrix well, and the WW-F conduit well that breaches the roof of an underwater ca ve. The WWSpg-ECK and WW-F wells are 920 m apart and b) simulated matrix and conduit water levels in the model cell that contains the WW-F well. 4.16.4. Model Statistics The residual mean, residual standar d deviation, and residual sum of squares for the 32 target wells are si milar among the groundwater flow models a b

PAGE 165

150 (Table 4-10) indicating that the calibrations are comparable among the 3 groundwater flow models. Table 4-10. Comparison of model statistics for the 32 target wells among the 3 steady-state/transient groundwater flow models for all 25 stress periods. Residuals calculated as obser ved minus simulated values. Residual mean Residual Standard De viation Residual Sum of Squares (m) (m) ( m2 ) ECM LAMINAR 0.15 0.98 779.28 ECM LAMINAR/TURBULENT 0.13 0.97 770.04 DUAL-CONDUCTIVITY LAMINAR/TURBULENT -0.27 0.97 812.49 4.17. DISCUSSION Based on the numerical results from this study, the Reynolds numbers may be highly uncertain when applying MODFLOW-2005 CFP Mode 1. The direction of flux from or into the matr ix was affected by turbulent flow, however, simulated discharges using MODFLOW-2005 CFP Mode 1 differed by less than 1% with laminar flow (Hagen-Poiseuille) or turbulent flow (Darcy-Weisbach) in the conduit network. The direction of flux into or out of the matrix is important for karstification processes (Bauer et al 2003) or contaminant transport, but karstification processes and contaminant tr ansport cannot currently be simulated with MODFLOW-2005 CFP (Shoemaker et al. 2008a). These findings are supported by previous numerical studies t hat did not account for turbulent flow. For example, improvements between t he match for simulated and observed discharges using a laminar, dual-poros ity model was documented in the GeoTrans, Inc. (1988a) study discussed in Chapter 1. The GeoTrans, Inc. (1988a) study did not account for turbulent flow in the conduit network. Moreover, Mohrlok and Sauter (1997) conclude that comparable dual-permeability and dualcontinua models used for a test-site in Swabian Alb, Germany were able to adequately simulate observed discharge and aquifer heads. Turbulent flow was

PAGE 166

151 simulated in the dual-permeability model and laminar flow was simulated in the dual-continua model (Mohrlok and Sauter, 1 997). The previous studies suggest that the conduit wall conductance, or linear exchange term, between the matrix and conduit networks is an important component of numerical simulations for dual-permeability karst aquifers, such as the Upper Floridan aquifer. In this study, widespread turbulent fl ow did not occur throughout the model domain (see Figure 4-12) using MODFLO W-2005 CFP Mode 2. The grid spacing used in this study may be a factor for the performance of the laminar/turbulent equivalent-continuum model. Although tur bulent flow was not a sensitive parameter for this test-site, Shoemaker et al. (2008b) found that turbulence affected simulated aquifer heads and ot her parameter sens itivities using MODFLOW-2005 CFP Mode 2 for the Biscayne aquifer, which consists of vuggy porosity. Widespread turbulent flow o ccurred throughout the model domain used for the Biscayne aquifer. Although widespread turbulent flow ma y not occur in the Upper Floridan aquifer underlying west-central Florida, it does occur where constrictions exist in the underling conduit network. Evidence for turbulent flow is provided in the analytical estimates fo r Reynolds numbers (105-106), which indicate turbulent flow, and descriptions of rigorous flow fr om cave divers (Karst Underwater Research, Inc., 2008b; Karst Underwate r Research, Inc., 2008c) discussed in Chapter 3. Quantitat ive dye-trace testing could confirm turbulent groundwater velocities. However, turbulent flow alone, in the absence of fluid exchange, will not likely improve the match between observed and simulated spring flows in transient simulations for the test-site, particularly for areas near the first magnitude spring, even Twin Ds Spring, a relatively smaller spring, showed an improvement in simulated discharges us ing the dual-conductivity model (Figure 4-13). Quantitative dye-tracing could also provide flow rates, or useful information regarding groundwater velocities at the test-site which could be used to constrain hydraulic conductivity values in the groundwater flow models. Based

PAGE 167

152 on the results of this study, the GeoTrans (1988a) and Mohrlok and Sauter (1997) studies, it is apparent that accounting for fluid exchange between the matrix and conduit networks is import ant for simulating discharge in dualpermeability karst aquifers. Adequate understanding of conduit locations, groundwater velocities, and the direction and magnitude of fluid exchange are needed, which requires collecting field data traditionally reserved for contaminant transport studies. Hydraulic conductivity, the second mo st sensitive model parameter, did vary among the groundwater flow models, ther efore it was important to verify that the differences in hydraulic conductivities were not the only cause for improving the simulated discharges. Indeed 8% of t he increased simulated spring flow at Weeki Wachee Spring can be attributed to changes in hydraulic conductivities upgradient of the springs and the northwe st portion of the dual-conductivity model. For Twin Ds Spring, reducing hy draulic conductivities resulted in an average simulated discharge value above the average observed value in the equivalent-continuum model. Permitting fluid exchange between matrix and conduit network in the dual-conductivity model permitted a better match with observed values for both springs. The improved match between simulated spring flows at Weeki Wachee and Twin Ds Springs using the dual-conductivity model is modest in comparison to model sensitivities associated with variations in the values of drain conductances (see Table 4-9), which produced a wider range of fluctuations in simulated discharges. However, a relatively conservative approach for incorporating the conduit networks into t he dual-conductivity model was used in this study. Only the conduit networks in the vicinity of the springs were explicitly incorporated into the dual-conductivity models. Although the conduits were extrapolated beyond the surveyed passages (Figure 4-8), the actual conduit network may extend further than the network incorporated into the dualconductivity models. Additional underwa ter cave locations within the model domain were not included in the dual-con ductivity model because they have not

PAGE 168

153 been surveyed, or survey data has not been released. Including these additional conduit locations in the dual-conductivi ty model would likely increase the simulated spring flows and provide a better match between observed and simulated aquifer water levels. Equivalent-continuum model s generally do not perform as well for a dualpermeability karst aquifer, such as the Upper Floridan aquifer, because it can be difficult to simulate discharge and aquifer water levels during periods of low net recharge and high net recharge using the same model, particularly in areas influenced by the underlying conduit netwo rk. Relatively larger hydraulic conductivities are typically needed to adequately simulate annual average discharge for first magnitude springs. Si mulated discharges can decrease when net recharge declines in transient simu lations that include drought conditions. Moreover, reducing bulk hydraulic conducti vities in equivalent-continuum models can produce bias in simulated aquifer water levels, whereas doing so with dualconductivity models mitigates bias because fluid exchange between the matrix and conduit networks reduces overpr essurization in the aquifer. This study quantifies the increased performance of dual-conductivity models over equivalent-continuum models fo r simulating discharge in the dualpermeability, Upper Floridan aquifer. The dual-conductivity model permits one to reduce bulk hydraulic conductivities while increasing both hydraulic conductivities in the conduit network and simulated spring flows. Drainage from the matrix into the conduit networks reduces aquifer heads that increase by simply reducing bulk hydraulic conductivities in equivalent-continuum models. Application of dual-conductivity models in areas influenced by conduits, particularly near discharge points where fairly large fluxes occur, will most likely outperform comparable equiva lent-continuum models because dual-conductivity models simulate the fluid exchange bet ween the matrix and conduit networks that is characteristic of both flashy and slow responding karst aquifers. As such, dual-conductivity models are conceptually more defensible than equivalentcontinuum models. Karst hydrologists hav e been very vocal about the limitations

PAGE 169

154 with using equivalent-conti nuum models in dual-perme ability karst aquifers, particularly when it comes to delineati on of springhead or wellhead protection areas. For example, Wilson (2 002) states emphatically t hat, hydrologic modelers . applying diffuse flow equations to cavernous aquife rs . should be immediately recognized as intellectually dishonest. Quinlan et al. (1995) further state that: Numerical flow models can be useful and can approach validity in unconfined carbonate aquifers at the regi onal scale. But they are not valid until or unless they have been history -matched with tracer test data . [a]ll to often, however, the anisotropic, dualor triple-porosity nature of the aquifer is either ignored or rationalized out of existence by the modelers (e.g., Blanford and Birdie, 1993).

PAGE 170

155 One can argue that equivalent-continuum models are conservative tools, for the Upper Floridan aquifer because they may simulate larger reductions in discharge during drought conditions due to inflated bulk hydraulic conductivities that are not supported by the observed dat a. Conversely, hydraulic conductivity values can be increased over large areas to maintain spring flows during drought conditions, however doing so can resu lt in simulated spring flows above observed values during average or high recharge conditions. However, ignoring the fluid exchange that occurs between t he matrix and conduit networks, which is an important component of discharge in dual-permeability aquifers, is conceptually incorrect. This also touches upon an important concept discussed in Hunt et al. (2007). Hunt et al. ( 2007) eloquently quote a theory of Albert Einsteins, that our approach to probl em solving should be as simple as possible, but not simpler, (Hunt et al. 2007). Results from this study and previous studies (Mohrlok and Sauter 1997; GeoTrans, Inc. 1988b) demonstrate the increased performance of using dual-co nductivity numerical models in terms of simulating discharge with the addition of the conduit wall conductance parameter. This parameter permits fluid exchange between the matrix and conduit networks. Ultimately, the purpose fo r the groundwater flow m odel must be taken into account before beginning any numerical effort, and while t he new parameters and characterization requirements introduced with MODFLOW-2005 CFP Mode 1 may seem extensive, the increas ed parameterization (i.e. conduit wall conductance) is justified conceptually and performance-wise. Moreover, with the release of MODFLOW-2005 CFP (Shoemaker et al., 2008a) in the public domain, the continued applic ation of equivalent-continuum models for simulating flow in dual-permeability karst aquifers in areas strongly influenced by fluid exchange between the matrix and conduit networks will become increasing less defensible.

PAGE 171

156 CHAPTER 5: CONCLUDING REMARKS 5.0. PROJECT CONCLUSIONS 1) Analytical estimates of Reynolds numbers (105 and 106) coupled with descriptions of rigorous flow by cave di vers (Karst Underwater Research, Inc., 2008b; Karst Underwater Research, Inc., 20 08c) indicate that turbulent flow occurs in portions of the conduit net works underlying Weeki Wachee and Twin Ds Springs, particularly where constricti ons occur in the aquifer. However, based on the numerical results from this study the Reynolds numbers may be highly uncertain when applying MODFLOW-2005 CFP Mode 1 to the test-site. Simulated discharges using MODFLOW2005 CFP Mode 1 differed by less than 1% with laminar flow (Hagen-Poiseuille) or turbulent flow (Darcy-Weisbach) in the conduit network. Based on the analytical estimates, laminar flow may occur in portions of the conduit net works underlying Weeki Wachee and Twin Ds Springs where large conduits occur in the aquifer. 2) A conceptual model of conduit locati ons and preferential flow pathways was developed for the test-site using multiple data types. Preferential flow pathways were inferred where multiple types of data coincide. The proposed conceptual model indicates that relatively higher hydraulic conductivities underlie the western portion of the st udy area and the breached por tion of the Brooksville Ridge located in the northeast corner of the study area w here minimum aquifer thickness occurs. Relatively lower hydr aulic conductivities underlie the southeast corner of the study ar ea where maximum aquifer thickness occurs.

PAGE 172

157 3) Probability estimates for in tercepting a cavity in t he Upper Floridan aquifer at the test-site (10%) are higher than probabili ty estimates reported for a Paleozoic limestone aquifer (from 0.4 to 3%; Worthington et al., 2000b), but are lower than estimates for the Floridan aquifer (25-50% ) reported in Wilson (2002). Plots with the modes of intercepted cavity elevations (Figure 3-18) indicate that the cavities are not regionally laterally extensive. This may be a reflection of several factors, including speleogenesis and the media that comprise the Upper Floridan aquifer, which can be friable (Table 3-1) and theref ore, subject to collapse. Therefore, relatively higher probability estimates for intercepting a cavity during drilling does not indicate that a laterally-continuous regionally-extensive, well integrated conduit network underlies the test-site. 4) Significant water level changes were not observed in the matrix or conduit networks in response to convective stor m activity, however noticeable changes occurred in response to tropical storm activity. Hydrographs for wells and springs across the study area shown in Figur es 3-23 through 3-24 and 3-27 provide empirical evidence showing that, while t he springs are conduit fed, the response in the matrix and conduit networks based on the shallow recession limbs following tropical storm activity show a sl ow or diffuse response. Moreover, the shallow recession limbs observed for hydrographs of discharge and from monitoring wells penetrating both the ma trix and conduit network following passage of the tropical storms indicate t hat the conduit network is not regionally interconnected. This also indicates that the matrix has a very large storage capacity. A similar response was observed in the Upper Floridan aquifer north of the study area (Florea and Vacher, 2007). 5) Water level data for a monitoring well t hat breaches the roof of an underwater cave (WW-F) and a matrix well (WWS pg-ECK) located 920 m from the WW-F well show that heads in the matrix netwo rk generally exceed those in the conduit network even during high recharge events. Although the magnitude of the fluxes

PAGE 173

158 between the matrix and conduit networks varied temporally, the observed direction of fluid exchange from the matrix into the conduit network was primarily unidirectional, indicating t hat the primary role of the conduit network in the study area is to drain the matrix. Head differences between the matrix and conduit network also suggest that the conduit network feeding the springs is not connected to point sources of recharge and that diffuse recharge dominates. 6) Hydrographs for monitoring wells penet rating the matrix and conduit networks closely mimic each other with water levels rising rapidly in both networks following the passage of two tropical storms which implies that a high leakance, or conductance, exists between the matr ix and conduit networks. A corollary is that the hydraulic conducti vities of each network do no t differ significantly from each other. This is an important observ ation because this data was used to constrain the conduit wall conductanc e in the dual-conductivity models. Moreover, simulated water levels and head differences for the matrix and conduit network in the vicinity of the WW-F conduit well, using the dual-conductivity model, agree favorably with observed data (Figure 4-15). 7) The dual-conductivity model best simula ted transient spring discharges as compared to the results of the la minar and laminar/turbulent equivalentcontinuum models. The dual-conductivi ty model improved the match between simulated and observed discharges by an average of 12% at Weeki Wachee Spring. Approximately 8% of the improvement was t he result of decreasing hydraulic conductivities upgradient of the springs and the northwest portion of the model domain and the remaining 4% was from fluid exchange between the matrix and conduit networks. The dual-cond uctivity model improved the match between simulated and observed discharges by an average of 40% at Twin Ds Spring. Accounting for fluid exchange bet ween the matrix and conduit networks improved the match between observed and simulated discharges at Twin Ds Spring over reductions in bulk hydraulic conductivities alone.

PAGE 174

159 8) The performance of the dual-conductivity model over the laminar, equivalentcontinuum models indicate that fluid exchange between the matrix and conduit networks is an important component for si mulating discharge in dual-permeability karst aquifers. 5.1. LIMITATIONS/RECO MMENDED PATHS FORWARD 1) The underwater cave survey data at Weeki Wachee Spring has not been verified by radiolocation and should ther efore be viewed as pr ovisional data that may be subject to change. 2) The spring basin boundaries need to be verfied with dye-tr ace testing during high and low flow conditions to provide def ensible interpretations of the spring basins. 3) A limitation with this study is the uncertainty associated with the hydraulic conductivity values used in the groundwater flow models. Dye-trace testing has not been performed to verify groundwater velocities, or hydraulic conductivity values used in the groundwater flow models. Quantitative dye-trace testing should be performed to estimate groundwat er velocities and to confirm nonDarcian flow in portions of the underly ing conduit network. The dual-conductivity model should be recalibrated when these data become available. 4) The well flow rates used in this study may need to be updated as published estimates for the SWFWMD were not available at the time of model development. Empirically based estimates of artificial recharge resulting from irrigation and septic tank leakage would also reduce uncertainty associated with the values of artificial recharge used in this study.

PAGE 175

160 5) The groundwater flow model was not ca librated with respect to the surficial aquifer or mantle layers and may need to be recalibrated if applied for other modeling purposes that differ from t he modeling purpose in this study. 6) While the dual-conductivity model best simulated observed spring discharge, it should not be used as is, to estimate springhead protection zones or minimum spring flows for Weeki Wachee or Twin Ds Springs until the limitations and uncertainties listed above are addressed. Future research should focus on co mpiling additional underwater cave survey data throughout the study area that is verified with radiolocation. These data should be incorporated into the dual-conductivity groundwater flow models when, and if, they become available. Moreover, the albino cave fauna documented in the course of this proj ect should be studied to determine if the species are endemic. Their biologic needs coupled with those of other fauna residing in the springs, could be used to assist regulators with setting appropriate minimum spring flows. 5.2. CONTRIBUTION OF THIS RESEARCH This research provides quantitative results evaluating the performance of the standard, Darcian, groundwater fl ow code (MOFLOW-2005) used with an equivalent-continuum model and a recent ly developed Darcian/non-Darcian groundwater flow code (MODFLOW-2005 CFP) used with a laminar/turbulent equivalent-continuum model and a laminar /turbulent dual-conductivity model. The application of these codes and their performance are quantified using 3 types of groundwater flow models with 3 di fferent conceptualizat ions of conduits: 1) an equivalent-continuum model where the bulk properti es of both the matrix and conduit networks are represented with laminar flow, 2) an equivalentcontinuum with both laminar and turbulent flow where the conduit network is

PAGE 176

161 represented by vuggy porosity, and 3) a dual-conductivity model where the locations for surveyed underwater caves ar e explicitly incorporated into the groundwater flow model. The conceptualizat ion of conduits, which varies among each of the 3 models, provi des insight into the limitat ions of these tools when applied to areas where the conduit locati ons and properties are not well known. The numerical analysis performed in this study demonstrated that accounting for turbulent flow alone, while ignori ng fluid exchange between the matrix and conduit networks, does not improve the match between observed and simulated discharges in areas strongl y influenced by the underlyin g conduit network at the test-site. Site specific information regarding the location, elevations, and approximate dimensions for conduits, the direction and magnitude of fluid exchange between the matrix and conduit net works, and groundwater velocities for constraining hydraulic conductivity values is needed to adequately defend conduit network interpretations and for constraining new model parameters introduced with MODFLOW-2005 CFP. The release of the USGS Darcian/non-Darcian dual-conductivity groundwater flow simulator in the public domain, coupled with the results of this and previous studies, makes it increasi ngly difficult to defend the continued application of laminar, equivalent-continuum models in karst aquifers where significant discrepancies exist between observed and simulated discharges or aquifer water levels for areas strongl y influenced by the underlying conduit network.

PAGE 177

162 REFERENCES CITED Armstrong, B., Chan, D., Collazos, A., and Mallams, J.L., 2003, Doline and aquifer characteristics within Hernando, Pasco, and northern Hillsborough Counties, in Florea, L.J., Vacher, H.L ., and Oches, E.A., eds., Karst studies in west-central Florida: US F seminar in karst environments: Southwest Florida Water Management District, p. 39-51. Bauer, S., Liedl, R., a nd Sauter, M., 2003, Modelin g of karst aquifer genesis: influence of exchange flow, Water Resources Research 39, 1285, doi:10.1029/2003WR002218. Blanchard, R.A. and Seidenfeld, A.V., 2005, Potentiometric surface of the Upper Floridan aquifer, west-central Florida, September 2004: U.S. Geological Survey Open File Report 2005-1222, 1 sheet. Blanford, T.N., and Birdie, T.R., 1993, Development of wellhead protection areas for the major public water supply well s in Hernando County, Florida, Final Technical Completion Report: Hydr oGeoLogic, Inc., variously p. Brinkmann, R., and Reeder, P., 1995, The re lationship between surface soils and cave sediments in west-central Flori da USA: Cave and Karst Science, 22, p. 95-102. Brinkmann, R., and Reeder, P., 1994, T he influence of sea-level change and geologic structure on cave development in west-central Florida: Physical Geography, 15, p. 52-61.

PAGE 178

163 Champion, K.M., and Star ks, R., 2001, The hydrol ogy and water quality of springs in west-central Florida: Southwest Florida Water Management District, 149 p. Court Complaint, 1976, Weeki Wachee Spring, Inc. vs. the Deltona Corporation, in the Circuit Court of the 5th Judicial Circuit of Flor ida, in and for Marion County No. 76-1173B, 11 p. Crawford, N.C., Lewis, M.A., Winter, S.A ., and Webster, J.A., 1999, Microgravity techniques for subsurface investigat ions of sinkhole collapses and for detection of groundwater flow paths through karst aquifers, in Beck, B.F., Pettit, A.J., and Herri ng, J.G., eds., Hydrogeology and Engineering Geology of Sinkholes and Karst 1999: A.A. Balkema, p. 203-218. Cunningham, K.J., Carlson, J.L., Wingard, G.L., Robi nson, E., and Wacker, M.A., 2004, Characterization of aquifer heterogeneity using cyclostratigraphy and geophysical methods in the upper part of the karstic Biscayne aquifer, southeastern Florida: U.S. Geol ogical Survey Water Resources Investigation Report 03-4208, 66 p. Curl, R.L., 1999, Entranceless a nd fractal caves revisited, in Palmer, A.N., Palmer, M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Special Publ ication 5, p. 183-185. Curl, R.L., 1986, Fractal dimensions a nd geometries of caves: Mathematical Geology 18, p. 765-783. Chemical Rubber Company Press, In c., 1995, CRC handbook of chemistry and physics, a ready-reference book of chemical and physical data, 76th ed.,

PAGE 179

164 Lide, D.R. and Frederikse, H.P.R., eds., Boca Raton: CRC Press, variously p. Davis, S.N., Campbell, D.J., Bentley, H.W., and Flynn, T. J., 1985, Ground water tracers: National Water Well Association, 200 p. Department of Environmental Protecti on, 2005, 2004 reuse inventory: Florida Department of Environmental Protection, 16 p. Diodato, D.M., 1994, A compendi um of fracture flow m odels 1994: Center for Environmental Restoration Systems, Argonne National Laboratory, 88 p. Dreiss, S., 1989, Regional scale transpor t in a karst aquifer 1. Component separation of spring flow hydrographs: Water Resources Research 25, p. 117-125. Duerr, A.D., 2001, Potentiometric surfac e of the upper Floridan aquifer, westcentral Florida, May 2001: U.S. Geological Survey Open-File Report 2001310, 1 sheet. Field, M.S., 2002, A lexicon of cave and karst terminology with special reference to environmental karst hydrology: U.S. Environmental Protection Agency, EPA/600/R-02/003, 214 p. Florea, L.J., 2006a, Architecture of ai r-filled caves within the karst of the Brooksville Ridge, west-central Florida: Journal of Cave and Karst Studies 68, p. 64-75. Florea. L.J., 2006b, The karst of west cent ral Florida, Ph.D. dissertation: Tampa, University of South Florida, 523 p.

