Groundwater Availability Model for the Hill Country Portion of the Trinity Aquifer System, Texas


previous item | next item

Citation
Groundwater Availability Model for the Hill Country Portion of the Trinity Aquifer System, Texas

Material Information

Title:
Groundwater Availability Model for the Hill Country Portion of the Trinity Aquifer System, Texas
Creator:
Jones, Ian C.
Anaya, Roberto
Wade, Shirley
Publication Date:
Language:
English

Subjects

Genre:
serial ( sobekcm )

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
This object is protected by copyright, and is made available here for research and educational purposes. Permission to reuse, publish, or reproduce the object beyond the bounds of Fair Use or other exemptions to copyright law must be obtained from the copyright holder.
Resource Identifier:
K26-05185 ( USFLDC: LOCAL DOI )
k26.5185 ( USFLDC: LOCAL Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Groundwater Availability Model for the Hill Country Portion of the Trinity Aquifer System , Texas By Ian C. Jones, Ph.D., P.G. Roberto Anaya, P.G. Shirley Wade, Ph.D., P.G.

PAGE 2

1

PAGE 3

2

PAGE 4

3 1.0 Executive Summary Mace and others (2000) constructed a groundwater availability model simulating groundwater flow through th e Hill Country portion of the Trinity Aquifer System as a groundwater resource management tool. The purpose of this report is to document update s to th is earlier model . Th e model is being updated by : (1) adding the Lower Trinity Aquifer as an additional layer to the model , (2) revising the spatial distribution of parameters, such as recharge and pumping, and (3) calibrating to steady state water level and river discharge conditions for 1980 and historical transient water level and discharge conditions for 1981 through 1997. The calibrated model can be used to predict future water level changes that may result from various projected pumping rates and/or changes in climatic conditions. Our c onceptual model subdivide s the Hill Country portion of the Trinity Aquifer System into three main components: the Upper, Middle, and Lower Trinity aquifers. T he Upper Trinity Aquifer is composed of the upper member of the Glen Rose Limestone . T he Middle Tri nity Aquifer is composed of the l ower member of the Glen Rose Limestone , Hensell Sand , and Cow Creek Limestone. T he Lower Trinity Aquifer is composed of the Sycamore Sand, Sligo Formation, and Hosston Formation. The Middle and Lower Trinity aquifers are se parated by the Hammett Shale which acts as a confining unit and is not explicitly included in the model. The model study area also includes easternmost parts of the Edwards Trinity (Plateau) Aquifer. Recharge in the updated model is a combination of infilt ration of precipitation that falls on the aquifer outcrop and infiltration from losing intermittent streams within the model area . Estimates of recharge due to infiltration of precipitation in this updated model vary spatially and are equivalent to 3.5 to 5 percent of average annual precipitation . T he highest of these recharge rates coincide with the Balcones Fault Zone . In addition to recharge from precipitation, there is also r echarge from streamflow losses in the downstream parts of the Cibolo Creek wat ershed to the underlying aquifers of about 70,000 acre feet per year . Groundwater in the aquifer generally flows towards the south and east . The Hill Country portion of Trinity Aquifer System discharges naturally as baseflow to gaining streams , such as t he Guadalupe, Blanco, and Medina rivers, and as cross formational flow to the adjacent Edwards ( Balcones Fault Zone ) Aquifer . This cross formation al flow accounts for about 100,000 acre feet per year of discharge. Pumping discharge from the Hill Country portion of the Trinity Aquifer System increas ed over the period 1980 through 1997. This increas e in pumping is most apparent in Bexar, Hays, Kendall , and Kerr counties — counties adjacent to the two largest metropolitan areas in the region, San Antonio and A ustin. Some of these counties have experienced a doubling of pumping over this period. The updated model does a good job of reproducing observed water level fluctuations. Comparison of measured and simulated 1997 water levels indicates a mean absolute erro r of 57 feet, or approximately 5.3 percent of the range of measured water levels. This is a slight improvement over the original model. Overall, the updated model also does a good job of mimicking baseflow fluctuations . The ability of the model to simulate spring discharge varies widely. Simulating discharge to springs using a regionalscale model is often difficult because of spatial and temporal scale issues. Of seventeen springs, six display a good comparison between measured and simulated discharge val ues.

PAGE 5

4 The main improvements in the updated model over the original model are due to the addition of the Lower Trinity Aquifer to the model and the revised recharge distribution. The addition of the Lower Trinity Aquifer is important because the Lower Trinit y Aquifer is an increasingly important source of groundwater in the study area. The revision of the recharge distribution in the updated model along with associated changes in the hydraulic conductivity distribution takes into consideration the major contr ibution to recharge from Cibolo Creek and will result in better simulation of groundwater flow in Bexar and surrounding counties. 2.0 Introduction This report describes update s to the earlier developed groundwater availability model for the Hill Country portion of the Trinity Aquifer System by Mace and others ( 2000). Th ese update s include : (1) addition of the Lower Trinity Aquifer to the model , (2) revis ions to the model layers ’ structural geometry , and recharge, hydraulic conductivity, and pumping distr ibution, and (3) changes to the model calibration periods to bring the model in line with Texas Water Development Board groundwater availability modeling standards that were developed after the earlier model was constructed ( http://www.twdb.state.tx.us/gam/GAM_documents/ GAM_RFQ_Oct2005.pdf ) . In this report, we use the term Trinity Aquifer System. The term aquifer system has not previously been used in Texas Water Development Board publications but is often used by the United States Geological Survey, for example, the Edwards Trinity Aquifer System (Barker and others, 1994), where multiple aquifers are grouped together. In this case, the Hill Country portion of the Trinity Aquifer System is subdivided into the Upper, Middle and Lower Trinity aquifers. The Trinity Aquifer System is an important source of groundwater to municipalities, industries, and landowners in the Hill Country. Rapid population growth and recent droughts have increased inter est in the Trinity Aquifer System and have increased the need for quantitative tools to assist in the estimation of groundwater availability in the area. Many groundwater conservation districts and the groundwater management area in the region need to assess the impacts of groundwater pumping and drought on the groundwater resources of the area. Regional w ater planning g roups are required to plan for future water needs under drought conditions and are similarly interested in the groundwater availability of the Hill Country. Several studies have noted the vulnerability of the Hill Country portion of the Trinity Aquifer System to drought and increased pumping. Ashworth (1983) concluded that heavy pumping is resulting in rapid water level declines in certain ar eas and that continued growth would result in continued water level declines. Bluntzer (1992) , Simpson and others (1993) , and Kalaswad and Mills (2000) noted that intense pumping ha s resulted in water level declines, decreased well yields, increased potent ial for the encroachment of saline groundwater into the aquifer, and depletion of baseflow in nearby streams. Calibrated g roundwater flow model s are simplified mathematical representation s of groundwater flow systems that can be used to refine and confirm the conceptual understanding of a groundwater flow system . Once the model is successfully calibrated , it can be used as a

PAGE 6

5 quantitative tool to investigate the effects of pumping, drought, and different water management scenarios on the groundwater flow s ystem . In this study, we enhanced and re calibrated the threedimensional finite difference groundwater flow model for the Hill Country portion of the Trinity Aquifer System to improve our conceptual understanding of groundwater flow in the region; and dev elop a m anagement tool to support water planning efforts for re gional w ater planning g roups, groundwater c onservation districts, groundwater management areas, and r iver a uthorities in the study area. This report describes the construction and re calibratio n of the numerical model owing to the addition of the Lower Trinity Aquifer and revisions to recharge, hydraulic conductivity, and pumping distribution to the earlier model . Our general approach involved (1) revis ing the conceptual groundwater flow model, (2) organizing and distributing aquifer parameter s for the model, (3) calibrating a steady state model for 1980 water level conditions , and (4) calibrating a transient model for the period 1981 through 1997. This report describes the study area, previous work, hydrogeologic setting used to develop the conceptual model , and model calibration results . 3.0 Study Area The study area is located in the Hill Country of south central Texas and includes all or parts of Bandera, Bexar, Blanco, Comal, Gillespie, H ays, Kendall, Kerr, Kimble, Medina, Travis, and Uvalde counties ( Figure 301). Hydrologic boundaries define the extent of the study area. These boundaries include (1) major faults of the Balcones Fault Zone in the east and south, (2) presumed groundwater f low paths in the west, and (3) aquifer outcrops and/ or rivers in the north ( Figure 3 01). Because we selected groundwater flow paths to the west to assign a model boundary , the study area does not include the entire Hill Country area , such as parts of west ern Bandera and northeastern Uvalde counties , and includes the easternmost parts of the Edwards Trinity (Plateau) Aquifer System (Ashworth and Hopkins, 1995) in Bandera, Gillespie, Kendall, and Kerr counties ( Figure 302). The study area includes parts of three regional water planning areas: the Lower Colorado Region (Region K), the South Central Texas Regi on (Region L), and the Plateau Region (Region J) ( Figure 3 03). The study area includes all or parts of several groundwater c onservation districts inclu ding: Bandera County River Authority and Ground W ater District, Blanco Pedernales Groundwater Conservation District, Cow Creek Groundwater Conservation District, Edwards Aquifer Authority, Hays Trinity Groundwater Conservation District, Headwaters Groundwa ter Conservation District, Hill Country Underground Water Conservation District, Kimble County Groundwater Conservation District, Medina County Groundwater Conservation District, Trinity Glen Rose Groundwater Conservation District, and Uvalde County Underground Water Conservation District ( Figure 304). The study area approximately coincides with Groundwater Management Area 9 (Figure 3 05). The study area also extends over four major river basins, the Colorado, Guadalupe, San Antonio and Nueces rivers, and five river a uthorities: the Lower Colorado River Authority ( that includes Blanco and Travis counties in the study area), the Guadalupe Blanco River Authority ( that includes Comal, Hays, and Kendall counties in the study area), the Upper Guadalupe River Aut hority ( that includes Kerr County), the Nueces River

PAGE 7

6 Authority (that includes Bandera, Medina, and Uvalde counties ), and the San Antonio River Authority that includes Bexar County in the study area (Figure 3 06). 3.1 Physiography and Climate The study ar ea is located along the southeastern margin of the Edwards Plateau region in a region commonly referred to as the Texas Hill Country ( Figure 307) . The Texas Hill Country is also known as the Balcones Canyonlands subregion, a deeply dissected terrain form ed by the head ward erosion of major streams between the Edwards P lateau and the Balcones Escarpment (Thornbury, 1965; Riskind and Diamond, 1986) . Land surface elevations across the study area range from 2,400 feet above sea level in the west to about 600 feet along eastern margin of the study area ( Figure 308). The more massive and resistant carbonate members of the Edwards Group form the nearly flat uplands of the Edwards Plateau in the west and the topographic divides in the central portion of the study area ( Figure 307) . The differential weathering of alternating beds of limestone and dolostone with soft marl and shale in the upper member of the Glen Rose Limestone form the characteristic stair step topography of the Balcones Canyonlands. In general, t he upper member of the Glen Rose Limestone is much less resistant to erosion than the overlying Edwards Group caprock. The study area is characterized by a sub humid to semi arid climate. A gradual decrease in average annual precipitation occurs from east to west (35 inches to 25 inches) due to increasing distance from the Gulf of Mexico (Carr, 1967; Figure 309). Additionally, local precipitation is highest in the central part of the study area and decreases to the north and south. Historical annual precipitation varies from less than 10 inches to more than 60 inches ( Figure 3 10). Precipitation has a bimodal distribution during the year with most of the rainfall occurring in the spring and fall ( Figure 311) . During the spring, weak cold fronts begin to st all and interact with warm moist air from the Gulf of Mexico. During the summer, sparse rainfall is due to infrequent convectional thunderstorms. In early fall, rainfall is due to more frequent convectional thunderstorms and occasional tropical cyclones that make landfall along the Texas coast. Rainfall frequency continues to increase in late fall as co ld fronts once again begin to strengthen and interact with the warm moist air masses of the Gulf of Mexico. The average annual maximum temperature ranges fro m 76F in the west to 78F in the east and south ( Figure 3 12 ). Average monthly temperatures range from about 60F during winter months to about 95F during summer months (Larkin and Bomar, 1983) . The average annual (19 50 to 1979) gross lake surface evapor ation is more than twice the average annual precipitation and ranges from 6 3 inches in the east to 68 inches in the west ( Figure 3 13 ). Seasonally , average monthly gross lake surface evaporation varies from about 2.5 inches during winter months to more tha n 9 inches during summer months (Larkin and Bomar, 1983). 3.2 Geology Lower Cretaceous rocks of the Trinity Group that compose the Hill Country portion of the Trinity Aquifer System overlie unconformably Paleozoic rocks in the study area ( Figure 314). T he se Lower Cretaceous rocks consist of ( from oldest to youngest ) , the Hosston Formation

PAGE 8

7 (known as Sycamore Sand where it outcrops at the surface) , Sligo Formation, Hammett Shale, Cow Creek Limestone, Hensell Sand, lo wer and upper member s of the Glen Rose L imestone , and the Fort Terrett and Segovia Formations of the Edwards Group ( Figure 314). The Trinity Group sediments are locally covered by Quaternary alluvium along streams and rivers and capped by Edwards Group sediments in the west . The stratigraphic units of the Hill Country portion of the Trinity Aquifer System were deposited during a period of rifting and subsidence in the ancestral Gulf of Mexico (Barker and others, 1994). These u nits were deposited on the landward margin of a broad continental she lf under shallow marine conditions. The Llano uplift was a dominant structural high, forming islands of Precambrian metamorphic and igneous rock and Paleozoic sedimentary rock that were sources of terri genous sediment occurring in the Trinity G roup (Figure 3 15) . The Hosston Formation is dominantly composed of siliciclastic siltstone and sandstone in updip areas and dolomitic mudstone and grainstone downdip derived from the Llano Uplift (Barker and others, 1994). This formation, which is up to 900 feet thick, grades upward into the Sligo Formation and where it is exposed at the surface is known as the Sycamore Sand. The Sycamore Sand is composed of quartz sand and gravel up to 50 feet thick (Barker and others, 1994). The Sycamore Sand also contains some feldspar and dolomite derived from the Llano Uplift. The Sligo Formation is composed of up to 250 feet of evaporites, limestone and dolostone (Barker and others, 1994). The evaporites were deposited in a supratidal environment while the limestone and dolostone were deposited in an intertidal environment. In the updip regions , the Sligo Formation sediments display greater a contribution of terrestrial sediment s from the Llano Uplift (Barker and others, 1994). The Hammett Shale is highly burrowed and is made up of mix ed clay, silt, and calcareous mud up to 130 feet thick (Barker and others, 1994). This stratigraphic unit inter finger s vertically with the overlying Cow Creek Limestone. The Cow Creek Limestone is a beach deposit on the southern flank of the Llano Uplift, up to 90 feet thick (Barker and others, 1994). The lower part of the Cow Creek Limestone is composed of fine to coarse grained calcareous sandstone. The middle part of the Cow Creek limestone is composed of silty calcareous sandstone, and the upper is composed of coarse grained fossiliferous calcareous sandstone with poorly sorted quartz grains and chert pebbles. The Hensell Sand crops out in the northern part of the study area in Gillespie County (Figure 316). The Hensell Sand is composed of poorly cemented clay, quartz and calcareous sand, and chert and dolomite gravel up to 200 feet thick(Barker and others, 1994) . The gravel beds occur at the base of this stratigraphic unit. The shallow marine deposits of the Bexar Shale Member of the Pearsall Formation are the downdip equivalent of the Hensell Sand (Barker and others, 1994). The Glen Rose Limestone is composed of sandy fossiliferous limestone and dolostone that is characterized by beds of calcareous marl , clay, and shale and includes thin la yers of gypsum and anhydrite (Barker and others, 1994). The Glen Rose Limestone has a maximum thickness of 1,500 feet. The lower member of the Glen Rose Limestone is composed of medium thick beds of limestone, dolostone and fossiliferous dolomitic limestone (Barker and others, 1994). The Glen Rose Limestone was deposited in a shallow marine to intertidal environment and grades northward into the terrestrial Hensell Sand. The upper member of the Glen Rose Limestone is exposed at land surface in most of the s tudy area except where it is (1) removed by erosion

PAGE 9

8 exposing the lower member of the Glen Rose Limestone , and (2) overlain by the Edwards Group in the Edwards Plateau to the west and in the Balcones Fault Zone to the south and east (Figure 316). The upper member of the Glen Rose Limestone is characterized by a thin to medium bedded sequence of alternating nonresistant marl and resistant limestone and dolostone. The alternating layers of resistant and nonresistant rock results in uneven erosion that produc es the “stair step” topography characteristic of much of the Hill Country. The basal parts of the Hosston Formation, the Sycamore Sand, and up dip parts of the Hensell Sand are mostly sand y and contain some of the most permeable sediments in the Hill Count ry portion of the Trinity Aquifer System (Barker and others, 1994). The Cow Creek Limestone is highly permeable in the outcrop due to carbonate dissolution and preservation of the pores but has relatively low permeability in the subsurface due to precipita tion of calcite cements (Barker and others, 1994). Similarly, the lower member of the Glen Rose Limestone is more permeable in the outcrop than at depth (Barker and others, 1994). The Sligo Formation may yield small to large quantities of water (Ashworth, 1983). The Lower Trinity Aquifer is not exposed at land surface with in the study area and exists only in the southern half of the study area (Figure s 3 14 and 316). The study area is completely underlain by sediments of the Middle Trinity A quifer. The Upper Trinity Aquifer exists in most of the study area except where it has been removed by erosion along and near the lower reaches of the Pedernales, Blanco, Guadalupe, Cibolo, and Medina rivers (Figure 316). In the western part of the study area, the Fort Terrett and Segovia f ormations of the Edwards Group (Figure 316) cap the Trinity Aquifer sediments. The Edwards Group may produce large amounts of water where it is saturated and has high transmissivity. The Llano Uplift is a regional dome formed by a mas sive Precambrian granitic pluton (Figure 315). The Llano Uplift remained a structural high throughout the Ouachita orogeny that folded and uplifted the Paleozoic rocks of this area and provided a source of sediments for terrigenous and near shore facies o f the Trinity Group (Ashworth, 1983; Barker and others, 1994). The San Marcos Arch is a broad anticlinal (upward folded ridge) extension of the Llano Uplift with a southeast plunging axis. The San Marcos Arch extends through central Blanco and southwest Ha ys counties (Ashworth, 1983) (Figure 315). This arch contributed to the formation of a carbonate platform with thinning sediments along the anticlinal axis. The Balcones Fault Zone is a northeast southwest trending system of highangle normal faults with downthrown blocks towards the Gulf of Mexico (Figure 315). The faulting occurred along the subsurface axis of the Ouachita fold belt as a result of extensional forces created by the subsidence of basin sediments in the Gulf of Mexico during the Tertiary Period. The last episode of movement in the fault zone is thought to have occurred in the late Early Miocene, approximately 15 million years ago (Young, 1972). The Balcones Fault Zone is a structural feature that laterally juxtaposes Trinity Group sediments against Edwards Group sediments of the Edwards (Balcones Fault Zone) Aquifer (Figures 3 15 and 317). The structural geometry of Lower Cretaceous sediments in the study area are characterized by (1) a southeast regional dip, (2) an uneven base of the T rinity Group, and (3) the occurrence of the San Marcos A rch in the southeast, Llano Uplift to the north, and Balcones Fault Zone to the south and east ( Figures 3 15 and 317). Both Trinity Group and Edwards Group sediments have a regional dip to the south and southeast. The dip increases from a rate of about 10 to 15 feet per mile near the Llano Uplift to about 100 feet per mile near the Balcones Fault Zone (Ashworth, 1983). These Lower Cretaceous sediments may be described as a series of stacked wedges that

PAGE 10

9 pinch out against the Llano Uplift and thicken downdip towards the Gulf of Mexico ( Figure 3 17). At the base of the Trinity Group sediments, underlying Paleozoic rocks have been moderately folded, uplifted, and eroded to form an unconformable surface upon which the Trinity Group sediments were deposited ( Figure 3 17). Along the northern margin of the study area, the Middle and Upper Trinity sediments directly overlay Paleozoic and Precambrian rocks (Figure 3 17). 4.0 Previous Work The Texas Water Development Board and the United States Geological Survey have conducted a number of hydrogeologic studies in the Hill Country area. Ashworth (1983), Bluntzer (1992), and Barker and others (1994) provide a thorough review of much of the previous geologic and h ydrogeologic work done in the area. A regional numerical groundwater flow model was developed and published for the area by the United States Geological Survey (Kuniansky and Holligan, 1994). Besides the Trinity Aquifer in the Hill Country, th is United Sta tes Geological Survey model include s the Edwards Trinity (Plateau) and Edwards ( Balcones Fault Zone) aquifers and extends almost 400 mile s across the s tate (Figure 4 01) . The purpose for the United States Geological Survey model was to better understand and describe the regional groundwater flow system. Using the model, Kuniansky and Holligan (1994) defined transmissivity ranges, estimated total flow through and recharge to the aquifer system , and simulated groundwater flow from the Trinity Aquifer into the Edwards ( Balcones Fault Zone ) Aquifer . The two dimensional, finite element, steady state model was developed as the simplest approximation of the regional flow system. The United States Geological Survey model is inappropriate for regional water planning b ecause: (1) it does not simulate water level changes with time , and (2) it simulates all of aquifers in the study area as a single layer . Subsequently, Anaya and Jones (200 9) developed a transient finite difference model covering a study area similar to the model by Kuniansky and Holligan (1994). The model by Anaya and Jones (2009) simulates the Trinity Aquifer System as a single layer (Figure 4-01). T he Texas Water Development Board developed a regional transient groundwater flow model for the Hill Countr y area of the Trinity Aquifer (Mace and others, 2000) (Figure 4 01) . They calibrated this mode l to 1975 steady state conditions and 1996 through 1997 transient conditions (Mace and others, 2000). This model simulated groundwater flow through the Edwards Gr oup and the Upper and Middle Trinity aquifer s . Th is update d model include s the Lower Trinity Aquifer previously excluded from the model by Mace and others (2000). 5.0 Hydro geo logic Setting The hydrogeologic setting describes the aquifer , hydrologic feat ures , and hydraulic properties that influence groundwater flow in the aquifer. We based the hydrogeologic setting for the Hill Country portion of the Trinity Aquifer System on previous work ( for example, Ashworth, 1983; Bluntzer, 1992; Barker and others, 1994; Kuniansky and Holligan, 1994) and additional studies

PAGE 11

10 we conducted in support of the modeling effort (Mace and others, 2000) . These additional studies included assembling structure maps, developing water level maps and hydrographs, estimating baseflow to streams, investigating recharge rates, conducting aquifer tests, and assembling pumping information. 5.1 Hydrostratigraphy The Hill Country portion of the Trinity Aquifer System is comprised of sediments of the Trinity Group and is divided into lower, middle, and upper aquifers ( Figure 314) based on hydraulic characteristics of the sediments (Barker and others, 1994). The Lower Trinity Aquifer consists of the Hosston (and the Sycamore Sand in outcrop) and Sligo Formations; the Middle Trinity Aquifer c onsists of the Cow Creek Limestone, Hensell Sand, and the lower member of the Glen Rose Limestone ; and the Upper Trinity Aquifer consists of the upper member of the Glen Rose Limestone . Low permeability sediments throughout the upper member of the Glen Ros e Limestone separate the Middle and Upper Trinity aquifers. The Lower and Middle Trinity aquifers are separated by the low permeability Hammett Shale except where the Hammett Shale pinches out in the northern part of the study area (Amsbury, 1974; Barker a nd Ardis, 1996) ( Figure 3 16). 5.2 Structure Building on the structural interpretations of Ashworth (1983) and us ing available drilling logs from the Hill Country Underground Water Conservation District, geophysical logs, and locations of outcrop areas, Mace and others (2000) developed structural elevation maps for the base s of the Edwards Group and the Upper and Middle Trinity aquifers ( Figures 501 through 5 04). Mace and others (2000) collected geophysical logs from Texas Water Development Board , Edwar ds Aquifer Authority, Bandera County River Authority and Groundwater District, and private collections and used natural gamma logs to locate (1) the base of the Edwards Group, (2) the contact between the upper and l ower member s of the Glen Rose Limestone ( as defined by the lower evaporite beds just above the “Corbula ” marker bed or correlated equivalent), and (3) the base of the Middle Trinity sediments. Mace and others (2000) used r esistivity logs to add control points in parts of the study area in the abs ence of gamma logs to complete the structure surfaces . To further enhance the control of structural elevation point data, Mace and others (2000) supplemented our well log based data with outcrop elevation points. Mace and others (2000) digitized the appropriate formation contacts for the base of the Edwards Group and Upper and Middle Trinity sediments from 1:250,000 scale maps of surface geology in the area (Brown and others, 1974; Proctor and others, 1974a, b; Barnes, 1981) using AutoCAD (Autodesk, 1997) and converted the digitized contacts into an Arc Info (ESRI, 1991) geographical information system line coverage. Mace and others (2000) t he n georeferenced the line coverage, converted it into a point coverage from the arc vertices, and intersected it wit h a T riangulated I r regular N etwork constructed from a United States Geological Survey 3arc second digital elevation model to determine their point elevations. Mace and others (2000) compiled t he structural elevation information and organized it into Arc In fo for the base of the Middle Trinity Aquifer , the base of the Upper Trinity Aquifer , and the base of the Edwards Group sediments. Mace and others (2000) then exported the point elevations from Arc Info into point coordinates and

