Uvalde Pool of the Edwards (Balcones Fault Zone) Aquifer Open Access

The Uvalde Pool is a statutorily defined subdivision of the Edwards (Balcones Fault Zone) Aquifer and constitutes the western portion of the San Antonio segment of the Edwards Aquifer (Senate Bill 1477, 1993). By rule, the boundary between the San Antonio and Uvalde Pools is defined by the Uvalde-Medina County boundary, but in a hydrogeological sense, the Uvalde Pool is bounded by a much more complex set of natural features. Early studies provided the foundation for the investigation of the western Edwards Aquifer (Figs. 1 and 2; Welder and Reeves, 1962). Subsequent studies by Maclay and Small (1986), Maclay and Land (1988), Maclay (1995), and Groschen (1996) provided assessments of the San Antonio segment of the Edwards Aquifer that included the Uvalde Pool. Rose (1972) synthesized the stratigraphy of the Edwards Aquifer in south Texas. Analysis by Small (1986) provided the structural geologic framework of the Edwards Aquifer. Subsequent studies by Clark and Small (1997) and Clark (2003) improved the understanding of the structural geology in Uvalde County. Rose (1972) and Hovorka et al. (1993, 1996) described the depositional environment of Edwards Aquifer strata in the study area, particularly in terms of how the depositional environment affects the hydraulic properties of the Edwards Aquifer. Significant analyses of the hydraulic properties of the San Antonio segment of the Edwards Aquifer have been performed (Small and Maclay, 1982; Mace, 1997; Mace and Hovorka, 2000; Hovorka et al., 1998), leading to a geostatistical assessment of the transmissivity by Painter et al. (2002). The prospect for conduit development in the Edwards Aquifer, including Uvalde County, was investigated by Worthington (2003) and Hovorka et al. (2004).

The Uvalde Pool is composed of a basin of layered carbonate aquifers (e.g., Austin Chalk, Buda Limestone, Edwards and Trinity Aquifers) formed by structural displacement and intense igneous activity (Ewing, 2004; Smith et al., 2002, 2008). In the vicinity of the city of Uvalde, the Balcones fault zone is manifested by a large number of faults with relatively small down-to-the-southeast displacements, together with several faults antithetic to the main trend (i.e., down-to-the-northwest displacement; Fig. 2). To the northeast of the city of Uvalde, fewer faults, one with a throw of 60 m, accomplish the same amount of total displacement. The geological formations and the systematic displacements associated with the Balcones fault zone are disrupted in central Uvalde County along a structural uplift referred to as the Uvalde salient (Welder and Reeves, 1962; Rives, 1967; Clark and Small, 1997; Clark, 2003). This structural high has the general shape of a ridge that is widest near the Edwards Aquifer recharge zone to the north and thins and plunges to the south (Green et al., 2006, 2009). The salient forms the eastern border of the Uvalde Pool.

The western boundary of the San Antonio segment was originally defined by what was believed to be a groundwater divide near Brackettville (Sayre, 1936; Sayre and Bennett, 1942; Bennett and Sayre, 1962; Bush et al., 1992; Hovorka et al., 1993; LBG-Guyton, 1994; Maclay, 1995; Groschen, 1996; Khorzad, 2003; Snyder, 2004), although the precise location of the divide varied among the studies. Subsequent analysis of the Edwards Aquifer in Kinney and Uvalde Counties indicated that the Edwards Aquifer in Kinney and Uvalde Counties forms separate pools, referred to respectively as the Kinney County Pool and the Uvalde Pool (Green et al., 2006). Justification for this separation and a description of the Kinney Pool are discussed in Green et al., Chapter 6, of this memoir. It is sufficient here to note that the western boundary of the Uvalde Pool is a hydraulic structural boundary in eastern Kinney County (Figs. 3 and 4; see also Green et al., this volume, Chapter 6).