PAGE 180

165 Florea, L.J., and Vacher, H.L., 2007, Eogene tic karst hydrology: insights from the 2004 hurricanes, peninsular Florida: Ground Water 45, p. 439-446. Florea, L.J., Vacher, H.L., Donahue, B ., and Naar, D., 2007, Quaternary cave levels in peninsular Florida: Quat ernary Science Reviews 26, p. 13341361. Florea, L.J. and Vacher, H.L., 2006, Springflow hydrographs: eogenetic vs. telegenetic karst: Ground Wa ter, 44, p. 352-361. Florida Geological Survey, 2008, Writt en communication, April 25, Brooksville, Florida. Freeze, R.A. and Cherry, J.A., 1979, Groundwater: New Jersey, Prentice-Hall, Inc., 604 p. GeoTrans, Inc., 1988a, Hydrol ogic investigation of the northern portion of the Southwest Florida Water Management District, Northern District model project phase II final report: GeoTrans, Inc., variously p. GeoTrans, Inc., 1988b, Addendum to hydrol ogic investigation of the northern portion of the Southwest Florida Wa ter Management District, Northern District model project phase II fina l report: GeoTrans, Inc., 9 p. Geurink, J.S., Nachabe, M., Ross, M. A., and Tara, P., 2000, Development of interfacial boundary conditions for t he southern District ground water model of the Southwest Florida Wate r Management District, final report surface water model development, calib ration, and recharge estimates: Center for Modeling Hyrologic and Aquat ic Systems, Univ ersity of South Florida, 34 p.

PAGE 181

166 Gilboy, T., and Moore, D., 1982a, A ddendum to hydrogeologic analysis Cross Bar Wellfield Pasco County: Sout hwest Florida Water Management District, variously p. Gilboy, T., and Moore, D., 1982b, Hydr ogeologic analysis Cross Bar Wellfield Pasco County: Southwest Florida Water Management District, 7 p. Goddard, M.L., 1998, The geoph ysical delineation of the fresh-water/salt-water interface in the Weeki Wachee Rive r basin, Hernando County, Florida, M.S. thesis: Tampa, Universi ty of South Florida, 116 p. Halihan, T., Sharp, J.M., Jr., and Mace R.E., 2000, Flow in the San Antonio segment of the Edwards aquifer: ma trix, fractures, or conduits?, in Sasowsky, I.D. and Wicks, C.M., eds., Groundwater Flow and Contaminant Transport in Carbonate Aquifers, Rotterdam, Netherlands, A.A. Balkema, p. 129-146. Harbaugh, A.W., 2005, MODFLO W-2005, The U.S. Geolog ical Survey modular ground-water Model -the Ground-Water Flow Process: U.S. Geological Survey Techniques and Met hods 6-A16, variously p. Hazlett, T.J., Kincaid, T., Loper, D.E., Da vies, G.J., Dehan, R., and Mckinlay, C., 2004, Realistic numerical modelin g of ground-water flow based on quantitative site characterization in the Woodville Karst Plain of north Florida, in Geological Society of America Abstracts with Programs, 36, p. 31. Hess, J.W. and White, W. B., 1988, Storm response of the karstic carbonate aquifer of southcentral Kentucky: Jour nal of Hydrology 99, p. 235-252.

PAGE 182

167 Hickey, J.J., 1984, Field testing the hy pothesis of Darcian flow through a carbonate aquifer: Ground Water 22, p. 544-547. Hill, M.E. and DeWitt, D.J., 2004, Drilling and testing re port for the ROMP WW-2 and ROMP WW-3 monitor well sites, Hernando County, Florida: Southwest Florida Water Management District, 12 p. Hyde, A.G., Law, L., Jr., Weatherspoon, R.L., Cheyney, M.D., and Eckenrode, J.J., 1977, Soil Survey of Hernando Count y, Florida: U.S. Department of Agriculture Soil Conservation Service, 152 p. Hoblit, B.C. and Curtis, D.C., 2005, R adar estimates + gauge data a perfect union: Southwest Hydrology 4, Tucson, SAHRA University of Arizona, variously p. Huang, C., 1994, Revised spring conductanc e coefficients Wekiva River Basin ground water flow model, Professional Paper SJ94-PP3: St. Johns River Water Management District, 16 p. Hunt, R.J., J. Doherty, and M.J. Tonk in, 2007, Are Models too simple? Arguments for increased parameterizati on: Ground Water 45, p. 254-262. Hydro-Environmental Associates, In c., 2003, Construction and testing of production wells PW-4 & PW-5, a nd monitoring well MW-4 Hernando County Utililties Southwest WellField Hernando County, Florida: HydroEnvironmental Associates, Inc., 21 p. HydroGeoLogic, Inc., 2007, Groundwater fl ow and saltwater intrusion model for the northern district water resour ces assessment project area, prepared

PAGE 183

168 for the Southwest Florida Water Management District: HydroGeologic, Inc., variously p. In-Situ, Inc., 2003a, MiniTROLL Model SSP-100 operator's manual: In-Situ, Inc., 93 p. In-Situ, Inc., 2003b, Multi-parameter TROLL 9000 WQP-100 oper ator's manual: In-Situ, Inc., 153 p. In-Situ, Inc., 2000, TROLL 8000 Multi-paramet er operator's manual: In-Situ, Inc., 52 p. In-Situ, Inc., 1995, TROLL Model SP4000-2 32 operator's manual: In-Situ, Inc., 22 p. Jaquet, O., Siegel, P., Klubertanz G., and Benabderrhamane, H., 2004, Stochastic discrete model of karstic networks: Advances in Water Resources 27, p. 751-760. Jiang, Y., Woodbury, A.D., and Painter, S. 2004, Full-Bayesian inversion of the Edwards Aquifer: Gro und Water 42, p. 724-733. Jones, G.W., Upchurch, S.B., Champi on, K.M., and DeWitt, D.J., 1997, Water quality and hydrology of The Ho mosassa, Chassahowitzka, Weeki Wachee, and Aripeka Spring Comple xes, Citrus and Hernando Counties, Florida origin of increasing nitrate concentrations: Southwest Florida Water Management District, 167 p. Jordan, C.L., 1984, Floridas weather and climate: Implications for water, in Fernald, E.A. and Patton, D.J., eds., Water Resources Atlas of Florida,

PAGE 184

169 Tallahassee: Institute of Science and Public Affairs, Florida State University, p. 18-35. Karst Underwater Research, Inc., 2008a, Written communication, January 7, Brooksville, Florida. Karst Underwater Research, Inc., 2008b, Weeki Wachee Spring Hernando County, Florida prelimi nary exploration of the underwater conduit system May-July 2007: Karst Underwa ter Research, Inc., 24 p. Karst Underwater Research, Inc., 2008c, http://karstunderwaterresearch.org/projects/twin_dees.asp Kincaid, T.R., Davies, G.J., Hazlett, T.J., Loper, D., Dehan, R., and Mckinlay, C., 2004, Groundbreaking characterization of th e karstified Floridan aquifer in the Woodville Karst Plain of north Florida in Geological Society of America Abstracts with Progr ams, 36, p. 31. Kiraly, L., 1975, Rapport sur ltat actuel des connaissances dans le domaine des caractres physiques des roches karstiques, in Burger, A. and Dubertret, L., eds., Hydrogeology of kars tic terrains, Paris, International Association of Hydrogeologists, Series B, 3, p. 53-67. Knochenmus, L.A. and Yobbi, D.K., 2001, Hydrology of the coastal springs ground-water basin and adjacent parts of Pasco, Hernando, and Citrus Counties, Florida: U.S. Geological Su rvey Water-Resources Investigations Report 01-4230, 88 p. Knowles, L. Jr., O'Reilly, A.M., and Adamski, J.C., 2002, Hydrogeology and simulated effects of ground-water withdr awals from the Floridan aquifer

PAGE 185

170 system in Lake County and in the Ocala National Forest and vicinity, North-Central Florida: U.S. G eological Survey Water-Resources Investigations Report 02-4207, 140 p. Kraemer, S.R., 1990, Modelin g of regional groundwater flow in fractured rock aquifers, Ph.D. dissertation: School of Public and Environmental Affairs, Indiana University, 134 p. Kurtzman, D., Nativ, R., and Adar, E. 2005, Correlating fracture trends and hydraulic head using semivariogram cloud analysis: Ground Water 43, p. 250-258. Labins, 2003, Land boundary information system on-line survey information and maps for the state of Florida, http://data.labins.org/2003/ Langevin, C.D., 2003, Stochastic ground water flow simulation with a fracture zone continuum model: Gr ound Water 41, p. 587-601. Lattman, L.H. and Parizek, R.R., 1964, Relationship bet ween fracture traces and the occurrence of ground water in ca rbonate rocks: Journal of Hydrology 2, p. 73-91. Mace, R.E., Marrett, R.A., and Hovorka, D.D., 2005, Fractal scaling of secondary porosity in karstic exposures of the Edwards aquifer, in Beck, B.F., ed., Sinkholes and the engineer ing and environmental impacts of karst, American Society of Civil Engineers, Geotechnical specia l publication 144, p. 129-146. Mace, R.E. and Hovorka, S.D., 2000, Estimating porosity and permeability in a karstic aquifer using core plugs, we ll tests, and outcr op measurements, in

PAGE 186

171 Sasowsky, I.D. and Wicks, C.M., eds., Groundwater Flow and Contaminant Transport in Carbonate Aquifers, Rotterdam, Netherlands, A.A. Balkema, p. 93-111. Marsh McBirney, Inc., 1990, FLO-MA TE Model 2000 portable flowmeter instruction manual: Marsh McBirney Inc., variously p. Martin. J.M., 2003, Quantificat ion of the matrix hydraulic conductivity in the Santa Fe River sink/rise system with implic ations for the exchange of water between the matrix and conduits, M.S. t hesis: Gainesville, University of Florida, 80 p. Martin, A., Jr. and Whiteman, C.D., Jr., 1990, Calibration and sensitivity analysis of a ground-water flow model of the coastal lowlands aquifer system in parts of Louisiana, Mississippi, Al abama, and Florida: U.S. Geological Survey Water-Resources Investigations Report 89-4189, 51 p. Martin, J.M., Screaton, E.J., and Martin, J.B., 2006, Monitoring well responses to karst conduit head fluctuations: implications for fluid exchange and matrix transmissivity in the Floridan aquifer, in Harmon R.S. and Wicks, C., eds., Perspectives on karst geomorphology, hydrology, and geochemistry-A tribute volume to Derek C. Ford and William B. White Boulder, Colorado: Geological Society of America, p. 209-217. Martin, J.B. and Screaton, E.J., 2001, Ex change of matrix and conduit water with examples from the Floridan aquifer, in Kuniansky, E.L., ed., U.S. Geological Survey Karst Interest Group Proceedings, Water-Resources Investigations Report 01-4011, p.38-44.

PAGE 187

172 Martin, J.B. and Dean, R.W., 2001, Ex change of water between conduits and matrix in the Floridan aquifer: Chemical Geology 179: p. 145-165. Martin, J.B. and Dean, R.W., 1999, Temperature as a natural tracer of short residence times for groundw ater in karst aquifers, in Palmer, A.N., Palmer, M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Special Publicati on 5, p. 236-242. McNeill, J.D., 1990, Use of Electrom agnetic Methods for Groundwater Studies, in Ward, S.H., ed., Geotechnical and Envi ronmental Geophysics Investigations in Geophysics 5: Societ y of Exploration Geophysicists, p. 191-218. Meinzer, O.E., 1927, Large springs in t he United States: U.S. Geological Survey Water-Supply Paper 557, 94 p. Miller, J.A., 1986, Hydrogeol ogic framework of the Floridan aquifer system in Florida and in parts of Georgia, Alabama, and South Ca rolina Regional Aquifer-System Analysis: U.S. Geol ogical Survey Professional Paper 1403-B, 91 p. Mohrlok, U., Kienle, J., Teutsch, G., 1997, Parameter identification in doublecontinuum models applied to karst aquifers, in Proceedings of the 12th International Congress on Speleology 2 and 6th Conference on Limestone Hydrology and Fissured Media, p. 163-166. Mohrlok, U. and Sauter, M., 1997, Modellin g groundwater flow in a karst terrane using discrete and double-continuum appr oaches importance of spatial and temporal distribution of recharge, in Proceedings of the 12th

PAGE 188

173 International Congress on Speleology 2 and 6th Conference on Limestone Hydrology and Fissured Media, p. 167-170. Moore, D.L., 1981, Geophysi cal signatures to photoli nears at the Cross-Bar Wellfield, Pasco County, Florida, M.S. thesis: Tampa, University of South Florida, 73 p. Mylroie, J.E. and Carew, J.L., 2000, Speleogenesis in coastal and oceanic settings, in Klimchouk, A.B., Ford, D.C., Palmer, A.N. and Dreybrodt, W., eds., Speleogenesis Evolution of Karst Aquifers: National Speleological Society, Inc., p. 226-233. National Geographic, 2007, 2004 hurricane season may be costliest on record, http://news.national geographic.com/news/2004/09/09/0927_040927_jean ne.html National Oceanographic and Atmos pheric Association, 2004, http://www.srh.noaa.gov/data/obhistory/KBKV.html Ortiz, A.G., 2007, Potentiometric surf ace of the upper Floridan aquifer, westcentral Florida, May 2006: U.S. Geological Survey Open-File Report 20071032, 1 sheet. Painter, S.L., Sun, A., and Green, R.T., 2007, Final technical report for enhanced characterization and representation of flow through karst aquifers-phase II, Revision 1: Southwest Research Institute, 99 p. Painter, S.L., Sun, A., and Green, R.T., 2004, Final technical report for enhanced characterization and representation of flow through karst aquifers: Southwest Research Institute, 77 p.

PAGE 189

174 Palmer, A.N., 1999a, Introduction, in Palmer, A.N., Palmer M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Special Publication 5, p 1-9. Palmer, A.N., 1999b, Anisotr opy in carbonate aquifers, in Palmer, A.N., Palmer, M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Special Publicati on 5, p. 223-227. Palmer, A.N., 1991, Origin and morphology of limestone caves: Geological Society of America Bu lletin 103, p. 1-21. Parasnis, D.S., 1986, Principles of applied geophysics 4th ed., New York: Chapman and Hall, 402 p. Pease, B., 1997, Determining depth by r adio-location: an extreme case: Cave Radio and Electronics Group 27, p. 22-25. Quinlan, J.F., Ray, J.A., and Schindel, G.M., 1995, Intrinsi c limitations of standard criteria and methods for delineation of groundwater-source protection areas (springhead and well head protection areas) in carbonate terranes: Critical review, technically -sound resolution of limitations, and case study in a Kentucky karst, in Beck, B.F., ed., Karst GeoHazards : A.A. Balkema, p. 525-536. Reeder, P. and Brinkmann, R., 1998, Pal eoenvironmental reconstruction of an Oligocene-aged island remnant in Florida, USA: Cave and Karst Science, 25, p. 7-13. Ross, M., 2006, Written communication, July 11, Brooksville, Florida.

PAGE 190

175 Safko, P.S. and Hickey, J.J. 1992, A preliminary approa ch to the use of borehole data, including television surveys, fo r characterizing secondary porosity of carbonate rocks in the Floridan aquife r system: U.S. Geological Survey Water Resources Investigat ions Report 91-4168, 70 p. Sasowsky, I.D. 2000. Carbonate aquifers: A review of thoughts and methods, in Sasowsky, I.D. and Wicks, C.M., eds., Groundwater Flow and Contaminant Transport in Carbonate Aquifers, Rotterdam, Netherlands, A.A. Balkema, p. 1-14. Scanlon, B.R., Mace, R.E., Barrett, M.E., and Smith, B., 2003, Can we simulate regional groundwater flow in a kars t system using equivalent porous media models? Case study, Barton Springs Edwards aquifer, USA: Journal of Hydrol ogy 276, p. 137-158. Schindel, G.M., 2003, Cavers and geologists what can they teach each other? in Significance of caves in wate rshed management and protection in Florida workshop proceedings: Florida Geological Survey Special Publication 53, 27 slides. Scott, T.M., 1997, Miocene to Holocene History of Florida, in Randazzo, A.F. and Jones, D.S., eds., The geology of Florida, Gainesville: University Press of Florida, p. 57-67. Scott, T.M., Means, G.H., Meegan, R.P. Means, R.C., Upchurch, S.B., Copeland, R.E., Jones, J., Roberts, T., and Willet, A., 2004, Springs of Florida: Florida Geological Survey Bulletin no. 66, 377 p.

PAGE 191

176 Screaton, E., Martin, J.B., Ginn, B., and Smith, L., 2004, Conduit properties and karstification in the unconfined Floridan aquifer: Ground Water 42, p. 338346. Shoemaker, W.B., Kuniansky, E.L., Birk S., Bauer, S., and Swain, E.D, 2008a, Documentation of a conduit flow pr ocess (CFP) for MODFLOW-2005: U.S. Geological Survey Techniques and Methods, Book 6, Chapter A24, 50 p. Shoemaker, W.B., Cunningham, K.J., and Kuniansky, E.L., and Dixon, J., 2008b, Effects of turbulence on hydraulic heads and parameter sensitivities in preferential groundwater flow layers: Water Resources Research, 44, W03501, doi:10.1029/2007WR006601. Sinclair, W.C., 1978, Preliminary evaluat ion of the water-supply potential of the spring-river system in the Weeki Wachee area and the lower Withlacoochee River, west-central Florid a: U.S. Geological Survey WaterResources Investigations 78-74, 40 p. Smith, B.A., Hunt, B.B., and Schindel G.M., 2005, Groundwater flow in the Edwards aquifer: Comparison of groundwater modeling and dye trace results, in Beck, B.F. ed., Sinkho les and the Engineering and Environmental Impacts of Karst, G eotechnical Special Publication 144: American Society of Civil Engineers, p. 131-141. Southwest Florida Water Management Distr ict, 2004, 2002 Estimated water use: Southwest Florida Water Management District, 37 p.

PAGE 192

177 Trommer, J.T., 1987, Potent ial for pollution of the Upper Floridan aquifer from five sinkholes and an internally drained basin in west-central Florida: U.S. Geological Survey Water-Resources Investigations Report 87-4013, 103 p. Upchurch, S.B. and Randazzo, A.F., 1997, Environmental geology of Florida, in Randazzo, A.F. and Jones, D.S., eds., The geology of Florida, Gainesville: University Press of Florida, p. 217-249. U.S. Geological Survey, 2007. Hurricane Frances impact study, http://coastal.er.usgs.gov/hurricanes/frances Vacher, H.L. and Mylroie, J.L., 2002, Eogenetic karst from the perspective of an equivalent porous medium: Carbonates and Evaporites 17, p. 182-196. Veni, G., 1999, A geomorphological strategy for conducting environmental impact assessments in karst areas: Geomorphology 31, p. 151-180. Wallace, R.E., III, 1993, Dye trace and bac teriological testing of sinkholes: Sulphur Springs, Tampa, Florida: En vironmental Geology 22, p. 362-366. Wetterhall, W.S., 1965, Re connaissance of springs and sinks in west-central Florida: Florida Geological Survey R eport of Investigations No. 39, 42 p. Williams, S.R., 1985, Relati onship of groundwater chemis try to photolinements in a karst aquifer, M.S. thesis: Tampa, Univ ersity of South Fl orida, 138 p. Wilson, W.L., 2002, Conduit morphology and hydrodynamics of the Floridan aquifer: moving to the next level conduit modeling, in Martin, J.B., Wicks, C.M., and Sasowsky, I.D., eds, Hydrogeology and biology of post-

PAGE 193

178 Paleozoic carbonate aquifers: Karst Waters Institute Special Publication 7, p. 5-8. White, W.A., 1970, The geomorphology of the Florida peninsula: Bureau of Geology, Geological Bu lletin No. 51, 164 p. White, W.B., 2002, Ground water flow in ka rst: matrix flow and conduit flow with implications for the Floridan aquifer, in Martin, J.B., Wicks, C.M., and Sasowsky, I.D., eds, Hydrogeology and biology of post-Paleozoic carbonate aquifers: Karst Waters Institut e Special Publication 7, p. 9-13. White, W.B., 1999, Conceptual models for karstic aquifers, in Palmer, A.N., Palmer, M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Special Publ ication 5, p. 11-16. White, W.B., 1988, Geomorphology and hydr ology of karst terrains: New York, Oxford University Press, 464 p. Wood, J.H. and Stewart, M.T., 1985, The geophysical and geologic characteristics of fracture zones in the carbonate Floridan aquifer: Florida Water Resources Center University of Florida, Gainesville, no. 88, 93 p. Worthington, S.R.H. and Sm art, C.C., 2003, Empirical determination of tracer mass for sink to spring tests in karst, in Beck, B.F., ed., Sinkholes and the Engineering Impacts of Karst, Geotec hnical Special Publication No. 122: American Society of Civil Engineers, p. 287-295. Worthington, S.R.H., Ford D.C., and Beddows, P.A., 2000a, Porosity and permeability enhancement in unconfined ca rbonate aquifers as a result of dissolution, in Klimchouk, A.B., Ford, D.C., Palmer, A.N., and Dreybrodt,

PAGE 194

179 W., eds., Speleogenesis Evolution of Karst Aquifers: National Speleological Society, Inc., p. 463-472. Worthington, S.R.H., Ford, D.C., and Davies, G.J., 2000b, Matrix, fracture and channel components of stor age and flow in a Paleozoic limestone aquifer, in Sasowsky, I.D. and Wicks, C.M., eds., Groundwater Flow and Contaminant Transport in Carbonate Aquifers, Rotterdam, Netherlands, A.A. Balkema, p. 113-128. Worthington, S.R.H., 1999, A comprehensive strategy for understanding flow in carbonate aquifers in Palmer, A.N., Palmer, M.V., and Sasowsky, I.D., eds., Karst Modeling: Karst Waters Institute Spec ial Publication 5, p. 3037. Worthington, S.R.H. and Ford, D.C ., 1997, Borehole tests for megascale channeling in carbonate aquifers, in Proceedings of the 12th International Congress on Speleology 2 and 6th Conference on Limestone Hydrology and Fissured Media, p. 195-198. Yobbi, D.K., 2004, Written communication, June 15, Weeki Wachee, Florida. Yobbi, D.K., 2000, Application of nonli near least-squares regression to groundwater flow modeling, west-central Florida: U.S. Geological Survey WaterResources Investigations Report 00-4094, 58 p. Yobbi, D.K., 1989, Simulation of steady-s tate ground water and spring flow in the Upper Floridan aquifer of coastal Cit rus and Hernando Counties, Florida: U.S. Geological Survey Water-Resource s Investigations Report 88-4036, 33 p.