PAGE 12

11 imported them into Surfer (Golden Software, 1995) for spatial interpolation ( Figures 5 01 through 504 ). As part of this project , we updated the model structure of Mace and others (2000) by revising the structure of the Middle Trinity Aquifer and adding the Lower Trinity Aquifer as a fo urth layer. These changes were aided by structural interpretations from the Hays Trinity Groundwater Conservation District . The base of the Lower Trinity Aquifer was taken from the base of the Edwards Trinity Aquifer System used in the groundwater ava ilability model for the Edwards T rinity (Plateau) Aquifer System by Anaya and Jones ( 2009). When we compared th e base elevation of the Middle Trinity Aquifer from the original model (Mace and others, 2000) with the base elevation of the Lower Trinity from the Edwards Trinity (Plateau) Aquifer System model (Anaya and Jones, 2009) , we noticed that the structure s were not consistent because the base of the Middle Trinity dipped below the base of the Lower Trinity in Blanco County. In order to resolve this inconsistency between the two structures w e revised the base of the Middle Trinity Aquifer using data from the T exas C ommission on E nvironmental Q uality Source W ater Assessment and Protection Geographical Information System database developed by the United Sta tes Geological Survey . We used the Source Water Assessment and Protection data for the base of the Middle Trinity in Blanco County and merged it with the structural surface from the original model (Mace and others, 2000) for the remainder of the model. The two surfaces were merged using a linear smoothing algorithm in ArcGIS version (ESRI, 2005) We developed thickness maps by subtracting elevations for the tops and bases of the respective model layers using ArcGIS 9.1 ( Figures 505 through 508 ). Th e thickness of the relatively flat lying beds of the Edwards Group is controlled by the dendritic erosional pattern of the surface topography ( Figures 501 and 505). Although mostly masked by the dendritic erosional pattern of the surface topography in the central and eastern portions of the study area, sediments of the Upper Trinity Aquifer thicken towards the Balcones Fault Zone ( Figure 50 6). Sediments of the Middle and Lower Trinity a quifer s also generally increase in thickness towards the Balcones Fau lt Zone ( Figures 5 07 and 508). 5.3 Water Levels and Regional Groundwater Flow We compiled water level measurements and developed generalized steady state water level maps for the Edwards Group, and the Upper , Middle , and Lower Trinity aquifers in the s tudy area. To increase the number of measurement points, we expanded our time interval to lie between 19 77 and 1985 to approximate steady state water levels for the period about 1980. If a well had multiple water level measurements, the average measurement was chosen for contouring the water level map . Water levels in the aquifers generally follow topography ( Figures 509 through 512 ). Kuniansky and Holligan (1994) noted that water levels in this area are a subdued representation of surface topography due to recharge in the uplands and discharge in the lowlands. Water level maps indicate that water levels are influenced by the location of rivers and springs. For example, the water level maps show that groundwater in the aquifer flows toward most of the rive rs in the study area ( Figures 509 through 512 ). In the case of the Edwards Group, ground water flows east toward the escarpment where there are numerous springs at the geologic contact between the Edwards Group and the u pper member of the Glen Rose Limest one ( Figure 509). Barker and

PAGE 13

12 Ardis (1996) also noted that water level elevations and the direction of groundwater flow in the Trinity Aquifer System are largely controlled by the position of springs and streams. Groundw ater flows from higher waterlevel elevations toward lower water level elevations. The water level maps show that regional groundwater flow is from the northwest toward the southeast and east ( Figures 509 through 512). Water level maps also show that groundwater in the Upper , Middle , and Lower Trinity aquifers flows out of the study area to the south and east in to the Edwards ( Balcones Fault Zone ) Aquifer ( Figures 510, 5 11, and 5 12). The 'Discharge' section of this report discusses the estimated amount of groundwater flow from the Hill Country portion of the Trinity Aquifer System into the Edwards ( Balcones Fault Zone) Aquifer . Water levels, especially in shallow wells ( less than 100feet deep), can seasonally vary up to 50feet (Barker and Ardis, 1996) in response to rainfall events . Some wells show relatively small changes in water level over time, for example, wells 6904502, 5648301, 5761803, and 5850120, while others show large fluctuations , for example, wells 68 19806 and 5663604 ( Figures 5 13 through 516) . Wells with detailed measurements , for example, wells 68 19806, 6802609, and 6801 314, show seasonal fluctuations ( Figures 515 and 516). Figures 5 13 through 516 suggest that overall there are no longterm trends of declining or rising water level s in the Hill Cou ntry portion of the Trinity Aquifer System and thus water levels in the 1990s will be similar to those in Figures 5 09 through 512. From 19801997, water levels generally rose in the Upper Trinity Aquifer of Bexar County (Figure 5 17). Over the same perio d of time, water levels generally declined in the Middle and Lower Trinity aquifer s in Bandera, Blanco, Kendall, and Kerr counties and r ose , at least locally, in Bexar and Comal c ount ies ( Figure 5 18). In other parts of the study area, water levels show se asonal fluctuations but have remained fairly constant since 1980. The area with the most significant water level decline is near the city of Kerrville in Kerr County. The largest water level d ecline is approximately 40 feet in the Middle Trinity Aquifer an d 85 feet in the Lower T rinity Aquifer (Figures 5 15 and 516) . The 1 28foot water level rise in Kerr County (Well 5663604) can be attributed to a reduction in pumping by the City of Kerrville. Well 68 08102, which is located near the city of Wimberley (Hays County), shows a water level decline of approximately 4 5 feet between 1980 and 2000 (Figure 5 15) . 5.4 Recharge The primary sources of inflow to the Hill Country portion of the Trinity Aquifer System are rainfall on the outcrop, seepage losses thr ough headwater creeks, and lakes during high stage levels. The outcrops in the study area are composed of the upper and lower member s of the Glen Rose Limestone , Hensell Sand, and Edwards Group and receive all of the direct recharge from rainfall . The Cow Creek Limestone and Lower Trinity Aquifer sediments are not exposed at land surface in the study area and re ceive water by vertical leakage from overlying strata (Ashworth, 1983). B eds containing relatively low permeability sediments within the u pper membe r of the Glen Rose Limestone impede downward percolation of interstream recharge and facilitate horizontal groundwater flow resulting in baseflow and springflow to the mostly gaining perennial streams that drain the Hill Country (Barker and Ardis, 1996; As hworth, 1983). Recharge in the Edwards Group limestones of the northwestern portion of the study area occurs as infiltration of rainfall and losing streams. Much of this water later emerges as springs and

PAGE 14

13 seeps along the geologic contact between the Edward s Group and the upper member of the Glen Rose Limestone . Sinkholes and caverns in the Glen Rose Limestone of southern Kendall, northern Bexar, and western Comal counties may transmit large quantities of water to the Hill Country portion of the Trinity Aqui fer System . Karst enhanced recharge is especially significant along Cibolo Creek between Boerne and Bulverde (Ashworth, 1983; Veni, 1994). However, because much of this recharge is quickly transmitted to the Edwards ( Balcones Fault Zone ) Aquifer , it has minimal effect on the Hill Country portion of the Trinity Aquifer System (Barker and Ardis, 1996; Veni, 1994) . Several investigators have estimated recharge rates for the Hill Country portion of the Trinity Aquifer System ( Table 5 01). Most of them used str eam baseflow to estimate recharge. Muller and Price (1979) assumed a recharge rate of 1.5 percent of average annual precipitation for their rough approximation of groundwater availability. This estimate of recharge was intended to minimize impacts of groun dwater production on baseflow and groundwater flow to the Edwards ( Balcones Fault Zone ) Aquifer . Based on a study of baseflow gains in the Guadalupe River between the Comfort and Spring Branch gauging stations during a 20year period between 1940 and 1960, Ashworth (1983) estimated a average annual effective recharge rate of 4 percent of average annual precipitation for the Hill Country. Kuniansky (1989) estimated baseflow for 11 drainage basins in our study area for a 28 month period between December 1974 and March 1977 and estimated an annual recharge rate of about 11 percent of average annual rainfall. However, Kuniansky and Holligan (1994) reduced this recharge rate to 7 percent of average annual precipitation to calibrate a groundwater model that includ ed the Hill Country portion of the Trinity Aquifer System . They suggested that the numerical model did not include all the local streams accepting discharge from the aquifer. Bluntzer (1992) calculated long term average annual baseflow from the Blanco, Gua dalupe, Medina, Pedernales, and Sabinal rivers and Cibolo and Seco creeks to be 369,100 acre feet per year . Using a long term average annual precipitation of 30 inches per year, the recharge estimate by Bluntzer (1992) is equivalent to a recharge rate of 6.7 percent of average annual precipitation ( Riggio and others, 1987). However, Bluntzer (1992) suggests that a recharge rate of 5 percent is more appropriate to account for human impacts on baseflow such as nearby groundwater pumping , streamflow diversions , municipal and irrigation return flows, and retention structures. Bluntzer (1992) also noted that baseflow was highly variable over time. Mace and others (2000) suggest ed that differences in recharge rates reflect biases in the record of analysis due to v ariation of precipitation. The higher recharge rate estimated by Kuniansky (1989) is likely due to the higher than normal precipitation between December 1974 and March 1977, her record of analysis. Ashworth's (1983) recharge rate is probably biased toward a lower value because his record of analysis includes the 1950's drought of record . Mace and others (2000) developed an automated digital hydrographseparation technique to estimate baseflow for the drainage basin defined by the Guadalupe River gauging sta tions between Comfort and Spring Branch. Mace and others (2000) developed this technique based on methods used by Nathan and McMahon (1990); and Arnold and others (1995). Mace and others (2000) used the program to estimate baseflow from 1940 to 1990 and adjusted parameters to attain the best fit with Ashworth’s (1983) and Kuniansky’s (1989) baseflow values for the same stream reach. Using this technique , Mace and others (2000) estimated a recharge rate of 6.6 percent of average annual precipitation. N ote th at the calibrated recharge rate by Mace and

PAGE 15

14 others (2000) is about 4 percent of average annual precipitation. A ll baseflow based estimates of recharge underestimate recharge because they do not consider the component of recharge that follows the regional f low paths and bypasses the local streams . There is additional error in this methodology associated with the implied assumption that each watershed is a closed system and thus all water that recharge s the aquifer discharges to the adjacent river. However, r egional groundwater flow between watersheds results in underestimation of recharge in upgradient watersheds and overestimation in downgradient watersheds. In the updated model, we spatially distributed recharge based on Parameter e levation Regressions on Independent Slopes Model ( PRISM ) data ( Daly and Taylor, 1998; Spatial Climate Analysis Service, 2004) . Parameter elevation Regressions on Independent Slopes Model is an analytical model that spatially distributes monthly, seasonal and annual precipitation . We assume d that recharge is a fraction of annual precipitation. This fraction or recharge coefficient is determine d during model calibration. In addition to precipitation, we assume that the aquifer receives recharge from streamflow losses in Cibolo Creek. This recharge is estimated based on watershed modeling of the Cibolo Creek watershed by the U nited S tates Geological Survey (Ockerman, 2007) . This watershed modeling indicates average annual recharge of approximately 72,000 acre feet to the Trinity Aqui fer System within the study area . The methodology used in the updated model is an improvement over the recharge estimation method used by Mace and others (2000) that was based on baseflow coefficients and precipitation distribution. In addition to the weak nesses in baseflow based recharge estimation methods stated above, the updated model was developed using data from a study of the Cibolo Creek watershed (Ockerman, 2007) that was not available for use by Mace and others (2000) . 5.5 Rivers, Streams, Sprin gs , and Lakes Most of the rivers in the study area arise along the eastern margin of the Edwards Plateau and descend with a steep gradient into the Hill Country ( Figure 3 06). Many of these streams have upper reaches contained within narrow canyons and bro aden into flat bottomed valleys further downstream (Barker and Ardis, 1996). Three major drainage basins, including the San Antonio, Guadalupe, and Colorado r ivers, traverse the study area and funnel flow towards the southeast. Most of the rivers in the st udy area gain water from the Hill Country portion of the Trinity Aquifer System (Ashworth, 1983; Slade and others, 2002; Figure 5 19) and are hydraulically connected to the regional flow system (Kuniansky, 1990). These streams receive g roundwater that disc harges through seeps and springs that occur along the tops of impermeable units where they appear at land surface (Barker and Ardis, 1996). Much of the ground water in local flow systems within the Hill Country portion of the Trinity Aquifer System discharg e s to adjacent deeply entrenched, perennial streams instead of flowing to deeper portions of the aquifer (Ashworth, 1983). Many springs issue from the Edwards Group along the margin of the Edwards P lateau in the western part of the study area (Ashworth, 1983). M ost of the rivers in the study area are perennial ( Figures 520 through 5 2 6) . Lower reaches of Cibolo Creek lose flow between Boerne and Bulverde where it flows over the l ower member of the Glen Rose Limestone (Ashworth, 1983) ( Figure 526) . U pstream of Boerne , Cibolo Creek gains water where it flows over the upper member of the Glen Rose Limestone (Guyton and Associates, 1958, 1970; Espey, Huston, and Associates, 1982; Stein and Klemt, 1995; Mace and others, 2000). Lower reaches of most of the strea ms in the study area lose significant quantities

PAGE 16

15 of flow where they cross the recharge zone of the Edwards ( Balcones Fault Zone ) Aquifer (Barker and others, 1994). Most perennial rivers in the study area experience extremely low flow for brief periods duri ng droughts ( Figures 5 21 through 523 ). The study area includes four major lakes : Lake Travis, Lake Austin, Canyon Lake, and Medina Lake ( Figure 3 01). Canyon Lake and Lake Travis have maintained approximately constant lake levels ( 20 feet ) , although La ke Travis had large declines during drought s in the 1950s and mid1960s ( Figure 5 27). Lake Medina has much more variation in water levels and has nearly been dry on a few occasions during the drought of the 1950s (Espey, Huston, and Associates, 1989) ( Fig ure 5 27). Numerous springs occur in the study area ( Figure 528). Most of these springs issue from topographically low lying areas below the base of bluffs along rivers and streams, discharg ing groundwater that flow s laterally along the tops of hard, mor e resistant Glen Rose Limestone beds. Other springs discharge along the margin of the Edwards P lateau and contribute significant flow to the headwaters of the major rivers in the study area. Many of the spring discharge zones are characterized by phreatic vegetation, such as marsh purslane, cattail s , ferns, and cypress trees , indicative of a constant supply of water (Brune, 1981). Springs that occur in the Edwards Group generally have higher discharge rates than those occurring in the l ower and upper member s of the Glen Rose Limestone and the Cow Creek Limestone ( Table 502) , presumably due to the cavernous nature of the Edwards Group. 5. 6 Hydraulic Properties Variations in well yields are generally a result in variation in hydraulic properties of aquifers . Well yields in the Hill Country portion of the Trinity Aquifer System are often controlled by the location of fractures and dissolution features and consequently, may vary considerably over short distances. Although the Hill Country portion of the Trini ty Aquifer System as a whole is recognized by the s tate as a major aquifer (Ashworth and Hopkins, 1995), well yields can be low compared to other major aquifers. Hydraulic conductivity is defined as the rate of movement of water through a porous medium under a unit gradient. For example, very porous limestone may have hydraulic conductivities greater than 1,000 feet per day, sandy limestone may range from 100 to 1,000 feet per day, while aquifers with moderate hydraulic conductivity values may range from 10 to 100 feet per day, and aquifers with low hydraulic conductivity may range from 0.1 to 10 feet per day. Transmissivity is defined as the hydraulic conductivity times the thickness of the aquifer, and thus is a measure of the rate of movement through a defined thickness of aquifer under a unit gradient. Pumping tests in wells are conducted to order to develop estimates of hydraulic conductivity and transmissivity. Based on 15 aquifer tests, Hammond (1984) determined that hydraulic conductivity rang es from 0.1 to 10 feet per day in the lower member of the Glen Rose Limestone. Barker and Ardis (1996) thought that hydraulic conductivity probably averages about 10 feet per day in the Hill Country portion of the Trinity Aquifer System. No one has investi gated vertical hydraulic conductivities, although vertical hydraulic conductivities are likely to be much lower than horizontal hydraulic conductivities, especially in the upper member of the Glen Rose Limestone. Barker and Ardis (1996) noted that rechargi ng water moves laterally more easily atop low permeability beds than vertically through them. Guyton and

PAGE 17

16 Associates (1993) estimated that the vertical hydraulic conductivity of the Hammett Shale, Bexar Shale, and the marls of the upper member of the Glen R ose Limestone was about 0.0001 to 0.003 feet per day. In their model that included the Hill Country portion of the Trinity Aquifer System, Kuniansky and Holligan (1994) considered part of the Hill Country portion of the Trinity Aquifer System along the Edw ards (Balcones Fault Zone) Aquifer to have anisotropic properties, with greater hydraulic conductivity in the direction of faulting. Ashworth (1983) reports average transmissivit ies of about 230 square feet per day and 1,300 square feet per day for the Mid dle and Lower Trinity aquifers, respectively, and that substantially lower transmissivities are expected for the Upper Trinity Aquifer . Kuniansky and Holligan (1994) determined that transmissivity for the Hill Country portion of the Trinity Aquifer System ranged from 100 to 58,000 square feet per day. Stein and Klemt (1995) summarized 53 aquifer tests in the Glen Rose Limestone along the Edwards ( Balcones Fault Zone) Aquifer and found a median transmissivity of about 220 square feet per day. The Glen Rose L imestone can be unusually permeable in outcrop and shallow subcrop in northern Bexar County and southwestern Comal County near Cibolo Creek (Kastning, 1986; Veni, 1994). Barker and Ardis (1996) developed a map of transmissivity for the Hill Country portion of the Trinity Aquifer System based on aquifer tests, geologic observation, and computer modeling. They determined that transmissivity is generally less than 5,000 square feet per day but increases from 5,000 to 50,000 square feet per day along the bounda ry between Comal and Bexar counties and through Kendall County and eastern Kerr County. The quar t zose clastic facies of the updip Hensell Sand include some of the most permeable sediments in the Hill Country portion of the Trinity Aquifer System (Barker a nd Ardis, 1996). Ardis and Barker (1993) and Barker and Ardis (1996) surmised that the variations in transmissivity in the Hill Country are probably due more to variations in aquifer thickness than to tectonic or diagene sis. However, Barker and Ardis (1996) note d that diagenesis of stable minerals has diminished permeability in most downgradient , subcropping strata and that the leaching of carbonate constituents has enhanced permeability in some of the outcrop. Storativity is the volume of water released f rom storage per decline of hydraulic head (water pressure) and is typically less than 0.01 for a confined aquifer. Specific storage is defined as the storativity divided by the aquifer thickness. Ashworth (1983) estimates that in the Trinity Group, the confined storativity ranges between 105 and 103 (a specific storage of about 106 per f oot) and that the unconfined storativity (specific yield) ranges between 0.1 and 0.3. Based on two aquifer tests, Hammond (1984) determined a storativity of 3 105 for the l ower member of the Glen Rose Limestone . Although we could not locate values for the Edwards Group in the plateau area, the specific yield for the Edwards Group in the Edwards ( Balcones Fault Zone ) Aquifer is 0.03 (Maclay and Small, 1986, p. 68–69). Specific yield is a ratio that describes the fraction of aquifer volume that will “yield” or be released when the water is allowed to drain out of the aquifer under gravity. To estimate hydraulic properties for the study area and expand upon previous studies, Mace and others (2000): (1) compiled available information on aquifer properties or tests from published reports and well records, (2) conducted and analyzed detailed aquifer tests in the study area, (3) used specific capacity information to es timate transmissivity, and (4) summarized the results using statistics. Mace and others (2000) compiled aquifer property data from : (1) available literature ( Meyers ,1969; Hammond, 1984; Simpson and others , 1993; LB G Guyton Associates , 1995; Bradley and others , 1997) , (2) aquifer tests that they conducted in the study area , analyzing the results using the methodologies of Theis (1935), Cooper and Jacob (1946), and Kruseman

PAGE 18

17 and de Ridder (1994), and (3) specificcapacity (well performance) tests from the Texas Water Development Board water well database and used an analytical technique (Theis, 1963) to estimate transmissivity. Mace and others (2000) developed a map of hydraulic conductivity for the Middle Trinity Aquifer , used the spatial distribution of hydra ulic conductivity in each unit of the Middle Trinity Aquifer (Cow Creek Limestone, Hensell Sand, and lower member of the Glen Rose Limestone ) and the relative thickness of each unit. To estimate the hydraulic conductivity of the Middle Trinity Aquifer at any given point, Mace and others (2000) weighted the hydraulic conductivity of each layer by the relative thickness of each respective layer at that point. As a result of the paucity of data from the Edwards Group and Upper Trinity A quifers , Mace and others (2000) distributed hydraulic conductivity uniformly through the study area. The hydraulic conductivity values used in the Edwards Group and Upper Trinity Aquifer , 7 feet per day and 5 feet per day, respectively, are derived from calibration of the model b y Mace and others (2000). In the updated model, we simplified the distribution of hydraulic conductivity in the model and adjusted it during model calibration. As a result, hydraulic conductivity in the Edwards Group is uniformly distributed value of 11 f eet per day, while hydraulic conductivity in the underlying Upper, Middle, and Lower Trinity aquifers was divided into two zones. One zone represents higher hydraulic conductivity values in the Balcones F ault Zone and along Cibolo Creek and the other zone represents the remainder of the aquifer (Figure 5 29). Hydraulic conductivity values for the Lower Trinity Aquifer obtained from the Texas Water Development Board groundwater database and Hays Trinity Groundwater Conservation District lie within the range 0.01 to 4.41 feet per day with a geometric mean of 0.52 feet per day. We calculated the hydraulic conductivity from specificcapacity data from the Texas Water Development Board well database using methods outlined in Mace (2001). 5.7 Discharge Discharge from the Upper and Middle Trinity aquifer s in the Hill Country portion of the Trinity Aquifer System is, from greatest to lowest, through (1) discharge to streams and springs (Ashworth, 1983), (2) lateral subsurface flow and diffuse upward leakage to the Edwards ( Balcones Fault Zone ) Aquifer (Veni, 1994), (3) pumping from the aquifer, and (4) vertical leakage to the Lower Trinity Aquifer . Discharge from the Lower Trinity Aquifer takes the form of pumping and vertical leakage to the overlying Middle Trinity Aquifer . The model by Kuniansky and Holligan (1994) indicates net discharge to streams from the Hill Country portion of the Trinity Aquifer System of 155,000 acrefeet per year . The volume of baseflow varies from year to year depending on precipitation. T he volume of water that moves laterally from the Hill Country portion of the Trinity Aquifer System into the Edwards ( Balcones Fault Zone ) Aquifer is not known, partially because of the difficulty in estimating the amount of flow. A number of studies have indicated , either through hydraulic or chemical analys i s, that groundwater likely flows from the Hill Country portion of the Trinity Aquifer System into the Edwards ( Balcones Fault Zone ) Aquifer (Long, 1962; Klemt and others, 1979; Walker, 1979; Senger and Kreitler, 1984; Slade and others, 1985; Maclay and Land, 1988; Waterreus, 1992; Veni, 1994, 1995). Most of the se studies have focused on the movement of groundwater from the Glen Rose Limestone into the Edwards ( Balcones Fault Zone ) Aquifer ; however, water levels ( Figures 51 0 through 512 ) suggest that groundwater from

PAGE 19

18 the entire Hill Country portion of the Trinity Aquifer System discharges to the south and east in the direction of the Edwards ( Balcones Fault Zone ) Aquifer . Some of this groundwater flows directly into the Edwards ( Balcones Fault Zone ) Aquifer along faults, while the remainder continues to flow in the Hill Country portion of the Trinity Aquifer System beneath the Edwards ( Balcones Fault Zone ) Aquifer . It is possible that groundwater that co ntinues to flow in the Hill Country portion of the Trinity Aquifer System eventually discharges upward in to the Edwards ( Balcones Fault Zone ) Aquifer . However, work by Hovorka and others (1996) suggest that this vertical cross formational flow is limited. The Glen Rose Limestone in the Cibolo Creek area has been argued to be a part of the Edwards ( Balcones Fault Zone ) Aquifer due to the hydraulic response and continuity of the formations (George, 1947; Pearson and others, 1975; Veni 1994, 1995). A few stud ies have estimated the volume of flow from the Hill Country portion of the Trinity Aquifer System into the Edwards ( Balcones Fault Zone ) Aquifer . Lowry (1955) attributed a five percent error between measured inflows and outflows in the Edwards ( Balcones Fault Zone ) Aquifer to cross formational flow from the Glen Rose Limestone . Woodruff and Abbott (1986), citing a personal communication with B ill Klemt, report that recharge from cross formational flow accounts for six percent of tot al recharge or about 41,000 acrefeet per year on average , to the Edwards ( Balcones Fault Zone ) Aquifer . Kuniansky and Holligan (1994) suggest pre development groundwater discharge of 360,000 acre feet per year from the Hill Country portion of the Trinity Aquifer System to the Edw ards ( Balcones Fault Zone) Aquifer . T his estimate is about 53 percent of average annual recharge to the Edwards ( Balcones Fault Zone ) Aquifer and is probably too high (Mace and others, 2000) . LBG Guyton Associates (1995) estimated cross formational flow fr om the Glen Rose Limestone to the Edwards ( Balcones Fault Zone ) Aquifer in the San Antonio area, excluding recharge from Cibolo Creek, to be about two percent of total recharge to the aquifer. Mace and others (2000) estimate net discharge from the Hill Cou ntry portion of the Trinity Aquifer System to the Edwards ( Balcones Fault Zone ) Aquifer of 64,000 acre feet per year. O f the numerical groundwater flow model s of the Edwards ( Balcones Fault Zone ) Aquifer , Klemt and others ( 1979), Maclay and Land ( 1988 ), Sl ade and others ( 1985), Wanakule and Anaya ( 1993), Barrett and Charbeneau ( 1996), and Lindgren and others ( 2004) , only Lindgren and others (2004) include s cross formational flow from the Hill Country portion of the Trinity Aquifer System . Maclay and Land (1988) recognize the occurrence of cross formational flow between the Hill Country portion of the Trinity Aquifer System and the Edwards (Balcones Fault Zone) Aquifer but only as a topic for future study. Kuniansky and Holligan (1994), estimated 1974 to 1975 cross formational flow from the Hill Country portion of the Trinity Aquifer System to be about 480,000 acre feet per year , an order of magnitude larger than calculated cross formational flow by Lindgren and others (2004) of about 40,000 acre feet per year . Groundwater also discharges from the aquifer through pumping of water wells. Lurry and Pavlicek (1991), Barker and Ardis (1996), and Kuniansky and Holligan (1994) estimate d pumping from the Hill Country portion of the Trinity Aquifer System to be between 10,000 and 15,000 acre feet per year in the 1970s. Based on information in Bluntzer (1992), about 14,000 acre feet per year was produced from the Hill Country portion of the Trinity and Edwards Trinity ( Plateau ) Aquifer System s in the study area. Guyton a nd Associates (1993) estimated that about 6,350 acre feet was pumped from the Hill Country portion of the Trinity Aquifer System in northern Bexar County in 1990 with 85 percent of production from the Middle Trinity Aquifer . Texas Water Development Board p umping data indicate that for the period 1980