The southern boundary of the Edwards Aquifer in the Uvalde Pool is modeled as a no-flow boundary aligned with the transition from fresh to saline water (typically set at 1000 mg/L total dissolved solids; Harden, 1968; Groschen and Buszka, 1997), with the exception of discharge occurring in the Leona River floodplain as underflow via paleo-streambeds and discharge from springs in the Nueces River channel south of the city of Uvalde (Green et al., 2008, 2009). The recharge zone of the Edwards Aquifer forms the northern boundary of the Uvalde Pool. The saturated thickness of the Edwards Aquifer thins in the recharge zone, which limits the northern extent of the Uvalde Pool. A schematic that illustrates the principal hydraulic features of the Uvalde Pool relative to the Kinney Pool to the west and the San Antonio Pool to the east is illustrated in Figure 3.

The principal geological features that define the structural and hydraulic relationships among the aquifers in Uvalde County are the Balcones fault zone, the Uvalde salient, a facies change in the Edwards Group, and the prevalence of igneous intrusions (Smith et al., 2002, 2008). The cumulative effect of these geologic factors is to restrict the eastward flow of groundwater in the Edwards Aquifer in eastern central Uvalde County. This restriction is referred to as the Knippa Gap (Maclay and Land, 1988). The hydraulic communication between the Uvalde and San Antonio Pools of the Edwards Aquifer via the Knippa Gap is variable and depends on groundwater elevation.

In addition to the Edwards Aquifer, there are significant secondary aquifers in the Uvalde Pool, including the Austin Chalk, Buda Limestone, and the Leona Formation (Green et al., 2008, 2009). Similarities in groundwater elevations to the west of the Uvalde salient support the conceptualization that the Edwards, Buda Limestone, and Austin Chalk Aquifers are in hydraulic communication with the Uvalde Pool. This hydraulic communication is due to the large offset associated with the high degree of faulting in the Uvalde Pool.

The Uvalde Pool portion of the Edwards Aquifer is classified as confined and is designated as the confined zone; however, since the Edwards Aquifer is in hydraulic communication with the Austin Chalk and Buda Limestone Aquifers, which are unconfined, the Edwards Aquifer in the Uvalde Pool is, in reality, unconfined (Fratesi et al., 2015). Using a similar argument, the Leona Aquifer is apparently hydraulically connected with the Edwards, Buda Limestone, and Austin Chalk Aquifers at the headwaters of the Leona River from Highway 90 in the city of Uvalde south to Fort Inge (Green et al., 2008, 2009). To the south of Fort Inge, a difference in groundwater elevations between the Leona and Edwards Aquifers suggests that the Leona Aquifer is not hydraulically connected with the Edwards Aquifer in this locality.

The principal source of recharge to the Uvalde Pool is the Nueces River, both as surface flow and as subsurface flow, which occurs mostly as underflow in the Trinity and Edwards bedrock associated with the Nueces River channel. Groundwater flow pathways with enhanced permeability are believed to have developed in the Nueces River channel (Woodruff and Abbott, 1979, 1986; Abbott and Woodruff, 1986; Green et al., 2014). There also is an opportunity for interformational flow from the Trinity Aquifer to the Edwards Aquifer to the north of the Uvalde Pool in interstream regions. The quantity of this interstream, interformational flow is not well defined, but it is not believed to be significant in Uvalde County (Maclay and Land, 1988). The Dry Frio River is interpreted to recharge the Edwards Aquifer in proximity to the Knippa Gap. The Frio River clearly appears to recharge the San Antonio Pool of the Edwards Aquifer. Based on this conceptualization, neither the Dry Frio River nor the Frio River recharges the Uvalde Pool. Both recharge the San Antonio Pool.