PAGE 195

180 YSI, Inc. 2007. 6-Series Envi ronmental monitoring systems, http:www.ysilifesciences.com/ext ranet/EPGKL.nsf/5992085488f9da9d852 56a550047c2a2/90a0378150c 2d2dd85256a1f0073f295 !OpenDocument

PAGE 196

181 APPENDICES

PAGE 197

Appendix A 182 Figure A-1. a) Caliper and b) gamma log for WWSpg-ECK well. a) b) Diameter (cm) 010203040506070 Depth (m bls) 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 c/s Depth (m bls)050100150200250300

PAGE 198

183 Appendix A (Continued) a) b) Figure A-1 cont. a) Spontaneous pot ential and b) resistance and re sistivity for WWSpg-ECK well. 0500100015002000 SP (mV) 0100200300400 Depth (m bls) 0 10 20 30 40 50 60 70 80 ( -m) 0500100015002000Depth (m bls) 16N m 16N m R R ( ) 0 10 20 30 40 50 60 70 80

PAGE 199

Appendix A (Continued) Diameter (cm) 184 a) b) Figure A-2. a) Caliper and b) gamma log WW-F well. 010203040506 070 Depth (m bls) 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 (c/s) Depth (m bls)050100150200250300

PAGE 200

Appendix A (Continued) a) b) 185 Figure A-3. a) Caliper and b) gamma log W eeki Wachee Deep well.

PAGE 201

Appendix B 20 40 60 80 100 120 140 2002200420062008Water Level (m NGVD) BKV East UFA BKV East Surficial Land Surface Elevation 60 80 100 120 140 160 180 20002002200420062008Water Level (m NGVD) ROMP BR2 UFA ROMP BR2 Surficial Land Surface Elevation 186 60 80 100 120 140 160 180 2001200320052007Water Level (m NGVD) 0 20 40 60 80 100 120 200220042006200 Water Level (m NGVD) 8 San Antonio Park UFA Spring Hill UFA San Antonio Park Surficial Land Surface Elevation Spring Hill Surficial Land Surface Elevation Figure B-1. Nested well sites in capped zone.

PAGE 202

Appendix B (Continued) 40 60 80 100 120 140 160 2004200520062007200 8 Water Level (m NGVD) Pless Park UFA Pless Park Surficial Land Surface Elevation 50 70 90 110 130 150 170 200320042005200620072008Water Level (m NGVD) McKendree UFA McKendree Surficial Land Surface Elevation Figure B-1 cont. Nested wells in capped zone. 187

PAGE 203

Appendix B (Continued) 0 20 40 60 80 100 120 1996199820002002200420062008Water Level (m NGVD) 2NE Barthle UFA 2NE Barthle Surficial Land Surface Elevation 0 20 40 60 80 100 120 2002200320042005200620072008Water Level (m NGVD) CR 581 North UFA CR 581 North Surficial Land Surface Elevation 188 0 20 40 60 80 100 120 1996199820002002200420062008Water Level (m NGVD) 120 3E Barthle UFA 3E Barthle Surficial Land Surface Elevation 0 20 40 60 80 100 200320042005200620072008Water Level (m NGVD) Tampa Bay Golf UFA Tampa Bay Golf Surficial Land Surface Elevation Figure B-1 cont. Nested wells in recharge zone.

PAGE 204

189 0 20 40 60 80 100 120 2005 2006 2007 200 8 Water Level (m NGVD ) WW-4 UFA WW-4 Surficial Land Surface Elevation 0 20 40 60 80 100 120 20032004200520062007200 8 Water Level (m NGVD) ROMP 18-1A UFA ROMP 18-1A Surficial Land Surface ElevationAppendix B (Continued) Figure B-1 cont. Nested wells in discharge zone.

PAGE 205

Appendix C 190Weeki Wachee Spring (WW) Conduit diameter = 0.9 m Date*Q WW-TD Area Q/A Density H2O at 24 C **R=2r Viscosity H2O at 24 C Reynolds Number (m^3/s)(m^2)(m/s)(kg/m^3)(m)(kg/m-s)(dimensionless) 6/15/20044.610.667.02 997 0.910.00092612 6877117 6/22/20044.590.666.99 997 0.910.00092612 6847727 6/29/20044.510.666.86 997 0.910.00092612 6720373 7/6/20044.540.666.92 997 0.910.00092612 6779152 7/13/20044.510.666.87 997 0.910.00092612 6730170 7/21/20044.520.666.89 997 0.910.00092612 6749763 7/28/20044.580.666.97 997 0.910.00092612 6828134 8/4/20044.660.667.09 997 0.910.00092612 6945692 8/10/20044.720.667.19 997 0.910.00092612 7043657 8/17/20044.770.667.26 997 0.910.00092612 7112232 9/10/20045.590.668.51 997 0.910.00092612 8336790 9/24/20046.280.669.57 997 0.910.00092612 9375215 10/8/20046.480.669.87 997 0.910.00092612 9669109 10/22/20046.450.669.83 997 0.910.00092612 9629923 11/5/20046.360.669.69 997 0.910.00092612 9492772 11/19/20046.170.669.39 997 0.910.00092612 9198878 3/3/20055.080.667.74 997 0.910.00092612 7582462 4/6/20054.860.667.40 997 0.910.00092612 7249382 5/13/20054.750.667.23 997 0.910.00092612 7082843 6/17/20054.650.667.08 997 0.910.00092612 6935896 2/2/20064.330.666.60 997 0.910.00092612 6465665 3/28/20064.240.666.45 997 0.910.00092612 6318718 4/6/20064.150.666.32 997 0.910.00092612 6191364 4/18/20064.030.666.14 997 0.910.00092612 6015028 5/2/20063.890.665.92 997 0.910.00092612 5799506 Average~ 10^6 Twin D's Spring (TD) Conduit diameter = 0.9 m Date*Q TD Area Q/A Density H2O at 24 C R=2r Viscosity H2O at 24 C Reynolds Number (m^3/s)(m^2)(m/s)(kg/m^3)(m)(kg/m-s)(dimensionless) 6/15/20040.110.660.16 997 0.910.00092612 156743 6/29/20040.100.660.15 997 0.910.00092612 146947 7/6/20040.110.660.17 997 0.910.00092612 166540 7/13/20040.100.660.15 997 0.910.00092612 146947 7/21/20040.140.660.22 997 0.910.00092612 215522 7/28/20040.190.660.28 997 0.910.00092612 274301 8/4/20040.200.660.31 997 0.910.00092612 303690 8/10/20040.210.660.32 997 0.910.00092612 313487 8/17/20040.240.660.37 997 0.910.00092612 362469 9/10/20040.510.660.77 997 0.910.00092612 754328 9/24/20040.530.660.81 997 0.910.00092612 793513 10/8/20040.630.660.96 997 0.910.00092612 940460 10/22/20040.570.660.86 997 0.910.00092612 842496 11/5/20040.500.660.77 997 0.910.00092612 754328 12/6/20040.500.660.77 997 0.910.00092612 754328 12/22/20040.400.660.61 997 0.910.00092612 597584 1/18/20050.390.660.59 997 0.910.00092612 577991 3/3/20050.300.660.45 997 0.910.00092612 440841 4/6/20050.260.660.40 997 0.910.00092612 391859 5/13/20050.220.660.34 997 0.910.00092612 333080 Average~ 10^5 *Dates reported vary for Weeki Wachee and Twin D's Springs because of missing data, or because pool stages at Twin D's dropped below water levels at the WW-F well as discussed in Chapter 3. **R=2r from White (1988)

PAGE 206

Appendix D 191

PAGE 207

Appendix D (C ontinued) 192

PAGE 208

Appendix D (C ontinued) 193

PAGE 209

Appendix D (C ontinued) 194

PAGE 210

Appendix D (C ontinued) 195

PAGE 211

Appendix D (C ontinued) 196

PAGE 212

Appendix D (C ontinued) 197

PAGE 213

Appendix E ECM LAMINAR COMPUTED CFP MODE 1 COMPUTED CFP MODE 2 COMPUTED TimeWeeki Wachee SpringTwin D's SpringWeeki Wachee SpringTwin D's SpringWeeki Wachee SpringTwin D's Spring (m3/s) (m3/s) (m3/s) (m3/s) (m3/s) (m3/s) SS* 4.2966 0.1696 4.9188 0.2741 4.1468 0.1731 Jun-04 4.4986 0.1857 5.1106 0.2862 4.3313 0.1904 Jul-04 4.5956 0.2052 5.1849 0.3040 4.4215 0.2092 Aug-04 4.3354 0.1645 4.9243 0.2641 4.1762 0.1674 Sep-04 5.2660 0.3635 5.8006 0.4597 5.0457 0.3671 Oct-04 4.9230 0.3596 5.5235 0.4696 4.7132 0.3658 Nov-04 4.3867 0.2577 5.0056 0.3689 4.2227 0.2620 Dec-04 4.0691 0.1667 4.6926 0.2747 3.9278 0.1700 Jan-05 3.8222 0.0864 4.4453 0.1902 3.6971 0.0891 Feb-05 3.5437 0.0284 4.1629 0.1299 3.4360 0.0301 Mar-05 3.3278 0.0000 3.9408 0.0930 3.2325 0.0000 Apr-05 3.1100 0.0000 3.7250 0.0638 3.0262 0.0000 May-05 3.1472 0.0000 3.7421 0.0588 3.0609 0.0000 Jun-05 3.5461 0.0553 4.1069 0.1476 3.4368 0.0548 Jul-05 3.8070 0.0996 4.3594 0.1911 3.6811 0.0995 Aug-05 3.7818 0.1013 4.3389 0.1946 3.6536 0.1019 Sep-05 3.5116 0.0427 4.0821 0.1365 3.4001 0.0432 Oct-05 3.6958 0.0680 4.2517 0.1593 3.5771 0.0684 Nov-05 3.6739 0.0799 4.2342 0.1718 3.5556 0.0809 Dec-05 3.7951 0.0612 4.3523 0.1491 3.6694 0.0627 Jan-06 3.4710 0.0473 4.0365 0.1386 3.3638 0.0482 Feb-06 3.5076 0.0826 4.0561 0.1733 3.3998 0.0827 Mar-06 3.1993 0.0558 3.7588 0.1490 3.1076 0.0557 Apr-06 2.9737 0.0219 3.5311 0.1139 2.8938 0.0216 May-06 2.7879 0.0267 3.3376 0.1189 2.7143 0.0263 *SS indicates steady-state 198

PAGE 214

Appendix E (Continued) 199TARGET NAME TIMEOBSERVEDECMCFP MODE1 CFP MODE 2 COMPUTEDCOMPUTEDCOMPUTED (m)(m)(m) (m) CBARWF_1E_PHILLIPS_DEEP SS*21.7120.686321.034220.7027 CBARWF_1E_PHILLIPS_DEEP Jun-0421.3420.843421.189720.8597 CBARWF_1E_PHILLIPS_DEEP Jul-0421.7521.107221.451121.1234 CBARWF_1E_PHILLIPS_DEEP Aug-0422.5021.351221.692121.3672 CBARWF_1E_PHILLIPS_DEEP Sep-0423.0421.856022.191121.8716 CBARWF_1E_PHILLIPS_DEEP Oct-0423.0822.009722.342922.0252 CBARWF_1E_PHILLIPS_DEEP Nov-0422.8221.859222.193621.8749 CBARWF_1E_PHILLIPS_DEEP Dec-0422.6821.669722.006221.6855 CBARWF_1E_PHILLIPS_DEEP Jan-0522.5921.483921.823221.4999 CBARWF_1E_PHILLIPS_DEEP Feb-0522.3721.308221.650421.3244 CBARWF_1E_PHILLIPS_DEEP Mar-0522.4821.184821.529921.2011 CBARWF_1E_PHILLIPS_DEEP Apr-0522.2921.019021.367521.0355 CBARWF_1E_PHILLIPS_DEEP May-0522.0120.835621.187820.8523 CBARWF_1E_PHILLIPS_DEEP Jun-0521.9920.795321.148420.8121 CBARWF_1E_PHILLIPS_DEEP Jul-0522.2220.857621.210320.8744 CBARWF_1E_PHILLIPS_DEEP Aug-0522.0920.855021.208620.8718 CBARWF_1E_PHILLIPS_DEEP Sep-0521.7320.703821.059120.7207 CBARWF_1E_PHILLIPS_DEEP Oct-0521.7120.595520.951820.6123 CBARWF_1E_PHILLIPS_DEEP Nov-0521.4420.492120.849520.5090 CBARWF_1E_PHILLIPS_DEEP Dec-0521.4920.390220.748220.4070 CBARWF_1E_PHILLIPS_DEEP Jan-0621.1620.258620.618220.2755 CBARWF_1E_PHILLIPS_DEEP Feb-0621.4520.244920.604720.2618 CBARWF_1E_PHILLIPS_DEEP Mar-0621.1620.138620.499820.1555 CBARWF_1E_PHILLIPS_DEEP Apr-0620.8119.943520.307119.9605 CBARWF_1E_PHILLIPS_DEEP May-0620.5419.728020.094419.7451 CBARWF_1ESE_BARTHLE_C_S_DPSS2 2.2622.000822 .229822.0111 CBARWF_1ESE_BARTHLE_C_S_DPJun-0 421.7622.113422 .337022.1232 CBARWF_1ESE_BARTHLE_C_S_DPJul-0 422.2122.361822 .582422.3714 CBARWF_1ESE_BARTHLE_C_S_DPAug-0 422.8222.632922 .849322.6422 CBARWF_1ESE_BARTHLE_C_S_DPSep-0 423.2723.112523 .318723.1210 CBARWF_1ESE_BARTHLE_C_S_DPOct-0 423.3023.194023 .404123.2030 CBARWF_1ESE_BARTHLE_C_S_DPNov-0 423.0522.978423 .193622.9879 CBARWF_1ESE_BARTHLE_C_S_DPDec-0 422.9722.749922 .969822.7599 CBARWF_1ESE_BARTHLE_C_S_DPJan-0 522.9522.539822 .764322.5501 CBARWF_1ESE_BARTHLE_C_S_DPFeb-0 522.6322.358622 .586722.3691 CBARWF_1ESE_BARTHLE_C_S_DPMar-0 522.7922.269022 .499622.2796 CBARWF_1ESE_BARTHLE_C_S_DPApr-0 522.5922.106322 .341522.1173 CBARWF_1ESE_BARTHLE_C_S_DPMay-0 522.3321.916022 .155321.9270 CBARWF_1ESE_BARTHLE_C_S_DPJun-0 522.5522.006922 .241622.0177 CBARWF_1ESE_BARTHLE_C_S_DPJul-0 522.8722.208722 .440322.2193 CBARWF_1ESE_BARTHLE_C_S_DPAug-0 522.7622.212322 .447222.2230 CBARWF_1ESE_BARTHLE_C_S_DPSep-0 522.3722.015822 .254822.0267 CBARWF_1ESE_BARTHLE_C_S_DPOct-0 522.2621.949122 .187821.9600 CBARWF_1ESE_BARTHLE_C_S_DPNov-0 521.9821.820422 .061421.8313 CBARWF_1ESE_BARTHLE_C_S_DPDec-0 522.0121.730721 .971821.7415 CBARWF_1ESE_BARTHLE_C_S_DPJan-0 621.7921.561421 .806621.5724 CBARWF_1ESE_BARTHLE_C_S_DPFeb-0 622.0121.524821 .767921.5357 CBARWF_1ESE_BARTHLE_C_S_DPMar-0 621.7121.364321 .611921.3753

PAGE 215

Appendix E (Continued) 200CBARWF_1ESE_BARTHLE_C_S_DPApr-0 621.3721.119021 .371521.1302 CBARWF_1ESE_BARTHLE_C_S_DPMay-0 621.0920.848421 .107720.8598 CBARWF_1N_FINEST_FARMS_DPSS10.7612.019112.825912.0571 CBARWF_1N_FINEST_FARMS_DPJun-0411.0312.049212.850312.0871 CBARWF_1N_FINEST_FARMS_DPJul-0410.7012.338113.131612.3758 CBARWF_1N_FINEST_FARMS_DPAug-0410.8612.549413.341912.5872 CBARWF_1N_FINEST_FARMS_DPSep-0412.4713.197813.962413.2347 CBARWF_1N_FINEST_FARMS_DPOct-0414.3113.313614.108013.3517 CBARWF_1N_FINEST_FARMS_DPNov-0414.1413.122213.935013.1610 CBARWF_1N_FINEST_FARMS_DPDec-0413.6112.844313.669612.8836 CBARWF_1N_FINEST_FARMS_DPJan-0513.0512.598713.435912.6384 CBARWF_1N_FINEST_FARMS_DPFeb-0512.5712.360813.205712.4006 CBARWF_1N_FINEST_FARMS_DPMar-0512.0812.146912.995112.1866 CBARWF_1N_FINEST_FARMS_DPApr-0511.5911.923712.776811.9634 CBARWF_1N_FINEST_FARMS_DPMay-0511.1511.722312.573711.7618 CBARWF_1N_FINEST_FARMS_DPJun-0510.7611.756012.589411.7945 CBARWF_1N_FINEST_FARMS_DPJul-0510.6911.904212.719511.9420 CBARWF_1N_FINEST_FARMS_DPAug-0510.8911.985712.798612.0232 CBARWF_1N_FINEST_FARMS_DPSep-0510.9211.833512.655911.8713 CBARWF_1N_FINEST_FARMS_DPOct-0510.7611.791412.604511.8287 CBARWF_1N_FINEST_FARMS_DPNov-0510.5611.682512.499811.7198 CBARWF_1N_FINEST_FARMS_DPDec-0510.3211.662312.471511.6992 CBARWF_1N_FINEST_FARMS_DPJan-0610.0611.483712.302411.5212 CBARWF_1N_FINEST_FARMS_DPFeb-069.8311.477312.282711.5140 CBARWF_1N_FINEST_FARMS_DPMar-069.6611.314312.130511.3514 CBARWF_1N_FINEST_FARMS_DPApr-069.4211.072111.892811.1093 CBARWF_1N_FINEST_FARMS_DPMay-069.1610.809611.633110.8468 CBARWF_1NW_KUKA_DEEP SS10.5110.830711.589610.8732 CBARWF_1NW_KUKA_DEEP Jun-0410.8410.888211.645310.9308 CBARWF_1NW_KUKA_DEEP Jul-0410.5211.148911.902611.1914 CBARWF_1NW_KUKA_DEEP Aug-0410.7511.350912.105411.3936 CBARWF_1NW_KUKA_DEEP Sep-0412.6511.936312.672811.9785 CBARWF_1NW_KUKA_DEEP Oct-0414.3312.062212.815412.1055 CBARWF_1NW_KUKA_DEEP Nov-0413.9811.865612.633411.9096 CBARWF_1NW_KUKA_DEEP Dec-0413.3611.604612.383011.6490 CBARWF_1NW_KUKA_DEEP Jan-0512.7411.358612.146011.4032 CBARWF_1NW_KUKA_DEEP Feb-0512.2011.128511.921711.1730 CBARWF_1NW_KUKA_DEEP Mar-0511.7010.917311.713310.9616 CBARWF_1NW_KUKA_DEEP Apr-0511.2510.707011.505310.7510 CBARWF_1NW_KUKA_DEEP May-0510.8510.535411.331210.5789 CBARWF_1NW_KUKA_DEEP Jun-0510.5010.565811.348310.6084 CBARWF_1NW_KUKA_DEEP Jul-0510.4110.704511.471710.7463 CBARWF_1NW_KUKA_DEEP Aug-0510.6110.787311.550210.8289 CBARWF_1NW_KUKA_DEEP Sep-0510.6910.661211.429710.7029 CBARWF_1NW_KUKA_DEEP Oct-0510.5110.609211.371310.6506 CBARWF_1NW_KUKA_DEEP Nov-0510.3110.511811.275010.5531 CBARWF_1NW_KUKA_DEEP Dec-0510.0810.475011.232410.5160 CBARWF_1NW_KUKA_DEEP Jan-069.8410.322511.084610.3637 CBARWF_1NW_KUKA_DEEP Feb-069.6310.305811.060410.3465 CBARWF_1NW_KUKA_DEEP Mar-069.4910.167210.926810.2079 CBARWF_1NW_KUKA_DEEP Apr-069.309.943110.70559.9836