PAGE 20

19 through 1997 pumping from the Hill Country portion of the Trinity Aquifer System ranged from 14,000 to 24,000 acre feet per year. The primary categories of water use in the Hill Country portion of the Trinity A quifer System are (1) municipal, (2) manufacturing, (3) livestock , ( 4 ) rural domestic , and ( 5) irrigation. Municipal and manufacturing water uses are based on reported values from the users. We associated these values with known well locations and aquifers by cross referencing the water use to the municipal and manufacturing well s through the Texas Commission on Environmental Quality municipal water well database, the Texas Water Development Board water well database, and through telephone interviews with w ater users ( Figure 530a) . We distributed livestock, rural domestic and irrigation pumping based on the spatial distribution of rangeland, nonurban population, and irrigated farm land, respectively ( Figures 5 30a through 5 3 0d) . Pumping from the Hill Count ry portion of the Trinity Aquifer System has been rising over time, from about 15,000 acre feet per year in 1981 to more than 20,000 acre feet per year by 1997 ( Figure 5 31). About twothirds of this pumping is for rural domestic and municipal uses with the remainder used by manufacturing, livestock and irrigation. The increasing pumping from the aquifer is mostly due to increasing rural domestic pumping that rose from 6,000 acre feet per year in 1980 to more than 10,000 acre feet per year by 1997 ( Figure 5 32). Municipal pumping rose gradually from 2, 500 acre feet per year in 1981 to about 5,000 acre feet per year in 1997. Livestock and irrigation have remained relatively constant over the period 1980 through 1997. Manufacturing pumping rose from about 2,500 acre feet per year to about 4,400 acre feet per year in the late 1980s and remained relatively constant after 1988. Pumping from the Hill Country portion of the Trinity Aquifer System has been progressively increasing in most counties within the study ar ea ( Figure 5 33; Tables 5 0 3 to 508). However , pumping has remained relatively constant in Comal, Kimble, Travis, and Uvalde counties. Over the period 1980 through 1997, pumping doubled in B lanco, Gillespie, Hays, and K endall counties. 5.8 Water Quality Total dissolved solids in groundwater are a measure of water salinity . F resh, slightly saline, moderately saline, and very saline water have total dissolved solids of less than 1,000, 1,000 to 3,000, 3,000 to 10,000, and 10,000 to 35,000 milligrams per li ter, respectively. M ost groundwater in the study area is fresh to slightly saline but in some parts of the Hill Country portion of the Trinity Aquifer System groundwater is moderately saline ( Figure 534). Although the groundwater in the Edwards Group gene rally has lower salinity than groundwater in the Upper, Middle, and Lower Trinity aquifers , the median total dissolved solids in groundwater is similar in the Edwards Group and Upper and Middle Trinity aquifers (Figure 534). The m edian total dissolved sol ids are 450, 470, and 410 m illig rams per l iter in the Edwards Group, Upper and Middle Trinity aquifers, respectively. In the Lower Trinity Aquifer, the median total dissolved solids is higher than the other aquifers at 760 m illig rams per l iter. F resh groundwater occur s throughout the Edwards Group in the study area ( Figure 535 ). In the Upper, Middle, and Lower Trinity aquifers, slightly to moderately saline groundwater t ypically occur s in eastern , downdip, parts of the aquifers, especially in Blanco, Coma l, Hays, Kendall, and Travis counties ( Figures 5 36 through 538). Groundwater in the Edwards Group is mainly calcium magnesium bicarbonate type ( Figure 5 39). Groundwater in the Upper Trinity Aquifer is also mainly calcium magnesium bicarbonate -

PAGE 21

20 type but progressively becomes calcium magnesium sulfate type in downdip parts of the aquifer ( Figure 5 40). Groundwater in the Middle and Lower Trinity aquifers display similar ranges of geo chemical compositions with the former displaying more sulfate dominated c ompositions and the latter displaying greater sodium and chloride ( Figures 541 and 542) . With increasing depth in the Hill Country portion of the Trinity Aquifer System , groundwater compositions c an be categorized into three groups : (1) calcium magnesiumbicarbonate type compositions, (2) groundwater compositions characterized by increasing m agnesium and sulfate , and (3) groundwater compositions characterized by increasing sodium and chloride ( Figure 543) . Groundwater compositions in the Edwards Group ar e characteristic of Group 1, groundwater in the Upper Trinity Aquifer display Groups 1 and 2, while groundwater in the Middle and Lower Trinity aquifer s display s compositions reflective of all three groups . These compositional trends can be explained by th e following processes: (1) groundwater interaction with the limestone of the Edwards Group and the u pper member of the Glen Rose Limestone producing the calcium magnesium bicarbonate type composition; (2) groundwater interaction with the dolostone and evap orites that occur within the Glen Rose Limestone , resulting in increased magnesium and sulfate in the groundwater ; and (3) mixing with sodium chloride brine migrating from depth. Distribution of total dissolved solids , chloride, and sulfate shows no specif ic trend with increasing well depth. Most of the samples from the Edwards Group show no significant changes in total dissolved solids , chloride, sulfate and nitrate from the ground surface to well depths of about 3,500 feet. In the Lower Trinity Aquifer , highest groundwater salinity occurs at depth greater than 500 feet. Nitrate concentrations progressive ly decrease with increasing well depth in the Edwards, Upper, Middle, and Lower Trinity aquifers. Groundwater in the Edwards Group ha s the least nitrate with the highest nitrate concentrations occurring in the Upper and Middle Trinity aquifers. 6.0 Conceptual Model of Regional Groundwater Flow in the Aquifer The conceptual model ( Figure 6 01) is our best understanding of regional groundwater flow in the Hill Country portion of the Trinity Aquifer System . The conceptual model does not treat the Hammett Shale confining unit that separates the Middle and Lower Trinity aquifers as a distinct layer of flow . Rather, t his confining unit is simulated as a zone of restricted vertical leakance between the two aquifers. When precipitation falls on the outcrop of the aquifer, much of the water evapo rates, is taken up and transpir ed by vegetation or runs off into local streams and eventually discharges through major st reams out side of the study area. A bout four to six percent of the precipitation infiltrates into and recharges the underlying aquifer s over most of the study area. This percentage is higher in the eastern portion of the study area where the fractures of the Balcones Fault Zone facilitate higher recharge rates. Losing streams contribute recharge to the Edwards Group in the headwater areas of the streams along the western margin of the study area (Figure 3 06a ) because the Edwards Group in the plateau area has high permeability . Most of the recharge to the Edwards Group in the study area discharges along the edge of the plateau through springs, seeps, and evapotranspiration. A small

PAGE 22

21 amount of the flow from the Edwards Group percolates downward into the underly ing Upper , Middle , and Lower Trinity aquifer s . Most of the precipitation that recharges the Upper and Middle Trinity a quifer s discharges to local and major streams through baseflow to these surfacewater features. An exception is Cibolo Creek, where karsti fication of the lower member of the Glen Rose Limestone changes the creek from a gaining stream to a losing stream between Boerne and Bulverde ( Figure 301) . Most of the remaining recharge in the aquifer discharges either through wells pumping from the aquifer or flows laterally into the Edwards ( Balcones Fault Zone ) Aquifer . There are likely several short flow paths along streams where the water table is shallow. In these areas recharged precipitation likely flows a short distance and is discharged via e vapotranspiration. Because of the localized nature of the flow paths and the limitations of the model grid, t his evapotranspiration discharge would likely be included in discharge to streams. Groundwater can perch on low permeability beds with in the Upper Trinity Aquifer and flow laterally to springs , however, some water percolates through the Upper Trinity Aquifer into the Middle Trinity Aquifer . The Lower Trinity Aquifer is not exposed at land surface. Consequently, groundwater flow enters the Lower Trinity Aquifer through downward cross formational flow from the Middle Trinity Aquifer and discharges by cross formation back to the Middle Trinity Aquifer in downdip portions of the aquifers . In general, groundwater in the Hill Country portion of the Trinity Aquifer System flows from areas of higher topography to areas of lower topography , from the west to the east. In general, lithology and local fracturing control permeability development and distributions in the Edwards Group and the Upper, Middle , and Low er Trinity aquifers . We believe that hydraulic conductivity is higher in the eastern portion of the study area , where they coincide with the Balcones Fault Zone, than in the remainder of the aquifer system. The Edwards Group in the plateau area has high vertical and horizontal permeability due to karstification. The Upper Trinity Aquifer generally has lower permeabilit y but can be locally very permeable, especially in the outcrop. Due to the occurrence of shaley beds, the Upper Trinity Aquifer has a much lo wer ratio of vertical to horizontal permeability than the overlying Edwards Group. The Middle Trinity Aquifer has moderate permeabilit y and greater ability to transmit water vertically than the Upper Trinity Aquifer . The Middle Trinity Aquifer is most permeable in the sandy outcrop area of Gillespie County. Specific yield in the limestone is primarily controlled by fractur es . The Lower Trinity Aquifer is on average less permeable than the overlying aquifers, with highest values occurring in the Kerrville ar ea. Pumping from the Hill Country portion of the Trinity Aquifer System has been progressively rising over the period 1980 through 1997. This increasing pumping is most apparent in counties adjacent to San Antonio and Austin, the two largest cities in the region, which are Bexar, Hays, Kendall , and Kerr counties. Some of these counties have experienced a doubling of pumping over the period of time covered by this study. 7.0 Model Design

PAGE 23

22 Model design includes (1) choice of code and processor, (2) discret ization of the aquifer into model layers and cells, and (3) assignment of model parameters in to the various model layers . The model design must agree as much as possible with the conceptual model of groundwater flow in the aquifer. 7.1 Code and Processor Groundwater flow through the Hill Country portion of the Trinity Aquifer System was simulated using MODFLOW 96, a widely used modular finite difference groundwater flow code written by the U nited S tates Geological Survey (Harbaugh and McDonald , 1996). This code was selected because of (1) its capabilities of simulating regionalscale groundwater processes in the Hill Country portion of the Trinity Aquifer System , (2) its documentation and wide use (McDonald and Harbaugh, 1988; Anderson and Woessner, 2002), (3) the availability of a number of thirdparty pre and post processors for facilitating easy use of the modeling software, and (4) its easy availab ility because it is public domain software. Processing MODFLOW Pro version 7.0.18 was used to load input da ta into the model and view model outputs ( Chiang , 2005). Other pre and post processors can read source files for MODFLOW 96. This model was developed and run on a Dell Precision 490 with a 3.0 G Hz Dual Core Xeon processor and 2 G B RAM running Microsoft Wi ndows XP Professional (v. 5). 7.2 Layers and Grid The lateral extent of the model corresponds to natural hydrologic boundaries, such as erosional limits of the aquifers , rivers, and the structural boundary with the Edwards (B alcones F ault Z one ) A quifer, a nd hydraulic boundaries to the west that coincide with groundwater divides. According to the hydrostratigraphy and conceptual model, we designed the model to have four layers. Layer 1 consists of the Edwards Group of the Edwards Trinity ( Plateau ) A quifer S ystem , and Layer s 2, 3 and 4 consist of the Upper , Middle and Lower Trinity a quifer s , respectively . We defined the active and inactive cells by first establishing the lateral extent of the formations in each layer using the geologic map ( Figure 3 16). We assigned a cell as active if the formation covered more than 50 percent of the cell area. Please note that the spatial extents of the respective aquifers were revised slightly during model calibration to address dry cell and numerical stability issues. We did not include the thin sliver s of the Edwards Group in the eastern part of the study area, for example in Blanco County, because: (1) our structure maps do not accurately represent the complexity of faulting in the area, (2) flow in some of these rocks i s associated with the Edwards ( Balcones Fault Zone ) aquifer , and (3) in many areas these rock are discontinuous and thus groundwater flow, if any, would be difficult to simulate at the regional scale. It should be noted that we did include a part of the Edwards Group that is not recognized by TWDB as part of the Edwards Trinity (Plateau) Aquifer in eastern Kerr County and western Kendall County. Each layer has 69 rows and 115 columns for a total of 31, 740 cells in the model. All the cells have uniform later al dimensions of 1 mile by 1 mile. We selected this cell size to be small enough to reflect the density of input data and the desired output detail and large enough for the model to be manageable. Cell thickness depended on differences in top and bottom el evation s of the model layers. After we made cells outside of the model area and outside the lateral extent of each layer inactive, the model had a total of 12,976 active cells: 1,107 in L ayer 1 ; 3,562 in L ayer 2 ; 4,517 in L ayer 3 ; and 3,790 in Layer 4 (F ig ure 7 01).

PAGE 24

23 7.3 Model Parameters We distributed model parameters, including (1) elevations of the top and bottom of each layer, (2) horizontal and vertical hydraulic conductivity, (3) specific storage, and (4) specific yield using Arc GIS 9.1. We defined top and bottom elevations for each layer from the structure maps and land surface elevations from digital elevation models downloaded from the United States Geological Survey . We used ArcGIS 9.1 to assign top and bottom elevations. For L ayer 1 (Edwards Group), we assigned the top as the landsurface elevation and the bottom according to the structure map of the base of the Edwards Group ( Figure 5 01). The top and base of Layer 2 (Upper Trinity A quifer) w ere assigned according to the structure map of the Up per Trinity Aquifer ( Figure 5 02) . W here covered by active cells in Layer 1 , the top of Layer 2 coincides with the base of Layer 1, otherwise it is defined by the land surface elevation. The bottom of Layer 2 was defined by the base of the Upper Trinity A quifer ( Figure 5 02 ). Similarly, the top of L ayer 3 (Middle Trinity A quifer) was defined as the bottom of Layer 2 and the landsurface elevation where exposed (Figure 5 03) . The bottom of layer 3 was assigned using the elevation of the base of the Middle Tr inity Aquifer ( Figure 5 03). The top of Layer 4 (Lower Trinity Aquifer) is defined as the base of the Hammett Shale , the confining unit separating the Middle and Lower Trinity aquifers (Figure 5 04). Groundwater flow through t he Hammett Shale is not explic itly simulated in the model. We initially assigned hydraulic conductivity values for Layers 1, 2, and 3 previously used in Mace and others (2000) and adjusted these values during calibration. These values were uniform values of 7 and 5 feet per in L ayers 1 and 2 based on geometric mean of hydraulic conductivity data, respectively, and a distributed range of values of 0.7 to 64 feet per day in Layer 3. The initial hydraulic conductivity value we assigned to Layer 4 was 0.6 feet per day, the geometric mean of the hydraulic conductivity data for the Lower Trinity Aquifer. We initially assigned vertical hydraulic conductivity to be one tenth the horizontal hydraulic conductivity. We simulated groundwater flow between L ayers 3 and 4, through the Hammett Shale, us ing vertical leakance values . These vertical leakance values were initially set to be proportional to the relative thickness of the Hammett Shale in each cell. The purpose for using vertical leakance is to simulate vertical flow through the Hammett Shale c onfining unit without the need to simulate horizontal flow through the unit which is assumed to be small. The range of vertical leakance values is 106 to 0.8 per day (Figure 7 02) . We assigned uniform values of specific storage and specific yield values i n each layer . Initially a ssigned specific storage values are 106, 107, 108, and 108 per foot in L ayers 1, 2, 3, and 4, respectively. Initially a ssigned specific yield values are 8 104, 5 105, 8 105, and 8 105 in L ayers 1, 2, 3, and 4, respectively. We assigned Layer 1 as unconfined and Layers 2 through 4 as confined/unconfined. We allowed the model to calculate transmissivity and storativity according to saturated thickness. We used units of feet for length and days for time for all input data to th e model. To solve the groundwater flow equation, we used the Slice Successive O ver R elaxation solver with a convergence criterion of 0.0001 f eet. 7.4 M odel Boundar y Conditions

PAGE 25

24 Model boundar y conditions are factors that control the inflow and outflow of groundwater in a numerical model. We assigned model boundar y conditions for (1) recharge, (2) pumping, (3) rivers and streams, (4) reservoirs , ( 5) outer model boundaries, and ( 6) initial head conditions. We used Arc GIS 9.1 to distribute values for model boundar y conditions spatially, such as drains, general head boundaries, recharge, and pumping. We assigned recharge based primarily on the spatial distribution of annual precipitation over the study area ( Figure 3 09). The initial recharge assigned to the model was 4.7 percent of annual precipitation. This value coincides with the value used in the groundwater availability model for the Edwards Trinity (Plateau) Aquifer (Anaya and Jones, 2009). We also included in the recharge distribution, recharge from st reamflow losses in Cibolo Creek. We assigned pumping values in the model according to our analysis of pumping as discussed in the 'Discharge' section of this report (Figure 5 30) . This model simulates the regional effects of pumping on water levels for rur al domestic, municipal, irrigation, industrial, and livestock uses (T ables 503 through 5 0 8). Municipal and manufacturing pumping was distributed based on known well locations and pumping data from the Texas Water Development Board Water Use Survey. The other uses (domestic, irrigation, and livestock) were distributed throughout the model grid, reflecting the spatial distribution of associated land use. Rural domestic pumping was distributed based on the spatial distribution of population outside major urban areas that lie within the model grid. Irrigation pumping was distributed based on 1:250,000scale land use and land c over data from United States Geological Survey . Irrigation was assumed to occur on all land classified as orchards, row crops, or small grains. Livestock pumping was also distri buted based on 1:250,000scale land u se and land c over data from United States Geological Survey . Livestock pumping was assumed on all rangeland. Figure 7 03 shows the spatial distribution of total pumping for the y ear 1980. We used the Drain Package of MODFLOW to represent rivers and streams in the model ( Fig ure 704). T his package only allow s the streams to gain water from the aquifer. The River Package, which is another possible approach for simulating rivers and streams, allows streams to gain and lose water. Mace and others ( 2000) found that the River Package could allow unrealistic amount s of water to move from the rivers and streams into the aquifer and thus underestimate potential water level declines due to pumping or drought. Observed streamflow losses in Cibolo Creek along the boundary between Bexar and Comal counties are simulated as recharge. The Drain Package requires a drain elevation and conductance. When the head in the aquifer is above the drain elevation , water flow s out of the model through the drain . If the head in the aquifer is equal to or below the drain elevation , no flow occurs from the drain to the aquifer. D rain conductance is a measure of hydraulic resistance to flow out of the drain. We def ined the drain elevation by intersecting stream bed location with the digital elevation model in Arc GIS 9.1. We assigned the drain conductance based on estimated width of the stream, a stream length of one mile (equivalent to the model cell size) , an assu med riverbed thickness of one foot , and an assumed vertical hydraulic conductivity of 0.1 f ee t per day. After Mace and others (2000) calibrated the model, they investigated the sensitivity of simulated water levels to different values of drain conductance. Except for very low values, the drain conductance generally has little effect on water levels in the model (Mace and others, 2000) . We also used drains to represent discharge to major spring s , seepage from the erosional edge of the Edwards Group in the pl ateau area, and flow out of the Middle Trinity A quifer in Gillespie County ( Figure 7 04). For the springs, we assigned the drain elevation as the land surface elevation at the spring location and an initial

PAGE 26

25 conductance based on an assumed one foot thicknes s and the geometric mean hydraulic conductivity of the layer. For the erosional edge of the Edwards Group and flow out of the Middle Trinity A quifer in Gillespie County, we assigned a drain elevation 10 feet above the base of Layer 1 and a drain conductanc e based on a one foot thickness and the geometric mean hydra ulic conductivity of the layer. We simulated the influence of Medina Lake , Canyon Lake , Lake Travis, and Lake Austin on the aquifer using MODFLOW ’s River Package (Figure 704). The River package r equires hydraulic conductance of riverbed, river stage , and bottom elevation of the river. We assigned the riverbed conductance according to estimated width of the stream, a stream length of one mile (equivalent to the model cell size) , riverbed thickness of one foot, and vertical hydraulic conductivity of 0.1 feet per day. We assigned the head in the river as the average lakelevel elevation for the respective lakes. We defined the elevation of the riverbed by intersecting stream bed location with the digital elevation model in ArcGIS 9.1. Outer model boundar y conditions define the spatial extent of active flow within the respective layers in the model. In this model, the outer boundar y conditions are defined by the use of noflow and general head boundari es. The model boundaries are generally simulated by no flow boundaries to the north and west and general head boundaries located in the south and east where the Hill Country portion of the Trinity Aquifer System bounds the Edwards (Balcones Fault Zone) Aquifer. The no flow boundary in the north coincides with surface water divides in the Pedernales and Colorado River basins. The no flow boundary in the west follows a flow path in the Edwards Trinity (Plateau) Aquifer. Layer 4 is also bound by noflow bounda ries in the south and east based on the assumption, in response to work by Hovorka and others (1996), that there is very little groundwater flow between the Hill Country portion of the Trinity Aquifer System and Trinity Group rocks underlying the Edwards ( Balcones Fault Zone) Aquifer. A no flow boundary also exists at the base of the Lower Trinity Aquifer based on the assumption that there is no cross formational flow between the Lower Trinity Aquifer and underlying Pre Cretaceous rocks. To model the flow of ground water between the Hill Country portion of the Trinity Aquifer System and the Edwards ( Balcones Fault Zone ) Aquifer, we used the General Head Boundary Package of MODFLOW. We placed general head boundary cells along the contact with the Edwards (B alcones F ault Z one ) A quifer in layers 2 and 3 ( Figure 7 04 ). The General Head Boundary Package requires values for hydraulic head and conductance. We assigned the hydraulic head according to the interpreted water level map ( Figure 5 03) in the area of the gen eral head boundary cells. We assigned the general head boundary conductance according to the hydraulic conductivity and geometry of the cell and an assumed one foot thickness. Conceptually, the general head boundary conductance represents the resistance to flow between a cell in the model and a constant head source or sink. In this case, we have used the general head boundary to represent flow out of the study area either into the Edwards (B alcones F ault Z one) A quifer across faults or continuing into the down dip parts of the Trinity A quifer System . For simplicity, we used an arbitrary thickness of unity ( one foot ) to define conductance. The updating of this model included c hanges to the boundary conditions. In addition to the addition of the Lower Trinity A quifer to the model, these changes include: (1) the constant head cells that were used by Mace and others (2000) to simulate reservoirs were replaced by river cells, (2) river cells simulating Lake Travis were removed from Layer 2 and now only appear in La yer 3, (3) the spatial extent of Medina Lake was revised , (4) the spatial distribution of recharge

PAGE 27

26 was revised to account for the effects of the Balcones Fault Zone and recharge from Cibolo Creek . The constant head cells were converted to river cells becau se constant head provide an unlimited, unrestricted source of water when impacted by nearby pumping and with therefore produce unrealistic high water level adjacent to the constant head cells. On the other hand, the River Package in MODFLOW include a conductance parameter that can be used to restrict flow and would therefore allow water levels to fall to more realistic values in response to pumping. Although the potential exists to produce unrealistically high flows from the River Package (similar to the u se of constant heads), amounts of water to the groundwater flow system under periods of high pumping, proper attention to boundary elevation and conductance can mitigate this effect. During model calibration, minor adjustments were made to the outer model boundar y conditons to address dry cell and numerical stability issues. 8.0 Modeling Approach Model calibration involves the adjustment of parameters until the model results of groundwater elevations and base flow discharge reasonably match measured field data. Our approach for calbrating the model included two major steps: (1) calibrating a steady state model and (2) calibrating a transient model. The steady state model was developed first to facilitate easier calibration because some parameters, such as aquifer storage and water level variations over time, do not need to be taken into consideration. In the steady state model, calibration only requires consideration of spatial variations of all input parameters within the aquifer. We calibrated the st eady state model to reproduce water levels for 1980, reproducing the 1977 through 1985 water level measurements (Figure 5 09 through 512) . We used the steady state model to investigate (1) recharge rates, (2) hydraulic properties, (3) boundary conditions, (4) discharge from the Hill Country portion of the Trinity A quifer System into the Edwards ( Balcones Fault Zone) A quifer, (5) groundwater flow budget, and (6) sensitivity of model results to different parameters. Our approach for calibrating the model was to match water levels and groundwater discharge to rivers (for steady state conditions) and water level and groundwater discharge fluctuations (for transient conditions) using our conceptual understanding of the flow system . We quantified the calibration, or goodness of fit between the simulated and measured water level values, using the mean absolute error: n i i s mh h n MAE11 , W here MAE is the mean absolute error, n is the number of calibration points, hm is the measured hydraulic head at point i , an d hs is the simulated hydraulic head at point i . The mean absolute error is the mean of the absolute value of the differences in measured and simulated hydraulic head (Anderson and Woessner, 2002). Our standards for calibration included: 1) t he mean absolute error must be less than 10 percent of the measured hydraulic head drop across the model area and 2) t he error shall not be biased by areas with considerably more control points than other areas . Once we completed the steady state model, we used the fram ework of the model to develop a transient model for the years 1980 through 1997 using annual stress periods .

PAGE 28

27 Please note that the first stress period in the transient model is 1,000,000 days long and represents the 1980 steady state model. Th e transi ent mo del allowed us to test how well the model could reproduce water level fluctuations in the aquifer. We calibrated the transient model by adjusting aquifer stora ge values to minimize the difference between simulated and measured water level variations. 9.0 Steady -State Model Once we assembled the input datasets and constructed the framework of the model, we calibrated the steady state model and assessed the sensitivity of the model to different hydrologic parameters. 9.1 Calibration We calibrated the mod el to measured water levels for 1977 through 1985 used to represent 1980 water levels . We chose the year 1980 for our steady state model because it fell within a period of relatively stable water level s in the Hill Country portion of the Trinity Aquifer Sy stem. We adjusted recharge and spatial distribution of hydraulic conductivity and general head boundary conductance to calibrate the steady state model. We assigned recharge into three zones based on varying aquifer characteristics and recharge pathways: ( 1) Balcones Fault Zone, (2) areas outside the fault zone, and (3) Cibolo Creek. We varied r echarge during the calibration process, resulting in a final recharge rate of 5 percent of average annual precipitation in the Balcones Fault Zone, along the easter n margin of the study area, and 3.5 percent of average annual precipitation throughout the remainder of the model area. A long Cibolo Creek, we set recharge equivalent to measured streamflow loss of about 70,300 acre feet per year (Figure 9 01) . We also ad justed hydraulic conductivity during model calibration. In the calibrated model, w e assigned a uniform hydraulic conductivity value of 11 feet per day to the Edwards Group. Assigned h ydraulic conductivity values in the Upper Trinity Aquifer are 150 feet pe r day along Cibolo Creek, 15 feet per day within the Balcones Fault Zone, and 9 feet per day in the remainder of the aquifer. The two lower hydraulic conductivities, within and outside the Balcones Fault Zone, fall within the range of measured hydraulic co nductivity in the Upper Trinity Aquifer. The highest hydraulic conductivities in the Upper Trinity Aquifer which lie along part of Cibolo Creek can be justified based on work done by Kastning (1986) and Veni (1994) that indicated very high hydraulic conduc tivity near the creek. In the Middle Trinity Aquifer, we assigned a uniform hydraulic conductivity of 7.64 feet per day, the geometric mean of the hydraulic conductivity values used by Mace and others (2000) , for the portion of the aquifer outside the Balc ones Fault Zone. In t he Balcones Fault Zone portion of the Middle Trinity Aquifer , we assigned a uniform hydraulic conductivity of 15 feet per day. In the Lower Trinity Aquifer, we assigned hydraulic conductivity values of 16.7 and 1.67 feet per to the Bal cones Fault Zone and the remainder of the aquifer, respectively.