Water is naturally discharged from the Uvalde Pool via the Knippa Gap, Leona Formation gravels in the Leona River floodplain, and the Nueces River (Fig. 3). The amount of water, if any, that is discharged at these particular locations is a function of the groundwater elevation. The highest point of discharge from the Uvalde Pool is the Nueces River via the Buda Limestone and Austin Chalk Aquifers where they crop out in the Nueces River channel. This is because of the hydraulic communication among the Austin Chalk, Buda Limestone, and Edwards Aquifers in the Uvalde Pool. Water from the Soldiers Camp (or Rose) Springs, located where the Austin Chalk crops out at ~261 m above mean sea level (amsl) in the Nueces River channel south of the city of Uvalde, is in hydraulic communication with the Edwards Aquifer due to the significant offset associated with faulting, which provides an opportunity for Edwards Aquifer water to discharge (indirectly via the Austin Chalk Aquifer) into the Nueces River (Fig. 4; Brune, 1981). There also are unnamed springs in the Nueces River downstream from Soldiers Camp Springs that allow discharge from the Uvalde Pool when groundwater elevations are low. These unnamed springs, at 250–256 m msl, appear to be the farthest downgradient opportunity for springs fed by the confined zone of the Edwards Aquifer to discharge into the Nueces River (Green et al., 2009).

It is important to note that rivers (such as the Nueces) that provide recharge in the upper reaches over the Edwards Aquifer recharge zone can also act as discharge points where the rivers exit the recharge zone. Downstream from the Uvalde Pool, flow is observed in the Nueces River beginning at Soldiers Camp Springs under most conditions, even when no flow is observed immediately upstream. U.S. Geological Survey gauging station 08192000 is located on the Nueces River ~10 km downstream from Soldiers Camp Springs and the southern limit of the Austin Chalk Aquifer outcrop. Measurements at this gauge allow calculation of the difference in flow in the Nueces River between Soldiers Camp Springs and the gauging station when there is no river flow in the reach upstream from Soldiers Camp Springs. The base flow of the Nueces River along this reach is believed to originate from the springs associated with the Austin Chalk Aquifer outcrop. Because of the hydraulic communication between the Austin Chalk and Edwards Aquifers, flow measurements at gauging station 08192000 are therefore a measure of discharge from the Edwards Aquifer to the Nueces River during times when there is no surface flow in the Nueces River upstream of the springs.

Mean monthly discharge has been recorded at this Nueces River gauging station since 1939 and has been zero only at times during the drought of the 1950s and during the drought that started in 2011. Comparison of Edwards Aquifer water levels recorded at J-27, the official groundwater index well in the Uvalde Pool, with discharge measured at U.S. Geological Survey gauge 08192000 indicates that river base flow ceased when the water level at J-27 dropped below ~250–253 m msl (i.e., the elevation of the southernmost springs where the Austin Chalk crops out on the Nueces River). Exceptions occur when river flow is due to excess surface runoff.

The Pleistocene-age Leona Formation gravels in a paleo–stream channel in the Leona River floodplain provide an opportunity for discharge from the Uvalde Pool of the Edwards Aquifer at elevations lower than that in the Nueces River. The basal elevation of the paleo–-stream channel is variable, with estimated elevations at 256 m msl near Highway 90 and decreasing to less than 244 m msl at Fort Inge (Fig. 5). According to the structural conceptualization by Green et al. (2008, 2009), the Leona Formation gravels are juxtaposed with the Edwards Aquifer near Highway 90 and with the Buda Limestone Aquifer near Fort Inge (Fig. 5). In reality, the structural geology is more complex than shown in Figure 5, and the Leona Formation gravels are probably juxtaposed with the Buda Limestone Aquifer near Highway 90 based on local well logs. However, the Buda Limestone and Edwards Aquifers are in hydraulic communication in this area, based on similar groundwater elevations and water chemistry, so that water that discharges into Leona Formation gravels near Highway 90 originates in the Edwards Aquifer. It is possible that the Leona Formation gravels are also in hydraulic communication with the Austin Chalk in the Leona River channel north of Fort Inge. If so, this could allow for an opportunity for the Edwards Aquifer to discharge to the Leona Formation gravel at an elevation lower than 244 m msl, possibly as low as 238 m msl. There is no evidence, however, that the Leona Formation directly overlies the Austin Chalk Aquifer in the Leona River floodplain south of Fort Inge (Green et al., 2009).