PAGE 216

Appendix E (Continued) Appendix E (Continued) 201CBARWF_1NW_KUKA_DEEP May-069.149.702110.46589.7424 CBARWF_1S_PASCO_TRAILS_DPSS2 2.6622.766522 .875922.7729 CBARWF_1S_PASCO_TRAILS_DPJun-0 422.3522.694222 .803622.7006 CBARWF_1S_PASCO_TRAILS_DPJul-0 422.6922.958223 .067622.9646 CBARWF_1S_PASCO_TRAILS_DPAug-0 423.0423.175523 .284823.1818 CBARWF_1S_PASCO_TRAILS_DPSep-0 423.2223.483023 .592423.4894 CBARWF_1S_PASCO_TRAILS_DPOct-0 423.3223.534823 .644023.5412 CBARWF_1S_PASCO_TRAILS_DPNov-0 422.9423.242523 .351723.2489 CBARWF_1S_PASCO_TRAILS_DPDec-0 422.7623.024723 .133823.0311 CBARWF_1S_PASCO_TRAILS_DPJan-0 522.7322.887322 .996322.8937 CBARWF_1S_PASCO_TRAILS_DPFeb-0 522.6122.742722 .851722.7491 CBARWF_1S_PASCO_TRAILS_DPMar-0 522.8322.781922 .890822.7883 CBARWF_1S_PASCO_TRAILS_DPApr-0 522.6422.621722 .730622.6281 CBARWF_1S_PASCO_TRAILS_DPMay-0522.5022.480722 .589622.4871 CBARWF_1S_PASCO_TRAILS_DPJun-0 522.5322.477522 .586522.4839 CBARWF_1S_PASCO_TRAILS_DPJul-0 522.7322.614222 .723222.6206 CBARWF_1S_PASCO_TRAILS_DPAug-0 522.8522.737122 .846222.7435 CBARWF_1S_PASCO_TRAILS_DPSep-0 522.5522.549222 .658422.5556 CBARWF_1S_PASCO_TRAILS_DPOct-0 522.6622.572222 .681522.5786 CBARWF_1S_PASCO_TRAILS_DPNov-0 522.4622.431422 .540822.4378 CBARWF_1S_PASCO_TRAILS_DPDec-0 522.5622.430722 .540222.4371 CBARWF_1S_PASCO_TRAILS_DPJan-0 622.3722.266422 .376022.2728 CBARWF_1S_PASCO_TRAILS_DPFeb-0 622.5722.300122 .409922.3065 CBARWF_1S_PASCO_TRAILS_DPMar-0 622.2722.107822 .217622.1142 CBARWF_1S_PASCO_TRAILS_DPApr-0 621.9221.852221 .962221.8586 CBARWF_1S_PASCO_TRAILS_DPMay-0621.6621.616121 .726321.6226 CBARWF_1SE_PHILLIPS_B1_DP SS22.1622.276422.486622.2867 CBARWF_1SE_PHILLIPS_B1_DPJun-0421.6322.420922.630822.4311 CBARWF_1SE_PHILLIPS_B1_DPJul-0421.8822.631522.840722.6417 CBARWF_1SE_PHILLIPS_B1_DPAug-0422.8322.852923.061122.8631 CBARWF_1SE_PHILLIPS_B1_DPSep-0423.0923.291223.497823.3012 CBARWF_1SE_PHILLIPS_B1_DPOct-0423.0423.401823.607023.4118 CBARWF_1SE_PHILLIPS_B1_DPNov-0422.9223.217823.422523.2277 CBARWF_1SE_PHILLIPS_B1_DPDec-0422.7923.014423.219223.0244 CBARWF_1SE_PHILLIPS_B1_DPJan-0522.7122.831723.036922.8417 CBARWF_1SE_PHILLIPS_B1_DPFeb-0522.5122.677422.883322.6875 CBARWF_1SE_PHILLIPS_B1_DPMar-0522.6922.592522.799422.6027 CBARWF_1SE_PHILLIPS_B1_DPApr-0522.4622.449222.657222.4594 CBARWF_1SE_PHILLIPS_B1_DPMay-0522.2222.288722.498122.2990 CBARWF_1SE_PHILLIPS_B1_DPJun-0522.3322.271022.481322.2813 CBARWF_1SE_PHILLIPS_B1_DPJul-0522.6622.388322.599022.3987 CBARWF_1SE_PHILLIPS_B1_DPAug-0522.6222.396922.608122.4073 CBARWF_1SE_PHILLIPS_B1_DPSep-0522.1922.238222.450322.2487 CBARWF_1SE_PHILLIPS_B1_DPOct-0522.1622.171922.384822.1824 CBARWF_1SE_PHILLIPS_B1_DPNov-0521.8722.073722.287322.0842 CBARWF_1SE_PHILLIPS_B1_DPDec-0521.9121.995322.209522.0058 CBARWF_1SE_PHILLIPS_B1_DPJan-0621.6721.852322.067321.8628 CBARWF_1SE_PHILLIPS_B1_DPFeb-0621.8921.823922.039521.8344 CBARWF_1SE_PHILLIPS_B1_DPMar-0621.5821.712421.928721.7230 CBARWF_1SE_PHILLIPS_B1_DPApr-0621.2321.519421.736821.5300 CBARWF_1SE_PHILLIPS_B1_DPMay-0621.0221.313921.532821.3246 201CBARWF_1NW_KUKA_DEEP May-069.149.702110.46589.7424 CBARWF_1S_PASCO_TRAILS_DPSS2 2.6622.766522 .875922.7729 CBARWF_1S_PASCO_TRAILS_DPJun-0 422.3522.694222 .803622.7006 CBARWF_1S_PASCO_TRAILS_DPJul-0 422.6922.958223 .067622.9646 CBARWF_1S_PASCO_TRAILS_DPAug-0 423.0423.175523 .284823.1818 CBARWF_1S_PASCO_TRAILS_DPSep-0 423.2223.483023 .592423.4894 CBARWF_1S_PASCO_TRAILS_DPOct-0 423.3223.534823 .644023.5412 CBARWF_1S_PASCO_TRAILS_DPNov-0 422.9423.242523 .351723.2489 CBARWF_1S_PASCO_TRAILS_DPDec-0 422.7623.024723 .133823.0311 CBARWF_1S_PASCO_TRAILS_DPJan-0 522.7322.887322 .996322.8937 CBARWF_1S_PASCO_TRAILS_DPFeb-0 522.6122.742722 .851722.7491 CBARWF_1S_PASCO_TRAILS_DPMar-0 522.8322.781922 .890822.7883 CBARWF_1S_PASCO_TRAILS_DPApr-0 522.6422.621722 .730622.6281 CBARWF_1S_PASCO_TRAILS_DPMay-0522.5022.480722 .589622.4871 CBARWF_1S_PASCO_TRAILS_DPJun-0 522.5322.477522 .586522.4839 CBARWF_1S_PASCO_TRAILS_DPJul-0 522.7322.614222 .723222.6206 CBARWF_1S_PASCO_TRAILS_DPAug-0 522.8522.737122 .846222.7435 CBARWF_1S_PASCO_TRAILS_DPSep-0 522.5522.549222 .658422.5556 CBARWF_1S_PASCO_TRAILS_DPOct-0 522.6622.572222 .681522.5786 CBARWF_1S_PASCO_TRAILS_DPNov-0 522.4622.431422 .540822.4378 CBARWF_1S_PASCO_TRAILS_DPDec-0 522.5622.430722 .540222.4371 CBARWF_1S_PASCO_TRAILS_DPJan-0 622.3722.266422 .376022.2728 CBARWF_1S_PASCO_TRAILS_DPFeb-0 622.5722.300122 .409922.3065 CBARWF_1S_PASCO_TRAILS_DPMar-0 622.2722.107822 .217622.1142 CBARWF_1S_PASCO_TRAILS_DPApr-0 621.9221.852221 .962221.8586 CBARWF_1S_PASCO_TRAILS_DPMay-0621.6621.616121 .726321.6226 CBARWF_1SE_PHILLIPS_B1_DP SS22.1622.276422.486622.2867 CBARWF_1SE_PHILLIPS_B1_DPJun-0421.6322.420922.630822.4311 CBARWF_1SE_PHILLIPS_B1_DPJul-0421.8822.631522.840722.6417 CBARWF_1SE_PHILLIPS_B1_DPAug-0422.8322.852923.061122.8631 CBARWF_1SE_PHILLIPS_B1_DPSep-0423.0923.291223.497823.3012 CBARWF_1SE_PHILLIPS_B1_DPOct-0423.0423.401823.607023.4118 CBARWF_1SE_PHILLIPS_B1_DPNov-0422.9223.217823.422523.2277 CBARWF_1SE_PHILLIPS_B1_DPDec-0422.7923.014423.219223.0244 CBARWF_1SE_PHILLIPS_B1_DPJan-0522.7122.831723.036922.8417 CBARWF_1SE_PHILLIPS_B1_DPFeb-0522.5122.677422.883322.6875 CBARWF_1SE_PHILLIPS_B1_DPMar-0522.6922.592522.799422.6027 CBARWF_1SE_PHILLIPS_B1_DPApr-0522.4622.449222.657222.4594 CBARWF_1SE_PHILLIPS_B1_DPMay-0522.2222.288722.498122.2990 CBARWF_1SE_PHILLIPS_B1_DPJun-0522.3322.271022.481322.2813 CBARWF_1SE_PHILLIPS_B1_DPJul-0522.6622.388322.599022.3987 CBARWF_1SE_PHILLIPS_B1_DPAug-0522.6222.396922.608122.4073 CBARWF_1SE_PHILLIPS_B1_DPSep-0522.1922.238222.450322.2487 CBARWF_1SE_PHILLIPS_B1_DPOct-0522.1622.171922.384822.1824 CBARWF_1SE_PHILLIPS_B1_DPNov-0521.8722.073722.287322.0842 CBARWF_1SE_PHILLIPS_B1_DPDec-0521.9121.995322.209522.0058 CBARWF_1SE_PHILLIPS_B1_DPJan-0621.6721.852322.067321.8628 CBARWF_1SE_PHILLIPS_B1_DPFeb-0621.8921.823922.039521.8344 CBARWF_1SE_PHILLIPS_B1_DPMar-0621.5821.712421.928721.7230 CBARWF_1SE_PHILLIPS_B1_DPApr-0621.2321.519421.736821.5300 CBARWF_1SE_PHILLIPS_B1_DPMay-0621.0221.313921.532821.3246

PAGE 217

Appendix E (Continued) Appendix E (Continued) 202CBARWF_2N_MASARYK_CNL_FLDNS S10.6311.598512 .390911.6373 CBARWF_2N_MASARYK_CNL_FLDNJun -0410.8611.6551 12.443811.6938 CBARWF_2N_MASARYK_CNL_FLDNJul -0410.6011.9207 12.700311.9592 CBARWF_2N_MASARYK_CNL_FLDNAug -0410.7712.1418 12.922212.1804 CBARWF_2N_MASARYK_CNL_FLDNSep -0412.3312.7837 13.535312.8214 CBARWF_2N_MASARYK_CNL_FLDNOct -0414.0412.8723 13.654512.9113 CBARWF_2N_MASARYK_CNL_FLDNNov -0413.8612.6571 13.456812.6969 CBARWF_2N_MASARYK_CNL_FLDNDec -0413.3512.3856 13.198312.4259 CBARWF_2N_MASARYK_CNL_FLDNJan -0512.7912.1270 12.949812.1676 CBARWF_2N_MASARYK_CNL_FLDNFeb -0512.2911.8869 12.716511.9277 CBARWF_2N_MASARYK_CNL_FLDNMar -0511.8111.6675 12.499111.7081 CBARWF_2N_MASARYK_CNL_FLDNApr -0511.3611.4470 12.282511.4876 CBARWF_2N_MASARYK_CNL_FLDNMay-0510.9411.2647 12.098111.3049 CBARWF_2N_MASARYK_CNL_FLDNJun -0510.5911.3087 12.122911.3478 CBARWF_2N_MASARYK_CNL_FLDNJul -0510.5711.4760 12.271411.5142 CBARWF_2N_MASARYK_CNL_FLDNAug -0510.7811.5436 12.336711.5816 CBARWF_2N_MASARYK_CNL_FLDNSep -0510.8211.3960 12.199911.4343 CBARWF_2N_MASARYK_CNL_FLDNOct -0510.6311.3515 12.144511.3894 CBARWF_2N_MASARYK_CNL_FLDNNov -0510.4111.2384 12.036011.2763 CBARWF_2N_MASARYK_CNL_FLDNDec -0510.1611.2091 11.997511.2466 CBARWF_2N_MASARYK_CNL_FLDNJan -069.9011.0332 11.831411.0713 CBARWF_2N_MASARYK_CNL_FLDNFeb -069.6911.0236 11.807511.0609 CBARWF_2N_MASARYK_CNL_FLDNMar -069.5310.8596 11.654110.8971 CBARWF_2N_MASARYK_CNL_FLDNApr -069.3110.6210 11.419410.6586 CBARWF_2N_MASARYK_CNL_FLDNMay-069.0410.3655 11.166010.4030 CBARWF_2NE_BARTHLE_A_N_DPSS17.2416.343716.936416.3690 CBARWF_2NE_BARTHLE_A_N_DPJun-0416.8116.523517.105816.5486 CBARWF_2NE_BARTHLE_A_N_DPJul-0416.4516.755917.323116.7805 CBARWF_2NE_BARTHLE_A_N_DPAug-0416.5317.116817.674517.1412 CBARWF_2NE_BARTHLE_A_N_DPSep-0417.8217.934218.447017.9573 CBARWF_2NE_BARTHLE_A_N_DPOct-0419.3117.830918.383817.8548 CBARWF_2NE_BARTHLE_A_N_DPNov-0419.5917.587118.162817.6117 CBARWF_2NE_BARTHLE_A_N_DPDec-0419.1917.305117.899117.3304 CBARWF_2NE_BARTHLE_A_N_DPJan-0518.7917.043917.652917.0697 CBARWF_2NE_BARTHLE_A_N_DPFeb-0518.4516.792017.412416.8183 CBARWF_2NE_BARTHLE_A_N_DPMar-0518.1016.606017.230516.6325 CBARWF_2NE_BARTHLE_A_N_DPApr-0517.8616.352216.987316.3791 CBARWF_2NE_BARTHLE_A_N_DPMay-0517.5516.132616.771816.1596 CBARWF_2NE_BARTHLE_A_N_DPJun-0517.7316.295616.915816.3221 CBARWF_2NE_BARTHLE_A_N_DPJul-0518.1016.522917.126216.5488 CBARWF_2NE_BARTHLE_A_N_DPAug-0518.0016.510317.119616.5362 CBARWF_2NE_BARTHLE_A_N_DPSep-0517.6716.270616.889316.2967 CBARWF_2NE_BARTHLE_A_N_DPOct-0517.2416.201616.814316.2274 CBARWF_2NE_BARTHLE_A_N_DPNov-0516.8416.043516.659316.0693 CBARWF_2NE_BARTHLE_A_N_DPDec-0516.5616.030016.637516.0555 CBARWF_2NE_BARTHLE_A_N_DPJan-0616.1815.808016.422915.8338 CBARWF_2NE_BARTHLE_A_N_DPFeb-0615.9615.846416.451915.8719 CBARWF_2NE_BARTHLE_A_N_DPMar-0615.7815.612016.223915.6376 CBARWF_2NE_BARTHLE_A_N_DPApr-0615.5215.345415.960915.3712 CBARWF_2NE_BARTHLE_A_N_DPMay-0615.2815.065215.682915.0910 CBARWF_2NW_SPG_HILL_12_DPSS9.799.340810.05479.3890 202CBARWF_2N_MASARYK_CNL_FLDNS S10.6311.598512 .390911.6373 CBARWF_2N_MASARYK_CNL_FLDNJun -0410.8611.6551 12.443811.6938 CBARWF_2N_MASARYK_CNL_FLDNJul -0410.6011.9207 12.700311.9592 CBARWF_2N_MASARYK_CNL_FLDNAug -0410.7712.1418 12.922212.1804 CBARWF_2N_MASARYK_CNL_FLDNSep -0412.3312.7837 13.535312.8214 CBARWF_2N_MASARYK_CNL_FLDNOct -0414.0412.8723 13.654512.9113 CBARWF_2N_MASARYK_CNL_FLDNNov -0413.8612.6571 13.456812.6969 CBARWF_2N_MASARYK_CNL_FLDNDec -0413.3512.3856 13.198312.4259 CBARWF_2N_MASARYK_CNL_FLDNJan -0512.7912.1270 12.949812.1676 CBARWF_2N_MASARYK_CNL_FLDNFeb -0512.2911.8869 12.716511.9277 CBARWF_2N_MASARYK_CNL_FLDNMar -0511.8111.6675 12.499111.7081 CBARWF_2N_MASARYK_CNL_FLDNApr -0511.3611.4470 12.282511.4876 CBARWF_2N_MASARYK_CNL_FLDNMay-0510.9411.2647 12.098111.3049 CBARWF_2N_MASARYK_CNL_FLDNJun -0510.5911.3087 12.122911.3478 CBARWF_2N_MASARYK_CNL_FLDNJul -0510.5711.4760 12.271411.5142 CBARWF_2N_MASARYK_CNL_FLDNAug -0510.7811.5436 12.336711.5816 CBARWF_2N_MASARYK_CNL_FLDNSep -0510.8211.3960 12.199911.4343 CBARWF_2N_MASARYK_CNL_FLDNOct -0510.6311.3515 12.144511.3894 CBARWF_2N_MASARYK_CNL_FLDNNov -0510.4111.2384 12.036011.2763 CBARWF_2N_MASARYK_CNL_FLDNDec -0510.1611.2091 11.997511.2466 CBARWF_2N_MASARYK_CNL_FLDNJan -069.9011.0332 11.831411.0713 CBARWF_2N_MASARYK_CNL_FLDNFeb -069.6911.0236 11.807511.0609 CBARWF_2N_MASARYK_CNL_FLDNMar -069.5310.8596 11.654110.8971 CBARWF_2N_MASARYK_CNL_FLDNApr -069.3110.6210 11.419410.6586 CBARWF_2N_MASARYK_CNL_FLDNMay-069.0410.3655 11.166010.4030 CBARWF_2NE_BARTHLE_A_N_DPSS17.2416.343716.936416.3690 CBARWF_2NE_BARTHLE_A_N_DPJun-0416.8116.523517.105816.5486 CBARWF_2NE_BARTHLE_A_N_DPJul-0416.4516.755917.323116.7805 CBARWF_2NE_BARTHLE_A_N_DPAug-0416.5317.116817.674517.1412 CBARWF_2NE_BARTHLE_A_N_DPSep-0417.8217.934218.447017.9573 CBARWF_2NE_BARTHLE_A_N_DPOct-0419.3117.830918.383817.8548 CBARWF_2NE_BARTHLE_A_N_DPNov-0419.5917.587118.162817.6117 CBARWF_2NE_BARTHLE_A_N_DPDec-0419.1917.305117.899117.3304 CBARWF_2NE_BARTHLE_A_N_DPJan-0518.7917.043917.652917.0697 CBARWF_2NE_BARTHLE_A_N_DPFeb-0518.4516.792017.412416.8183 CBARWF_2NE_BARTHLE_A_N_DPMar-0518.1016.606017.230516.6325 CBARWF_2NE_BARTHLE_A_N_DPApr-0517.8616.352216.987316.3791 CBARWF_2NE_BARTHLE_A_N_DPMay-0517.5516.132616.771816.1596 CBARWF_2NE_BARTHLE_A_N_DPJun-0517.7316.295616.915816.3221 CBARWF_2NE_BARTHLE_A_N_DPJul-0518.1016.522917.126216.5488 CBARWF_2NE_BARTHLE_A_N_DPAug-0518.0016.510317.119616.5362 CBARWF_2NE_BARTHLE_A_N_DPSep-0517.6716.270616.889316.2967 CBARWF_2NE_BARTHLE_A_N_DPOct-0517.2416.201616.814316.2274 CBARWF_2NE_BARTHLE_A_N_DPNov-0516.8416.043516.659316.0693 CBARWF_2NE_BARTHLE_A_N_DPDec-0516.5616.030016.637516.0555 CBARWF_2NE_BARTHLE_A_N_DPJan-0616.1815.808016.422915.8338 CBARWF_2NE_BARTHLE_A_N_DPFeb-0615.9615.846416.451915.8719 CBARWF_2NE_BARTHLE_A_N_DPMar-0615.7815.612016.223915.6376 CBARWF_2NE_BARTHLE_A_N_DPApr-0615.5215.345415.960915.3712 CBARWF_2NE_BARTHLE_A_N_DPMay-0615.2815.065215.682915.0910 CBARWF_2NW_SPG_HILL_12_DPSS9.799.340810.05479.3890