PAGE 29

28 The calibration process resulted in only minor changes to drain conductance values in individual cells. We increased general head boundary conductance values by factors of 5 and 2.5 in layer s 2 to 3, respectively, to facilitate increased inter aquifer flow between the Hill Country portion of the Trinity Aquifer System and the Edwards (Balcones Fault Zone) Aquifer due to the large amounts of recharge flowing from the Cibolo Creek. Inter aquif er flow between the Middle and Lower Trinity aquifer through the Hammett Shale are simulated using vertical leakance . We varied vertical leakance spatially based on the Hammett Shale thickness . Vertical leakance values decrease with increasing Hammett Shal e thickness reach ing a maximum value where the Hammett Shale is absent. Vertical leakance values lie in the range 106 to 0.8 per day. Simulated water levels from the calibrated steady state model are fairly close to measured water levels , and display no a pparent spatial biases ( Figure 902). The mean absolute error of the calibrated model is 54 feet, which is approximately 4 percent of the 1,700foot range of measured water levels ( Figure 9 03). This indicates that the average difference between measured and simulated water levels in the model is 54 feet —acceptable because the result lies within the 10percent target for model calibration. Water balance discrepancies are also acceptable, approaching 0 percent. In addition to the comparison of measured and simulated water levels, we compar ed measured streamflow and simulated drain discharge to indicate how well the model reproduces groundwater discharge to major streams in the study area ( Figure s 904 and 9 05). There is general agreement between measured st ream discharge of Barton Creek, Blanco River, Guadalupe River, Hondo Creek, Medina River, Onion Creek, and Pedernales River indicating that the steady state model does a reasonable job of reproducing baseflow to streams. The water budget of the steady stat e model indicates that total groundwater flow through the model is approximat ely 321,000 acre feet per year (T able 9 01). Of this flow, about 60 percent discharges to streams , springs, and reservoirs , and 35 percent discharges through cross formational flo w to the Edwards (Balcones Fault Zone) Aquifer . About 5 percent of groundwater discharge is due to well pumping , mostly for municipal and rural domestic uses. We used the calibrated model to investigate the volume of recharge to and ground water moving bet ween the different aquifers (Table 9 02) . The total volume of recharge to the aquifer due to precipitation falling on the land surface and streamflow loss from Cibolo Creek is about 304,000 acre f ee t per y ea r. About 50 percent of the recharge in the study area occurs in the Upper Trinity A quifer while 20 and 30 percent of recharge occurs in the Edwards Group and Middle Trinity A quifer , respectively . Recharge to the Lower Trinity Aquifer is insignificant. In the model, very small amounts of recharge to the L ower Trinity Aquifer occur along the Pedernales River where the overlying Middle Trinity Aquifer is thin and may not be saturated. About 20 percent of the water that recharges the Edwards Group flows into the Upper Trinity A quifer. The total inflow of water to the Upper Trinity A quifer , including infiltration of precipitation and cross formational flow , is about 166,000 acre f eet per y ea r. About 40 percent of the total in flow into the Upper Trinity A quifer flows into the Middle Trinity A quifer. Total inflow in to the Middle Trinity A quifer is about 153,000 acre f ee t per y ea r. According to the model, slightly less water enters the Middle Trin ity A quifer through cross formational flow than through direct infiltration on the outcrop. Based on our conceptual mode l, total groundwater circulation in the Lower Trinity Aquifer is a relatively minor component of the total groundwater budget of the Hill Country

PAGE 30

29 portion of the Trinity Aquifer System. In this steady state model, net cross formational flow from the Middle Trinity Aquifer to the Lower Trinity Aquifer is approximately equal to total pumping from the Lower Trinity Aquifer. The model shows that over 100,000 acre feet per year of groundwater flows out through the general head boundary along the eastern and southern margins of the model. This groundwater flows from the Upper and Middle Trinity aquifer s in to the Edwards (B alcones F ault Z one ) A quifer. Some of this groundwater flows directly from the Trinity A quifer System into the Edwards (B alcones F ault Z one ) A quif er and some continues to flow in the portion of the Trinity A quifer System that underlies the Edwards (Balcones Fault Zone) Aquifer (Ashworth and Hopkins, 1995). Presumably, groundwater moves downdip in the Trinity A quifer System and eventually discharges upward into the Edwards (B alcones F ault Z one ) A quifer. The model results show that the flow of ground water across the general head boundary is much less in the northeastern part of the boundary than the central and southwestern parts (Table 9 03) . The gro undwater flow across the general head boundary is 260 acref eet per y ea r per mile for the boundary within Travis and Hays counties, reaches a maximum of 1,700 acref eet per y ea r per mile in Comal and Bexar counties, and is 490 acre f eet per y ea r per mile w ithin Medina , Bandera, and Uvalde c ount ies . This numerical result is qualitatively supported by the measured potentiometric surface which shows groundwater generally flowing perpendicular to the boundary in Comal, Bexar, and Medina counties and subparallel to the boundary in Travis and Hays counties ( Figure 9 02). The spatial distribution of groundwater flow between the Trinity Aquifer System and the Edwards (Balcones Fault Zone) Aquifer is likely influenced by the large amounts of recharge taking place al ong Cibolo Creek in Bexar and Comal counties. Faults also have greater displacements to the east and therefore may act as more effective barriers to flow. 9.2 Sensitivity Analysis After we completed calibration of the steady state model, we analyzed the input parameters to assess the sensitivity of model results to respective input parameters : vertical and horizontal hydraulic conductivity, general head boundary conductance, drain conductance, river conductance, pumping, and recharge. Sensitivity analysis is a method of quantifying uncertainty of the calibrated model related to uncertainty in the estimates of respective aquifer parameters, stresses, and boundary conditions (Anderson and Woessner, 2002). Determining the sensitivity of the model to specific parameters offers insights into the uniqueness of the calibrated model. Sensitivity analysis identifies which parameters have the greatest influence on water levels and groundwater discharge to springs and streams. A model is sensitive to a specified input parameter if relatively small changes in that parameter result in relatively large changes in simulated water levels. In other words, calibration is possible only over a narrow range of values and, consequently, model uncertainties are relatively low. A m odel is insensitive if relatively large changes of a specific input parameter produce small water level changes. Insensitivity results in higher uncertainties because the model will remain calibrated over a large range of input parameter values. Sensitivit y is analyzed by systematically varying parameter values and noting changes in water levels over the calibrated model. The water level changes are quantified by calculating the Mean Difference ( MD ) as follows:

PAGE 31

30 n i cal senh h n MD11 where: n is the number of points, hsen is the simulated water level for the sensitivity analysis, and hcal is the calibrated water level. The Mean Difference is positive if water levels are higher than calibrated values and negative if they are lower than calibrated values. Wat er levels in the model are most sensitive to recharge and horizontal hydraulic conductivity , and to a lesser extent, to vertical hydraulic conductivity ( Figure 906). The model is insensitive to pumping, and general head boundary, drain, and river conducta nce. The insensitivity to pumping can be attributed to the fact that pumping is a relatively minor component of the overall aquifer water budget. Insensitivity to drain and general head boundary conductance can be attributed to high conductance values of up to 109 square feet per day. Consequently, in order to have much of an effect on water levels, drain and general head boundary conductance would probably have to be lowered by several orders of magnitude. Additionally, the effects of drain and general hea d boundary conductance are local. As a result, varying drain and general head boundary conductance only produces water level changes close the boundaries and does not have widespread effects throughout the model. 10.0 Transient Model Once we calibrated the steady state model to 1980 conditions, we proceeded to calibrate the model for transient conditions for the period 1980 through 1997 (Table 10 01) . 10.1 Calibration We simulated water level fluctuations during the period 1980 through 1997 using annua l stress periods for 1981 through 1997. Calibration was achieved by adjusting storage parameter values, specific storage, and specific yield until the model responses approximated water level fluctuations observed in wells in the model area. Specific yield is applicable to the unconfined part s of the aquifer and is defined as the volume of water that an unconfined aquifer releases from storage per unit surface area of aquifer per unit decline in the water level (Domenico and Schwartz, 1990) . Specific storag e is applicable to the confined part s of the aquifer and is defined as a measure of the volume of water per unit volume of aquifer rock that enters or leaves storage per unit change in water level (Domenico and Schwartz, 1990). Specific storage and specifi c yield are important factors in transient calibration because they influence water level responses to changes in recharge and discharge. Low specific storage or specific yield values result in water level fluctuations that are larger and more rapid than t hose associated with higher specific storage or specific yield values. This difference occurs because less water is required to produce a given water level change. Using annual stress periods, we simulated water level fluctuations due to recharge and pumpi ng variations during the period 1980 through 1997. We found that specific storage values of 105, 106, 107, and 108 per f oot for the Edwards Group, and the Upper, Middle, and Lower Trinity aquifers , respectively, and specificyield values of 0.008, 0.0005, 0.0008, and 0.0008 for t he

PAGE 32

31 Edwards Group, and the Upper, Middle, and Lower Trinity aquifers , respectively, worked best for reproducing observed water level fluctuations (Table 10 02) . The model does a good job of reproducing observed water level fluctu ations in some areas but not as well in other areas ( Figures 10 01 through 10 05). N ote that baseline shifts in water levels in Figure 1002 are often due to the influence of local scale conditions not represented in the regional model or errors in our par ameterization of the aquifer data. Although there are limitations, the model does a good job of reproducing year to year water level variations in most wells. Comparison of measured and simulated 1990 and 1997 water levels indicate mean absolute errors of 52 and 57 feet , respectively, or approximately 3.5 and 5.3 percent of the range of measured water levels ( Table 10 03; Figure 1004). Table 10 04 shows the water budgets for the respective model layers in 1980, 1990, and 1997. Simulating discharge to sprin gs using a regional scale model is often difficult because of spatial and temporal scale issues. Table 10 05 shows simulated and measured discharge for selected springs in the study area. It should be noted that the measured discharge values represent sing le snapshots in time that: (1) in most cases did not fall within the 1980 through 1997 transient model period, and (2) may not be representative of average discharge from the spring during the transient modeling period because spring discharge varies widel y over time . Simulated discharge values represent discharge averaged over each annual stress period. Additionally, springs are often discharge sites for highly localized flow systems that can not be simulated in regional models. The result is that apparent ability of the model to simulate spring discharge varies widely. Of seventeen springs, six display a good comparison between measured and simulated discharge values. Simulated spring discharge from springs with the highest measured discharge values differ from measured values by about an order of magnitude. Most springs in the study area represent discharge from highly localized flow systems within the aquifer system that are characterized by short flow paths. The localized nature of these flow paths and t he limitations of the regional model grid, result in much of the spring discharge being included in baseflow discharge to streams. Overall, the model also does a good job of mimicking baseflow fluctuations (Figure 1006). 10.2 Sensitivity Analysis Upon c ompletion of transient model calibration, we assessed the storage parameters to determine the sensitivity of the model to variation of specific yield and specific storage values . Sensitivity analy sis involves systematically varying specific yield and speci fic storage to determine associated changes in aquifer response over the transient model run. We ran the model multiple times , lowering and then raising the calibrated specific yield and specific storage values by an order of magnitude. Sensitivity analysis indicates that the unconfined Edwards Group (Layer 1) is sensitive to increasing specific yield input values and insensitive to specific storage input values ( Figure s 1007 and 1008). This is not surprising because MODFLOW only utilizes specific yield input values when simulating groundwater flow through an unconfined aquifer. Overall, the model is much more sensitive to specific yield than specific storage .

PAGE 33

32 11.0 Limitations of the Model All numerical groundwater flow models have limitations. These limitations are usually associated with (1) the extent of current understanding of the workings of the aquifer, (2) availability and accuracy of input data, (3) assumptions and simplifications used in developing the conceptual and numerical models , and (4) t he scale of application of the model . The limitations determine the spatial and temporal variation of uncertainties in the model because calibration uncertainty decreases with increased availability of input data. Additionally, many of the assumptions, deg ree of simplification, and spatial resolution of groundwater flow models are influenced by availability of input data. 1 1 . 1 Input D ata Several of the input data sets for the model are based on limited information. These include structural geology, recharg e, water level data, hydraulic conductivity, specific storage, and specific yield. Although this model’s representation of aquifer hydraulic propoerties may be adequate for the regional model, it may not be appropriate for local scale conditions. The same problem occurs in the assigning of specific storage and specific yield values in the model. The paucity of measured specific storage and specific yield values is overcome partially by calibrating the model based on observed water level responses in the we lls in the model area with the most water level measurements over the model period. There is no published information on the spatial distribution of recharge throughout the Hill Country portion of the Trinity A quifer Sy s tem . Calibrat ion of recharge rates i s obtained by trial and error during construction of the steady state model . Application of these recharge rates to the transient model assumes that (1) a linear relationship exists between precipitation and recharge and (2) there is no threshold that must be exceeded before recharge occurs. This assumption suggests the possibility of overestimating recharge during dry periods, when all precipitation may be taken up by evapotranspiration or absorbed by dry soils. The relatively good correlation between obse rved and simulated water levels and stream discharge suggests that, despite uncertainties, the model water budget reasonably represents the regional groundwater budget. Our structur al maps simplify faulting along the southeastern margin of the model and s mooth out the base of the Middle Trinity A quifer in the northern part of the model. This simplification causes the model represent the regional structural control s and regional groundwater flow, but limits the ability of simulating local groundwater flow in these areas. Greater structural control may be attained with more detailed maps and a finer model grid in this area. However, this increased complexity would come at the cost of the requirement of a finer model grid and consequently much longer run time s and increased computational complexity resulting increased instability of the model with no guarantee of increased model accuracy. Water level maps, and therefore the calibration of the model, are affected by limited information, especially in layer 1 wh ere there are few measurements. Limited availability of wells with multiple water levels measurements affect s calibration of the transient model. Limited water level measurements bias model calibration to areas where water levels have been measured. The di fference between measured and simulated water levels can be accounted for by factors such as

PAGE 34

33 unavoidable simplifications incorporated in to the model, and water levels measurements not representative of the average water level for a specific period of time simulated by the model. 1 1 . 2 Assumptions We used several assumptions to simplify construction of the model. The most important assumptions are: (1) there is no flow between the Lower Trinity A quifer and underlying Paleozoic units , (2) the Drain Package o f MODFLOW can be used to simulate discharge to streams and rivers , (3) the G eneral H ead B oundary package of MODFLOW can be used to simulate cross formational flow between the Hill Country portion of the Trinity Aquifer System and the Edwards (Balcones Faul t Zone) Aquifer , and (4) recharge from Cibolo Creek is constant over time . We assumed that the vertical leakance between the Middle and Lower Trinity aquifers is a function of the thickness of the Hammett shale. Most of the base of the Middle Trinity Aqui fer is underlain by the Hammett Shale (Amsbury, 1974; Barker and Ardis, 1996), and restricts flow between the Middle and Lower Trinity aquifers (Ashworth, 1983). We used the Drain Package of MODFLOW to simulate streams and rivers in the study area. The Dr ain Package only allows water to move from the aquifer to the streams and rivers , thus implying that the streams and rivers in the study area are gaining streams and will remain so in the future. We used t he G eneral H ead B oundary package to simulate cross formational flow between the Hill Country portion of the Trinity Aquifer System and the Edwards (Balcones Fault Zone) Aquifer . The spatial distribution of general head boundary cells in the model is based on the assumption that cross formational flow take s place where the two aquifer s juxtapose along the Balcones Fault Zone. We also assumed that there is no groundwater flow from the Lower Trinity Aquifer to the Trinity rocks underlying the Edwards (Balcones Fault Zone) Aquifer. Annual fluctuations in recha rge from Cibolo Creek are small enough during the transient model period to not affect calibration, thus allowing the use of constant recharge. However, during periods of extreme drought, it is likely that recharge from Cibolo Creek will decline and eventu ally cease. Consequently, predictive model runs that include periods of lower precipitation and streamflow (e.g. drought of record ) should include reduced recharge in this area. 1 1 . 3 Scale of Application The limitations described earlier and the nature o f regional groundwater flow models affect s the scale of application of the model. As calibrated , t his model is most accurate in assessing regional scale groundwater issues, such as predicting aquifer wide water level changes and trends in the groundwater budget that may result from different proposed water management strategies , on an annual timescale. Accuracy and applicability of the model decreases when moving from addressing regional to local scale issues because of limitations of the information used in model construction and the model cell size that determines spatial resolution of the model. Consequently, this model is not likely to accurately predict water level declines associated with a single well or spring because (1) these water level declines depend on site specific hydrologic properties not included in detail in regional scale models and (2) the cell size

PAGE 35

34 used in the model is too large to resolve changes in water levels that occur over relatively short distances. Addressing local scale issues requires a more detailed model, with local estimates of hydrologic properties, or an analytical model. This model is more useful in determining the impacts of groups of wells or well fields distributed over a few square miles. The model can be used to pred ict changes in ambient water levels rather than actual water level changes at specific location s , such as an individual well. 12.0 Future Improvements The Texas Water Development Board plans periodically to update, and thus improve, its groundwater avai lability models. This model may be improved by incorporating greater complexity or hydrologic information that was not available when it was updated . Model uncertainty may be reduced with additional information on streamflow, hydraulic properties, water le vel elevations, and recharge. Additional hydraulic head measurements and aquifer test data are required for the Hill Country portion of the Trinity A quifer Sy s tem . This information can be used to improve calibration of the model by increasing the number a nd spatial distribution of sites and the frequency of measurements for comparing measured and simulated water levels. Aquifer tests will facilitate determination of whether improving the model by more complex spatial distribution of hydraulic conductivity, specific storage, and specific yield can be justified. Future updates of this model might include using the Stream flow Routing Package (Prudic, 1989) to simulate streams. Using the Streamflow Routing Package would simulate two way interaction between t he aquifer and rivers or streams. This is a potentially superior alternative to the Drain P ackage and may allow better simulation of recharge from Cibolo Creek. 13.0 Conclusions We updated a finite difference groundwater flow model that can be used to predict water level changes in response to specified pumping and drought s cenarios . The updated model has four layers — the Edwards Group, and the Upper , Middle, and Lower Trinity aquifer s — and 12,976 active cells , each with a uniform grid size of 1 mile b y 1 mile . We developed the conceptual model of groundwater flow and defined aquifer properties based on a review of previous work and studies we conducted on water levels, structure, recharge, and hydraulic properties. The process of updating the model inc luded: (1) adding the Lower Trinity Aquifer as an additional layer to the model, (2) revising the structure and spatial distribution of parameters, such as recharge and pumping, and (3) calibrating to steady state conditions for 1980 and historical transie nt conditions for the period 1980 through 1997. The calibrated model does a reasonable job of matching the water level distribution and water level fluctuations in the aquifer . The steady state model has an overall mean absolute error of 54 f eet, about 3.5 percent of the hydraulic head drop across the study area. Calibration of the steady state model indicates in an average recharge rate of about 5 percent of average annual

PAGE 36

35 precipitation in the Balcones Fault Zone portion of the aquifer and 3.5 percent in the remainder of the aquifer. Estimated recharge from Cibolo Creek averages about 70,000 acre feet per year. Calibrated hydraulic conductivity is 11 feet per day in the Edwards Group, 9 to 150 feet per day in the Upper Trinity Aquifer , 7.6 to 15 feet per d ay in the Middle Trinity Aquifer , and 1.7 to 17 f eet per day in the Lower Trinity A quifer. Water levels in the model are most sensitive to changes in (1) recharge, (2) horizontal hydraulic conductivity, and (3) vertical hydraulic conductivity. We also cali brated values of vertical hydraulic conductivity, specific storage, and specific yield for the aquifer. We found that over 300,000 acre feet per year of water flows through the aquifer, mostly in the Upper and Middle Trinity aquifers. Of the total flow, al most all is derived from infiltration of precipitation, with minor amounts from inflow from reservoirs and the adjacent Edwards (Balcones Fault Zone) Aquifer. The model estimates that about 100,000 acre f eet per y ea r of ground water flow s from the Upper and Middle Trinity aquifers to the Edwards (B alcones F ault Z one ) A quifer. 14.0 Acknowledgements In this modeling effort, many people should be acknowledged for their assistance. I would like to thank the Bandera County River Authority and Groundwater Distr ict, Blanco Pedernales Groundwater Conservation District, Cow Creek Groundwater Conservation District, Edwards Aquifer Authority, Hays Trinity Groundwater Conservation District, Headwaters Groundwater Conservation District, Hill Country Underground Water C onservation District, and Trinity Glen Rose Groundwater Conservation District for their support and interest during the model development. A number of current and previous Texas Water Development Board staff members have also been helpful in providing assi stance and advice, including Roberto Anaya, Ali Chowdhury, Scott Hamlin, Cindy Ridgeway, and Shirley Wade. The support of Texas Water Development Board management, including Robert Mace and Bill Hutchison has been helpful in ensuring the successful complet ion of this project. 15.0 References Amsbury, D. L., 1974, Stratigraphic petrology of Lower and Middle Trinity rocks on the San Marcos platform, southcentral Texas: in Perkins, B. F., ed., Aspects of Trinity division geology a symposium: Louisiana Sta te University, Geoscience and Man, v. 8, p. 135. Anaya, R., and Jones, I. C., 2009, Groundwater availability model for the Edwards Trinity (Plateau) and Pecos Valley Aquifer System s, Texas: Texas Water Development Board report 373, 103 p. Anderson, M. P., and Woessner, W. W., 2002, Applied groundwater modeling: simulation of flow and advective transport: Academic Press, San Diego, CA, 381 p.

PAGE 37

36 Ardis, A. F., and Barker, R. A., 1993, Historical saturated thickness of the Edwards Trinity Aquifer System and sele cted contiguous hydraulically connected units, west central Texas: U.S. Geological Survey Water Resources Investigation Report 924125, 2 plates. Arnold, J. G., Allen, P. M., Muttiah, R., and Bernhardt, G., 1995, Automated base flow separation and recessio n analysis techniques: Ground Water, v. 33, no. 6, p. 10101018. Ashworth, J. B., 1983, Groundwater availability of the Lower Cretaceous formations in the Hill Country of southcentral Texas: Texas Department of Water Resources Report 273, 173 p. Ashworth , J. B., and Hopkins, J ., 1995, Aquifers of Texas: Texas Water Development Board Report 345, 69 p. Autodesk, 1997, AutoCad Map 2.0, Autodesk, Inc., San Rafael, CA. Barker, R. A., and Ardis, A. F., 1996, Hydrogeologic framework of the Edwards Trinity Aquife r System , west central Texas: U.S. Geological Survey Professional Paper 1421 B, 61 p. with plates. Barker, R. A., Bush, P. W., and Baker, E. T., Jr., 1994, Geologic history and hydrogeologic setting of the Edwards Trinity Aquifer System , west central Texas : U.S. Geological Survey Water Resources Investigation Report 944039, 50p. Barnes, V. E., 1981, Geologic Atlas of Texas Llano Sheet: Bureau of Economic Geology, The University of Texas at Austin. Barrett, M. E., and Charbeneau, R. J., 1996, A parsimonious model for simulation of flow and transport in a karst aquifer: Center for Research in Water Resources, The University of Texas at Austin, Technical Report CRWR 269, 149 p. Bluntzer, R. L., 1992, Evaluation of Groundwater Resources of the Paleozoic and Cretaceous Aquifers in the Hill Country of Central Texas: Texas Water Development Board Report 339, 130 p. Bradley, R. G., Coker, D. B., and Moore, S. W., 1997, Data and results from an aquifer test performed at the Medina Water Supply Corporation well fie ld, Medina, Texas: Texas Water Development Board OpenFile Report 9702, 25p. Brown, T. E., Waechter, N. B., and Barnes, V. E., 1974, Geologic Atlas of Texas San Antonio Sheet: Bureau of Economic Geology, The University of Texas at Austin. Brune, Gunnar, 1981, Springs of Texas, Volume 1: BranchSmith Inc., Fort Worth, Texas, 566 p. Carr, J. T., Jr., 1967, The climate and physiography of Texas: Texas Water Development Board Report 53, 27 p. Chiang, W., 2005, 3D groundwater modeling with PMWIN: A Simulation System for Modeling Groundwater Flow and Transport Processes : Springer, New York, 2nd ed., 398 p. Cooper, H. H., Jr., and Jacob, C. E., 1946, A generalized graphical method for evaluating formation constants and summarizing well field history: American Ge ophysical Union Transactions, v. 27, p. 526534. Daly, C. and G. Taylor, 1998. PRISM Briefing Book and Questionnaire, A Description of the PRISM Model for Spatially Distributing Observed Precipitation: Natural Resources

PAGE 38

37 Conservation Service Water and Clima te Center and Oregon Climate Service, Oregon State University. ( http://www.ocs.orst.edu/prism/nrcs_brief.html ). Domenico, P. A., and Schwartz, F. W., 1990, Physical and chemical hydrogeology: J ohn Wiley & Sons, Inc., New York, NY, 824 p. Espey, Huston, and Associates, 1982, Feasibility study of recharge facilities on Cibolo Creek: Draft consulting report to the Edwards Underground Water District, 56 p. Espey, Huston, and Associates, 1989, Medina Lake hydrology study: Report to the Edwards Underground Water District, 56 p. ESRI, 1991, ARC/INFO, Version 6.0: Environmental Systems Research Institute, Inc., Redlands, California. ESRI, 2005, A rcGIS , Version 9.1: Environmental Systems Research Institute, Inc., Redlands, California. George, W. O., 1947, Geology and groundwater resources of Comal County, Texas: Texas Board of Water Engineers, 142 p. Golden Software, 1995, Surfer for Windows, Version 6, Contouring and 3D surface mapping: Golden Software, Inc., Golden, Colorado, variously paginated. Guyton, W. F., and Associates, 1958, Memorandum on groundwater gains in upper Cibolo Creek area: Consulting report to the San Antonio City Water Board, 8 p. Guyton, W. F., and Associates, 1970, Memorandum on Ci bolo Creek studies: Consulting report to the San Antonio City Water Board, 17 p. Guyton, W. F. and Associates, 1993, Northern Bexar County water resources study for the Edwards Underground Water District Volume 1: Ground water: final report to the Edwards Underground Water District, San Antonio, Texas, 66 p., tables and figures. Hammond, W. W., 1984, Hydrogeology of the Lower Glen Rose Aquifer, SouthCentral Texas: Ph.D. dissertation, The University of Texas at Austin. 243 p. Harbaugh, A . W.; and McDonald, M. G. , 1996, User’s documentation for MODFLOW 96, an update to the U.S. Geological Survey modular finite difference groundwater flow model: U.S. Geological Survey Open File Report 96485, 56 p. Hovorka, S.D., Dutton, A.R., Ruppel, S.C., and Yeh, J.S., 1996, Edwards aquifer groundwater resources —Geologic controls on porosity development in platform carbonates, South Texas: The University of Texas at Austin, Bureau of Economic Geology Report of Investigations 238, 75 p. Kalaswad, S ., and Mills, K. W., 2000, Evaluation of northern Bexar County for inclusion in the Hill Country Priority Groundwater Management Area: Austin, Texas, Priorit y Groundwater Managemen t Area Report, Texas Natural Resources Conservation Commission, 82 p. Kastning, E. H., 1986, Cavern development in the New Braunfels area, central Texas: in Abbott, P. L., and Woodruff, C. M., Jr., eds., The Balcones escarpment, geology, hydrology, ecology and social development in central Texas: Geological Society of America, p. 91 100.