Some wells in the Leona River floodplain near Zavala County did not go dry during the 1950s drought of record (Vic Hildebran, 2005, personal commun.). This provides evidence that the gravels are recharged at elevations lower than 247 m msl, i.e., the historic low water level recorded at the Uvalde Pool index well (J-27) on 13 April 1957. Because the base of the gravel is encountered at a depth of no greater than 20 m below the Leona River floodplain ground surface (Cary Spurgeon, 2005, personal commun.), the floodplain topographic elevation cannot be higher than ~264–267 m msl to allow for recharge when groundwater at J-27 is at historic lows (i.e., 247 m msl). This ground elevation occurs near Fort Inge, which suggests the lowest point of discharge to the Leona Formation gravel is near Fort Inge. An inspection of the geologic map of the area (Clark, 2003; Fisher, 1983) suggests that the Buda Limestone, or possibly the Austin Chalk, directly underlies the Leona Formation gravels at this point (Green et al., 2009).

Location of the Knippa Gap, as illustrated in Figure 6, is consistent with the location originally designated by Maclay and Land (1988). The structure of the Knippa Gap is defined by multiple geological factors, including the facies change of the Edwards Formation from the Maverick Basin to the Devils River Trend, the Balcones fault zone, the Uvalde salient, and volcanic intrusions. Maclay and Land (1988) correctly identified the Knippa Gap as a constriction to groundwater flow in the Edwards Aquifer. Clark et al. (2013) presented a similar conceptualization of the geology in this area; however, they designated the Knippa Gap location as downgradient and to the southeast of the location noted in Figure 6. This constriction is bordered on the north by an extension of Cooks fault located at the southern extent of the Edwards Aquifer recharge zone (Fig. 6; Clark, 2003; Adkins, 2014).

The Knippa Gap is bordered on the south by an unnamed fault that is coincident with the east-west–oriented portion of Ranch Road 2690 where it intersects U.S. Highway 83. The terrain south of this fault is dominated by either igneous intrusions or areas of the Uvalde salient where the Salmon Peak Formation (i.e., the uppermost and most permeable unit in the Maverick Basin facies of the Edwards Aquifer) is elevated such that much or most of the permeable portion of the Edwards Aquifer is above the water table. Two of the larger intrusions to the south of the unnamed fault that defines the southern border of the Knippa Gap are Blue Mountain and Black Mountain (Fig. 6). The uplifted area between these two intrusive bodies summits at 378 m msl and is composed of Cretaceous limestones, with the Edwards Limestone exposed at the ground surface. The thickness of the Salmon Peak Formation averages 107–116 m in the area (Clark and Small, 1997; Blome et al., 2005), and so the base of the Salmon Peak Formation in this uplifted area would be ~262 m msl. Because groundwater elevations rarely exceed 268 m msl in this area, the Salmon Peak Formation of the Edwards Aquifer is mostly dewatered at this location (Green et al., 2009).

Using the geologic constraints discussed above, the narrowest width of the Knippa Gap is ~4 km near the intersection of U.S. Highway 83 and Ranch Road 2690 (Fig. 6). The top and bottom of the Salmon Peak Formation have been identified using driller logs and an assumed thickness of 107–116 m. With a base elevation of 136–179 m msl, the Knippa Gap is the lowest point of discharge from the Uvalde Pool (Green et al., 2009).