PAGE 218

Appendix E (Continued) Appendix E (Continued) 203CBARWF_2NW_SPG_HILL_12_DPJun-049.919.428710.14459.4771 CBARWF_2NW_SPG_HILL_12_DPJul-049.699.654010.36879.7026 CBARWF_2NW_SPG_HILL_12_DPAug-0410.049.842810.56229.8918 CBARWF_2NW_SPG_HILL_12_DPSep-0411.8510.405611.109510.4542 CBARWF_2NW_SPG_HILL_12_DPOct-0413.2910.493311.213510.5437 CBARWF_2NW_SPG_HILL_12_DPNov-0412.9610.263710.995710.3146 CBARWF_2NW_SPG_HILL_12_DPDec-0412.3810.007310.747510.0584 CBARWF_2NW_SPG_HILL_12_DPJan-0511.809.753510.49869.8045 CBARWF_2NW_SPG_HILL_12_DPFeb-0511.279.529610.27729.5802 CBARWF_2NW_SPG_HILL_12_DPMar-0510.799.323910.07089.3739 CBARWF_2NW_SPG_HILL_12_DPApr-0510.349.12609.87149.1752 CBARWF_2NW_SPG_HILL_12_DPMay-059.938.98759.72669.0356 CBARWF_2NW_SPG_HILL_12_DPJun-059.639.06009.78679.1071 CBARWF_2NW_SPG_HILL_12_DPJul-059.709.22749.94179.2738 CBARWF_2NW_SPG_HILL_12_DPAug-059.979.307610.01869.3540 CBARWF_2NW_SPG_HILL_12_DPSep-0510.029.17819.89439.2247 CBARWF_2NW_SPG_HILL_12_DPOct-059.799.12099.82949.1671 CBARWF_2NW_SPG_HILL_12_DPNov-059.539.02499.73359.0710 CBARWF_2NW_SPG_HILL_12_DPDec-059.278.98949.69289.0353 CBARWF_2NW_SPG_HILL_12_DPJan-069.018.84389.55048.8897 CBARWF_2NW_SPG_HILL_12_DPFeb-068.838.82239.52208.8677 CBARWF_2NW_SPG_HILL_12_DPMar-068.688.68539.38788.7304 CBARWF_2NW_SPG_HILL_12_DPApr-068.438.46989.17318.5145 CBARWF_2NW_SPG_HILL_12_DPMay-068.248.24748.95008.2916 CBARWF_2W_CREWS_ LAKE_DP SS9.158.75999.14468.7932 CBARWF_2W_CREWS_L AKE_DPJun-049.468.77009.15478.8033 CBARWF_2W_CREWS_L AKE_DPJul-049.279.05789.44279.0912 CBARWF_2W_CREWS_L AKE_DPAug-049.879.21839.60369.2518 CBARWF_2W_CREWS_L AKE_DPSep-0410.959.60249.98779.6359 CBARWF_2W_CREWS_L AKE_DPOct-0411.719.741410.12719.7751 CBARWF_2W_CREWS_L AKE_DPNov-0411 .409.59279.97999.6267 CBARWF_2W_CREWS_L AKE_DPDec-0410 .969.41279.80179.4469 CBARWF_2W_CREWS_L AKE_DPJan-0510.509.22719.61809.2614 CBARWF_2W_CREWS_L AKE_DPFeb-0510.119 .06029.45269.0945 CBARWF_2W_CREWS_L AKE_DPMar-059.778.91069.30458.9449 CBARWF_2W_CREWS_L AKE_DPApr-059.488.76439.15938.7985 CBARWF_2W_CREWS_L AKE_DPMay-059.268.61909.01478.6530 CBARWF_2W_CREWS_L AKE_DPJun-059.038.64559.04098.6792 CBARWF_2W_CREWS_L AKE_DPJul-059.068.89179.28628.9251 CBARWF_2W_CREWS_L AKE_DPAug-059.439.05629.45009.0894 CBARWF_2W_CREWS_L AKE_DPSep-059.398.98339.37719.0164 CBARWF_2W_CREWS_L AKE_DPOct-059.158.92049.31418.9533 CBARWF_2W_CREWS_L AKE_DPNov-058.928.83929.23258.8719 CBARWF_2W_CREWS_L AKE_DPDec-058.698.78539.17828.8179 CBARWF_2W_CREWS_L AKE_DPJan-068.458.66759.06018.7000 CBARWF_2W_CREWS_L AKE_DPFeb-068.398.62929.02148.6615 CBARWF_2W_CREWS_L AKE_DPMar-068.448.52628.91808.5584 CBARWF_2W_CREWS_L AKE_DPApr-068.118.35258.74398.3845 CBARWF_2W_CREWS_L AKE_DPMay-067.908.16778.55878.1995 CBARWF_3E_BARTHLE_B_DEEP SS21.3120.400620.770620.4171 CBARWF_3E_BARTHLE_B_DEEPJun-0421.2220.569120.924220.5845 203CBARWF_2NW_SPG_HILL_12_DPJun-049.919.428710.14459.4771 CBARWF_2NW_SPG_HILL_12_DPJul-049.699.654010.36879.7026 CBARWF_2NW_SPG_HILL_12_DPAug-0410.049.842810.56229.8918 CBARWF_2NW_SPG_HILL_12_DPSep-0411.8510.405611.109510.4542 CBARWF_2NW_SPG_HILL_12_DPOct-0413.2910.493311.213510.5437 CBARWF_2NW_SPG_HILL_12_DPNov-0412.9610.263710.995710.3146 CBARWF_2NW_SPG_HILL_12_DPDec-0412.3810.007310.747510.0584 CBARWF_2NW_SPG_HILL_12_DPJan-0511.809.753510.49869.8045 CBARWF_2NW_SPG_HILL_12_DPFeb-0511.279.529610.27729.5802 CBARWF_2NW_SPG_HILL_12_DPMar-0510.799.323910.07089.3739 CBARWF_2NW_SPG_HILL_12_DPApr-0510.349.12609.87149.1752 CBARWF_2NW_SPG_HILL_12_DPMay-059.938.98759.72669.0356 CBARWF_2NW_SPG_HILL_12_DPJun-059.639.06009.78679.1071 CBARWF_2NW_SPG_HILL_12_DPJul-059.709.22749.94179.2738 CBARWF_2NW_SPG_HILL_12_DPAug-059.979.307610.01869.3540 CBARWF_2NW_SPG_HILL_12_DPSep-0510.029.17819.89439.2247 CBARWF_2NW_SPG_HILL_12_DPOct-059.799.12099.82949.1671 CBARWF_2NW_SPG_HILL_12_DPNov-059.539.02499.73359.0710 CBARWF_2NW_SPG_HILL_12_DPDec-059.278.98949.69289.0353 CBARWF_2NW_SPG_HILL_12_DPJan-069.018.84389.55048.8897 CBARWF_2NW_SPG_HILL_12_DPFeb-068.838.82239.52208.8677 CBARWF_2NW_SPG_HILL_12_DPMar-068.688.68539.38788.7304 CBARWF_2NW_SPG_HILL_12_DPApr-068.438.46989.17318.5145 CBARWF_2NW_SPG_HILL_12_DPMay-068.248.24748.95008.2916 CBARWF_2W_CREWS_ LAKE_DP SS9.158.75999.14468.7932 CBARWF_2W_CREWS_L AKE_DPJun-049.468.77009.15478.8033 CBARWF_2W_CREWS_L AKE_DPJul-049.279.05789.44279.0912 CBARWF_2W_CREWS_L AKE_DPAug-049.879.21839.60369.2518 CBARWF_2W_CREWS_L AKE_DPSep-0410.959.60249.98779.6359 CBARWF_2W_CREWS_L AKE_DPOct-0411.719.741410.12719.7751 CBARWF_2W_CREWS_L AKE_DPNov-0411 .409.59279.97999.6267 CBARWF_2W_CREWS_L AKE_DPDec-0410 .969.41279.80179.4469 CBARWF_2W_CREWS_L AKE_DPJan-0510.509.22719.61809.2614 CBARWF_2W_CREWS_L AKE_DPFeb-0510.119 .06029.45269.0945 CBARWF_2W_CREWS_L AKE_DPMar-059.778.91069.30458.9449 CBARWF_2W_CREWS_L AKE_DPApr-059.488.76439.15938.7985 CBARWF_2W_CREWS_L AKE_DPMay-059.268.61909.01478.6530 CBARWF_2W_CREWS_L AKE_DPJun-059.038.64559.04098.6792 CBARWF_2W_CREWS_L AKE_DPJul-059.068.89179.28628.9251 CBARWF_2W_CREWS_L AKE_DPAug-059.439.05629.45009.0894 CBARWF_2W_CREWS_L AKE_DPSep-059.398.98339.37719.0164 CBARWF_2W_CREWS_L AKE_DPOct-059.158.92049.31418.9533 CBARWF_2W_CREWS_L AKE_DPNov-058.928.83929.23258.8719 CBARWF_2W_CREWS_L AKE_DPDec-058.698.78539.17828.8179 CBARWF_2W_CREWS_L AKE_DPJan-068.458.66759.06018.7000 CBARWF_2W_CREWS_L AKE_DPFeb-068.398.62929.02148.6615 CBARWF_2W_CREWS_L AKE_DPMar-068.448.52628.91808.5584 CBARWF_2W_CREWS_L AKE_DPApr-068.118.35258.74398.3845 CBARWF_2W_CREWS_L AKE_DPMay-067.908.16778.55878.1995 CBARWF_3E_BARTHLE_B_DEEP SS21.3120.400620.770620.4171 CBARWF_3E_BARTHLE_B_DEEPJun-0421.2220.569120.924220.5845

PAGE 219

Appendix E (Continued) Appendix E (Continued) 204CBARWF_3E_BARTHLE_B_DEEPJul-0420.2220.823521.172520.8386 CBARWF_3E_BARTHLE_B_DEEPAug-0420.8121.127221.464921.1416 CBARWF_3E_BARTHLE_B_DEEPSep-0423.0921.725822.035921.7385 CBARWF_3E_BARTHLE_B_DEEPOct-0422.8421.791722.128921.8064 CBARWF_3E_BARTHLE_B_DEEPNov-0422.6821.576421.929021.5922 CBARWF_3E_BARTHLE_B_DEEPDec-0422.4721.329321.693221.3458 CBARWF_3E_BARTHLE_B_DEEPJan-0522.3021.095521.468721.1125 CBARWF_3E_BARTHLE_B_DEEPFeb-0522.0820.893821.273820.9112 CBARWF_3E_BARTHLE_B_DEEPMar-0522.0620.754421.136720.7718 CBARWF_3E_BARTHLE_B_DEEPApr-0521.9520.552820.944820.5708 CBARWF_3E_BARTHLE_B_DEEPMay-0521.6620.332320.730220.3505 CBARWF_3E_BARTHLE_B_DEEPJun-0521.8220.431020.812520.4483 CBARWF_3E_BARTHLE_B_DEEPJul-0522.1420.603320.978120.6202 CBARWF_3E_BARTHLE_B_DEEPAug-0521.9820.566020.950620.5834 CBARWF_3E_BARTHLE_B_DEEPSep-0521.6820.383320.775120.4009 CBARWF_3E_BARTHLE_B_DEEPOct-0521.3120.304220.692220.3215 CBARWF_3E_BARTHLE_B_DEEPNov-0520.6820.174220.567220.1917 CBARWF_3E_BARTHLE_B_DEEPDec-0520.9420.108220.497620.1255 CBARWF_3E_BARTHLE_B_DEEPJan-0620.8119.925320.325019.9429 CBARWF_3E_BARTHLE_B_DEEPFeb-0620.8319.906020.294619.9232 CBARWF_3E_BARTHLE_B_DEEPMar-0620.6019.723120.125219.7408 CBARWF_3E_BARTHLE_B_DEEPApr-0620.0219.456619.866419.4745 CBARWF_3E_BARTHLE_B_DEEPMay-0619.0719.163719.579519.1818 CBARWF_4N_AP_MASARYK_FLDNSS10.8210.568711.391610.6057 CBARWF_4N_AP_MASARYK_FLDNJun-0410.9810.661611.480810.6984 CBARWF_4N_AP_MASARYK_FLDNJul-0410.7810.938111.747610.9743 CBARWF_4N_AP_MASARYK_FLDNAug-0410.9411.156411.973211.1930 CBARWF_4N_AP_MASARYK_FLDNSep-0412.3711.862312.646711.8971 CBARWF_4N_AP_MASARYK_FLDNOct-0413.9611.808312.638811.8459 CBARWF_4N_AP_MASARYK_FLDNNov-0413.8411.501412.350211.5401 CBARWF_4N_AP_MASARYK_FLDNDec-0413.3711.190412.049111.2297 CBARWF_4N_AP_MASARYK_FLDNJan-0512.8710.891711.755310.9312 CBARWF_4N_AP_MASARYK_FLDNFeb-0512.3810.630511.495310.6700 CBARWF_4N_AP_MASARYK_FLDNMar-0511.9210.382811.244510.4220 CBARWF_4N_AP_MASARYK_FLDNApr-0511.4910.158611.016610.1973 CBARWF_4N_AP_MASARYK_FLDNMay-0511.1010.039210.884810.0770 CBARWF_4N_AP_MASARYK_FLDNJun-0510.8010.124610.947510.1612 CBARWF_4N_AP_MASARYK_FLDNJul-0510.8510.383311.181410.4189 CBARWF_4N_AP_MASARYK_FLDNAug-0511.0510.402311.204510.4379 CBARWF_4N_AP_MASARYK_FLDNSep-0511.0410.218411.032210.2544 CBARWF_4N_AP_MASARYK_FLDNOct-0510.8210.209111.007510.2445 CBARWF_4N_AP_MASARYK_FLDNNov-0510.5810.080110.885910.1157 CBARWF_4N_AP_MASARYK_FLDNDec-0510.3310.053610.850110.0889 CBARWF_4N_AP_MASARYK_FLDNJan-0610.059.853510.65929.8893 CBARWF_4N_AP_MASARYK_FLDNFeb-069.859.866910.65599.9020 CBARWF_4N_AP_MASARYK_FLDNMar-069.689.665810.46619.7010 CBARWF_4N_AP_MASARYK_FLDNApr-069.439.410810.21349.4461 CBARWF_4N_AP_MASARYK_FLDNMay -069.209.14969.95139.1847 CR_581_NORTH_FLDN SS22.9122.955223.150522.9638 CR_581_NORTH_FLDN Jun-0422.1023.049823.235623.0575 CR_581_NORTH_FLDN Jul-0422.8523.252123.437523.2598 204CBARWF_3E_BARTHLE_B_DEEPJul-0420.2220.823521.172520.8386 CBARWF_3E_BARTHLE_B_DEEPAug-0420.8121.127221.464921.1416 CBARWF_3E_BARTHLE_B_DEEPSep-0423.0921.725822.035921.7385 CBARWF_3E_BARTHLE_B_DEEPOct-0422.8421.791722.128921.8064 CBARWF_3E_BARTHLE_B_DEEPNov-0422.6821.576421.929021.5922 CBARWF_3E_BARTHLE_B_DEEPDec-0422.4721.329321.693221.3458 CBARWF_3E_BARTHLE_B_DEEPJan-0522.3021.095521.468721.1125 CBARWF_3E_BARTHLE_B_DEEPFeb-0522.0820.893821.273820.9112 CBARWF_3E_BARTHLE_B_DEEPMar-0522.0620.754421.136720.7718 CBARWF_3E_BARTHLE_B_DEEPApr-0521.9520.552820.944820.5708 CBARWF_3E_BARTHLE_B_DEEPMay-0521.6620.332320.730220.3505 CBARWF_3E_BARTHLE_B_DEEPJun-0521.8220.431020.812520.4483 CBARWF_3E_BARTHLE_B_DEEPJul-0522.1420.603320.978120.6202 CBARWF_3E_BARTHLE_B_DEEPAug-0521.9820.566020.950620.5834 CBARWF_3E_BARTHLE_B_DEEPSep-0521.6820.383320.775120.4009 CBARWF_3E_BARTHLE_B_DEEPOct-0521.3120.304220.692220.3215 CBARWF_3E_BARTHLE_B_DEEPNov-0520.6820.174220.567220.1917 CBARWF_3E_BARTHLE_B_DEEPDec-0520.9420.108220.497620.1255 CBARWF_3E_BARTHLE_B_DEEPJan-0620.8119.925320.325019.9429 CBARWF_3E_BARTHLE_B_DEEPFeb-0620.8319.906020.294619.9232 CBARWF_3E_BARTHLE_B_DEEPMar-0620.6019.723120.125219.7408 CBARWF_3E_BARTHLE_B_DEEPApr-0620.0219.456619.866419.4745 CBARWF_3E_BARTHLE_B_DEEPMay-0619.0719.163719.579519.1818 CBARWF_4N_AP_MASARYK_FLDNSS10.8210.568711.391610.6057 CBARWF_4N_AP_MASARYK_FLDNJun-0410.9810.661611.480810.6984 CBARWF_4N_AP_MASARYK_FLDNJul-0410.7810.938111.747610.9743 CBARWF_4N_AP_MASARYK_FLDNAug-0410.9411.156411.973211.1930 CBARWF_4N_AP_MASARYK_FLDNSep-0412.3711.862312.646711.8971 CBARWF_4N_AP_MASARYK_FLDNOct-0413.9611.808312.638811.8459 CBARWF_4N_AP_MASARYK_FLDNNov-0413.8411.501412.350211.5401 CBARWF_4N_AP_MASARYK_FLDNDec-0413.3711.190412.049111.2297 CBARWF_4N_AP_MASARYK_FLDNJan-0512.8710.891711.755310.9312 CBARWF_4N_AP_MASARYK_FLDNFeb-0512.3810.630511.495310.6700 CBARWF_4N_AP_MASARYK_FLDNMar-0511.9210.382811.244510.4220 CBARWF_4N_AP_MASARYK_FLDNApr-0511.4910.158611.016610.1973 CBARWF_4N_AP_MASARYK_FLDNMay-0511.1010.039210.884810.0770 CBARWF_4N_AP_MASARYK_FLDNJun-0510.8010.124610.947510.1612 CBARWF_4N_AP_MASARYK_FLDNJul-0510.8510.383311.181410.4189 CBARWF_4N_AP_MASARYK_FLDNAug-0511.0510.402311.204510.4379 CBARWF_4N_AP_MASARYK_FLDNSep-0511.0410.218411.032210.2544 CBARWF_4N_AP_MASARYK_FLDNOct-0510.8210.209111.007510.2445 CBARWF_4N_AP_MASARYK_FLDNNov-0510.5810.080110.885910.1157 CBARWF_4N_AP_MASARYK_FLDNDec-0510.3310.053610.850110.0889 CBARWF_4N_AP_MASARYK_FLDNJan-0610.059.853510.65929.8893 CBARWF_4N_AP_MASARYK_FLDNFeb-069.859.866910.65599.9020 CBARWF_4N_AP_MASARYK_FLDNMar-069.689.665810.46619.7010 CBARWF_4N_AP_MASARYK_FLDNApr-069.439.410810.21349.4461 CBARWF_4N_AP_MASARYK_FLDNMay -069.209.14969.95139.1847 CR_581_NORTH_FLDN SS22.9122.955223.150522.9638 CR_581_NORTH_FLDN Jun-0422.1023.049823.235623.0575 CR_581_NORTH_FLDN Jul-0422.8523.252123.437523.2598

PAGE 220

Appendix E (Continued) Appendix E (Continued) 205CR_581_NORTH_FLDN Aug-0423.6723.531423.711323.5387 CR_581_NORTH_FLDN Sep-0423.9524.019824.183724.0257 CR_581_NORTH_FLDN Oct-0423.8824.054324.230524.0616 CR_581_NORTH_FLDN Nov-0423.5823.824024.008023.8320 CR_581_NORTH_FLDN Dec-0423.3923.574123.764223.5827 CR_581_NORTH_FLDN Jan-0523.2523.338523.533923.3476 CR_581_NORTH_FLDN Feb-0523.0323.137423.336323.1466 CR_581_NORTH_FLDN Mar-0523.1623.039823.239523.0489 CR_581_NORTH_FLDN Apr-0523.0922.864323.070322.8738 CR_581_NORTH_FLDN May-0522.8622.644222.854922.6537 CR_581_NORTH_FLDN Jun-0523.2422.812223.009522.8211 CR_581_NORTH_FLDN Jul-0523.6523.140623.330623.1494 CR_581_NORTH_FLDN Aug-0523.6223.137523.338123.1465 CR_581_NORTH_FLDN Sep-0522.9622.899723.108422.9090 CR_581_NORTH_FLDN Oct-0522.9122.867723.071622.8769 CR_581_NORTH_FLDN Nov-0522.6522.700822.911322.7101 CR_581_NORTH_FLDN Dec-0522.5422.570322.780922.5796 CR_581_NORTH_FLDN Jan-0622.4522.357022.574322.3665 CR_581_NORTH_FLDN Feb-0622.6022.302322.515022.3116 CR_581_NORTH_FLDN Mar-0622.3322.100922.320122.1103 CR_581_NORTH_FLDN Apr-0621.8221.828522.051621.8381 CR_581_NORTH_FLDN May-0621.4021.532521.760121.5423 CSPR-7_COASTAL_SPG_U_FLD N SS2.602.53562.63132.5455 CSPR-7_COASTAL_SPG_U_FLDNJun-042.552.61032.70582.6197 CSPR-7_COASTAL_SPG_U_FLDNJul-042.782.78382.88082.7931 CSPR-7_COASTAL_SPG_U_FLDNAug-042.762.80112.90022.8103 CSPR-7_COASTAL_SPG_U_FLDNSep-043.153.42893.52423.4380 CSPR-7_COASTAL_SPG_U_FLDNO ct-042.883.44083.54623.4500 CSPR-7_COASTAL_SPG_U_FLDNN ov-042.823.28423.39253.2935 CSPR-7_COASTAL_SPG_U_FLDND ec-042.673.13503.24473.1443 CSPR-7_COASTAL_SPG_U_FLDNJan-052.563.00443.11453.0137 CSPR-7_COASTAL_SPG_U_FLDNF eb-052.532.90223.01242.9115 CSPR-7_COASTAL_SPG_U_FLDNM ar-052.572.83452.94392.8437 CSPR-7_COASTAL_SPG_U_FLDNA pr-052.652.76102.87022.7701 CSPR-7_COASTAL_SPG_U_FLDNM ay-052.602.74832.85602.7571 CSPR-7_COASTAL_SPG_U_FLDNJun-052.822.86722.97502.8759 CSPR-7_COASTAL_SPG_U_FLDNJul-052.903.08533.19333.0939 CSPR-7_COASTAL_SPG_U_FLDNAug-052.653.07673.18773.0852 CSPR-7_COASTAL_SPG_U_FLDNSep-052.652.99663.10903.0051 CSPR-7_COASTAL_SPG_U_FLDNO ct-052.602.90653.01812.9149 CSPR-7_COASTAL_SPG_U_FLDNN ov-052.452.88162.99142.8900 CSPR-7_COASTAL_SPG_U_FLDND ec-052.552.91873.02832.9271 CSPR-7_COASTAL_SPG_U_FLDNJan-062.432.84222.95282.8505 CSPR-7_COASTAL_SPG_U_FLDNF eb-062.832.87132.98092.8795 CSPR-7_COASTAL_SPG_U_FLDNM ar-062.552.79142.90182.7996 CSPR-7_COASTAL_SPG_U_FLDNA pr-062.402.69302.80412.7016 CSPR-7_COASTAL_SPG_U_FLDNM ay-062.262.59742.71112.6061 ENGLE_PARK_FLDN SS5.114.07824.26874.0943 ENGLE_PARK_FLDN Jun-044.864.17084.36124.1868 ENGLE_PARK_FLDN Jul-045.074.41694.60804.4328 ENGLE_PARK_FLDN Aug-045.864.50744.69984.5233 205CR_581_NORTH_FLDN Aug-0423.6723.531423.711323.5387 CR_581_NORTH_FLDN Sep-0423.9524.019824.183724.0257 CR_581_NORTH_FLDN Oct-0423.8824.054324.230524.0616 CR_581_NORTH_FLDN Nov-0423.5823.824024.008023.8320 CR_581_NORTH_FLDN Dec-0423.3923.574123.764223.5827 CR_581_NORTH_FLDN Jan-0523.2523.338523.533923.3476 CR_581_NORTH_FLDN Feb-0523.0323.137423.336323.1466 CR_581_NORTH_FLDN Mar-0523.1623.039823.239523.0489 CR_581_NORTH_FLDN Apr-0523.0922.864323.070322.8738 CR_581_NORTH_FLDN May-0522.8622.644222.854922.6537 CR_581_NORTH_FLDN Jun-0523.2422.812223.009522.8211 CR_581_NORTH_FLDN Jul-0523.6523.140623.330623.1494 CR_581_NORTH_FLDN Aug-0523.6223.137523.338123.1465 CR_581_NORTH_FLDN Sep-0522.9622.899723.108422.9090 CR_581_NORTH_FLDN Oct-0522.9122.867723.071622.8769 CR_581_NORTH_FLDN Nov-0522.6522.700822.911322.7101 CR_581_NORTH_FLDN Dec-0522.5422.570322.780922.5796 CR_581_NORTH_FLDN Jan-0622.4522.357022.574322.3665 CR_581_NORTH_FLDN Feb-0622.6022.302322.515022.3116 CR_581_NORTH_FLDN Mar-0622.3322.100922.320122.1103 CR_581_NORTH_FLDN Apr-0621.8221.828522.051621.8381 CR_581_NORTH_FLDN May-0621.4021.532521.760121.5423 CSPR-7_COASTAL_SPG_U_FLD N SS2.602.53562.63132.5455 CSPR-7_COASTAL_SPG_U_FLDNJun-042.552.61032.70582.6197 CSPR-7_COASTAL_SPG_U_FLDNJul-042.782.78382.88082.7931 CSPR-7_COASTAL_SPG_U_FLDNAug-042.762.80112.90022.8103 CSPR-7_COASTAL_SPG_U_FLDNSep-043.153.42893.52423.4380 CSPR-7_COASTAL_SPG_U_FLDNO ct-042.883.44083.54623.4500 CSPR-7_COASTAL_SPG_U_FLDNN ov-042.823.28423.39253.2935 CSPR-7_COASTAL_SPG_U_FLDND ec-042.673.13503.24473.1443 CSPR-7_COASTAL_SPG_U_FLDNJan-052.563.00443.11453.0137 CSPR-7_COASTAL_SPG_U_FLDNF eb-052.532.90223.01242.9115 CSPR-7_COASTAL_SPG_U_FLDNM ar-052.572.83452.94392.8437 CSPR-7_COASTAL_SPG_U_FLDNA pr-052.652.76102.87022.7701 CSPR-7_COASTAL_SPG_U_FLDNM ay-052.602.74832.85602.7571 CSPR-7_COASTAL_SPG_U_FLDNJun-052.822.86722.97502.8759 CSPR-7_COASTAL_SPG_U_FLDNJul-052.903.08533.19333.0939 CSPR-7_COASTAL_SPG_U_FLDNAug-052.653.07673.18773.0852 CSPR-7_COASTAL_SPG_U_FLDNSep-052.652.99663.10903.0051 CSPR-7_COASTAL_SPG_U_FLDNO ct-052.602.90653.01812.9149 CSPR-7_COASTAL_SPG_U_FLDNN ov-052.452.88162.99142.8900 CSPR-7_COASTAL_SPG_U_FLDND ec-052.552.91873.02832.9271 CSPR-7_COASTAL_SPG_U_FLDNJan-062.432.84222.95282.8505 CSPR-7_COASTAL_SPG_U_FLDNF eb-062.832.87132.98092.8795 CSPR-7_COASTAL_SPG_U_FLDNM ar-062.552.79142.90182.7996 CSPR-7_COASTAL_SPG_U_FLDNA pr-062.402.69302.80412.7016 CSPR-7_COASTAL_SPG_U_FLDNM ay-062.262.59742.71112.6061 ENGLE_PARK_FLDN SS5.114.07824.26874.0943 ENGLE_PARK_FLDN Jun-044.864.17084.36124.1868 ENGLE_PARK_FLDN Jul-045.074.41694.60804.4328 ENGLE_PARK_FLDN Aug-045.864.50744.69984.5233