PAGE 39

38 Klemt, W. B., Knowles, T. R., Elder, G. R., and Sieh, T. W., 1979, Groundwater resources and model applications for the Edwards (Balcones Fault Zone) Aquifer in the San Antonio region, Texas: Texas Department of Water Resources Report 239, 88 p. Kruseman, G. P., and de Ri dder, N. A., 1994, Analysis and evaluation of pumping test data, second edition: International Institute for Land Reclamation and Improvement, The Netherlands, 377 p. Kuniansky, E.L., 1989, Precipitation, streamflow, and baseflow, in west central Texas, De cember 1974 through March 1977: U.S. Geological Survey Water Resources Investigations Report 894208, 2 sheets. Kuniansky, E. L., 1990, Potentiometric surface of the Edwards Trinity Aquifer System and contiguous hydraulically connected units, west central Texas, winter 197475: U.S. Geological Survey Water Resources Report 894208, 2 sheets. Kuniansky, E. L., and Holligan, K. Q., 1994, Simulations of flow in the Edwards –Trinity Aquifer System and contiguous hydraulically connected units, west central Texas: U. S. Geological Survey Water Resources Investigations Report 934039, 40 p. LBG Guyton Associates, 1995, Edwards/Glen Rose hydrologic communication, San Antonio region, Texas: final report submitted to the Edwards Underground Water District, 38p. with 7 tables, 13 figures and 4 appendices. Larkin, T. J., and Bomar, G. W., 1983, Climatic atlas of Texas: Texas Water Development Board, Limited Publication 192, 151 p. Lindgren, R.J., Dutton, A.R., Hovorka, S.D., Worthington, S.R.H., and Painter, Scott, 2004, Conceptualization and simulation of the Edwards Aquifer , San Antonio region, Texas: U.S. Geological Survey Scientific Investigations Report 2004–5277, 143 p. Long, A. T., 1962, Groundwater geology of Edwards County, Texas: Texas Water Commission Bulletin 6208, 123 p. Lowry, R. L., 1955, Recharge to Edwards ground water reservoir: Consulting report to the San Antonio City Water Board, 66 p. Lurry, D. L., and Pavlicek, D. J., 1991, Withdrawals from the Edwards Trinity Aquifer System and contiguous hydraulica lly connected units, west central Texas, December 1974 through March 1977: U.S. Geological Survey Water Resources Investigations Report 914021, 1 sheet. Mace, R. E., 2001, Estimating transmissivity using specific capacity data: The University of Texas at Austin, Bureau of Economic Geology, Geological Circular No. 01 2, 44 p. Mace, R. E., Chowdhury, A. H., Anaya, R., and Way, S.C., 2000, Groundwater availability of the Trinity Aquifer , Hill Country area, Texas: numerical simulations through 2050: Texas Wat er Development Board Report 353, 117 p. Maclay, R. W., and Land, L. F., 1988, Simulation of flow in the Edwards Aquifer , San Antonio Region, Texas, and refinements of storage and flow concepts: U. S. Geological Survey Report Water Supply Paper 2336, 48 p. Maclay, R. W., and Small, T. A., 1986, Carbonate geology and hydrology of the Edwards Aquifer in the San Antonio area, Texas: Texas Water Development Board Report 296, 90 p.

PAGE 40

39 McDonald, M. G., and Harbaugh, A. W., 1988, A modular three dimensional finite dif ference ground water flow model: U.S. Geological Survey, Techniques of Water Resources Investigations of the United States Geological Survey, Book 6: Model Techniques, Chapter A1. Meyers, B.N., 1969, Compilation of results of aquifer tests in Texas: Texas Water Development Board Report 98, 532 p. Muller, D. A. and Price, R. D., 1979, Groundwater Availability in Texas, Estimates and Projections Through 2030: Texas Department of Water Resources Report 238, 77 p. Nathan, R. J., and McMahon, T. A., 1990, Evaluation of automated techniques for baseflow and recession analysis: Water Resources Research, v. 26, no. 7, p. 14651473. Ockerman, D. J., 2007, Simulation of streamflow and estimation of ground water recharge in the upper Cibolo Creek watershed, SouthCent ral Texas, 19922004: U.S. Geological Survey Scientific Investigations Report 20075202, 34 p. Pearson, F. J., Jr., Rettman, P. L., and Wyerman, T. A., 1975, Environmental tritium in the Edwards Aquifer, central Texas, 196371: U.S. Geological Survey OpenFile Report 74362, 32 p. Proctor, C. V., Jr., Brown, T. E., McGowen, J. H., Waechter, N. B., and Barnes, V. E., 1974a, Geologic Atlas of Texas Austin Sheet: Bureau of Economic Geology, The University of Texas at Austin , 1 sheet . Proctor, C. V., Jr., Brown, Waechter, N. B., Aronow, S., and Barnes, V. E., 1974b, Geologic Atlas of Texas Seguin Sheet: Bureau of Economic Geology, The University of Texas at Austin. Prudic, D. E., 1989, Documentation of a computer program to simulate streamaquifer relations using a modular, finite difference, groundwater flow model : U.S. Geological Survey OpenFile Report 88729, 113 p. Riggio, R. F., Bomar, G. W., and Larkin, T. J., 1987, Texas drought its recent history (19311985): Texas Water Commission LP 8704, 74 p. Riskind, D. H., and Diamond, D. D., 1986, Plant communities of the Edwards Plateau of Texas: an overview emphasizing the Balcones Escarpment zone between San Antonio and Austin with special attention to landscape contrasts and natural diversity. In: Abbott , P. L. and Woodruff, C. M., Jr. (eds.), The Balcones Escarpment, central Texas: Geological Society of America, p. 2132. Senger, R. K., and Kreitler, C. W., 1984, Hydrogeology of the Edwards Aquifer, Austin area, Central Texas: The University of Texas at Austin, Bureau of Economic Geology Report of Investigations No. 141, 35 p. Simpson, W. E., Company Limited, and Guyton, W. F., and Associates, 1993, North Bexar County water resources study for the Edwards Underground Water District Executive summary: San Antonio, Texas, final report by W. E. Simpson Company, Inc. in association with William F. Guyton and Associates, Inc. for the Edwards Underground Water District, variously paginated. Slade, R. M., Jr., Ruiz, Linda, and Slagle, Diana, 1985, Simulation of the flow system of Barton Springs and associated Edwards Aquifer in the Austin area, Texas: U.S. Geological Survey Water Resources Investigations Report 854299.

PAGE 41

40 Slade, R. M., Jr., 2002, Results of streamflow gain loss studies in Texas, with emphasis on ga ins from and losses to major and minor aquifers: U.S. Geological Survey OpenFile report 02 068. Spatial Climate Analysis Service, 2004, Near real time monthly high resolution precipitation climate data set for the conterminous United States (2.5 arc minut e 19712000 average monthly precipitation grids for the conterminous United States): Corvallis, OR, Spatial Climate Analysis Service, Oregon State University, arc grid file, http://www.ocs.oregonstate.edu/prism/ , accessed March 2006. Stein, E. G., and Klem t, W. B., 1995, Edwards/Glen Rose hydrologic communication, San Antonio region, Texas: report prepared for the Edwards Underground Water District by LBG Guyton Associates, Austin, TX, variously paginated. Theis, C. V., 1935, The relation between the loweri ng of the piezometric surface and the rate and duration of discharge of a well using groundwater storage: American Geophysical Union Transaction, v. 16, p. 519 –524. Theis, C. V., 1963, Estimating the transmissivity of a water table aquifer from the specifi c capacity of a well: U.S. Geological Survey Water Supply Paper 1536I, p. 332336. Thornbury, W. D., 1965, Regional geomorphology of the United States : New York, Wiley , 609 p. Veni, G., 1994, Geomorphology, hydrology, geochemistry, and evolution of karsti c Lower Glen Rose Aquifer , southcentral Texas: Pennsylvania State University, Ph.D. Dissertation, 712 p. Veni, G., 1995, Revising the boundaries of the Edwards (Balcones Fault Zone) Aquifer recharge zone: Water for Texas, Proceedings of the 24th Water for Texas Conference, p. 99107. Walker, L. E., 1979, Occurrence, availability, and chemical quality of ground water in the Edwards Plateau Region of Texas: Texas Department of Water Resources Report 235, 336 p. Wanakule, Nisai, and Anaya, Roberto, 1993, A lu mped parameter model for the Edwards Aquifer : Texas A&M University, Texas Water Resources Institute, Technical Report No. 163, 84 p. Waterreus, P. A., 1992, Hydrogeology of the Camp Bullis area, northern Bexar County, Texas: Master’s thesis, The University of Texas at San Antonio, 186 p. Wet Rock Groundwater Services, 2008, An evaluation of the Trinity Aquifer within Kendall County and analysis of the Trinity (Hill Country ) GAM: Unpublished report prepared for Cow Creek Groundwater Conservation District, 47 p. Woodruff, C. M., Jr., and Abbott, P. L., 1986, Stream piracy and evolution of the Edwards Aquifer along the Balcones escarpment, Central Texas: in Abbott, P. L., and Woodruff, C. M., Jr., The Balcones Escarpment, Central Texas: Geological Society of A merica, p. 7790. Young, K., 1972, Mesozoic history, Llano region: in Barnes, V. E., Bell, W. C., Clabaugh, S. E., and Cloud, P. E., eds., Geology of the Llano region and Austin area, field excursion, The University of Texas at Austin, Bureau of Economic G eology Guidebook 13, 77 p.

PAGE 42

41 Table 5 01. Estimates of recharge rates to the Hill Country portion of the Trinity Aquifer System as a percentage of average annual precipitation . Literature s ource Recharge rate (inches per year) Percent value Muller and Pric e (1979) 0.5 1.5 Ashworth (1983) 1.3 4.0 Kuniansky (1989) 3.6 11.0 Kuniansky and Holligan (1994) 2.3 7.0 Bluntzer (1992, calc.) 2.2 6.7 Bluntzer (1992, est.) 1.7 5.0 Mace (200 1 ) 2.2 6.6 Mace and others (2000) 1.3 4.0 Wet Rock Groundwater Services (2008) 3.1 9.5 Anaya and Jones (2009) 1.4 4.7

PAGE 43

42 Table 5 02. Estimated flow for selected springs in the study area (see Figure 528 ) (from Mace and others, 2000) . Spring Est imated f low (g allons per m inute ) Formation Remarks 1 150 Edwards Grou p and associated limestone Measured on 4/13/67 2 100 Edwards Group and associated limestone Measured on 4/12/67, reported flow never ceased 3 100 Edwards Group and associated limestone 4 2,500 Edwards Group and associated limestone Measured on 3/31/66, reported flow never ceased 5 310 Edwards Group and associated limestone Measured on 3/11/70 6 480 Edwards Group and associated limestone Measured on 3/11/70, owner’s trough spring 7 100 Edwards Group and associated limestone Measured on 6/15/66, never ceased flowing 8 20 Upper member of the Glen Rose Limestone Measured on 7/13/76 9 75 Lower member of the Glen Rose Limestone Measured on 7/10/75, ceased flowing in 1956 10 50 Lower member of the Glen Rose Limestone Measured on 1/17/40 11 150 Lower memb er of the Glen Rose Limestone Measured on 7/17/75, owners well #9 12 300 Lower member of the Glen Rose Limestone 13 300 Cow Creek Limestone Measured on 7/11/75 14 500 Cow Creek Limestone Measured on 8/31/76, estimated flow 1,070 gpm, Jan. 1955 15 25 L ower member of the Glen Rose Limestone Measured on 1/1/66 16 50 Upper member of the Glen Rose Limestone Measured on 12/30/88, Bassett springs 17 50 Upper member of the Glen Rose Limestone Measured on 5/25/73 18 9,000 Edwards Group and associated limesto ne Measured on 12/20/60 19 5,000 Lower member of the Glen Rose Limestone Measured on 8/20/91, springs discharge into Medina River

PAGE 44

43 Table 5 03. Total pumping from the Hill Country portion of the Trinity Aquifer System for each county for the period 1980 through 1997 (All values are acrefeet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall 1980 1,084 4,120 195 1,135 1,223 1,621 1,585 1981 1,077 4,280 234 1,076 1,235 1,788 1,690 1982 1,120 4,486 230 998 1,248 1,903 1,663 1983 1,129 3, 875 224 978 1,260 2,046 1,829 1984 1,182 4,359 217 916 1,273 2,059 2,115 1985 1,175 3,892 261 918 1,289 2,087 1,781 1986 1,154 4,165 312 949 1,332 2,018 1,793 1987 1,290 4,775 333 987 1,273 1,817 1,518 1988 1,374 5,774 350 1,035 1,289 1,865 2,337 198 9 1,441 5,900 367 1,058 1,421 2,116 2,343 1990 1,462 7,372 386 1,080 1,440 2,093 2,185 1991 1,529 6,098 388 1,128 1,484 2,096 1,751 1992 1,528 6,227 422 1,200 1,558 2,125 1,728 1993 1,784 6,249 432 1,125 1,633 2,506 2,414 1994 1,684 6,609 413 1,199 2, 308 2,539 2,482 1995 1,723 6,767 453 1,214 2,329 2,719 2,823 1996 1,709 6,814 465 1,112 2,615 2,935 3,092 1997 1,785 6,832 472 1,268 2,297 2,923 3,738 Year Kerr Kimble Medina Travis Uvalde Total 1980 5,994 7 63 111 11 17,148 1981 3,463 7 60 108 11 1 5,027 1982 3,176 6 57 101 11 15,000 1983 2,954 6 53 100 11 14,466 1984 3,517 5 50 96 11 15,799 1985 3,529 5 45 100 11 15,093 1986 3,104 7 45 110 10 14,999 1987 2,727 6 49 111 10 14,896 1988 3,135 6 49 116 10 17,342 1989 3,433 5 49 116 10 18,259 19 90 3,263 5 50 117 10 19,461 1991 3,282 5 51 125 10 17,945 1992 3,787 5 57 127 11 18,775 1993 4,161 5 66 139 11 20,525 1994 3,962 5 60 134 11 21,406 1995 3,886 6 64 138 11 22,133 1996 4,439 6 62 200 12 23,460 1997 4,095 5 59 146 11 23,631

PAGE 45

44 Table 5 04. Total pumping from the Hill Country portion of the Trinity Aquifer System by use category for each county for the period 1980 through 1997 (All values are acre feet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Municipal 1980 190 157 0 0 0 573 380 3,491 0 0 0 0 4,791 1981 168 177 0 0 0 732 404 1,042 0 0 0 0 2,523 1982 198 245 0 0 0 834 424 735 0 0 0 0 2,436 1983 193 220 0 0 0 965 500 538 0 0 0 0 2,416 1984 232 380 0 0 0 964 700 1,036 0 0 0 0 3,312 1985 199 360 0 0 0 1,150 553 1,248 0 0 0 0 3,510 1986 222 612 0 0 0 1,062 582 925 0 0 0 0 3,403 1987 204 645 0 0 0 825 449 506 0 0 0 0 2,629 1988 227 761 0 0 0 834 712 830 0 0 0 0 3,364 1989 297 869 0 0 0 1,076 737 1,02 3 0 0 0 0 4,002 1990 269 719 0 0 0 1,019 632 720 0 0 0 0 3,359 1991 275 612 0 0 0 979 378 658 0 0 0 0 2,902 1992 219 719 0 0 0 962 322 1,035 0 0 0 0 3,257 1993 298 719 0 0 0 1,220 412 1,178 0 0 0 0 3,827 1994 340 1,071 0 0 0 1,281 474 924 0 0 0 0 4,09 0 1995 322 1,213 0 0 0 1,317 566 867 0 0 0 0 4,285 1996 299 1,213 0 0 0 1,485 746 1,363 0 0 0 0 5,106 1997 331 1,213 0 0 0 1,432 999 965 0 0 0 0 4,940 Manufacturing 1980 0 2,449 0 0 0 0 0 0 0 0 0 0 2,449 1981 0 2,449 0 0 0 0 0 0 0 0 0 0 2,449 1982 0 2,449 0 0 0 0 0 0 0 0 0 0 2,449 1983 0 1,727 0 0 0 0 0 0 0 0 0 0 1,727 1984 0 1,912 0 0 0 0 0 0 0 0 0 0 1,912 1985 0 2,516 0 0 0 0 0 0 0 0 0 0 2,516 1986 0 2,516 0 0 0 0 0 0 0 0 0 0 2,516 1987 0 3,085 0 0 0 0 0 0 0 0 0 0 3, 085 1988 0 3,949 0 0 1 0 0 0 0 0 0 0 3,950 1989 0 3,949 0 0 0 0 0 0 0 0 0 0 3,949 1990 0 5,549 0 0 0 0 0 0 0 0 0 0 5,549 1991 0 4,363 0 0 0 0 0 0 0 0 0 0 4,363 1992 0 4,363 0 0 0 0 0 4 0 0 0 0 4,367 1993 0 4,363 0 0 0 0 0 7 0 0 0 0 4,370 1994 0 4,37 0 0 0 0 0 0 7 0 0 0 0 4,377 1995 0 4,370 0 0 0 0 0 7 0 0 0 0 4,377 1996 0 4,370 0 0 0 0 0 6 0 0 0 0 4,376 1997 0 4,370 0 0 0 0 0 7 0 0 0 0 4,377

PAGE 46

45 Table 5 04. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis U valde Total Pumpage Rural Domestic 1980 570 878 39 557 832 624 564 1,654 0 21 34 7 5,780 1981 598 897 85 581 854 663 652 1,619 0 21 36 7 6,013 1982 626 915 88 587 877 705 613 1,687 0 22 35 7 6,162 1983 654 930 87 650 899 747 710 1,709 0 22 39 7 6,454 1984 683 948 87 672 922 791 803 1,820 0 22 40 7 6,795 1985 710 966 138 697 945 832 770 1,813 0 23 41 7 6,942 1986 739 984 177 728 967 874 808 1,844 0 23 48 7 7,199 1987 766 1,001 198 755 989 916 643 1,865 0 23 54 7 7,217 1988 794 1,019 210 778 1,012 959 909 1,916 0 24 54 8 7,683 1989 822 1,036 213 803 1,035 997 963 1,969 0 24 55 8 7,925 1990 850 1,054 215 828 1,057 1,031 968 2,108 0 25 54 8 8,198 1991 908 1,073 214 870 1,080 1,073 779 2,179 0 26 61 8 8,271 1992 964 1,091 225 916 1,102 1 ,132 722 2,222 0 27 67 8 8,476 1993 1,022 1,110 235 843 1,124 1,249 787 2,266 0 28 70 8 8,742 1994 1,078 1,128 245 905 1,146 1,217 904 2,309 0 29 77 8 9,046 1995 1,135 1,147 268 909 1,168 1,361 1,075 2,352 0 30 81 8 9,534 1996 1,193 1,165 304 859 1,190 1,418 1,234 2,396 0 31 82 8 9,880 1997 1,249 1,184 307 1,016 1,213 1,462 1,632 2,439 0 32 91 8 10,633 Irrigation 1980 62 611 47 368 52 102 200 500 4 0 0 0 1,946 1981 58 734 45 279 70 89 221 469 4 0 0 0 1,969 1982 54 857 43 190 88 76 241 437 4 0 0 0 1,990 1983 50 979 40 101 105 63 262 406 4 0 0 0 2,010 1984 47 1,102 38 12 123 50 282 374 3 0 0 0 2,031 1985 68 0 28 0 111 64 132 204 4 0 0 0 611 1986 10 0 28 0 93 44 176 136 5 0 0 0 492 1987 124 0 28 0 30 35 176 136 5 0 0 0 5 34 1988 124 0 28 0 8 29 440 136 4 0 0 0 769 1989 95 0 41 0 127 0 369 191 3 0 0 0 826 1990 115 0 47 0 113 0 274 187 3 0 0 0 739 1991 115 0 47 0 127 0 274 187 3 0 0 0 753 1992 115 0 47 0 127 0 274 187 3 0 0 0 753 1993 248 0 51 0 170 0 808 396 3 0 0 0 1 ,676 1994 15 0 51 10 845 0 718 406 3 0 0 0 2,048 1995 14 0 54 9 841 0 808 355 4 0 0 0 2,085 1996 15 0 54 10 957 0 808 396 4 0 0 0 2,244 1997 15 0 54 9 782 0 808 396 3 0 0 0 2,067

PAGE 47

46 Table 5 04. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Ke ndall Kerr Kimble Medina Travis Uvalde Total Pumpage Livestock 1980 262 25 109 210 339 322 441 349 3 42 78 4 2,184 1981 252 23 104 216 311 305 413 333 3 39 72 4 2,075 1982 241 21 100 221 283 288 386 318 3 35 66 4 1,966 1983 231 18 96 227 256 271 358 302 2 32 61 3 1,857 1984 221 16 92 232 228 254 330 286 2 28 55 3 1,747 1985 198 50 96 221 232 41 326 264 2 22 59 3 1,514 1986 184 53 108 221 272 38 228 199 2 22 62 2 1,391 1987 197 44 106 232 254 40 249 219 2 26 58 2 1,429 1988 229 46 112 257 268 43 276 253 2 25 62 2 1,575 1989 227 46 113 255 259 43 274 250 2 25 61 2 1,557 1990 228 50 124 252 269 42 312 248 2 25 62 2 1,616 1991 231 50 126 258 278 44 319 258 2 25 64 2 1,657 1992 231 54 150 284 330 31 410 338 2 30 60 3 1,923 1993 216 57 146 282 339 37 407 314 2 38 69 3 1,910 1994 251 40 118 284 317 41 386 317 2 31 57 3 1,847 1995 251 37 131 296 321 41 374 305 2 34 57 3 1,852 1996 203 66 107 243 468 32 303 278 2 31 118 4 1,855 1997 190 65 111 243 302 28 298 288 2 27 55 3 1,612

PAGE 48

47 Tabl e 5 05. Total pumping from the Edwards Group by use category for each county for the period 1980 through 1997 (All values are acre feet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Mun icipal 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 1992 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0 Manufacturing 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 19 84 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 19 92 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0

PAGE 49

48 Table 5 05. (cont.) Year Bandera Bexar Blanco Comal Gillespie H ays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Rural Domestic 1980 47 0 0 0 262 0 77 448 0 0 0 0 834 1981 49 0 0 0 269 0 89 439 0 0 0 0 846 1982 52 0 0 0 276 0 83 457 0 0 0 0 868 1983 54 0 0 0 283 0 96 463 0 0 0 0 896 1984 5 6 0 0 0 290 0 109 493 0 0 0 0 948 1985 59 0 0 0 297 0 104 492 0 0 0 0 952 1986 61 0 0 0 304 0 110 500 0 0 0 0 975 1987 63 0 0 0 311 0 87 506 0 0 0 0 967 1988 66 0 0 0 318 0 123 519 0 0 0 0 1,026 1989 68 0 0 0 326 0 131 534 0 0 0 0 1,059 1990 70 0 0 0 333 0 131 572 0 0 0 0 1,106 1991 75 0 0 0 340 0 106 591 0 0 0 0 1,112 1992 80 0 0 0 347 0 98 603 0 0 0 0 1,128 1993 84 0 0 0 354 0 107 614 0 0 0 0 1,159 1994 89 0 0 0 361 0 123 626 0 0 0 0 1,199 1995 94 0 0 0 368 0 146 638 0 0 0 0 1,246 1996 99 0 0 0 375 0 167 650 0 0 0 0 1,291 1997 103 0 0 0 382 0 221 661 0 0 0 0 1,367 Irrigation 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 1992 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0

PAGE 50

49 Table 5 05. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Livestock 1980 16 0 0 0 0 0 0 157 3 0 0 0 176 1981 16 0 0 0 0 0 0 150 3 0 0 0 169 1982 15 0 0 0 0 0 0 143 3 0 0 0 161 1983 15 0 0 0 0 0 0 136 2 0 0 0 153 1984 14 0 0 0 0 0 0 129 2 0 0 0 145 1985 12 0 0 0 0 0 0 119 2 0 0 0 133 1986 11 0 0 0 0 0 0 89 2 0 0 0 102 1987 12 0 0 0 0 0 0 98 2 0 0 0 112 1988 14 0 0 0 0 0 0 113 2 0 0 0 129 1989 14 0 0 0 0 0 0 112 2 0 0 0 128 1990 14 0 0 0 0 0 0 112 2 0 0 0 128 1991 15 0 0 0 0 0 0 116 2 0 0 0 13 3 1992 15 0 0 0 0 0 0 152 2 0 0 0 169 1993 14 0 0 0 0 0 0 141 2 0 0 0 157 1994 17 0 0 0 0 0 0 143 2 0 0 0 162 1995 17 0 0 0 0 0 0 137 2 0 0 0 156 1996 13 0 0 0 0 0 0 125 2 0 0 0 140 1997 12 0 0 0 0 0 0 130 2 0 0 0 144

PAGE 51

50 Table 5 06. Total pumping fr om the Upper Trinity Aquifer by use category for each county for the period 1980 through 1997 (All values are acre feet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Municipal 1980 0 0 0 0 0 0 33 0 0 0 0 0 33 1981 0 0 0 0 0 0 38 0 0 0 0 0 38 1982 0 0 0 0 0 0 38 0 0 0 0 0 38 1983 0 0 0 0 0 0 43 0 0 0 0 0 43 1984 0 0 0 0 0 0 67 0 0 0 0 0 67 1985 0 0 0 0 0 0 48 0 0 0 0 0 48 1986 0 0 0 0 0 0 46 0 0 0 0 0 46 1987 0 0 0 0 0 0 32 0 0 0 0 0 32 1988 0 0 0 0 0 0 67 0 0 0 0 0 67 1989 0 0 0 0 0 0 69 0 0 0 0 0 69 1990 0 0 0 0 0 0 57 0 0 0 0 0 57 1991 0 0 0 0 0 0 22 0 0 0 0 0 22 1992 0 0 0 0 0 0 10 0 0 0 0 0 10 1993 0 0 0 0 0 0 22 0 0 0 0 0 22 1994 0 0 0 0 0 0 31 0 0 0 0 0 31 1995 0 0 0 0 0 0 38 0 0 0 0 0 38 1996 0 0 0 0 0 0 65 0 0 0 0 0 65 1997 0 0 0 0 0 0 103 0 0 0 0 0 103 Manufacturing 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 1992 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0

PAGE 52

51 Table 5 06. (cont.) Year Bandera Bexar Bl anco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Rural Domestic 1980 409 865 25 345 79 559 375 1,205 0 21 32 7 3,922 1981 429 884 54 360 81 593 434 1,180 0 21 34 7 4,077 1982 449 902 56 363 84 632 407 1,229 0 22 33 7 4,184 1983 469 917 56 402 86 669 472 1,246 0 22 38 7 4,384 1984 490 934 55 416 88 708 534 1,327 0 22 39 7 4,620 1985 509 952 88 431 90 745 512 1,322 0 23 39 7 4,718 1986 530 969 113 450 92 782 537 1,344 0 23 46 7 4,893 1987 549 987 126 467 94 821 428 1,360 0 23 51 7 4,913 1988 570 1,004 134 482 96 859 604 1,396 0 24 52 8 5,229 1989 590 1,021 136 497 99 892 640 1,435 0 24 53 8 5,395 1990 610 1,038 137 512 101 923 643 1,536 0 25 52 8 5,585 1991 651 1,058 136 539 103 961 518 1,588 0 26 58 8 5,646 1992 692 1,075 143 567 105 1,013 480 1,620 0 27 64 8 5,794 1993 733 1,094 149 521 107 1,118 523 1,651 0 28 67 8 5,999 1994 773 1,112 156 560 109 1,089 601 1,683 0 29 73 8 6,193 1995 814 1,130 170 563 111 1,218 714 1,715 0 30 77 8 6,550 1996 855 1,148 193 532 113 1,269 821 1,746 0 31 78 8 6,794 1997 896 1,166 195 629 115 1,309 1,085 1,778 0 32 87 8 7,300 Irrigation 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 1992 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0

PAGE 53

52 Table 5 06. (cont.) Year Bandera Bex ar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Livestock 1980 227 25 95 155 257 298 299 192 0 42 74 4 1,668 1981 218 23 91 158 236 281 280 183 0 39 69 4 1,582 1982 209 21 88 161 215 264 261 175 0 35 63 4 1,496 1983 200 18 84 165 194 247 242 166 0 32 58 3 1,409 1984 192 16 80 168 173 230 223 157 0 28 53 3 1,323 1985 172 50 83 155 176 37 221 145 0 22 56 3 1,120 1986 160 53 94 155 206 35 154 109 0 22 60 2 1,050 1987 171 44 93 163 192 36 168 121 0 26 55 2 1,071 1988 199 46 98 181 203 39 187 140 0 25 59 2 1,179 1989 197 46 99 179 196 39 185 138 0 25 58 2 1,164 1990 197 50 108 177 204 38 211 136 0 25 59 2 1,207 1991 200 50 110 181 210 40 216 142 0 25 61 2 1,237 1992 200 54 131 200 250 28 277 186 0 30 57 3 1,416 1993 187 57 128 198 257 34 276 173 0 38 66 3 1,417 1994 217 40 103 200 240 37 261 174 0 31 54 3 1,360 1995 217 37 114 208 243 37 253 168 0 34 54 3 1,368 1996 175 66 94 171 354 29 205 153 0 31 113 4 1,395 1997 164 65 97 171 229 26 202 158 0 27 53 3 1,195

PAGE 54

53 Table 5 07. Total pumping from the Middle T rinity Aquifer by use category for each county for the period 1980 through 1997 (All values are acre feet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Municipal 1980 0 157 0 0 0 510 346 293 0 0 0 0 1,306 1981 0 177 0 0 0 666 366 200 0 0 0 0 1,409 1982 0 245 0 0 0 756 386 250 0 0 0 0 1,637 1983 0 220 0 0 0 869 457 262 0 0 0 0 1,808 1984 0 355 0 0 0 827 595 37 2 0 0 0 0 2,149 1985 0 341 0 0 0 1,003 469 355 0 0 0 0 2,168 1986 0 581 0 0 0 988 492 373 0 0 0 0 2,434 1987 0 613 0 0 0 724 353 318 0 0 0 0 2,008 1988 0 723 0 0 0 745 576 370 0 0 0 0 2,414 1989 0 830 0 0 0 981 596 409 0 0 0 0 2,816 1990 0 689 0 0 0 928 508 349 0 0 0 0 2,474 1991 0 587 0 0 0 882 293 347 0 0 0 0 2,109 1992 0 689 0 0 0 875 240 384 0 0 0 0 2,188 1993 0 691 0 0 0 1,098 316 441 0 0 0 0 2,546 1994 0 1,030 0 0 0 1,149 370 400 0 0 0 0 2,949 1995 0 1,166 0 0 0 1,218 442 349 0 0 0 0 3,175 1996 0 1,168 0 0 0 1,368 597 435 0 0 0 0 3,568 1997 0 1,169 0 0 0 1,313 817 356 0 0 0 0 3,655 Manufacturing 1980 490 0 0 0 0 0 0 0 0 0 0 0 490 1981 490 0 0 0 0 0 0 0 0 0 0 0 490 1982 490 0 0 0 0 0 0 0 0 0 0 0 490 1983 345 0 0 0 0 0 0 0 0 0 0 0 345 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 419 0 0 0 0 0 0 0 0 0 0 0 419 1986 359 0 0 0 0 0 0 0 0 0 0 0 359 1987 441 0 0 0 0 0 0 0 0 0 0 0 441 1988 564 0 0 0 1 0 0 0 0 0 0 0 565 1989 564 0 0 0 0 0 0 0 0 0 0 0 564 1990 793 0 0 0 0 0 0 0 0 0 0 0 793 1991 623 0 0 0 0 0 0 0 0 0 0 0 623 1992 623 0 0 0 0 0 0 4 0 0 0 0 627 1993 623 0 0 0 0 0 0 7 0 0 0 0 630 1994 624 0 0 0 0 0 0 7 0 0 0 0 631 1995 624 0 0 0 0 0 0 7 0 0 0 0 631 1996 624 0 0 0 0 0 0 6 0 0 0 0 630 1997 624 0 0 0 0 0 0 7 0 0 0 0 631

PAGE 55

54 Table 5 07. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Rural Domestic 1980 114 13 14 212 491 65 113 0 0 0 1 0 1,023 1981 120 13 31 222 504 69 130 0 0 0 1 0 1,090 1982 125 13 32 224 517 74 122 0 0 0 1 0 1,108 1983 131 14 32 248 531 78 142 0 0 0 1 0 1,177 1984 137 14 32 256 544 83 160 0 0 0 1 0 1,227 1985 142 14 50 266 557 87 154 0 0 0 1 0 1,271 1986 148 14 64 277 571 91 161 0 0 0 1 0 1,327 1987 153 15 72 288 584 96 128 0 0 0 1 0 1,337 1988 159 15 76 297 597 100 181 0 0 0 1 0 1,426 1989 165 15 77 306 611 104 192 0 0 0 1 0 1,471 1990 170 15 78 316 624 108 193 0 0 0 1 0 1,505 1991 182 16 78 332 637 112 155 0 0 0 2 0 1,514 1992 193 16 82 349 650 119 14 4 0 0 0 2 0 1,555 1993 204 16 85 321 663 131 157 0 0 0 2 0 1,579 1994 216 17 89 345 676 127 180 0 0 0 2 0 1,652 1995 227 17 97 347 689 142 214 0 0 0 2 0 1,735 1996 239 17 111 328 702 148 246 0 0 0 2 0 1,793 1997 250 17 112 387 715 153 325 0 0 0 2 0 1, 961 Irrigation 1980 16 385 47 257 52 102 200 335 4 0 0 0 1,398 1981 15 462 45 196 70 89 221 314 4 0 0 0 1,416 1982 15 540 43 135 88 76 241 293 4 0 0 0 1,435 1983 14 617 40 73 105 63 262 272 4 0 0 0 1,450 1984 14 694 38 12 1 23 50 282 251 3 0 0 0 1,467 1985 20 0 28 0 111 64 132 137 4 0 0 0 496 1986 0 0 28 0 93 44 176 91 5 0 0 0 437 1987 36 0 28 0 30 35 176 91 5 0 0 0 401 1988 36 0 28 0 8 29 440 91 4 0 0 0 636 1989 26 0 41 0 127 0 369 128 3 0 0 0 694 1990 33 0 47 0 113 0 274 125 3 0 0 0 595 1991 33 0 47 0 127 0 274 125 3 0 0 0 609 1992 33 0 47 0 127 0 274 125 3 0 0 0 609 1993 77 0 51 0 170 0 808 265 3 0 0 0 1,374 1994 0 0 51 7 845 0 718 272 3 0 0 0 1,896 1995 0 0 54 7 841 0 808 238 4 0 0 0 1,952 1996 0 0 54 8 957 0 8 08 265 4 0 0 0 2,096 1997 0 0 54 7 782 0 808 265 3 0 0 0 1,919

PAGE 56

55 Table 5 07. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Livestock 1980 18 0 14 55 82 24 142 0 0 0 3 0 338 1981 18 0 13 58 76 24 133 0 0 0 3 0 325 1982 17 0 13 60 69 24 125 0 0 0 3 0 311 1983 16 0 12 62 62 24 116 0 0 0 3 0 295 1984 15 0 12 64 55 24 107 0 0 0 2 0 279 1985 14 0 12 66 56 4 105 0 0 0 3 0 260 1986 13 0 14 66 66 3 74 0 0 0 3 0 239 1987 14 0 13 69 62 4 81 0 0 0 3 0 246 1988 16 0 14 76 65 4 89 0 0 0 3 0 267 1989 16 0 14 76 63 4 89 0 0 0 3 0 265 1990 16 0 16 75 65 4 101 0 0 0 3 0 280 1991 16 0 16 77 67 4 103 0 0 0 3 0 286 1992 16 0 19 84 80 3 133 0 0 0 3 0 338 1993 15 0 18 84 82 3 131 0 0 0 3 0 336 1994 17 0 15 84 77 4 125 0 0 0 3 0 325 1995 17 0 16 88 78 4 121 0 0 0 3 0 327 1996 14 0 13 72 113 3 98 0 0 0 5 0 318 1997 13 0 14 72 73 2 96 0 0 0 2 0 272

PAGE 57

56 Table 5 08. Total pumping from the Lower T rinity Aquifer by use category for each count y for the period 1980 through 1997 (All values are acre feet per year). Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Municipal 1980 190 0 0 0 0 63 0 3,198 0 0 0 0 3,451 1981 168 0 0 0 0 66 0 841 0 0 0 0 1,075 1982 198 0 0 0 0 77 0 485 0 0 0 0 760 1983 193 0 0 0 0 97 0 276 0 0 0 0 566 1984 232 25 0 0 0 137 39 665 0 0 0 0 1,098 1985 199 19 0 0 0 147 36 893 0 0 0 0 1,294 1986 222 31 0 0 0 74 43 551 0 0 0 0 921 1987 204 32 0 0 0 10 1 64 188 0 0 0 0 589 1988 227 38 0 0 0 89 69 460 0 0 0 0 883 1989 297 40 0 0 0 95 73 614 0 0 0 0 1,119 1990 269 30 0 0 0 91 67 371 0 0 0 0 828 1991 275 26 0 0 0 98 63 311 0 0 0 0 773 1992 219 30 0 0 0 87 71 651 0 0 0 0 1,058 1993 298 28 0 0 0 122 75 737 0 0 0 0 1,260 1994 340 41 0 0 0 132 73 524 0 0 0 0 1,110 1995 322 47 0 0 0 99 87 518 0 0 0 0 1,073 1996 299 45 0 0 0 117 84 927 0 0 0 0 1,472 1997 331 43 0 0 0 119 79 609 0 0 0 0 1,181 Manufacturing 1980 0 1,959 0 0 0 0 0 0 0 0 0 0 1,959 1981 0 1,959 0 0 0 0 0 0 0 0 0 0 1,959 1982 0 1,959 0 0 0 0 0 0 0 0 0 0 1,959 1983 0 1,382 0 0 0 0 0 0 0 0 0 0 1,382 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 2,097 0 0 0 0 0 0 0 0 0 0 2,097 1986 0 2,157 0 0 0 0 0 0 0 0 0 0 2,157 1987 0 2,644 0 0 0 0 0 0 0 0 0 0 2,644 1988 0 3,385 0 0 0 0 0 0 0 0 0 0 3,385 1989 0 3,385 0 0 0 0 0 0 0 0 0 0 3,385 1990 0 4,756 0 0 0 0 0 0 0 0 0 0 4,756 1991 0 3,739 0 0 0 0 0 0 0 0 0 0 3,739 1992 0 3,739 0 0 0 0 0 0 0 0 0 0 3,739 1993 0 3,739 0 0 0 0 0 0 0 0 0 0 3,739 1994 0 3,746 0 0 0 0 0 0 0 0 0 0 3,746 1995 0 3,746 0 0 0 0 0 0 0 0 0 0 3,746 1996 0 3,746 0 0 0 0 0 0 0 0 0 0 3,746 1997 0 3,746 0 0 0 0 0 0 0 0 0 0 3,746

PAGE 58

57 Table 5 08. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendal l Kerr Kimble Medina Travis Uvalde Total Pumpage Rural Domestic 1980 0 0 0 0 0 0 0 0 0 0 1 0 1 1981 0 0 0 0 0 0 0 0 0 0 1 0 1 1982 0 0 0 0 0 0 0 0 0 0 1 0 1 1983 0 0 0 0 0 0 0 0 0 0 1 0 1 1984 0 0 0 0 0 0 0 0 0 0 1 0 1 1985 0 0 0 0 0 0 0 0 0 0 1 0 1 1986 0 0 0 0 0 0 0 0 0 0 1 0 1 1987 0 0 0 0 0 0 0 0 0 0 1 0 1 1988 0 0 0 0 0 0 0 0 0 0 1 0 1 1989 0 0 0 0 0 0 0 0 0 0 1 0 1 1990 0 0 0 0 0 0 0 0 0 0 1 0 1 1991 0 0 0 0 0 0 0 0 0 0 1 0 1 1992 0 0 0 0 0 0 0 0 0 0 1 0 1 1993 0 0 0 0 0 0 0 0 0 0 1 0 1 1994 0 0 0 0 0 0 0 0 0 0 2 0 2 1995 0 0 0 0 0 0 0 0 0 0 2 0 2 1996 0 0 0 0 0 0 0 0 0 0 2 0 2 1997 0 0 0 0 0 0 0 0 0 0 2 0 2 Irrigation 1980 46 226 0 111 0 0 0 165 0 0 0 0 548 1981 43 271 0 83 0 0 0 155 0 0 0 0 552 1982 40 317 0 55 0 0 0 144 0 0 0 0 556 1983 36 362 0 28 0 0 0 134 0 0 0 0 560 1984 33 408 0 0 0 0 0 123 0 0 0 0 564 1985 48 0 0 0 0 0 0 67 0 0 0 0 115 1986 10 0 0 0 0 0 0 45 0 0 0 0 55 1987 88 0 0 0 0 0 0 45 0 0 0 0 133 1988 88 0 0 0 0 0 0 45 0 0 0 0 133 1989 68 0 0 0 0 0 0 63 0 0 0 0 131 1990 81 0 0 0 0 0 0 62 0 0 0 0 143 1991 81 0 0 0 0 0 0 62 0 0 0 0 143 1992 81 0 0 0 0 0 0 62 0 0 0 0 143 1993 171 0 0 0 0 0 0 131 0 0 0 0 302 1994 15 0 0 3 0 0 0 134 0 0 0 0 152 1995 14 0 0 2 0 0 0 117 0 0 0 0 133 1996 15 0 0 2 0 0 0 131 0 0 0 0 148 1997 15 0 0 2 0 0 0 131 0 0 0 0 148

PAGE 59

58 Table 5 08. (cont.) Year Bandera Bexar Blanco Comal Gillespie Hays Kendall Kerr Kimble Medina Travis Uvalde Total Pumpage Livestock 1980 0 0 0 0 0 0 0 0 0 0 0 0 0 1981 0 0 0 0 0 0 0 0 0 0 0 0 0 1982 0 0 0 0 0 0 0 0 0 0 0 0 0 1983 0 0 0 0 0 0 0 0 0 0 0 0 0 1984 0 0 0 0 0 0 0 0 0 0 0 0 0 1985 0 0 0 0 0 0 0 0 0 0 0 0 0 1986 0 0 0 0 0 0 0 0 0 0 0 0 0 1987 0 0 0 0 0 0 0 0 0 0 0 0 0 1988 0 0 0 0 0 0 0 0 0 0 0 0 0 1989 0 0 0 0 0 0 0 0 0 0 0 0 0 1990 0 0 0 0 0 0 0 0 0 0 0 0 0 1991 0 0 0 0 0 0 0 0 0 0 0 0 0 1992 0 0 0 0 0 0 0 0 0 0 0 0 0 1993 0 0 0 0 0 0 0 0 0 0 0 0 0 1994 0 0 0 0 0 0 0 0 0 0 0 0 0 1995 0 0 0 0 0 0 0 0 0 0 0 0 0 1996 0 0 0 0 0 0 0 0 0 0 0 0 0 1997 0 0 0 0 0 0 0 0 0 0 0 0 0

PAGE 60

59 Table 9 01. Water budget for the calibrated steady state model for 1980. All values are acre feet per year, negative values indicate net discharge from the aquifer. (The numbers are rounded to hundreds of a crefeet.) In Out Net Wells 0 1 6,7 00 1 6,7 00 Streams and springs 0 1 6 4,5 00 1 6 4,5 00 Reservoirs 9,0 00 2 8,8 00 19,800 Edwards (Balcones Fault Zone) Aquifer 8,1 0 0 110,6 00 102,5 0 0 Recharge 30 3,5 00 0 30 3,5 0 0 Table 9 02. Water budget for the respectiv e layers in the calibrated steady state model for 1980. All values are acrefeet per year, negative values indicate net discharge from the aquifer. (The numbers are rounded to hundreds of acre feet.) Edwards Group Upper Trinity Aquifer Middle Trinity Aqu ifer Lower Trinity Aquifer Total Inter Aquifer Flow (Upper) 0 9,800 64,100 5,80 0 79,700 Inter Aquifer Flow (Lower) 9,800 64,100 5,800 0 79,700 Wells 1,000 5, 100 4, 600 6,000 16,700 Streams and springs 47,700 60,900 55,900 0 164,500 Reservoi rs 0 2,50 0 1 7,300 0 19,800 Edwards (Balcones Fault Zone) Aquifer 0 33, 300 69, 200 0 102,500 Recharge 5 8 , 500 1 56, 200 88,700 100 303,500

PAGE 61

60 Table 9 03. Water budget for the respective counties in the calibrated steady state model for 1980. All values are acre feet per year, negative values indicate net discharge from the aquifer. (The numbers are rounded to hundreds of acre feet.) County Wells Streams and springs Recharge Reservoirs Edwards (Balcones Fault Zone) Aquifer Lateral i nflo w Lateral o utflow Bandera 1, 10 0 3 4 , 3 00 3 6,9 00 1, 0 00 1 ,8 00 2 5,5 00 24 , 2 00 Bexar 3,9 00 9,900 39, 000 0 37, 2 00 36, 2 00 24, 3 00 Blanco 200 1 4 , 2 00 1 9 , 000 0 0 6,9 00 1 1,5 00 Comal 1, 0 00 3,7 00 40 , 3 00 5,9 00 3 7,9 00 3 7,6 00 29 , 5 00 Gillespie 1,2 00 1 4 , 3 00 28, 3 00 0 0 9 00 1 3,7 00 Hays 1,6 00 1 8,8 00 2 1,8 0 0 0 6 , 7 00 14, 2 00 9,0 00 Kendall 1,60 0 2 8,5 00 51 , 000 0 0 9 ,6 00 3 0 , 5 0 0 Kerr 6 , 000 3 2,6 00 47, 1 00 0 0 1 0,5 00 1 9 ,0 00 Kimble 0 0 40 0 0 0 20 0 5 0 0 Medina 0 2,4 00 5, 8 00 2, 6 00 1 4 , 3 00 20, 4 00 6, 9 00 Travis 10 0 5, 2 00 1 1,9 00 10 , 3 00 2,1 00 6,1 0 0 40 0 Uvalde 0 5 0 0 1,8 00 0 2 , 5 00 2, 0 0 0 80 0 Total 1 6,7 0 0 164,5 00 30 3,50 0 19, 8 00 102, 5 00 170, 2 00 1 70,20 0

PAGE 62

61 Table 10 01. Stress periods of the transient model. Stress Period Year Length (Days) 1 Steady state (1980) 100 , 000 2 1981 365 3 1982 365 4 1983 365 5 1984 365 6 1985 365 7 1986 365 8 1987 365 9 1988 365 10 1989 365 11 1990 365 12 1991 365 13 1992 365 14 1993 365 15 1994 365 16 1995 365 17 1996 365 18 1997 365

PAGE 63

62 Table 10 02. Calibrated specific yield, specific storage, and hydraulic conductivity data for the respective model layers. Model Layer Aquifer Specific Yield Specific Storage (Per foot) Hydraulic Conductivity (Feet per Day) Range Mean 1 Edwards Group 0.008 1 .0E05 11 11.0 2 Upper Trinity Aquifer 0.0005 1.0E 06 9 to 150 10.4 3 Middle Trinity Aquifer 0.0008 1.0E 07 7.6 to 15 8.8 4 Lower Trinity Aquifer 0.0008 1.0E 07 1.67 to 16.7 4.4

PAGE 64

63 Table 10 03. Calibration statistics for the transient model for the yea rs 1980, 1990, and 1997. The percentage represents the mean absolute error relative to the range of measured water levels. 1980 Mean Error Mean Absolute Error Mean Absolute Error (Percent) Overall 14 59 4% Edwards Group 23 31 17% Upper Trinity Aquifer 23 68 6% Middle Trinity Aquifer 14 53 5% Lower Trinity Aquifer 17 58 5% 1990 Mean Error Mean Absolute Error Mean Absolute Error (Percent) Overall 6 52 4% Edwards Group 34 34 -Upper Trinity Aquifer 81 99 9% Middle Trinity Aquifer 6 54 7% Lo wer Trinity Aquifer 17 45 4% 1997 Mean Error Mean Absolute Error Mean Absolute Error (Percent) Overall 15 57 4% Edwards Group 26 26 -Upper Trinity Aquifer 44 82 7% Middle Trinity Aquifer 10 66 7% Lower Trinity Aquifer 26 48 5% -indicates too few water level measurements to calculate percent mean absolute error.