PAGE 221

Appendix E (Continued) Appendix E (Continued) 206ENGLE_PARK_FLDN Sep-046.935.12075.31195.1365 ENGLE_PARK_FLDN Oct-047.085.31815.51355.3340 ENGLE_PARK_FLDN Nov-046.675.15885.35725.1748 ENGLE_PARK_FLDN Dec-046.254.98175.18244.9978 ENGLE_PARK_FLDN Jan-055.944.81385.01614.8299 ENGLE_PARK_FLDN Feb-055.664.67054.87384.6866 ENGLE_PARK_FLDN Mar-055.464.55474.75854.5707 ENGLE_PARK_FLDN Apr-055.324.43774.64184.4536 ENGLE_PARK_FLDN May-055.054.36764.57134.3834 ENGLE_PARK_FLDN Jun-054.964.47384.67714.4894 ENGLE_PARK_FLDN Jul-055.184.75634.95934.7717 ENGLE_PARK_FLDN Aug-055.444.86545.07004.8808 ENGLE_PARK_FLDN Sep-055.474.78054.98624.7957 ENGLE_PARK_FLDN Oct-055.114.67804.88414.6932 ENGLE_PARK_FLDN Nov-054.794.59564.80144.6107 ENGLE_PARK_FLDN Dec-054.634.59864.80404.6136 ENGLE_PARK_FLDN Jan-064.674.52144.72724.5364 ENGLE_PARK_FLDN Feb-064.884.51344.71874.5283 ENGLE_PARK_FLDN Mar-064.684.43064.63604.4455 ENGLE_PARK_FLDN Apr-064.334.29214.49734.3069 ENGLE_PARK_FLDN May-063.944.15134.35704.1661 MASARYKTOWN_DEEP SS10.71 11.470512.2 58311.5105 MASARYKTOWN_DEEP Jun-0411.0 911.510112.2 93811.5501 MASARYKTOWN_DEEP Jul-0410.7 511.788712.5 67211.8286 MASARYKTOWN_DEEP Aug-0410.8 711.995812.7 74112.0358 MASARYKTOWN_DEEP Sep-0412.7 112.615013.3 70812.6544 MASARYKTOWN_DEEP Oct-0414.5 512.737313.5 15712.7776 MASARYKTOWN_DEEP Nov-0414.1 912.543213.3 38212.5843 MASARYKTOWN_DEEP Dec-0413.6 012.271413.0 78112.3130 MASARYKTOWN_DEEP Jan-0513.0 312.025912.8 43112.0678 MASARYKTOWN_DEEP Feb-0512.5211.790712.6 14611.8327 MASARYKTOWN_DEEP Mar-0512.0 311.576812.4 03711.6186 MASARYKTOWN_DEEP Apr-0511.5 611.359612.1 90111.4012 MASARYKTOWN_DEEP May-0511.1 511.171912.0 00311.2132 MASARYKTOWN_DEEP Jun-0510.7 811.201712.0 14411.2420 MASARYKTOWN_DEEP Jul-0510.6 211.344612.1 40311.3842 MASARYKTOWN_DEEP Aug-0510.7 711.427112.2 19411.4664 MASARYKTOWN_DEEP Sep-0510.8 611.288112.0 87811.3276 MASARYKTOWN_DEEP Oct-0510.7 111.241912.0 33811.2810 MASARYKTOWN_DEEP Nov-0510.5 411.138411.9 33011.1775 MASARYKTOWN_DEEP Dec-0510.3 311.109611.8 97311.1483 MASARYKTOWN_DEEP Jan-0610.1 110.943711.7 38410.9829 MASARYKTOWN_DEEP Feb-069. 9010.932311.7 16910.9708 MASARYKTOWN_DEEP Mar-069. 7510.781611.5 73710.8202 MASARYKTOWN_DEEP Apr-069. 5610.547411.3 43110.5860 MASARYKTOWN_DEEP May-069. 2310.294211.0 91810.3327 PLESS_PARK_FLDN SS27.56 24.920125.0 92424.9270 PLESS_PARK_FLDN Jun-0425.4 025.144125.2 98925.1507 PLESS_PARK_FLDN Jul-0425.7 525.398525.5 47125.4048 PLESS_PARK_FLDN Aug-0426.0 725.711825.8 44825.7177 PLESS_PARK_FLDN Sep-0427.4 626.311326.4 21726.3166 206ENGLE_PARK_FLDN Sep-046.935.12075.31195.1365 ENGLE_PARK_FLDN Oct-047.085.31815.51355.3340 ENGLE_PARK_FLDN Nov-046.675.15885.35725.1748 ENGLE_PARK_FLDN Dec-046.254.98175.18244.9978 ENGLE_PARK_FLDN Jan-055.944.81385.01614.8299 ENGLE_PARK_FLDN Feb-055.664.67054.87384.6866 ENGLE_PARK_FLDN Mar-055.464.55474.75854.5707 ENGLE_PARK_FLDN Apr-055.324.43774.64184.4536 ENGLE_PARK_FLDN May-055.054.36764.57134.3834 ENGLE_PARK_FLDN Jun-054.964.47384.67714.4894 ENGLE_PARK_FLDN Jul-055.184.75634.95934.7717 ENGLE_PARK_FLDN Aug-055.444.86545.07004.8808 ENGLE_PARK_FLDN Sep-055.474.78054.98624.7957 ENGLE_PARK_FLDN Oct-055.114.67804.88414.6932 ENGLE_PARK_FLDN Nov-054.794.59564.80144.6107 ENGLE_PARK_FLDN Dec-054.634.59864.80404.6136 ENGLE_PARK_FLDN Jan-064.674.52144.72724.5364 ENGLE_PARK_FLDN Feb-064.884.51344.71874.5283 ENGLE_PARK_FLDN Mar-064.684.43064.63604.4455 ENGLE_PARK_FLDN Apr-064.334.29214.49734.3069 ENGLE_PARK_FLDN May-063.944.15134.35704.1661 MASARYKTOWN_DEEP SS10.71 11.470512.2 58311.5105 MASARYKTOWN_DEEP Jun-0411.0 911.510112.2 93811.5501 MASARYKTOWN_DEEP Jul-0410.7 511.788712.5 67211.8286 MASARYKTOWN_DEEP Aug-0410.8 711.995812.7 74112.0358 MASARYKTOWN_DEEP Sep-0412.7 112.615013.3 70812.6544 MASARYKTOWN_DEEP Oct-0414.5 512.737313.5 15712.7776 MASARYKTOWN_DEEP Nov-0414.1 912.543213.3 38212.5843 MASARYKTOWN_DEEP Dec-0413.6 012.271413.0 78112.3130 MASARYKTOWN_DEEP Jan-0513.0 312.025912.8 43112.0678 MASARYKTOWN_DEEP Feb-0512.5211.790712.6 14611.8327 MASARYKTOWN_DEEP Mar-0512.0 311.576812.4 03711.6186 MASARYKTOWN_DEEP Apr-0511.5 611.359612.1 90111.4012 MASARYKTOWN_DEEP May-0511.1 511.171912.0 00311.2132 MASARYKTOWN_DEEP Jun-0510.7 811.201712.0 14411.2420 MASARYKTOWN_DEEP Jul-0510.6 211.344612.1 40311.3842 MASARYKTOWN_DEEP Aug-0510.7 711.427112.2 19411.4664 MASARYKTOWN_DEEP Sep-0510.8 611.288112.0 87811.3276 MASARYKTOWN_DEEP Oct-0510.7 111.241912.0 33811.2810 MASARYKTOWN_DEEP Nov-0510.5 411.138411.9 33011.1775 MASARYKTOWN_DEEP Dec-0510.3 311.109611.8 97311.1483 MASARYKTOWN_DEEP Jan-0610.1 110.943711.7 38410.9829 MASARYKTOWN_DEEP Feb-069. 9010.932311.7 16910.9708 MASARYKTOWN_DEEP Mar-069. 7510.781611.5 73710.8202 MASARYKTOWN_DEEP Apr-069. 5610.547411.3 43110.5860 MASARYKTOWN_DEEP May-069. 2310.294211.0 91810.3327 PLESS_PARK_FLDN SS27.56 24.920125.0 92424.9270 PLESS_PARK_FLDN Jun-0425.4 025.144125.2 98925.1507 PLESS_PARK_FLDN Jul-0425.7 525.398525.5 47125.4048 PLESS_PARK_FLDN Aug-0426.0 725.711825.8 44825.7177 PLESS_PARK_FLDN Sep-0427.4 626.311326.4 21726.3166

PAGE 222

Appendix E (Continued) 207PLESS_PARK_FLDN Oct-0428.5 226.557626.6 74626.5630 PLESS_PARK_FLDN Nov-0428.2 126.508626.6 33626.5141 PLESS_PARK_FLDN Dec-0427.7 226.387226.5 22226.3931 PLESS_PARK_FLDN Jan-0527.6 526.234326.3 76926.2404 PLESS_PARK_FLDN Feb-0527.4 626.094226.2 44926.1005 PLESS_PARK_FLDN Mar-0527.4 025.987626.1 43525.9942 PLESS_PARK_FLDN Apr-0527.2 325.836426.0 01225.8433 PLESS_PARK_FLDN May-0526.8 825.623125.7 96825.6302 PLESS_PARK_FLDN Jun-0526.8 325.709925.8 76225.7169 PLESS_PARK_FLDN Jul-0527.5 625.999826.1 56026.0067 PLESS_PARK_FLDN Aug-0527.7 726.139926.2 97626.1468 PLESS_PARK_FLDN Sep-0527.7 426.063726.2 29026.0706 PLESS_PARK_FLDN Oct-0527.5 625.939826.1 08625.9469 PLESS_PARK_FLDN Nov-0527.0 925.828126.0 01725.8353 PLESS_PARK_FLDN Dec-0526.8 225.738025.9 15025.7453 PLESS_PARK_FLDN Jan-0626.5 025.545925.7 30225.5534 PLESS_PARK_FLDN Feb-0625.8 825.414125.5 99325.4216 PLESS_PARK_FLDN Mar-0625.3 925.221625.4 13925.2293 PLESS_PARK_FLDN Apr-0624.6 624.946225.1 46124.9540 PLESS_PARK_FLDN May-0624.1 624.641224.8 49124.6490 ROMP_105_AVPK_490_F T SS11.6910.3668 11.095110.3814 ROMP_105_AVPK_490_FT Jun-041 1.6210.573511 .307810.5881 ROMP_105_AVPK_490_FT Jul-041 1.4610.731911 .473810.7465 ROMP_105_AVPK_490_FT Aug-041 1.6410.937311 .688110.9519 ROMP_105_AVPK_490_FT Sep-041 2.6412.011612 .770412.0262 ROMP_105_AVPK_490_FT Oct-041 3.6211.724412 .488711.7389 ROMP_105_AVPK_490_F T Nov-0413.6811.33 1812.096 211.3466 ROMP_105_AVPK_490_F T Dec-0413.4210.96 9411.731 110.9843 ROMP_105_AVPK_490_FT Jan-051 3.1310.615711 .372410.6307 ROMP_105_AVPK_490_F T Feb-0512.8310.31 3411.063 410.3285 ROMP_105_AVPK_490_F T Mar-0512.5310.05 0610.791 010.0657 ROMP_105_AVPK_490_FT Apr-0512.259.802710.53229.8176 ROMP_105_AVPK_490_F T May-0512.009.788010.50609.8028 ROMP_105_AVPK_490_FT Jun-051 1.799.809210.51639.8237 ROMP_105_AVPK_490_FT Jul-051 1.7610.010010 .706010.0243 ROMP_105_AVPK_490_FT Aug-051 1.8610.185310 .874710.1994 ROMP_105_AVPK_490_FT Sep-051 1.839.977310.66419.9913 ROMP_105_AVPK_490_FT Oct-051 1.699.988110.671310.0019 ROMP_105_AVPK_490_FT Nov-0511.609.931610.61369.9454 ROMP_105_AVPK_490_FT Dec-0511.459.909510.59129.9232 ROMP_105_AVPK_490_FT Jan-061 1.309.623010.30379.6366 ROMP_105_AVPK_490_F T Feb-0611.209.666210.34389.6798 ROMP_105_AVPK_490_FT Mar-0611.079.385410.05889.3989 ROMP_105_AVPK_490_F T Apr-0610.959.08849.75599.1018 ROMP_105_AVPK_490_F T May-0610.848.78429.44348.7975 ROMP_93_SWNN/AVPK SS22.23 22.763522.8 35122.7668 ROMP_93_SWNN/AVPK Jun-0421.6 622.642222.7 12722.6454 ROMP_93_SWNN/AVPK Jul-0422.0 922.933223.0 02922.9363 ROMP_93_SWNN/AVPK Aug-0422.6 923.245223.3 13823.2482 ROMP_93_SWNN/AVPK Sep-0423.0 823.567123.6 33523.5699 ROMP_93_SWNN/AVPK Oct-0423.0 623.640423.7 06723.6432

PAGE 223

Appendix E (Continued) 208ROMP_93_SWNN/AVPK Nov-0422.8 123.326423.3 93523.3293 ROMP_93_SWNN/AVPK Dec-0422.6 423.121823.1 90023.1249 ROMP_93_SWNN/AVPK Jan-0522.5 723.010523.0 79823.0136 ROMP_93_SWNN/AVPK Feb-0522.4 022.866922.9 37222.8701 ROMP_93_SWNN/AVPK Mar-0522.6 522.955523.0 26722.9588 ROMP_93_SWNN/AVPK Apr-0522.4 522.789022.8 61322.7924 ROMP_93_SWNN/AVPK May-0522.2422.637122.7 10622.6405 ROMP_93_SWNN/AVPK Jun-0522.4 022.708222.7 81222.7116 ROMP_93_SWNN/AVPK Jul-0522.7 322.923222.9 95522.9266 ROMP_93_SWNN/AVPK Aug-0522.7 022.994923.0 67822.9982 ROMP_93_SWNN/AVPK Sep-0522.2 922.732622.8 06422.7360 ROMP_93_SWNN/AVPK Oct-0522.2 322.754822.8 28922.7582 ROMP_93_SWNN/AVPK Nov-0521.9 622.590022.6 64722.5934 ROMP_93_SWNN/AVPK Dec-0522.0 022.601922.6 76922.6053 ROMP_93_SWNN/AVPK Jan-0621.7 722.427922.5 03822.4314 ROMP_93_SWNN/AVPK Feb-0622.0 022.515622.5 91422.5190 ROMP_93_SWNN/AVPK Mar-0621.6 722.288422.3 65222.2919 ROMP_93_SWNN/AVPK Apr-0621.3 021.992122.0 70421.9956 ROMP_93_SWNN/AVPK May-0621.0421.723821.8 04821.7275 ROMP_97_AVPK SS6.085.20855.51915.2460 ROMP_97_AVPK Jun-046.08 5.20505.51555.2431 ROMP_97_AVPK Jul-046.165.32995.64085.3682 ROMP_97_AVPK Aug-046.54 5.34825.66045.3867 ROMP_97_AVPK Sep-047.60 5.92766.23535.9663 ROMP_97_AVPK Oct-047.83 6.03796.35726.0789 ROMP_97_AVPK Nov-047.525.82636.15195.8675 ROMP_97_AVPK Dec-047.245.62115.95085.6620 ROMP_97_AVPK Jan-056.91 5.42125.75265.4616 ROMP_97_AVPK Feb-056.67 5.26035.59205.2999 ROMP_97_AVPK Mar-056.46 5.12865.45955.1672 ROMP_97_AVPK Apr-056.27 4.99105.31985.0284 ROMP_97_AVPK May-056.06 4.91415.23754.9501 ROMP_97_AVPK Jun-055.90 4.98975.30565.0248 ROMP_97_AVPK Jul-056.105.21175.52395.2466 ROMP_97_AVPK Aug-056.32 5.30545.62075.3406 ROMP_97_AVPK Sep-056.26 5.20755.52635.2430 ROMP_97_AVPK Oct-056.08 5.17615.49345.2113 ROMP_97_AVPK Nov-055.855.12275.43905.1579 ROMP_97_AVPK Dec-055.725.10735.42235.1426 ROMP_97_AVPK Jan-065.53 4.99895.31515.0340 ROMP_97_AVPK Feb-065.65 4.98195.29585.0166 ROMP_97_AVPK Mar-065.45 4.87275.18674.9070 ROMP_97_AVPK Apr-065.14 4.71205.02534.7458 ROMP_97_AVPK May-064.84 4.55904.87214.5922 ROMP_98_FLDN SS11.0911.390912.179111.4267 ROMP_98_FLDN Jun-0411.2411.478612.262911.5142 ROMP_98_FLDN Jul-0410.9911.747712.520811.7828 ROMP_98_FLDN Aug-0411.2411.977212.754012.0126 ROMP_98_FLDN Sep-0412.7212.684113.428912.7180 ROMP_98_FLDN Oct-0414.2712.680913.465212.7169 ROMP_98_FLDN Nov-0413.9012.407913.210312.4449