PAGE 65

64 Table 10 04. Water budget for the respective layers in the calibrated transient model for 1980, 1990 and 1997. (All values are acrefeet per year, negative values indicate net disc harge from the aquifer). 1980 Edwards Group Upper Trinity Aquifer Middle Trinity Aquifer Lower Trinity Aquifer Inter Aquifer Flow (Upper) 0 9,773 6 4,138 5,82 5 Inter Aquifer Flow (Lower) 9,773 6 4,138 5,82 5 0 Wells 1,007 5,1 57 4,556 5,961 Streams and springs 47,735 60,879 56,013 0 Reservoirs 0 2,519 17,329 0 Edwards (Balcones Fault Zone) Aquifer 0 33,224 69,293 0 Recharge 58,516 156,135 88,910 155 1990 Edwards Group Upper Trinity Aquifer Middle Trinity Aquifer Lower Trinity Aquifer S torage 7,960 9,839 5,788 232 Inter Aquifer Flow (Upper) 0 10,087 6 8,750 5, 793 Inter Aquifer Flow (Lower) 10,087 6 8,750 5, 7 93 0 Wells 1,229 6,2 53 5,650 5,732 Streams and springs 51,290 70,642 64,676 0 Reservoirs 0 3,097 18,990 0 Edward s (Balcones Fault Zone) Aquifer 0 37,821 68,783 0 Recharge 70,567 186,292 1 00,916 180 1997 Edwards Group Upper Trinity Aquifer Middle Trinity Aquifer Lower Trinity Aquifer Storage 12,380 1 6,923 1 1,852 8 447 Inter Aquifer Flow (Upper) 0 10,329 77, 150 5, 297 Inter Aquifer Flow (Lower) 10,329 77, 150 5, 297 0 Wells 1,504 7,90 1 8,448 5,079 Streams and springs 54,343 85,266 75,397 0 Reservoirs 0 4,408 23,563 0 Edwards (Balcones Fault Zone) Aquifer 0 45,1623 7 0,962 0 Recharge 78,557 226,464 1 18,348 240

PAGE 66

65 Table 1005. Estimated spring discharge and simulated average spring discharge rates from the calibrated transient model. The locations of these springs can be found in Figure 528 (All values are expressed in gallons per minute). Ple ase note that: (1) the spring discharge measurements are single measurements collected over a wide range of conditions and time periods, (2) only two of the spring discharge measurements coincide with the calibration period, and (3) due to scale issues, th e model results may not reflect the more localized flow systems that influence discharge at specific springs . Spring 1 2 3 4 5 8 9 10 11 12 13 14 15 16 17 18 19 Bee Caves Spring Lynx Haven Springs Ellebracht Springs Cave Without A Name Kenmor e Ranch Spring #9 Edge Falls Springs Rebecca Springs Jacob's Well Spring Bassett Springs Cold Springs Estimated Flow (g allons p er m inute ) 15 0 10 0 10 0 250 0 31 0 20 75 50 15 0 30 0 30 0 50 0 25 50 50 900 0 500 0 Measure Date 4/13/67 4/12/67 3/31/66 3/11/70 7/13/76 7/10/75 1/17/40 7/17/75 7/11/75 8/31/76 1/1/66 12/30/88 5/25/73 12/20/60 8/20/91 Simulated Flow (g allon p er m inute ) 198 0 13 9 75 82 225 3 3 0 36 6 33 1 1 9 0 0 0 0 6 0 0 4 07 441 198 1 14 2 83 86 2 38 35 8 47 4 40 12 7 81 0 0 0 9 0 0 4 23 516 198 2 1 4 0 7 8 84 2 17 33 1 35 0 33 11 5 0 0 0 0 8 0 0 407 437 198 3 13 9 75 82 213 31 7 3 4 6 36 1 1 9 0 0 0 0 9 0 0 400 448 198 4 13 9 74 82 21 8 32 1 3 2 2 32 11 3 0 0 0 0 7 0 0 408 419 198 5 14 0 76 83 2 26 3 3 2 38 8 42 13 2 11 3 0 0 0 9 0 0 4 13 489 198 6 14 2 84 86 2 41 36 0 46 6 46 13 4 1 5 2 0 0 0 11 0 0 4 29 542 198 7 1 4 5 92 90 2 55 39 3 5 0 0 46 13 2 1 4 0 0 0 0 12 0 0 4 46 558 198 8 14 2 8 7 88 22 8 35 8 3 6 8 32 11 1 0 0 0 0 7 0 0 416 442 198 9 1 4 0 81 85 2 22 33 8 3 0 8 32 11 0 0 0 0 0 6 0 0 410 414 199 0 14 2 85 87 2 36 3 5 9 39 2 40 12 5 1 0 0 0 8 0 0 4 28 474

PAGE 67

66 199 1 1 4 5 93 91 2 44 38 2 5 0 8 50 13 9 19 5 0 0 0 12 0 0 4 36 568 199 2 1 4 6 98 94 2 50 4 0 4 52 8 56 15 0 35 1 83 0 0 13 0 0 4 47 626 199 3 1 4 2 88 89 219 35 5 3 5 9 40 12 4 5 9 0 0 0 10 0 0 415 473 199 4 14 4 92 91 242 3 7 8 4 2 6 44 12 9 70 0 0 0 10 0 0 4 32 518 199 5 1 4 2 88 89 2 27 3 6 3 38 6 37 11 8 0 0 0 0 9 0 0 4 25 471 199 6 14 2 86 88 2 2 4 3 5 0 3 3 5 31 11 0 0 0 0 0 7 0 0 4 20 419 199 7 14 4 90 90 2 47 38 8 44 6 47 13 2 35 0 0 0 11 0 0 4 46 522

PAGE 68

67 Figure 3 01. Location of the study area relative to roads, major cities and towns, lakes , and rivers (modified from Mace and others, 2000) .

PAGE 69

68 Figure 3 02. Map of outcrops of the major aquifers in the study area. Trinity sediments in the study area include sediments that are part of the Edwards Trinity (Plateau) Aquifer System to the w est and underlie the Edwards ( Balcones Fault Zone) Aquifer to the south and east (modified from Mace and others, 2000) .

PAGE 70

69 Figure 3 03. Regional water planning groups in the study area (modified from Mace and others, 2000) .

PAGE 71

70 Figure 3 04. Groundwater conservation districts in the study area as of January 2008 (Area with diagonal hatch lines represent the Edwards Aquifer Authority).

PAGE 72

71 Figure 3 05. Groundwater management areas in the study area.

PAGE 73

72 Figure 3 06. (a) Major perennial and intermittent rive rs and streams in the study area. (b) River authorities in the study area.

PAGE 74

73 Figure 3 07. Physiographic provinces in the study area (modified from Anaya and Jones , 2009) .

PAGE 75

74 Figure 3 08. Land surface elevation in the study area (modified from Mace and others, 2000) .

PAGE 76

75 Figure 3 09. Average annual r ainfall distribution for the period 1960 through 1996 ( data from National Climate Data Center ).

PAGE 77

76 Figure 3 10. Historic annual precipitation for three rain gauge stations in the study area (modified from Mace and others, 2000) .

PAGE 78

77 Figure 3 11. Average monthly precipitation for three rain gauge s in the study area for the period 1960 through 1996 (data from National Climate Data Center) .

PAGE 79

78 Figure 3 12. Average annual maximum temperature for 1971 through 2000. The contours are expressed in degrees Fahrenheit ( modified from data from S patial C limate A nalysis S ervice, 2004).

PAGE 80

79 Figure 3 13. Average annual gross lake evaporation for 1950 through 1979. Contours are expressed in inches (modified from Larki n and Bomar, 1983).

PAGE 81

80 Figure 3 14. Stratigraphic and hydrostratigraphic column of the Hill Country area.

PAGE 82

81 Figure 3 15. The main geologic structures in the study area (modified from Mace and others, 2000).

PAGE 83

82 Figure 3 16. Surface geology of the stu dy area (modified from Mace and others, 2000) . Please note that this map excludes isolated outliers of the Edwards Group that overly the Upper member of the Gl en Rose Limestone, some of which are included in the original and updated models.

PAGE 84

83 Figure 3 17. Geologic cross sections through the study area (modified from Ashworth, 1983; Mace and others, 2000). Inset map shows cross section line AA.

PAGE 85

84 Figure 4 01. The approximate extents of previous model grids for models used for simulating groundwater flow through the study area.

PAGE 86

85 Figure 5 01. Elevation s of (a) the top and (b) the base of the Edwards Group. The gray and white circles indicate control points from well logs and outcrop, respectively. The contour interval is 100 feet (modified from Ashworth, 1983; Mace and others, 2000) .

PAGE 87

86 Figu re 5 02. Elevation of (a) the top and (b) the base of the Upper Trinity Aquifer. The gray and white circles indicate control points from well logs and outcrop, respectively. The contour interval is 100 feet (modified f rom Ashworth, 1983; Mace and others, 2000) .

PAGE 88

87 Figure 5 03. Elevation of (a) the top and (b) the base of the Middle Trinity Aquifer. The gray and white circles indicate control points from well logs and outcrop, respectively. The contour interval is 100 fee t (modified from Ashworth, 1983; Mace and others, 2000) .

PAGE 89

88 Figure 5 04. Elevation of (a) the top (modified from Ashworth, 1983; Mace and others, 2000) and (b) the base of the Lower Trinity Aquifer. The gray and white circles indicate control points from we ll logs and outcrop, respectively. The contour interval is 100 feet. Please note: the top of the Lower Trinity Aquifer coincides with the base of the Hammett Shale and thus differs from the base of the Middle Trinity Aquifer.

PAGE 90

89 Figure 5 05. The approxima te thickness of the Edwards Group in the study area. The contour interval is 100 feet.

PAGE 91

90 Figure 5 06. The approximate thickness of the Upper Trinity Aquifer in the study area. The contour interval is 100 feet.

PAGE 92

91 Figure 5 07. The approximate thickne ss of the Middle T rinity Aquifer in the study area. The contour interval is 100 feet.

PAGE 93

92 Figure 5 08. The approximate thickness of the Lower Trinity Aquifer in the study area. The contour interval is 100 feet.

PAGE 94

93 Figure 5 09. Average water level elev ations in the Edwards Group in the study area for the period 1977 through 1985. The contour interval is 100 feet.

PAGE 95

94 Figure 5 10. Average water level elevations in the Upper Trinity Aquifer in the study area for the period 1977 through 1985. The contour interval is 100 feet.

PAGE 96

95 Figure 5 11. Average water level elevation in the Middle Trinity Aquifer in the study area for the period 1977 through 1985. The contour interval is 100 feet.

PAGE 97

96 Figure 5 12. Average water level elevation in the Lower Trinit y Aquifer in the study area for the period 1977 through 1985. The contour interval is 100 feet.

PAGE 98

97 Figure 5 13. Hydrographs from selected Edwards Group wells in the study area.

PAGE 99

98 Figure 5 14. Hydrographs from selected Upper Trinity Aquifer wells in the study area.

PAGE 100

99 Figure 5 15. Hydrographs from selected Middle Trinity Aquifer wells in the study area.

PAGE 101

100 Figure 5 16. Hydrographs from selected Lower Trinity Aquifer wells in the study area.

PAGE 102

101 Figure 5 17. Net water level change in the Upper Trin ity Aquifer between 1980 and 1997 at selected well locations . Positive values (blue points) indicate rise in water level and nega tive values (red points) decline in water levels.

PAGE 103

102 Figure 5 18. Net water level change in (a) the Middle Trinity Aquifer and (b) Lower Trinity Aquifer between 1980 and 1997 at selected well locations . Positive values (blue points) indicate rise in water level and negative values (red points) decline in water levels.

PAGE 104

103 Figure 5 19. Streamflow gain (positive values) and loss (n egative values) from Slade and others (2002).

PAGE 105

1 04 Figure 5 20. Location of stream gauges for the streamflow hydrographs shown in Figures 521 through 526 (from Mace and others, 2000).

PAGE 106

105 Figure 5 21. Average monthly streamflow for the United States Geo logical Survey gauging 08153500 on the Pedernales River near Johnson City for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 107

106 Figure 5 22. Average monthly streamflow for the United States Geological Survey gauging 08167000 on the Guadalupe River at Comfort for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 108

107 Figure 5 23. Average monthly streamflow for the Uni ted States Geological Survey gauging 08167500 on the Guadalupe River near Spring Branch for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 109

108 Figure 5 24. Average monthly streamflow for the United States Geological Survey gauging 08171000 on the Blanco River at Wimberley for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 110

109 Figure 5 25. Average monthly streamflo w for the United States Geological Survey gauging 08179000 on the Medina River near Pipe Creek for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 111

110 Figure 5 26. Average monthly stre amflow for the United States Geological Survey gauging 08184000 on Cibolo Creek near Bulverde for (a) linear and (b) logarithmic scales. The station location can be found in Figure 529 (from Mace and others, 2000) .

PAGE 112

111 Figure 5 27. Lake level elevations i n (a) Lake Travis, (b) Canyon Lake, and (c) Medina Lake. Lake Travis water levels are from the Lower Colorado River Authority. Canyon Lake water levels are from the U.S. Army Corps of Engineers. Medina Lake water levels for the period 1940 through 1986 are from Espey, Huston, and Associates (1989). Water levels for the periods January 1987 through September 1994 and October 1997 through September 1999 are from the U.S. Geological Survey. Mace and others (2000) calculated lake levels for the period October 1994 through September 1997 by relating lake volumes from a Texas Water Development Board database to lake level using the rating curve by Espey, Huston, and Associates (1989).

PAGE 113

112 Figure 5 28. Location and estimated spring discharge in the study area. Sp ringflow and geological formations where the numbered springs occur are included in Table 502 (from Mace and others, 2000) .

PAGE 114

113 Figure 5 29. Distribution of hydraulic conductivity in the Upper, Middle, and Lower Trinity a quifer s .

PAGE 115

114 Figure 5 29. (Cont.)

PAGE 116

115 Figure 5 30. The spatial distribution of pumping throughout the 1980 through 1997 model period for manufacturing, municipal, livestock, rural domestic, and irrigation uses are based on the spatial distribution of (a) industrial and public supply wells, (b) rangeland, (c) rural population, and (d) irrigated farmland, respectively.

PAGE 117

116 Figure 5 30. (Cont.)

PAGE 118

117 Figure 5 31. Total annual groundwater pumping from the Hill Country portion of the Trinity Aquifer System , 1980 through 1997.

PAGE 119

118 Figure 5 32. Annua l groundwater pumping from the Hill Country portion of the Trinity Aquifer System for livestock, rural domestic, manufacturing, municipal, and irrigation uses, 1980 through 1997.

PAGE 120

119 Figure 5 33. Total annual pumping from the Hill Country portion of the Tr inity Aquifer System for each county in the study area.

PAGE 121

120 Figure 5 34. The ranges of total dissolved solids found in groundwater in the Edwards Group, and the Upper, Middle, and Lower Trinity aquifers. The black line indicates the median value for each aquifer.

PAGE 122

121 Figure 5 35. Map of total dissolved solids in the Edwards Group. mg/l = milligrams per liter.

PAGE 123

122 Figure 5 36. Map of total dissolved solids in the Upper Trinity Aquifer . mg/l = milligrams per liter .

PAGE 124

123 Figure 5 37. Map of total dissolved solids in the Middle Trinity Aquifer . mg/l = milligrams per liter.

PAGE 125

124 Figure 5 38. Map of total dissolved solids in the Lower Trinity Aquifer . mg/l = milligrams per liter.

PAGE 126

125 Figure 5 39. Piper diagram of groundwater from the Edwards Group Aquifer showing the relative concentrations of the major ions present in the groundwater . Ca = calcium, Mg = magnesium, Na = sodium, K = potassium, HCO3 = bicarbonate, SO4 = sulfate, Cl = chloride.

PAGE 127

126 Figure 5 40. Piper diagram of groundwater from the Uppe r Trinity Aquifer showing the relative concentrations of the major ions present in the groundwater . Ca = calcium, Mg = magnesium, Na = sodium, K = potassium, HCO3 = bicarbonate, SO4 = sulfate, Cl = chloride.

PAGE 128

127 Figure 5 41. Piper diagram of groundwater from the Middle Trinity Aquifer showing the relative concentrations of the major ions present in the groundwater . Ca = calcium, Mg = magnesium, Na = sodium, K = potassium, HCO3 = bicarbonate, SO4 = sulfate, Cl = chloride.

PAGE 129

128 Figure 5 42. Piper diagram o f groundwater from the Lower Trinity Aquifer showing the relative concentrations of the major ions present in the groundwater . Ca = calcium, Mg = magnesium, Na = sodium, K = potassium, HCO3 = bicarbonate, SO4 = sulfate, Cl = chloride.

PAGE 130

129 Figure 5 43. Gro undwater geochemical trends that are apparent in the Hill Country portion of the Trinity Aquifer System . Ca = calcium, Mg = magnesium, Na = sodium, K = potassium, HCO3 = bicarbonate, SO4 = sulfate, Cl = chloride.

PAGE 131

130 Figure 6 01. Conceptual model of the Hi ll Country portion of the Trinity Aquifer System . (a) Schematic cross section through the aquifer system. (b) diagram showing the boundary conditions at the outer edge of the model, flows between the layers, and how the conceptual model translates into the numerical model (modified from Mace and others, 2000) .

PAGE 132

131 Figure 7 01. Active and inactive cells in model grid for (a) Layer 1 (Edwards G roup), (b) Layer 2 (Upper Trinity Aquifer), (c) Layer 3 (Middle Trinity Aquifer), and (d) Layer 4 (Lower Trinity aquifer).

PAGE 133

132 Figure 7 01. (Cont.).

PAGE 134

133 Figure 7 02. Vertical leakance between the Middle and Lower Trinity aquifers. Values expressed in per day.

PAGE 135

134 Figure 7 03. The spatial distribution of total pumping for 1980 for (a) Layer 1, (b) Layer 2, (c) Layer 3, and (d) Layer 4 .

PAGE 136

135 Figure 7 03. (Cont.).

PAGE 137

136 Figure 7 04. Boundary cells in model grid for (a) Layer 1, (b) Layer 2, (c) Layer 3, and (d) Layer 4.

PAGE 138

137 Figure 7 04. (Cont.).

PAGE 139

138 Figure 9 01. Spatial distribution of recharge for 1980. This is estimated based preci pitation data for the study area and Cibolo Creek streamflow loss studies . All value s are expressed in inches per year.

PAGE 140

139 Figure 9 02. Comparison of measured and calculated water levels from the steady state model for (a) Layer 1, (b) Layer 2, (c) Layer 3, and (d) Layer 4. The contours represent calculated water levels while the points indicate the difference between measured and simulated water levels relative to the measured water levels .

PAGE 141

140 Figure 9 02. (Cont.).

PAGE 142

141 Figure 9 03. Comparison of m easured and calculated water levels from the steady state model .

PAGE 143

142 Figure 9 04. Locations of stream gauges used to compare measured streamflow and calculated discharge to streams from the model.

PAGE 144

143 Figure 9 05. Comparison of the calculated groundwater discharge rate to perennial streams from the 1980 steady state model ( gray line) and measured streamflow data. Stream gauge l ocations are shown in Figure 904.

PAGE 145

144 Figure 9 05. (Cont.).

PAGE 146

145 Figure 9 05. (Cont.).

PAGE 147

146 Figure 9 05. (Cont.).

PAGE 148

147 Figure 9 06. Sensitivity of calculated water levels in the steady state model to changes in model parameters .

PAGE 149

148 Figure 1001. Locations of wells used to compare measured water levels over the transient period (1980 through 1997) and calculated water levels .

PAGE 150

149 Figure 1002. Compar ison of simulated water level fluctuations to measured water levels . Well locations are shown in Figure 1001.

PAGE 151

150 Figure 1002. (Cont.).

PAGE 152

151 Figure 1002. (Cont.).

PAGE 153

152 Figure 1002. (Cont.).

PAGE 154

153 Figure 1002. (Cont.).

PAGE 155

154 Figure 1002. (Cont.).

PAGE 156

155 Figure 1002. (Cont.).

PAGE 157

156 Figure 1003. Comparison of measured and calculated water levels for 1990 and 1997 from the transient model .

PAGE 158

157 Figure 10 04. Comparison of 1990 measured and calculated water levels from the transient model for (a) Layer 1, (b) Layer 2, (c) Layer 3, and (d) Layer 4 . The contours represent calculated water levels while the points indicate the difference between measured and simulated water levels relative to the measured water levels.

PAGE 159

158 Figure 10 04. (Cont.).

PAGE 160

159 Figure 10 05. Comparison of 1997 measured and calculated water levels from the transient model for (a) Layer 1, (b) Layer 2, (c) Layer 3, and (d) Layer 4 . The contours represent calculated water levels while the points indicate the difference between measured and simulated water level s relative to the measured water levels.

PAGE 161

160 Figure 10 05. (Cont.).

PAGE 162

161 Figure 1006. Comparison of calculated annual groundwater discharge rates to perennial streams from the transient model ( gray line) and measured streamflow data. Stream gauge l ocations are shown in Figure 904.

PAGE 163

162 Figure 1006. (Cont.).

PAGE 164

163 Figure 1006. (Cont.).

PAGE 165

164 Figure 1006. (Cont.).

PAGE 166

165 Figure 1007. Sensitivity of the transient calibration to specific yield . T he red and blue lines represent one order of magnitude lower and higher th an the calibrated values, respectively , relative to calibrated specific yield values (black line) .

PAGE 167

166 Figure 1007. (Cont.).

PAGE 168

167 Figure 1007. (Cont.).

PAGE 169

168 Figure 1007. (Cont.).

PAGE 170

169 Figure 1007. (Cont.).

PAGE 171

170 Figure 1007. (Cont.).

PAGE 172

171 Figure 1007. (Cont.).

PAGE 173

172 Figure 1008. Sensitivity of the transient calibration to specific storage. The red and blue lines represent one order of magnitude lower and higher than the calibrated values, respectively , relative to calibrated specific yield values (black line).

PAGE 174

173 F igure 1008. (Cont.).

PAGE 175

174 Figure 1008. (Cont.).

PAGE 176

175 Figure 1008. (Cont.).

PAGE 177

176 Figure 1008. (Cont.).

PAGE 178

177 Figure 1008. (Cont.).

PAGE 179

178 Figure 1008. (Cont.).

PAGE 180

179 A ppendix: Comments and Responses

PAGE 181

180 General Comments The report is well organized, well written, and co ntains many quality figures. However, this report may have been prepared prior to the report writing guidelines for the Water Science and Conservation Group and therefore, some of the figures may need to be updated to meet the requirements and all acronyms except for TWDB and MODFLOW need to be spelled out. However, these editorial issues, specific to figures and clarifications or expansions of materials in the text, can be more readily addressed. No response. More important issues are inbred in modeling and their supporting documentation. For example, water level residuals to match simulated and measured water levels reported for the Upper Trinity Aquifer exceed 100 feet for several wells in the steady state model. Per the program guidelines, please inc lude reporting mean absolute errors by aquifers rather than lumping all aquifer together in the calculation. This will help determine whether there is any bias in the calibration by aquifer. Some of the key springs, such as the Jacob’s Well springs —often c onsidered to be indicative of the health of the aquifer — do not flow at all throughout the simulation period. This may be a significant concern for stakeholders given the importance of the springs. Underestimation of spring flow is plausible given the regi onal scale of the model, but to have no flow at all, is a serious concern. This may indicate that the simulated water levels are lower than the elevation head of the drains assigned to simulate springflow. No general head boundary is assigned along the Bal cones Fault Zone in the east for the Lower Trinity Aquifer even though the aquifer is juxtaposed along the permeable part of the Middle Trinity Aquifer (Lower Glen Rose Limestone) in this area. Figures 903 and 1003 report calibration statistics for the respective layers in the model. The issue of the difficulty of using a regional model to simulate discharge to springs that are often part of localized flow systems is discussed in the text in Section 10.1. In the previous version of the model attempts we re made to force the model to match spring discharge measurements that may not have been representative of average spring discharge. This effort produced a model with large water balance errors. To guarantee discharge to all of these springs would require much smaller grid spacing and much shorter stress periods which is difficult to justify given: (1) the absence of spring discharge data to calibrate to, (2) the available discharge measurements are as low as 25 gallons per minute, and (3) the springs are m inor, accounting for less than 3 percent of total discharge to surface water bodies and less than 2 percent of total discharge from the aquifer. The decision to place a noflow boundary along the southern and eastern boundary of the Lower Trinity Aquifer w as based on previous work by Hovorka and others (1996) that indicated little groundwater flow crossing that boundary. We have made numerous technical and editorial comments in the report. These comments need to be addressed to improve the content and qual ity of the report. For example, some structural control points are shown in Gillespie County that was used for developing the top and base of the Upper Glen Rose Limestone where the aquifer does not exist. A time frame of 1977 to 1985 selected for construction of water level map for the steady state model may leave out many water level measurements in the north and the west thus preventing an opportunity to measure

PAGE 182

181 quality of the calibration in these areas. Hydraulic conductivity values assigned for calibra tion in the aquifers are considerably higher than the previous model and therefore, need adequate explanations on how these different values were derived. For example, hydraulic conductivity assigned in the Upper Trinity Aquifer is 5 feet per day in the pr evious model. In the current model, hydraulic conductivity assigned in the Upper Trinity Aquifer is 9 feet per day for most of the model area and 15 feet per day along the Balcones Fault Zone. In the Lower Trinity Aquifer, hydraulic conductivity assigned f or most part of the aquifer is 1.67 feet per day, an order of magnitude lower than 16.7 feet per day assigned along the Balcones Fault Zone. Calibrated hydraulic conductivity values should always be kept close to the measured values unless there is a stron ger hydrogeological reasoning to change this. We suggest strengthening the arguments for using higher hydraulic conductivity values for calibration along with presentation of maps showing such measured values, if available. The calibrated hydraulic conduc tivity values used in the model lie within the range of the measured data for the respective layers. These calibrated hydraulic conductivity values represent the overall effect of the actual range of hydraulic conductivity values that occur in the aquifer system. The Trinity Aquifer System is highly heterogeneous both vertically and laterally within the stratigraphic units that make up each model layer, thus it is erroneous to assume that any one value is representative of the model layer or that any two me asured values are related to each other. Especially when hydraulic conductivity data: (1) are few in number, (2) are widely scattered throughout the aquifer, and (3) there is high uncertainty over the stratigraphic or hydrostratigraphic is represented by t he data. In the previous model simulated discharges to the Guadalupe, Medina, and Blanco Rivers were within 25 percent of estimated values. In the current model, simulated stream discharges may show a higher difference than this. We suggest reporting simulated discharges for 1980, 1990, and 1997 in tables. It is not possible to do this comparison because baseflow analysis was not part of this project. Specific Comments 1. Page 1: Executive summary does not adequately capture important elements of the repor t. We suggest inclusion of short descriptions on structure development, conceptual flow system, rationale for assignment of model boundaries, calibration statistics, and flows. Please compare to previous version of the model. Text added that is appropriat e to the executive summary (Page 1) . 2. Page 1, paragraph 2: Please add the Edwards Group in the write up as it forms layer 1 of the model. Text added to paragraph (Page 1, paragraph 2) .

PAGE 183

182 3. Page 1, paragraph 3: Please verify the statement “Preliminary estimates of recharge equal four to six percent of annual precipitation over most parts of the aquifers”. Recharge through the large swath of the Balcones Fault Zone is much higher, as much as 15 percent of precipitation. Text revised to reflect full range of r echarge rates (Page 1, paragraph 3) . 4. Page 5, paragraph 2: Please replace the term “carbonate sandstone” with “calcareous sandstone”. Done (Page 5, paragraph 6) . 5. Page 6, end of Section 3.2: Per GAM program guidelines, moved discussion of structural and t ectonic features and cross sections from Section 5.2 to Geology Section, please update numbering of figures as needed throughout the report. Done. 6. Page 8, Section 5.2: Please clarify and discuss how contributions from groundwater conservation districts w ere used to adjust structure in the updated model. At minimum, please discuss that information was reviewed and in agreement with our interpretation of structure. Text revised as appropriate. 7. Section 5.3: Per GAM program guidelines, for each model layer please include potentiometric surface maps 1980, 1990, and 1997. Please update report with potentiometric surface maps for 1990 for each layer and add a map for 1997 for the Edwards Group. Figures 513 through 516 along with model results indicate that potentiometric surfaces in 1990 and 1997 are essentially the same as those in 1980. In the absence of regional water level changes it is unnecessary to show potentiometric maps for 1990 and 1997. Text added to Page 10, paragraph 2 to discuss this. 8. Section 5.3, Page 9, second to last paragraph: I suggest commenting on what may have caused the 130 foot water level rise in Kerr County shown in well number 5663604, in Figure 5 18. Text revised (Page 10, paragraph 3). 9. Page 11, paragraph 1: Recharge estimates by various authors are included in this section. Recently, Wet Rock Consultants estimated groundwater recharge for Kendall and adjacent counties that has not been discussed. Discussion of this report may be relevant given the higher recharge used in the current model in the east.

PAGE 184

183 Added recharge estimate by Wet Rock Groundwater Services to Table 501. 10. Page 11, paragraph 2: We could not entirely agree with these statements “ The recharge estimation method used by Mace and others (2000) is less accurate tha n recharge estimation based on precipitation because regional groundwater flow results in underestimation of recharge in upstream areas and overestimation downstream. The method used by Mace and others (2000) assumes that watersheds are closed systems with all recharge discharging to adjacent streams and does not take into account regional groundwater circulation that results in groundwater leaving or entering the watershed” . This is because Mace and others (2000) used baseflow coefficients a s well as preci pitation distribution to estimate recharge. Perhaps, t he more important difference in recharge estimation between the current and the previous model is that the previous model did not include additional recharge in the Cibolo Creek watershed and along a se lected zone in the BFZ that has never been reported in the literature before . Revised the text but retained the criticism of the Mace and others (2000) methodology which also appears in Mace and others (2000) on Page 34, paragraph 2. 11. Section 5.5, Page 12, second paragraph: Since the stream flow plots are log plots, it’s not clear that the flows ever go to zero. Therefore instead of saying perennial rivers cease flowing, I suggest “Most.experience significant decrease in flow during droughts”. Text revis ed as requested (Page 12, paragraph 5). 12. Page 13, paragraph 1: Please verify whether the Lower Trinity aquifer is more transmissive as stated. These are the numbers reported by Ashworth in his 1983 report. They do not agree with the hydraulic property data collected for this project. 13. Section 5.6, Page 14, second paragraph: Please include a figures showing the distribution of hydraulic conductivity data for the lower Trinity, per GAM checklist of deliverables page 4 of 11, Hydraulic Properties, “A map of hydraulic conductivity for each model layer” A map showing the distribution of hydraulic conductivity data for the lower Trinity appears in Figure 529. 14. Page 15, paragraph 3: This section discusses pumpage reported by various authors. It may be pertinent to report the current pumpage information available from recent GAM Run reports. Current pumpage lies outside of the calibration period for this model and is therefore not relevant .

PAGE 185

184 15. Section 6.0, page 17: Please include discussion on Hammett Shale as a c onfining unit and adjust figure appropriately. In addition, please note no flow boundary along Balcones Fault Zone for Lower Trinity or note in model development section that your conceptual model changed for this boundary. Done (Page 18, paragraph 1). 16. S ection 6.0, page 17, fourth paragraph: I suggest mentioning that evapotranspiration is accounted for by the drain package since the loss is discussed, but the EVT package is not used. Done (Page 18, paragraph 3). 17. Section 7.0 page 18, last paragraph: Text notes that MODFLOW 2000 was used. Please clarify whether MODFLOW 96 or 2000 was used. “MODFLOW 200” changed to “MODFLOW 96” and associated reference changed (Page 19, paragraph 5). 18. Page 19, paragraph 1: Please also add that the thin slivers in the Edwar ds Aquifer in the east were not included in the model due to physical discontinuity between the units. Done (Page 20, paragraph 2). 19. Page 19, paragraph 2: Please explain in details on how the Hammett Shale was assigned in the model. This is an important e lement and should be discussed in some detail. This comment is already addressed in the following paragraph (Page 21, Paragraph 2). 20. Page 20, paragraph 3: Please define what model boundaries are before providing a list of the boundaries. Added a definiti on of model boundaries to the paragraph (Page 21, paragraph 4). 21. Page 21, paragraph 1: Please explain what are outer boundaries —areas outside the footprint of the aquifer. A definition of outer boundaries was added to Page 22, paragraph 4. 22. Section 7.4, page 21: Please discuss base of the model boundary as a noflow and other assignment of no flow boundaries. Please discuss why Colorado River was changed from constant head as it was modeled in previous version. Discussion of noflow boundaries appears on Page 22, paragraph 4. Discussion of why reservoirs, including those on the Colorado River were changed from constant head to river boundaries appears on Page 23, paragraph 2.

PAGE 186

185 23. Page 21, paragraph 3: Please explain why no General Head Boundary was assigned a long the Balcones Fault Zone in the east in the Lower Trinity Aquifer? The Hosston and the Sligo formations of the Lower Trinity Aquifer are in contact with appreciable segments of the Upper and Middle Trinity aquifers potentially allowing significant disc harge. This is an important conceptual issue that has ramifications on the calibration of the rest of the model. This discussion was added to Page 23, paragraph 1. 24. Section 8.0, pages 2223: Per GA M program guidelines, please discuss mean absolute error b etween measured hydraulichead and simulated hydraulic head shall be less than 10 percent of the measured hydraulichead drop across the model area and better if possible; the error shall not be biased by areas with considerably more control points than ot her areas (that is, not spatially biased); please discuss statistics and what they mean final calibration results shall report the mean absolute error and the mean error (Anderson and Woessner, 1992, p. 238241); please discuss difference between the tot al simulated inflow and the total simulated outflow (that is, the water balance) shall be less than one percent and ideally less than 0.1 percent as a modeling target; and please discuss calibration targets: wells(water levels/hydrographs), springs, lakes/reservoirs, rivers/streams, et cetera. See Page 24, paragraph 2; Sections 9.1 and 10.1, and Table 1003. 25. Section 8.0, page 23: please clarify if transient begins in 1980 or 1981. As it is written it appears we repeat 1980 as the steady state and then model 1980 again as the second stress period of the transient. Discussion on Page 24, paragraph 3. 26. Section 8.0, page 21, last paragraph: I suggest adding a note that the steady state model is a long stress period at the beginning of the transient model per GAM checklist page 6 of 11 Modeling Approach -“ ..discuss including the steady state model as part of the transient model ” Done (Page 24, paragraph 3). 27. Page 22, paragraph 6: This section discusses how recharge was assigned in the steady state model. For example, 3.5 percent of precipitation recharge for most of the model area, 5 percent of precipitation over a large swath of the Balcones Fault Zone in the east, and about 70,300 acre feet of streamflow loss through the Cibolo Creek. After all this, t he recharge amount matches the previous version of the model where precipitation was set at about 4 inches of precipitation. Need also more detailed discussion on how the recharge was assigned. Was it done using the Recharge package? Paragraph revised (Page 24, paragraph 5).

PAGE 187

186 28. Page 23, paragraph 1: Hydraulic conductivity assigned in the Upper Trinity Aquifer is considerably higher than the previous version of the model. For example, in the previous version a uniform value of 5 was assigned for the Upper Tri nity Aquifer. In the current version, a ssigned hydraulic conductivity values in the Upper Trinity Aquifer are 150 feet per day along Cibolo Creek, 15 feet per day within the Balcones Fault Zone, and 9 feet per day in the remainder of the aquifer . Please ex plain how were these values determined to be appropriate. Report should also contain measured values for comparison of the deviation of calibrated values from the measured values. Done (Page 25, paragraph 1). 29. Section 9.1 page 24, fourth paragraph: Please clarify statement that recharge to the Lower Trinity Aquifer is insignificant. Per the conceptual model, recharge to the Lower Trinity Aquifer should be zero. Done (Page 26, paragraph 1). 30. Section 10.0, page 25, third paragraph: Please include contour m aps and residuals of water levels in 1990 and 1997, per GAM checklist, page 8 of 11 Calibration, “Contour maps comparing simulated water levels to maps of the measure water levels shall be discussed for 1990 and 1997” . See Figures 1004 and 1005. 31. Sectio n 10.0, page 25, last paragraph: I suggest using scientific notation for the specific storage values to make them easier to read. Done (Page 28, paragraph 2). 32. Section 10.0, page 26, first paragraph: Please include a table listing final values for calibrated parameters for each layer per GAM checklist, page 8 of 11 Calibration, “ table of range and mean of horizontal and vertical hydraulic conductivity and storativity as used in the calibrated model” See Table 1002. 33. Page 26, paragraph 2: This section di scusses the specific yields and specific storage values used for transient calibration. These values are considerably low but was kept the same as in the previous version of the model. Low storage parameters were one of the main criticisms of the previous model. The extremely low storage values in the Trinity Aquifer System are the result of very low porosity of the aquifer rocks which have almost no inter granular porosity and at the regional scale have very low fracture porosity.

PAGE 188

187 34. Page 26, paragraph 2: S ome of the springs never flow throughout the simulation period. This may indicate that simulated water levels in part of the aquifers are considerably underestimated resulting in water level in the aquifer lying at lower than the elevation head in the drai n(s). Most of the springs in the study area are: (1) very small with measured discharge rates as low as 25 gallons per minute, (2) reflect discharge from local scale flow systems, and (3) are a small fraction of total discharge to surface water bodies for the aquifer. A small deviation of a few feet from measured water levels, well within model standards, is all that is needed for these springs not to flow during the simulation period. 35. Section 10.1, page 26: Please include a table with calibration statist ics for each model layer and include a map with the location of all wells used for the statistics per GAM checklist, page 8 of 11 Calibration, “a table listing mean absolute error and mean error for the transient calibration per layer for 1990 and 1997 and maps showing the locations of the wells used to develop the above scatter plots.” See Table 1003. 36. Section 10.1, page 26: Please provide tables of water budgets for 1990 and 1997 and a table listing stress periods and corresponding years per GAM checkli st page 8 of 11, “water budget for 1990 and 1997; a table showing stress periods and corresponding time periods for combined transient model” . See Table 1004. 37. Section 10.2, page 26: Please also consider transient sensitivity analyses for recharge, pumping, and vertical K per GAM checklist Page 8 of 11 Sensitivity Analyses, ”Model parameters include 2. vertical hydraulic conductivity, 5. recharge, 6. pumping” Measuring the sensitivity of the model to recharge, pumping, vertical hydraulic conductivity was conducted as part of steady state model construction and it is therefore unnecessary to repeat it again. Analysis of model sensitivity to storage parameters of primary importance in transient models. 38. Tables: Please adjust font to larger than 8. GAM guidelines require fonts no smaller than 6 points. 39. Page 37, Table 5 01: Please insert current recharge estimate from this model calibration. Also, include Anaya and Jones (2004) estimate mentioned in the report. Anaya and Jones (2009) recharge was added to Table 501. It is not an appropriate for calibration data to appear in the supporting data section of the report.

PAGE 189

188 40. Page 38, Table 5 02. Please add a column to sum up total pumpage by year. A column to sum up total pumpage by year was added to Table 503. 41. Page 55, Table 9 01: Please report values for springs. Discharge into the springs in the previous version was about 45,000 acre feet/yr. Does the estimate for streams include springs? If so, then stream discharges are lower by as much as 33 percent tha n the previous version. One can not compare results from the previous model with Table 901 because they cover different time periods. Spring discharge is included with streams. The table has been revised to reflect this. 42. Page 55, Tables 9 01, 9 02, and 903: The model wide, by layer, and by county water budgets should agree when summed. Please adjust or discuss reasons for differences in Section 9.1. The differences are due to rounding off of the numbers. The tables have been revised to remove the appa rent errors. 43. Page 56, Table 1001: According to this table, some key springs including Jacob’s Well Springs never flowed – an unreasonable result. It is perhaps OK to underestimate flow given the scale issue, but to have no flow at all, is a concern. Spri ngs flowed, some less, equal or more than estimated, in Mace and others (2000). Most of the springs in the study area are: (1) very small with measured discharge rates as low as 25 gallons per minute, (2) reflect discharge from local scale flow systems, and (3) are a small fraction of total discharge to surface water bodies for the aquifer. A small deviation of a few feet from measured water levels, well within model standards, is all that is needed for these springs not to flow during the simulation period. 44. Page 62, Table 1004: I calculated the recharge by aquifer for 1980 (from output.dat). I compared with the numbers in Table 1004 of the report, it seems to me they are different (in Upper and Middle Trinity Aquifers). Table 1004 has been revised to reflect the results of the latest version of the model. 45. Figures: Please update legends to include county boundary, model extent, and/or contour intervals, as appropriate. Done. 46. Page 57, Figure 301: The roads appear way too prominent in the figure. Sugg est use form of grays for roads so that model boundary stands out. Figure revised using color.

PAGE 190

189 47. Page 59, Figure 302: Figure does not show island of Edwards Group along Kendall, Kerr county boundary. Please update figure so it agrees with model. This fig ure, taken from Mace and others (2000), shows the official aquifers in the study area. The Edwards Group outlier mentioned above does not fall, despite being included in both the previous and updated versions of the model, is not included in any of the off icial aquifers. 48. Page 61, Figure 304: Please check whether the extension of the hatched areas to represent the EAA is correct. The area also covers Medina County, Uvalde County, and Trinity Glen Rose GCDs. Figure revised. 49. Page 65, Figure 309: Study are a symbol in the legend does not match the figure content. What do the dots signify? Please add explanations in the legend. Why use PG seal in one map only and not in others for consistency? Figure revised. 50. Page 67, Figure 311: Please report the time per iod for mean monthly precipitation data. Figure caption revised as requested. 51. Page 71, Figure 314: Please add Bexar Shale to figure since figure 3 17 shows this unit within the study/model area. Also suggest noting unconformities on column. Figure revi sed adding Bexar Shale. 52. Page 71, Figure 315: Only the extent of the Hammet Shale is shown on the map. Why not for the Sligo, Sycamore, and the Hosston formations? This figure was taken from Mace and others (2000). The updip extent is important because of its influence on cross formational flow between the Middle Trinity Aquifer (outcrop shown in this figure) and the Lower Trinity Aquifer that has no outcrop and therefore does not appear in this figure that shows surface geology. 53. Page 72, Figures 316: Where the caption for Figure 3 16 mentions things not on the surface geology map, you ought to add that the map also exclude s all the outcrops of Hensell and Cow Creek in the Guadalupe River valley as well as along the Pedernales and Colorado Rivers in the nort heastern most part of the study area. Springs from the Cow Creek and uppermost He nsell supply water to the Guadalupe River in Kendall and Comal Counties . East of Johnson City there are significant springs issuing from the C ow Creek and even the sand a nd limestone units of the Hammett.

PAGE 191

190 The updip limit of the Hammett Shale is shown because it is an important factor in the model. The outcrops of the Hensell Sand and the Cow Creek Limestone in the study area are very small outcrops along the Guadalupe, Colorado, and Pedernales rivers. These outcrops are too small to appear on the map in Figure 316. 54. Page 72, Figures 317: The northern part of the A A' cross section in Figure 3 17 is illogical and does not reflect the geology th at is exposed at the surf ace. I doubt that there is any Lower Glen Rose in well KKL 5741 301. The rock equival ent to the Lower Glen Rose and most of the Upper Glen Rose in that area is the Hens ell. North of Fredericksburg there is only a t hin unit of uppermost Glen Rose unde rlyi ng the Walnut and Edwards. It makes no sense to have a wedge of Lower Glen Rose in the northern well on this cross section. Figure 317 is a cross section by Ashworth (1983) based on interpretations of logs. 55. Page 76, Figure 504: There are some discrepancies in these diagrams. For the top of the Upper Trinity Aquifer, we expect to see use of outcrop control points but they don’t show up in figure (a). However, these control points show up in the figure for the base of the Upper Trinity Aquifer (b). Also, note if there are structure control points from well logs in northern parts of Gillespie County then there should be outcrop points farther north from it but they are not reported. This also brings the questions whether any control points should exist in northern parts of Gillespie County as the Upper Trinity Aquifer pinches out north of the Pedernales River. Figure 502(a) was revised to show the outcrop control points. The control points in northern Gillespie County already lie outside of the study area, beyond the northern model boundary. It is unnecessary to use outcrop control point that are even farther outside of the study area to interpolate the structure. 56. Page 78, Figure 506: Please insert structure control points for construction of the base of the Lower Trinity Aquifer (b) as was done in (a). Also please clarify gap between base of Middle Trinity [Figure 5 05 (b)] and parts of the Lower Trinity [Figure 506 (a)] are the Hammett Shale in caption and in text. See revised caption in Figure 504. The base of the Lower Trinity Aquifer was taken from the groundwater availability model for the Edwards Trinity (Plateau) Aquifer and therefore it is not appropriate to show control points. 57. Page 79, Figures 507 through 510: Please report units of measur ement . See revised captions in Figures 505 through 508. 58. Page 83, Figures 511 through 513: Figure shows water levels for 1977 to 1985. Please explain in the text why this time window was selected. If this is to show how the water levels remained under steady state then should we not have go ne back in time, e.g., 1965 or something like that as was done before . This would have also allowed addition of more water level control points for calibration in the north and the west.

PAGE 192

191 See Page 9, paragraph 3. Th e period 1977 through 1985 coincides with the steady state model period and the beginning of transient model period. 59. Page 83, Figure 513: No water level measurements reported for most of Gillespie and western part of Kerr Counties. There was no water le vel data available for those areas. 60. Figures 5 19 and 520, Net water level change: Please clarify if difference was for 1980 to 2000 as noted in caption or 1980 to 1997 as is modeled. Please update to 1980 to 1997 as necessary and also include maps for Ed wards and Upper Trinity. Figure 517 and 518 revised as requested. Limited water level data prevented construction of a map for the Edwards Group. 61. Page 91, Figure 519 and 520: Please explain the two shades for measurement points in the legend. Captions revised to explain the different colors. 62. Page 93, Figure 521: Please describe the figure in the caption. What does the ‘+’ and ‘ – ‘signify with respect to baseflow discharge? Figure 519 c aption revised as requested. 63. Page 103, Figure 531: It appear s that no hydraulic conductivity values were assigned within a few active cell areas in the Upper Trinity Aquifer. Please explain if these are inactive cells or cells turned off to enable convergence? Figure 29 revised. 64. Page 119, Figure 545: Trend (3) c ould simply be a mixture with water containing higher dissolved solids but not necessarily brine. Saturated brine has as much as 319 g/l total dissolved solids. Even a small fraction of mixing with this brine would have resulted in much higher concentratio ns of Na and Cl and Cl/Br . Saturated brine is an extreme case that it is unlikely that the groundwater in the Trinity Aquifer System will encounter . 65. Page 120, Figure 601: Please describe figures (a) and (b) in the caption. Also, altitude scale to the left is in blue where the other lines are in black. Also, please check whether there is discordance between fig 5 02 and this figure. In Figure 502, Hosston (LT) is juxtaposed against the Lower Glen Rose Limestone (MT) along the BFZ but here it is not. Please adjust cross formational flow from Lower Trinity to no flow or explain in text. Please account for Hammett Shale as a confining unit. Please clarify what a drain

PAGE 193

192 represents since the conceptual model also has symbols for springs, pumpage, and surface wa ter groundwater interaction. Possibly re label surface water groundwater interaction to include “reservoirs” and relabel drains to “gaining rivers”? Caption revised to describe (a) and (b). The altitude scale has been remove because it is a schematic cro sssection. One can not compare the Figure 601 cross section with Figure 502 because Figure 601 is an East West crosssection and Figure 502 is oriented NorthSouth. The text has numerous references to cross formational flow to/from the Lower Trinity A quifer as shown in the figure. The figure has been revised to show the Hammett Shale. “Drains” has been removed from the figure. 66. Page 121, Figure 701: Please explain why a few cells in Comal County were turned inactive when the aquifer is present. Some of these cells may also represent river cells? A few cells along the edge of the Upper Trinity Aquifer outcrop were turned off to address dry cell or stability issues (Page 20, paragraph 2). In the event that the inactive cell was a river/drain, the featur e was transferred to the underlying active cell. 67. Page 123, Figure 702: Please explain why there is so much difference between vertical leakance values in the north and the rest of the model area . How w ere the leakance zones determined? Was this zone assi gned to allow vertical communication between the Middle and Lower Trinity aquifers where the Lower Trinity Aquifer is absent? It is explained in the text that vertical leakance is inversely proportional to the thickness of the Hammett Shale confining uni t and reaches a maximum in the north where the confining unit is absent (Page 21, paragraph 2). 68. Page 123, Figure 901: Please report the range for category >5 inches. Is this for the steady state model? Please cite time period for recharge. Figure revised as requested. 69. Page 129, Figure 902: Please remove the simulated water level contours because they are misleading. These are water level difference maps . These can be displayed by show ing the differences either by points or contours made using these poi nts . Please provide the range for category more than 100 feet . The contours show model results and the points indicate the difference between simulated and measured water levels as indicated in the figure caption. No changes made to the figure. 70. Page 131 , Figure 903: Please report mean error and absolute error by aquifer which will better show any bias in calibration by aquifer . Per GAM guidelines, please reference and/or provide a map showing target wells used for cross plots by layer. Figure revised t o show results for each aquifer on a separate graph.

PAGE 194

193 71. Page 146, Figure 1003: Please group the targets by aquifers as in the steady state model and report the mean absolute error by aquifers . Per GAM guidelines, please reference and/or provide a map showing target wells used for cross plots by layer. Figure revised to show results for each aquifer on a separate graph. See Figures 1004 and 1005 for the maps. Comments on the model files for the Hill Country portion of the Trinity Aquifer GAM (1) Reviewed the recharge amount assigned along the Cibolo Creek. It was suggested in the report that an amount of 70,300 acre feet per year of water was assigned through the Cibolo Creek as per recent USGS investigation. Cibolo Creek flows over the Upper and Lower Glen R ose Limestone. Reviewed the recharge in the Cibolo Creek for the steady state model and found that the actual recharge through the Cibolo Creek is about 51,000 acre feet per year. We suggest changes to the recharge rate assigned in the model to reflect the text or provide appropriate justification for assignment of the lower recharge in the model. The 70,300 acre feet per year represents total recharge, diffuse and stream channel recharge, to the Cibolo Creek watershed where it overlies the Trinity Aquifer System, while the 51,000 acre feet per year value represents stream channel recharge only. (2) Compared various parameters reported in Table 901 with results obtained from running the steady state model. Small differences were noted between reported and mo del run results for several parameters (see table below). If the numbers were rounded to the next thousandth, please mention that in the table caption. Parameter Table 9 01 (ac ft/yr) Model run result (ac ft/yr) Streams 165,000 164,494 Well 17,000 16,6 68 Recharge 304,000 303,466 Edwards Aquifer (Balcones Fault Zone) 103,000 102,489 Table 901 was revised, rounding the numbers to hundreds of acre feet. The caption was revised to indicate this. (3) No simulated water level maps were reported in the text for the transient run. We suggest inclusion of some of the simulated water level maps for 1990 and 1997 of the transient period. Simulated water level maps show spatial distribution of water level contours and their position with respect to the streams an d lakes/reservoirs. Some dry cells were noted adjacent to cells that were turned off in the Upper Trinity Aquifer. Also, it was surprising to note that the simulated water level contours in the Lower Trinity Aquifer mimic the simulated water

PAGE 195

194 levels produce d in the overlying aquifers even though no streams run through the Lower Trinity Aquifer. In other words, streams assigned in the overlying aquifers have a strong effect in shaping the simulated water level contours in the Lower Trinity Aquifer which shoul d not have been the case. See Figures 1004 and 1005. It should NOT be surprising that water level contours in the Lower Trinity aquifer mimic the simulated water levels produced in the overlying aquifers even though no streams run through the Lower Tri nity Aquifer because the lowered water levels in the Middle Trinity Aquifer along the Guadalupe River will induce upward groundwater flow from the underlying aquifer. This occurs wherever there is vertical flow between aquifers. Similar relationships can be seen in the central Gulf Coast Aquifer GAM results. (4) Water budget results for various flow parameters have not been presented for the transient period. We suggest inclusion of water budget data for 1990 and 1997 to observe changes in the various flow par ameters during progression of the model calibration. See Table 1004. (5) There are some differences between Table 5 03 showing pumpage data used in the model compared to data retrieved from the well file. For example, the model has 10 percent more pumping in 1984 than what is reported in Table 5 03. All other years have 3 to 4 percent less, probably due to inactive or dry cells with pumping. Please clarify in text. Year Total pumpage from Table 503 (ac ft/yr) Total pumpage retrieved from well file (ac f t/yr) Difference (ac ft/yr) Percent difference 1980 17,149 16,678 471 3 1981 15,029 14,543 486 3 1982 14,999 14,508 491 3 1983 14,465 13,952 513 4 1984 13,888 15,274 1,386 10 1985 15,093 14,564 529 4 1986 14,999 14,458 541 4 1987 14,896 14,339 557 4 1988 17,340 16,766 574 3 1989 18,259 17,671 588 3 1990 19,463 18,861 602 3 1991 17,947 17,325 622 3 1992 18,775 18,132 643 3 1993 20,525 19,890 635 3 1994 21,406 20,749 657 3 1995 22,133 21,461 672 3 1996 23,461 22,773 688 3 1997 23,631 22, 927 704 3 Revised value 15,799 acre feet per year. With this value, the percentage difference is consistent with the rest of the pumpage data. Tables 503, 504, and 508 have been revise to reflect the change.

PAGE 196

195 (6) Suggest pulling all targets together into one location with a description of the procedure for calculating RMSE, MAE, etc. so that when the model is updated again, the statistics can easily be reproduced. There is a spreadsheet that has that data. (7) Figure 10 02, well 5761 507 observed data in tr ans_heads2.txt does not match figure. Model results do match figure. Please revisit and update text as needed. Revised Figure 10 02 to reflect the correct observed data.


printinsert_linkshareget_appmore_horiz

Download Options

close


  • info Info

    There are only PDFs associated with this resource.

  • link PDF(s)



Cite this item close

APA

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

MLA

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

CHICAGO

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

WIKIPEDIA

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