PAGE 224

Appendix E (Continued) Appendix E (Continued) 209ROMP_98_FLDN Dec-0413.4012.113112.927812.1507 ROMP_98_FLDN Jan-0513.1311.824712.647311.8626 ROMP_98_FLDN Feb-0512.6511.567012.394011.6051 ROMP_98_FLDN Mar-0512.2011.327812.154811.3658 ROMP_98_FLDN Apr-0511.7711.096611.924111.1343 ROMP_98_FLDN May-0511.3710.940611.762510.9779 ROMP_98_FLDN Jun-0511.0611.017011.816411.0532 ROMP_98_FLDN Jul-0511.1211.253012.029311.2882 ROMP_98_FLDN Aug-0511.3411.286712.065111.3217 ROMP_98_FLDN Sep-0511.3211.108811.899711.1441 ROMP_98_FLDN Oct-0511.0911.079811.856211.1146 ROMP_98_FLDN Nov-0510.8410.947211.730910.9820 ROMP_98_FLDN Dec-0510.6210.916811.690610.9514 ROMP_98_FLDN Jan-0610.3110.721311.504310.7564 ROMP_98_FLDN Feb-0610.1110.722111.489110.7566 ROMP_98_FLDN Mar-069.9310.531411.309410.5661 ROMP_98_FLDN Apr-069.6710.281311.062710.3161 ROMP_98_FLDN May-069.3910.018410.800910.0532 ROMP_CENTRALIA_OCAL SS3.714.93075.53524.9396 ROMP_CENTRALIA_OCAL Jun-043.364.66045.23374.6694 ROMP_CENTRALIA_OCAL Jul-043.384.72565.27954.7346 ROMP_CENTRALIA_OCAL Aug-043.734.68025.20844.6892 ROMP_CENTRALIA_OCAL Sep-045.045.59076.18355.5993 ROMP_CENTRALIA_OCAL Oct-045.535.41965.99055.4288 ROMP_CENTRALIA_OCAL Nov-045.115.17375.71715.1831 ROMP_CENTRALIA_OCAL Dec-044.674.88735.40024.8968 ROMP_CENTRALIA_OCAL Jan-054.284.61505.09414.6242 ROMP_CENTRALIA_OCAL Feb-054.014.40894.85694.4180 ROMP_CENTRALIA_OCAL Mar-053.794.26164.68284.2704 ROMP_CENTRALIA_OCAL Apr-053.644.13244.52704.1409 ROMP_CENTRALIA_OCAL May-053.614.17844.56984.1865 ROMP_CENTRALIA_OCAL Jun-053.584.31554.71724.3234 ROMP_CENTRALIA_OCAL Jul-053.924.63525.06654.6430 ROMP_CENTRALIA_OCAL Aug-054.094.66905.09704.6767 ROMP_CENTRALIA_OCAL Sep-053.924.44364.85114.4514 ROMP_CENTRALIA_OCAL Oct-053.714.58595.03114.5935 ROMP_CENTRALIA_OCAL Nov-053.544.47094.91514.4786 ROMP_CENTRALIA_OCAL Dec-053.544.39974.83534.4074 ROMP_CENTRALIA_OCAL Jan-063.414.15504.55864.1627 ROMP_CENTRALIA_OCAL Feb-063.544.19254.59084.2001 ROMP_CENTRALIA_OCAL Mar-063.453.98714.35953.9946 ROMP_CENTRALIA_OCAL Apr-063.193.78864.13473.7959 ROMP_CENTRALIA_OCAL May-063.053.60823.92823.6153 ROMP_TR_17-3_SWNN SS0.842.67632.79892.6845 ROMP_TR_17-3_SWNN Jun-040.812.78992.91332.7980 ROMP_TR_17-3_SWNN Jul-040.843.03023.15533.0383 ROMP_TR_17-3_SWNN Aug-040.973.12213.24933.1302 ROMP_TR_17-3_SWNN Sep-040.713.91174.03873.9198 ROMP_TR_17-3_SWNN Oct-040.863.96894.10313.9770 ROMP_TR_17-3_SWNN Nov-040.753.77703.91293.7851 ROMP_TR_17-3_SWNN Dec-040.853.59573.73253.6038 209ROMP_98_FLDN Dec-0413.4012.113112.927812.1507 ROMP_98_FLDN Jan-0513.1311.824712.647311.8626 ROMP_98_FLDN Feb-0512.6511.567012.394011.6051 ROMP_98_FLDN Mar-0512.2011.327812.154811.3658 ROMP_98_FLDN Apr-0511.7711.096611.924111.1343 ROMP_98_FLDN May-0511.3710.940611.762510.9779 ROMP_98_FLDN Jun-0511.0611.017011.816411.0532 ROMP_98_FLDN Jul-0511.1211.253012.029311.2882 ROMP_98_FLDN Aug-0511.3411.286712.065111.3217 ROMP_98_FLDN Sep-0511.3211.108811.899711.1441 ROMP_98_FLDN Oct-0511.0911.079811.856211.1146 ROMP_98_FLDN Nov-0510.8410.947211.730910.9820 ROMP_98_FLDN Dec-0510.6210.916811.690610.9514 ROMP_98_FLDN Jan-0610.3110.721311.504310.7564 ROMP_98_FLDN Feb-0610.1110.722111.489110.7566 ROMP_98_FLDN Mar-069.9310.531411.309410.5661 ROMP_98_FLDN Apr-069.6710.281311.062710.3161 ROMP_98_FLDN May-069.3910.018410.800910.0532 ROMP_CENTRALIA_OCAL SS3.714.93075.53524.9396 ROMP_CENTRALIA_OCAL Jun-043.364.66045.23374.6694 ROMP_CENTRALIA_OCAL Jul-043.384.72565.27954.7346 ROMP_CENTRALIA_OCAL Aug-043.734.68025.20844.6892 ROMP_CENTRALIA_OCAL Sep-045.045.59076.18355.5993 ROMP_CENTRALIA_OCAL Oct-045.535.41965.99055.4288 ROMP_CENTRALIA_OCAL Nov-045.115.17375.71715.1831 ROMP_CENTRALIA_OCAL Dec-044.674.88735.40024.8968 ROMP_CENTRALIA_OCAL Jan-054.284.61505.09414.6242 ROMP_CENTRALIA_OCAL Feb-054.014.40894.85694.4180 ROMP_CENTRALIA_OCAL Mar-053.794.26164.68284.2704 ROMP_CENTRALIA_OCAL Apr-053.644.13244.52704.1409 ROMP_CENTRALIA_OCAL May-053.614.17844.56984.1865 ROMP_CENTRALIA_OCAL Jun-053.584.31554.71724.3234 ROMP_CENTRALIA_OCAL Jul-053.924.63525.06654.6430 ROMP_CENTRALIA_OCAL Aug-054.094.66905.09704.6767 ROMP_CENTRALIA_OCAL Sep-053.924.44364.85114.4514 ROMP_CENTRALIA_OCAL Oct-053.714.58595.03114.5935 ROMP_CENTRALIA_OCAL Nov-053.544.47094.91514.4786 ROMP_CENTRALIA_OCAL Dec-053.544.39974.83534.4074 ROMP_CENTRALIA_OCAL Jan-063.414.15504.55864.1627 ROMP_CENTRALIA_OCAL Feb-063.544.19254.59084.2001 ROMP_CENTRALIA_OCAL Mar-063.453.98714.35953.9946 ROMP_CENTRALIA_OCAL Apr-063.193.78864.13473.7959 ROMP_CENTRALIA_OCAL May-063.053.60823.92823.6153 ROMP_TR_17-3_SWNN SS0.842.67632.79892.6845 ROMP_TR_17-3_SWNN Jun-040.812.78992.91332.7980 ROMP_TR_17-3_SWNN Jul-040.843.03023.15533.0383 ROMP_TR_17-3_SWNN Aug-040.973.12213.24933.1302 ROMP_TR_17-3_SWNN Sep-040.713.91174.03873.9198 ROMP_TR_17-3_SWNN Oct-040.863.96894.10313.9770 ROMP_TR_17-3_SWNN Nov-040.753.77703.91293.7851 ROMP_TR_17-3_SWNN Dec-040.853.59573.73253.6038

PAGE 225

Appendix E (Continued) Appendix E (Continued) 210ROMP_TR_17-3_SWNN Jan-050.613.43433.57163.4424 ROMP_TR_17-3_SWNN Feb-050.523.30733.44473.3154 ROMP_TR_17-3_SWNN Mar-050.533.23873.37623.2468 ROMP_TR_17-3_SWNN Apr-050.973.13893.27643.1470 ROMP_TR_17-3_SWNN May-051.373.10713.24383.1152 ROMP_TR_17-3_SWNN Jun-050.923.21333.34973.2214 ROMP_TR_17-3_SWNN Jul-051.083.54733.68513.5553 ROMP_TR_17-3_SWNN Aug-050.753.61033.75183.6183 ROMP_TR_17-3_SWNN Sep-051.043.49933.64173.5072 ROMP_TR_17-3_SWNN Oct-050.843.38213.52433.3901 ROMP_TR_17-3_SWNN Nov-050.683.30253.44393.3105 ROMP_TR_17-3_SWNN Dec-050.483.31443.45523.3223 ROMP_TR_17-3_SWNN Jan-060.483.22453.36573.2323 ROMP_TR_17-3_SWNN Feb-060.703.24453.38493.2523 ROMP_TR_17-3_SWNN Mar-060.673.15353.29423.1613 ROMP_TR_17-3_SWNN Apr-060.713.03563.17563.0433 ROMP_TR_17-3_SWNN May-060.802.92053.05952.9282 ROMP_TR_18-1A_DP_UP_FLDN SS3.983.13103.22093.1481 ROMP_TR_18-1A_DP_UP_FLDNJun-044.053.16383.25403.1808 ROMP_TR_18-1A_DP_UP_FLDN Jul-044.193.27123.36123.2880 ROMP_TR_18-1A_DP_UP_FLDNAug-044.253.25613.34593.2728 ROMP_TR_18-1A_DP_UP_FLDNSep-044.773.90113.98913.9178 ROMP_TR_18-1A_DP_UP_FLDN Oct-044.613.90464.00433.9224 ROMP_TR_18-1A_DP_UP_FLDNNov-044.523.71563.81913.7335 ROMP_TR_18-1A_DP_UP_FLDNDec-044.463.53973.64463.5575 ROMP_TR_18-1A_DP_UP_FLDNJan-054.283.38423.48963.4018 ROMP_TR_18-1A_DP_UP_FLDNFeb-054.213.26553.37083.2825 ROMP_TR_18-1A_DP_UP_FLDNMar-054.173.18783.29273.2043 ROMP_TR_18-1A_DP_UP_FLDN Apr-054.123.10803.21303.1239 ROMP_TR_18-1A_DP_UP_FLDNMay-054.093.08653.18953.1018 ROMP_TR_18-1A_DP_UP_FLDNJun-054.003.19643.29703.2112 ROMP_TR_18-1A_DP_UP_FLDN Jul-054.203.43403.53513.4487 ROMP_TR_18-1A_DP_UP_FLDNAug-054.083.41613.51923.4310 ROMP_TR_18-1A_DP_UP_FLDNSep-054.003.30993.41343.3248 ROMP_TR_18-1A_DP_UP_FLDN Oct-053.983.24453.34593.2593 ROMP_TR_18-1A_DP_UP_FLDNNov-053.833.24003.34093.2548 ROMP_TR_18-1A_DP_UP_FLDNDec-053.853.26693.36883.2818 ROMP_TR_18-1A_DP_UP_FLDNJan-063.803.17283.27583.1876 ROMP_TR_18-1A_DP_UP_FLDNFeb-063.993.19443.29643.2090 ROMP_TR_18-1A_DP_UP_FLDNMar-063.793.10223.20523.1166 ROMP_TR_18-1A_DP_UP_FLDN Apr-063.642.98603.08993.0005 ROMP_TR_18-1A_DP_UP_FLDNMay-063.432.87522.98172.8896 ROMP_TR_18-2_UPPER_AVPK SS1.412.44052.48162.4516 ROMP_TR_18-2_UPPER_AVPK Jun-041.352.48902.53012.4995 ROMP_TR_18-2_UPPER_AVPK Jul-041.522.60612.64782.6165 ROMP_TR_18-2_UPPER_AVPK Aug-041.512.58902.63092.5993 ROMP_TR_18-2_UPPER_AVPK Sep-043.023.22433.26613.2346 ROMP_TR_18-2_UPPER_AVPK Oct-042.833.14383.19143.1545 ROMP_TR_18-2_UPPER _AVPK Nov-042.812.98263.03282.9935 ROMP_TR_18-2_UPPER _AVPK Dec-042.732.83822.88972.8491 ROMP_TR_18-2_UPPER_AVPK Jan-051.552.71552.76772.7262 210ROMP_TR_17-3_SWNN Jan-050.613.43433.57163.4424 ROMP_TR_17-3_SWNN Feb-050.523.30733.44473.3154 ROMP_TR_17-3_SWNN Mar-050.533.23873.37623.2468 ROMP_TR_17-3_SWNN Apr-050.973.13893.27643.1470 ROMP_TR_17-3_SWNN May-051.373.10713.24383.1152 ROMP_TR_17-3_SWNN Jun-050.923.21333.34973.2214 ROMP_TR_17-3_SWNN Jul-051.083.54733.68513.5553 ROMP_TR_17-3_SWNN Aug-050.753.61033.75183.6183 ROMP_TR_17-3_SWNN Sep-051.043.49933.64173.5072 ROMP_TR_17-3_SWNN Oct-050.843.38213.52433.3901 ROMP_TR_17-3_SWNN Nov-050.683.30253.44393.3105 ROMP_TR_17-3_SWNN Dec-050.483.31443.45523.3223 ROMP_TR_17-3_SWNN Jan-060.483.22453.36573.2323 ROMP_TR_17-3_SWNN Feb-060.703.24453.38493.2523 ROMP_TR_17-3_SWNN Mar-060.673.15353.29423.1613 ROMP_TR_17-3_SWNN Apr-060.713.03563.17563.0433 ROMP_TR_17-3_SWNN May-060.802.92053.05952.9282 ROMP_TR_18-1A_DP_UP_FLDN SS3.983.13103.22093.1481 ROMP_TR_18-1A_DP_UP_FLDNJun-044.053.16383.25403.1808 ROMP_TR_18-1A_DP_UP_FLDN Jul-044.193.27123.36123.2880 ROMP_TR_18-1A_DP_UP_FLDNAug-044.253.25613.34593.2728 ROMP_TR_18-1A_DP_UP_FLDNSep-044.773.90113.98913.9178 ROMP_TR_18-1A_DP_UP_FLDN Oct-044.613.90464.00433.9224 ROMP_TR_18-1A_DP_UP_FLDNNov-044.523.71563.81913.7335 ROMP_TR_18-1A_DP_UP_FLDNDec-044.463.53973.64463.5575 ROMP_TR_18-1A_DP_UP_FLDNJan-054.283.38423.48963.4018 ROMP_TR_18-1A_DP_UP_FLDNFeb-054.213.26553.37083.2825 ROMP_TR_18-1A_DP_UP_FLDNMar-054.173.18783.29273.2043 ROMP_TR_18-1A_DP_UP_FLDN Apr-054.123.10803.21303.1239 ROMP_TR_18-1A_DP_UP_FLDNMay-054.093.08653.18953.1018 ROMP_TR_18-1A_DP_UP_FLDNJun-054.003.19643.29703.2112 ROMP_TR_18-1A_DP_UP_FLDN Jul-054.203.43403.53513.4487 ROMP_TR_18-1A_DP_UP_FLDNAug-054.083.41613.51923.4310 ROMP_TR_18-1A_DP_UP_FLDNSep-054.003.30993.41343.3248 ROMP_TR_18-1A_DP_UP_FLDN Oct-053.983.24453.34593.2593 ROMP_TR_18-1A_DP_UP_FLDNNov-053.833.24003.34093.2548 ROMP_TR_18-1A_DP_UP_FLDNDec-053.853.26693.36883.2818 ROMP_TR_18-1A_DP_UP_FLDNJan-063.803.17283.27583.1876 ROMP_TR_18-1A_DP_UP_FLDNFeb-063.993.19443.29643.2090 ROMP_TR_18-1A_DP_UP_FLDNMar-063.793.10223.20523.1166 ROMP_TR_18-1A_DP_UP_FLDN Apr-063.642.98603.08993.0005 ROMP_TR_18-1A_DP_UP_FLDNMay-063.432.87522.98172.8896 ROMP_TR_18-2_UPPER_AVPK SS1.412.44052.48162.4516 ROMP_TR_18-2_UPPER_AVPK Jun-041.352.48902.53012.4995 ROMP_TR_18-2_UPPER_AVPK Jul-041.522.60612.64782.6165 ROMP_TR_18-2_UPPER_AVPK Aug-041.512.58902.63092.5993 ROMP_TR_18-2_UPPER_AVPK Sep-043.023.22433.26613.2346 ROMP_TR_18-2_UPPER_AVPK Oct-042.833.14383.19143.1545 ROMP_TR_18-2_UPPER _AVPK Nov-042.812.98263.03282.9935 ROMP_TR_18-2_UPPER _AVPK Dec-042.732.83822.88972.8491 ROMP_TR_18-2_UPPER_AVPK Jan-051.552.71552.76772.7262

PAGE 226

Appendix E (Continued) 211ROMP_TR_18-2_UPPER _AVPK Feb-051.512.62252.67522.6330 ROMP_TR_18-2_UPPER _AVPK Mar-051.582.56592.61852.5762 ROMP_TR_18-2_UPPER _AVPK Apr-051.492.50472.55792.5147 ROMP_TR_18-2_UPPER _AVPK May-051.492.50512.55792.5147 ROMP_TR_18-2_UPPER_AVPK Jun-051.492.60092.65262.6103 ROMP_TR_18-2_UPPER_AVPK Jul-051.642.81562.86752.8248 ROMP_TR_18-2_UPPER_AVPK Aug-051.492.75942.81142.7686 ROMP_TR_18-2_UPPER_AVPK Sep-051.472.67372.72622.6829 ROMP_TR_18-2_UPPER_AVPK Oct-051.412.60672.65802.6159 ROMP_TR_18-2_UPPER _AVPK Nov-051.352.61572.66702.6249 ROMP_TR_18-2_UPPER _AVPK Dec-051.262.65252.70462.6617 ROMP_TR_18-2_UPPER_AVPK Jan-061.282.56912.62212.5783 ROMP_TR_18-2_UPPER _AVPK Feb-061.402.60542.65832.6145 ROMP_TR_18-2_UPPER _AVPK Mar-061.252.52172.57532.5307 ROMP_TR_18-2_UPPER _AVPK Apr-061.202.43182.48722.4411 ROMP_TR_18-2_UPPER _AVPK May-061.052.34562.40522.3549 ROMP_TR_19-2_OCAL SS1.542.13112.18602.1349 ROMP_TR_19-2_OCAL Jun-041.812.09712.15192.1011 ROMP_TR_19-2_OCAL Jul-041.982.11282.16452.1167 ROMP_TR_19-2_OCAL Aug-041.652.08872.13892.0923 ROMP_TR_19-2_OCAL Sep-041.782.46422.51162.4679 ROMP_TR_19-2_OCAL Oct-041.592.40692.46252.4112 ROMP_TR_19-2_OCAL Nov-041.692.27722.33212.2813 ROMP_TR_19-2_OCAL Dec-041.332.17552.22852.1792 ROMP_TR_19-2_OCAL Jan-051.352.09192.14302.0955 ROMP_TR_19-2_OCAL Feb-051.432.03292.08202.0361 ROMP_TR_19-2_OCAL Mar-052.001.99822.04522.0010 ROMP_TR_19-2_OCAL Apr-051.861.97292.01921.9755 ROMP_TR_19-2_OCAL May-051.551.99102.03571.9934 ROMP_TR_19-2_OCAL Jun-051.882.06852.11282.0708 ROMP_TR_19-2_OCAL Jul-051.672.20032.24502.2027 ROMP_TR_19-2_OCAL Aug-052.052.15962.20462.1621 ROMP_TR_19-2_OCAL Sep-051.742.08102.12562.0835 ROMP_TR_19-2_OCAL Oct-051.542.08122.12492.0837 ROMP_TR_19-2_OCAL Nov-051.542.09322.13812.0959 ROMP_TR_19-2_OCAL Dec-051.362.08442.12952.0871 ROMP_TR_19-2_OCAL Jan-061.422.01682.06232.0194 ROMP_TR_19-2_OCAL Feb-061.612.03772.08232.0402 ROMP_TR_19-2_OCAL Mar-061.711.98222.02751.9846 ROMP_TR_19-2_OCAL Apr-061.511.91891.96281.9210 ROMP_TR_19-2_OCAL May-061.881.86371.90721.8660 SAWYER_CB_4E_FLDN_REPL SS22.1922.236922.535822.2499 SAWYER_CB_4E_FLDN_REPL Jun-0421.2522.489522.777922.5017 SAWYER_CB_4E_FLDN_REPL Jul-0421.4222.627322.912122.6394 SAWYER_CB_4E_FLDN_REPL Aug-0422.1122.902123.178922.9137 SAWYER_CB_4E_FLDN_REPL Sep-0422.8023.598523.853023.6087 SAWYER_CB_4E_FLDN_REPL Oct-0423.4323.469223.734423.4803 SAWYER_CB_4E_FLDN_REPL Nov-0423.3623.160723.438623.1726 SAWYER_CB_4E_FLDN_REPL Dec-0423.0422.862623.151722.8752 SAWYER_CB_4E_FLDN_REPL Jan-0522.8022.606622.905022.6198 SAWYER_CB_4E_FLDN_REPL Feb-0522.5822.378622.683522.3922

PAGE 227

Appendix E (Continued) Appendix E (Continued) 212SAWYER_CB_4E_FLDN_REPL Mar-0522.4622.261222.569222.2749 SAWYER_CB_4E_FLDN_REPL Apr-0522.4822.017622.332222.0315 SAWYER_CB_4E_FLDN_REPL May-0522.3121.769722.090221.7839 SAWYER_CB_4E_FLDN_REPL Jun-0522.3522.020722.332822.0345 SAWYER_CB_4E_FLDN_REPL Jul-0522.9022.326522.630722.3401 SAWYER_CB_4E_FLDN_REPL Aug-0523.0722.263222.570522.2768 SAWYER_CB_4E_FLDN_REPL Sep-0522.4022.018222.331622.0319 SAWYER_CB_4E_FLDN_REPL Oct-0522.1922.081422.394222.0951 SAWYER_CB_4E_FLDN_REPL Nov-0521.8521.850922.167021.8647 SAWYER_CB_4E_FLDN_REPL Dec-0521.6221.738122.054921.7518 SAWYER_CB_4E_FLDN_REPL Jan-0621.4321.478121.798821.4919 SAWYER_CB_4E_FLDN_REPL Feb-0621.3621.460021.779221.4737 SAWYER_CB_4E_FLDN_REPL Mar-0621.2021.218121.540021.2319 SAWYER_CB_4E_FLDN_REPL Apr-0620.8720.947421.272520.9614 SAWYER_CB_4E_FLDN_REPL May-0620.5720.659320.987520.6734 SPRING_HILL_FLDN SS7.128.09119.11818.1202 SPRING_HILL_FLDN Jun-046.938.13449.17418.1636 SPRING_HILL_FLDN Jul-046.708.38189.43568.4110 SPRING_HILL_FLDN Aug-046.948.50429.57558.5335 SPRING_HILL_FLDN Sep-047.729.381110.45159.4099 SPRING_HILL_FLDN Oct-0410.329.007110.08609.0373 SPRING_HILL_FLDN Nov-049.948.57699.64698.6078 SPRING_HILL_FLDN Dec-049.298.20929.26448.2402 SPRING_HILL_FLDN Jan-058.657.87808.91567.9087 SPRING_HILL_FLDN Feb-058.127.62178.64157.6519 SPRING_HILL_FLDN Mar-057.787.38008.37667.4095 SPRING_HILL_FLDN Apr-057.387.21568.19227.2443 SPRING_HILL_FLDN May-056.957.23198.19037.2597 SPRING_HILL_FLDN Jun-056.767.39988.34507.4269 SPRING_HILL_FLDN Jul-056.997.68538.61907.7119 SPRING_HILL_FLDN Aug-057.427.67938.60527.7059 SPRING_HILL_FLDN Sep-057.407.45508.38047.4817 SPRING_HILL_FLDN Oct-057.127.57128.49427.5977 SPRING_HILL_FLDN Nov-056.847.45638.38197.4828 SPRING_HILL_FLDN Dec-056.677.43298.35877.4594 SPRING_HILL_FLDN Jan-066.377.16458.08777.1909 SPRING_HILL_FLDN Feb-066.227.21678.13407.2428 SPRING_HILL_FLDN Mar-066.176.95697.86706.9827 SPRING_HILL_FLDN Apr-065.986.68957.58676.7149 SPRING_HILL_FLDN May-065.816.43917.31956.4640 WEEKI_WACHEE_DEEP SS5.405.53766.24715.5729 WEEKI_WACHEE_DEEP Jun-04 5.065.43766.15045.4737 WEEKI_WACHEE_DEEP Jul-04 5.015.60516.32305.6414 WEEKI_WACHEE_DEEP Aug-04 5.285.58846.31365.6246 WEEKI_WACHEE_DEEP Sep-04 6.426.35567.07136.3920 WEEKI_WACHEE_DEEP Oct-04 7.066.10206.83986.1413 WEEKI_WACHEE_DEEP Nov-046.785.77406.49815.8124 WEEKI_WACHEE_DEEP Dec-046.445.50306.20815.5401 WEEKI_WACHEE_DEEP Jan-05 6.095.25735.94175.2929 WEEKI_WACHEE_DEEP Feb-05 5.785.07475.73945.1087 WEEKI_WACHEE_DEEP Mar-05 5.534.92835.57094.9606 212SAWYER_CB_4E_FLDN_REPL Mar-0522.4622.261222.569222.2749 SAWYER_CB_4E_FLDN_REPL Apr-0522.4822.017622.332222.0315 SAWYER_CB_4E_FLDN_REPL May-0522.3121.769722.090221.7839 SAWYER_CB_4E_FLDN_REPL Jun-0522.3522.020722.332822.0345 SAWYER_CB_4E_FLDN_REPL Jul-0522.9022.326522.630722.3401 SAWYER_CB_4E_FLDN_REPL Aug-0523.0722.263222.570522.2768 SAWYER_CB_4E_FLDN_REPL Sep-0522.4022.018222.331622.0319 SAWYER_CB_4E_FLDN_REPL Oct-0522.1922.081422.394222.0951 SAWYER_CB_4E_FLDN_REPL Nov-0521.8521.850922.167021.8647 SAWYER_CB_4E_FLDN_REPL Dec-0521.6221.738122.054921.7518 SAWYER_CB_4E_FLDN_REPL Jan-0621.4321.478121.798821.4919 SAWYER_CB_4E_FLDN_REPL Feb-0621.3621.460021.779221.4737 SAWYER_CB_4E_FLDN_REPL Mar-0621.2021.218121.540021.2319 SAWYER_CB_4E_FLDN_REPL Apr-0620.8720.947421.272520.9614 SAWYER_CB_4E_FLDN_REPL May-0620.5720.659320.987520.6734 SPRING_HILL_FLDN SS7.128.09119.11818.1202 SPRING_HILL_FLDN Jun-046.938.13449.17418.1636 SPRING_HILL_FLDN Jul-046.708.38189.43568.4110 SPRING_HILL_FLDN Aug-046.948.50429.57558.5335 SPRING_HILL_FLDN Sep-047.729.381110.45159.4099 SPRING_HILL_FLDN Oct-0410.329.007110.08609.0373 SPRING_HILL_FLDN Nov-049.948.57699.64698.6078 SPRING_HILL_FLDN Dec-049.298.20929.26448.2402 SPRING_HILL_FLDN Jan-058.657.87808.91567.9087 SPRING_HILL_FLDN Feb-058.127.62178.64157.6519 SPRING_HILL_FLDN Mar-057.787.38008.37667.4095 SPRING_HILL_FLDN Apr-057.387.21568.19227.2443 SPRING_HILL_FLDN May-056.957.23198.19037.2597 SPRING_HILL_FLDN Jun-056.767.39988.34507.4269 SPRING_HILL_FLDN Jul-056.997.68538.61907.7119 SPRING_HILL_FLDN Aug-057.427.67938.60527.7059 SPRING_HILL_FLDN Sep-057.407.45508.38047.4817 SPRING_HILL_FLDN Oct-057.127.57128.49427.5977 SPRING_HILL_FLDN Nov-056.847.45638.38197.4828 SPRING_HILL_FLDN Dec-056.677.43298.35877.4594 SPRING_HILL_FLDN Jan-066.377.16458.08777.1909 SPRING_HILL_FLDN Feb-066.227.21678.13407.2428 SPRING_HILL_FLDN Mar-066.176.95697.86706.9827 SPRING_HILL_FLDN Apr-065.986.68957.58676.7149 SPRING_HILL_FLDN May-065.816.43917.31956.4640 WEEKI_WACHEE_DEEP SS5.405.53766.24715.5729 WEEKI_WACHEE_DEEP Jun-04 5.065.43766.15045.4737 WEEKI_WACHEE_DEEP Jul-04 5.015.60516.32305.6414 WEEKI_WACHEE_DEEP Aug-04 5.285.58846.31365.6246 WEEKI_WACHEE_DEEP Sep-04 6.426.35567.07136.3920 WEEKI_WACHEE_DEEP Oct-04 7.066.10206.83986.1413 WEEKI_WACHEE_DEEP Nov-046.785.77406.49815.8124 WEEKI_WACHEE_DEEP Dec-046.445.50306.20815.5401 WEEKI_WACHEE_DEEP Jan-05 6.095.25735.94175.2929 WEEKI_WACHEE_DEEP Feb-05 5.785.07475.73945.1087 WEEKI_WACHEE_DEEP Mar-05 5.534.92835.57094.9606

PAGE 228

Appendix E (Continued) Appendix E (Continued) 213WEEKI_WACHEE_DEEP Apr-05 5.334.81805.44444.8489 WEEKI_WACHEE_DEEP May-05 5.204.84945.46224.8791 WEEKI_WACHEE_DEEP Jun-05 5.145.04235.65485.0720 WEEKI_WACHEE_DEEP Jul-05 5.595.28905.90135.3190 WEEKI_WACHEE_DEEP Aug-05 5.695.25805.86665.2885 WEEKI_WACHEE_DEEP Sep-05 5.585.05365.65865.0840 WEEKI_WACHEE_DEEP Oct-05 5.405.15545.75845.1856 WEEKI_WACHEE_DEEP Nov-055.205.07085.67965.1013 WEEKI_WACHEE_DEEP Dec-055.035.03185.63955.0623 WEEKI_WACHEE_DEEP Jan-06 4.874.82695.43034.8568 WEEKI_WACHEE_DEEP Feb-06 4.914.87785.47564.9071 WEEKI_WACHEE_DEEP Mar-06 4.754.67215.26184.7006 WEEKI_WACHEE_DEEP Apr-06 4.614.46875.04354.4961 WEEKI_WACHEE_DEEP May-06 4.254.29024.84834.3166 WEEKI_WELL_11_DEEP SS15.1914.939115.4 85114.9605 WEEKI_WELL_11_DEEP Jun-0415.0615.109115 .653915.1305 WEEKI_WELL_11_DEEP Jul-0414 .6415.154515 .695615.1757 WEEKI_WELL_11_DEEP Aug-0414.4015.589116 .126715.6103 WEEKI_WELL_11_DEEP Sep-0415.3816.543817 .072516.5647 WEEKI_WELL_11_DEEP Oct-0416.9016.445716 .973816.4665 WEEKI_WELL_11_DEEP Nov-0416 .9916.193116 .726716.2141 WEEKI_WELL_11_DEEP Dec-0416 .8815.857216 .398215.8784 WEEKI_WELL_11_DEEP Jan-0516.7215.552916 .101615.5744 WEEKI_WELL_11_DEEP Feb-0516 .3315.264515 .819715.2862 WEEKI_WELL_11_DEEP Mar-0516 .0215.041615 .602315.0636 WEEKI_WELL_11_DEEP Apr-0515 .4914.764315 .329414.7865 WEEKI_WELL_11_DEEP May-0515.3314.564915 .133114.5872 WEEKI_WELL_11_DEEP Jun-0515.6914.748615 .315114.7708 WEEKI_WELL_11_DEEP Jul-0516 .1315.114315 .675015.1363 WEEKI_WELL_11_DEEP Aug-0516.1715.147915 .703915.1697 WEEKI_WELL_11_DEEP Sep-0515.4814.837015 .391814.8587 WEEKI_WELL_11_DEEP Oct-0515.1914.791115 .343614.8127 WEEKI_WELL_11_DEEP Nov-0514 .8314.568015 .119114.5895 WEEKI_WELL_11_DEEP Dec-0514 .5514.528715 .077514.5500 WEEKI_WELL_11_DEEP Jan-0614.0114.258514 .806714.2798 WEEKI_WELL_11_DEEP Feb-0613 .7514.266214 .812614.2874 WEEKI_WELL_11_DEEP Mar-0613 .5114.015914 .561314.0371 WEEKI_WELL_11_DEEP Apr-0613 .2413.733914 .279413.7550 WEEKI_WELL_11_DEEP May-0612.9813.445613 .991313.4667 WOLFE_WELL_CBAR_1W_DEEP SS15.8615.512315.992815.5444 WOLFE_WELL_CBAR_1W_DEEPJun-0416.0415.514215.994815.5463 WOLFE_WELL_CBAR_1W_DEEPJul-0416.4515.722916.203515.7550 WOLFE_WELL_CBAR_1W_DEEPAug-0416.7915.845916.326415.8780 WOLFE_WELL_CBAR_1W_DEEPSep-0417.4116.243816.724216.2760 WOLFE_WELL_CBAR_1W_DEEPOct-0417.7416.372416.852416.4045 WOLFE_WELL_CBAR_1W_DEEPNov-0417.4716.224016.703916.2562 WOLFE_WELL_CBAR_1W_DEEPDec-0417.2816.045216.525416.0775 WOLFE_WELL_CBAR_1W_DEEPJan-0517.0415.851416.332015.8837 WOLFE_WELL_CBAR_1W_DEEPFeb-0516.7615.666516.147715.6989 WOLFE_WELL_CBAR_1W_DEEPMar-0516.6915.505415.987415.5378 WOLFE_WELL_CBAR_1W_DEEPApr-0516.5315.337815.820715.3703 213WEEKI_WACHEE_DEEP Apr-05 5.334.81805.44444.8489 WEEKI_WACHEE_DEEP May-05 5.204.84945.46224.8791 WEEKI_WACHEE_DEEP Jun-05 5.145.04235.65485.0720 WEEKI_WACHEE_DEEP Jul-05 5.595.28905.90135.3190 WEEKI_WACHEE_DEEP Aug-05 5.695.25805.86665.2885 WEEKI_WACHEE_DEEP Sep-05 5.585.05365.65865.0840 WEEKI_WACHEE_DEEP Oct-05 5.405.15545.75845.1856 WEEKI_WACHEE_DEEP Nov-055.205.07085.67965.1013 WEEKI_WACHEE_DEEP Dec-055.035.03185.63955.0623 WEEKI_WACHEE_DEEP Jan-06 4.874.82695.43034.8568 WEEKI_WACHEE_DEEP Feb-06 4.914.87785.47564.9071 WEEKI_WACHEE_DEEP Mar-06 4.754.67215.26184.7006 WEEKI_WACHEE_DEEP Apr-06 4.614.46875.04354.4961 WEEKI_WACHEE_DEEP May-06 4.254.29024.84834.3166 WEEKI_WELL_11_DEEP SS15.1914.939115.4 85114.9605 WEEKI_WELL_11_DEEP Jun-0415.0615.109115 .653915.1305 WEEKI_WELL_11_DEEP Jul-0414 .6415.154515 .695615.1757 WEEKI_WELL_11_DEEP Aug-0414.4015.589116 .126715.6103 WEEKI_WELL_11_DEEP Sep-0415.3816.543817 .072516.5647 WEEKI_WELL_11_DEEP Oct-0416.9016.445716 .973816.4665 WEEKI_WELL_11_DEEP Nov-0416 .9916.193116 .726716.2141 WEEKI_WELL_11_DEEP Dec-0416 .8815.857216 .398215.8784 WEEKI_WELL_11_DEEP Jan-0516.7215.552916 .101615.5744 WEEKI_WELL_11_DEEP Feb-0516 .3315.264515 .819715.2862 WEEKI_WELL_11_DEEP Mar-0516 .0215.041615 .602315.0636 WEEKI_WELL_11_DEEP Apr-0515 .4914.764315 .329414.7865 WEEKI_WELL_11_DEEP May-0515.3314.564915 .133114.5872 WEEKI_WELL_11_DEEP Jun-0515.6914.748615 .315114.7708 WEEKI_WELL_11_DEEP Jul-0516 .1315.114315 .675015.1363 WEEKI_WELL_11_DEEP Aug-0516.1715.147915 .703915.1697 WEEKI_WELL_11_DEEP Sep-0515.4814.837015 .391814.8587 WEEKI_WELL_11_DEEP Oct-0515.1914.791115 .343614.8127 WEEKI_WELL_11_DEEP Nov-0514 .8314.568015 .119114.5895 WEEKI_WELL_11_DEEP Dec-0514 .5514.528715 .077514.5500 WEEKI_WELL_11_DEEP Jan-0614.0114.258514 .806714.2798 WEEKI_WELL_11_DEEP Feb-0613 .7514.266214 .812614.2874 WEEKI_WELL_11_DEEP Mar-0613 .5114.015914 .561314.0371 WEEKI_WELL_11_DEEP Apr-0613 .2413.733914 .279413.7550 WEEKI_WELL_11_DEEP May-0612.9813.445613 .991313.4667 WOLFE_WELL_CBAR_1W_DEEP SS15.8615.512315.992815.5444 WOLFE_WELL_CBAR_1W_DEEPJun-0416.0415.514215.994815.5463 WOLFE_WELL_CBAR_1W_DEEPJul-0416.4515.722916.203515.7550 WOLFE_WELL_CBAR_1W_DEEPAug-0416.7915.845916.326415.8780 WOLFE_WELL_CBAR_1W_DEEPSep-0417.4116.243816.724216.2760 WOLFE_WELL_CBAR_1W_DEEPOct-0417.7416.372416.852416.4045 WOLFE_WELL_CBAR_1W_DEEPNov-0417.4716.224016.703916.2562 WOLFE_WELL_CBAR_1W_DEEPDec-0417.2816.045216.525416.0775 WOLFE_WELL_CBAR_1W_DEEPJan-0517.0415.851416.332015.8837 WOLFE_WELL_CBAR_1W_DEEPFeb-0516.7615.666516.147715.6989 WOLFE_WELL_CBAR_1W_DEEPMar-0516.6915.505415.987415.5378 WOLFE_WELL_CBAR_1W_DEEPApr-0516.5315.337815.820715.3703

PAGE 229

Appendix E (Continued) 214WOLFE_WELL_CBAR_1W_DEEPMay-05 WOLFE_WELL_CBAR_1W_DEEPJun-05 WOLFE_WELL_CBAR_1W_DEEPJul-05 WOLFE_WELL_CBAR_1W_DEEPAug-05 WOLFE_WELL_CBAR_1W_DEEPSep-05 WOLFE_WELL_CBAR_1W_DEEPOct-05 WOLFE_WELL_CBAR_1W_DEEPNov-05 WOLFE_WELL_CBAR_1W_DEEPDec-05 WOLFE_WELL_CBAR_1W_DEEPJan-06 WOLFE_WELL_CBAR_1W_DEEPFeb-06 WOLFE_WELL_CBAR_1W_DEEPMar-06 WOLFE_WELL_CBAR_1W_DEEPApr-06 WOLFE_WELL_CBAR_1W_DEEPMay-06 WW_F SS WW_F Jun-04 WW_F Jul-04 WW_F Aug-04 WW_F Sep-04 WW_F Oct-04 WW_F Nov-04 WW_F Dec-04 WW_F Jan-05 WW_F Feb-05 WW_F Mar-05 WW_F Apr-05 WW_F May-05 WW_F Jun-05 WW_F Jul-05 WW_F Aug-05 WW_F Sep-05 WW_F Oct-05 WW_F Nov-05 WW_F Dec-05 WW_F Jan-06 WW_F Feb-06 WW_F Mar-06 WW_F Apr-06 WW_F May-06 *SS indicates steady-state 16.3415.171115.655015.2036 16.1415.131215.616015.1637 16.2215.255215.740615.2877 16.2815.351515.837315.3840 16.0315.261815.747915.2941 15.8615.203615.690015.2359 15.6415.110615.597215.1429 15.5115.046915.533515.0791 15.3014.899815.386514.9319 15.3314.835615.322214.8676 15.0714.706315.192914.7383 14.8114.505114.991614.5370 14.4214.294414.781014.3263 3.663.90683.95803.9135 3.453.80543.85003.8136 3.493.85903.89653.8662 3.583.86413.90483.8700 4.194.31634.32994.3233 4.284.32474.35604.3356 4.154.16014.20434.1682 3.984.00234.05344.0089 3.803.85493.91083.8604 3.693.75373.81403.7578 3.653.67973.74293.6824 3.603.61213.68103.6136 3.573.59923.66433.5999 3.553.68973.73993.6905 3.803.86313.90683.8647 3.793.87173.91593.8742 3.723.76813.81963.7705 3.663.76953.81603.7717 3.553.72073.76873.7237 3.483.67923.72443.6828 3.423.58323.63523.5860 3.493.58423.63223.5858 3.383.50363.55903.5049 3.273.38183.44103.3826 3.133.26583.32693.2665

PAGE 230

Appendix F 215

PAGE 231

Appendix F (Continued) 216

PAGE 232

Appendix F (Continued) 217

PAGE 233

Appendix F (Continued) 218

PAGE 234

Appendix F (Continued) 219

PAGE 235

Appendix F (Continued) 220

PAGE 236

Appendix F (Continued) 221

PAGE 237

Appendix F (Continued) 222

PAGE 238

Appendix F (Continued) 223

PAGE 239

Appendix F (Continued) 224

PAGE 240

Appendix F (Continued) 225

PAGE 241

Appendix F (Continued) 226

PAGE 242

Appendix F (Continued) 227

PAGE 243

Appendix F (Continued) 228

PAGE 244

Appendix F (Continued) 229

PAGE 245

Appendix F (Continued) 230

PAGE 246

ABOUT THE AUTHOR Melissa Hill earned her bachelors degree in Geology from St. Marys University in San Antonio, Texas. She received her masters degree in Geology from the University of Te xas at San Antonio while working as a scientist at Southwest Research Instit ute. In 2002, Melissa mov ed to Tampa, Florida to study springs of the dual-permeability Upper Floridan aquifer. She earned her doctorate degree in geology from the Univer sity of South Florida while working as a hydrologist at the Southwest Flori da Water Management District. She is a licensed professional geologist in both Texas and Florida.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200397Ka 4500
controlfield tag 001 002000477
005 20090422141105.0
007 cr mnu|||uuuuu
008 090422s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002497
035
(OCoLC)318990617
040
FHM
c FHM
049
FHMM
090
QE26.2 (Online)
1 100
Hill, Melissa Estelle.
3 245
An evaluation of conduit conceptualizations and model performance
h [electronic resource] /
by Melissa Estelle Hill.
260
[Tampa, Fla] :
b University of South Florida,
2008.
500
Title from PDF of title page.
Document formatted into pages; contains 230 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2008.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
520
ABSTRACT: The karst research community has known that traditional numerical groundwater flow codes ignore the non-Darcian, dual-permeability components of flow that can occur in karst aquifers. In this study, the potential limitations of using such tools are quantified by evaluating the relative performances of 3 groundwater flow models at a test-site near Weeki Wachee, Florida, in the dual-permeability Upper Floridan aquifer. MODFLOW-2005 and MODFLOW-2005 Conduit Flow Process (CFP), a Darcian/non-Darcian, dual-permeability groundwater flow code recently developed by the U.S. Geological Survey, are used in this study. A monitoring program consisting of discharge measurements and high frequency data from 2 springs and monitoring wells penetrating the matrix and conduit networks of a karst aquifer was initiated to characterize the test-site and constrain new parameters introduced with MODFLOW-2005 CFP.The monitoring program spanned conditions prior to, during, and following convective and tropical storm activity, and a drought. Analytical estimates for Reynolds numbers, ranging from 10 to 10, suggest that turbulent flow occurs in portions of the underlying conduit network. The direction and magnitude of fluid exchange observed between the matrix and conduit network indicate the conduit network underlying the test-site drains the matrix. Head differences and observed responses in monitoring wells penetrating the matrix and conduit network indicate that the hydraulic conductivities between the 2 networks do not significantly differ from each other. A conceptual model for the spatial distribution of preferential flow pathways using multiple data types, including shallow recession limbs observed in discharge hydrographs indicate a slow responding aquifer with a high storage capacity, and a poorly integrated conduit drainage network with little to no point recharge.Model performances were evaluated by comparing observed hydrographs for discharge and monitoring wells penetrating the matrix and conduit network following convective and tropical storm events, and drought conditions, to simulated values from transient simulations. Model statistics for 32 target wells and sensitivity analysis were included in the evaluation. The dual-permeability model using the MODFLOW-2005 CFP Mode 1 displayed the highest performance with improved matches ranging from 12 to 40% between simulated and observed discharges relative to the laminar and laminar/turbulent equivalent-continuum models.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Mark T. Stewart, Ph.D.
653
Dual-permeability
Conduit-matrix fluid exchange
Non-Darcian flow
Karst hydrogeology
Model parameters
0 690
Dissertations, Academic
z USF
x Geology
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2497