Wells in the Knippa Gap tend to have high pumping capacity and do not have to penetrate the entire Salmon Peak Formation to provide sufficient capacity, thereby attesting to the high permeability of the formation. Ground elevations of the wells in this area vary from 319 to 329 m msl. According to their well logs (Green et al., 2009), the top of the Salmon Peak Formation is estimated at ~252–285 m msl in the central portion of the Knippa Gap, with higher elevations observed to the north and lower elevations observed to the south. Based on this conceptualization, the Salmon Peak Formation may not be fully saturated in the Knippa Gap.

Elevations of groundwater measured at the index wells in Uvalde (J-27), Medina (Hondo), and Bexar (J-17) Counties during 2006 are presented in Figure 7. This graph illustrates two distinct differences between elevations measured at the Uvalde County index well and those elevations measured at the Medina and Bexar County wells. The first difference is seen in the elevation at the Uvalde County index well (J-27) relative to the elevation at the Medina County (Hondo) index well, which is much greater than the difference between the Medina County and Bexar County index wells, even though the distance between the Uvalde County and Medina County wells is comparable to the distance between the Medina County and Bexar County index wells. The higher groundwater elevations at the Uvalde County index well are attributed to the constricting effect of the Knippa Gap on groundwater flow from the Uvalde Pool to the San Antonio Pool.

The second observation is the lower-amplitude changes in groundwater elevations observed at the Uvalde County well relative to either the Medina County or Bexar County index wells. The maximum rate of change in the groundwater level observed in the Uvalde Pool has been no greater than ~0.15 m/mo (Fig. 8). This contrasts with groundwater elevation changes in the confined portion of the San Antonio Pool (index well J-17), which can change more than a meter per day. This modulating effect is attributed to two factors: (1) The Edwards Aquifer in the Uvalde Pool is essentially at water-table conditions due to the fact that the Edwards, Buda Limestone, and Austin Chalk Aquifers are in hydraulic communication, and (2) there is an increased capacity for discharge from the Uvalde Pool when the groundwater elevation in the Uvalde Pool is raised. This explanation accounts for the subtle changes in groundwater elevations that are observed within the Uvalde Pool compared with much larger changes observed in the San Antonio Pool to the east of the Knippa Gap.

Once water exits at the Nueces River, through the Leona River channel gravels, or at the Knippa Gap discharge points, that water is lost from the Uvalde Pool of the Edwards Aquifer. Surface flow from the Nueces, Frio, Dry Frio, and the Leona Rivers and subsurface flow through the Leona Formation gravel in the Leona River floodplain where they exit the Edwards Aquifer recharge zone provide recharge to the Carrizo-Wilcox Aquifer and other aquifers to the south, or it continues as surface flow in rivers that eventually discharge into the Gulf of Mexico. Discharge via the Knippa Gap flows into the San Antonio Pool and continues to the east.

The Uvalde Pool has relatively limited areal extent over the area in which the Edwards Limestone is fully saturated. The Uvalde Pool spans from approximately west of the Nueces River to the west, possibly extending slightly into eastern Kinney County, the recharge zone to the north, Knippa Gap to the east, and the saline-water zone and Uvalde salient to the south, for a total area of ~259 km². As discussed in the previous section, the Uvalde Pool, including the Knippa Gap, is characterized as unconfined. If a porosity of 10% is assigned to the Devils River Trend facies (Hovorka et al., 1996), then the change in storage from the record high of 271 m msl in 1987 to the record low of 247 m msl in 1957 at J-27 equates to ~6.17 × 10⁸ m³ of storage. In summary, the Uvalde Pool is characterized as being highly transmissive, mostly due to karst development of the Edwards Aquifer, but having relatively limited storage capacity.

Recharge (and discharge) of the Uvalde Pool is (are) highly variable. Over the 1934–2013 period of record, recharge of the Edwards Aquifer via the West Nueces and Nueces Rivers was calculated by the U.S. Geological Survey to vary from a low of 1.06 × 10⁷ m³ in 1934 to a high of 5.94 × 10⁸ m³ in 2004 (Tremallo et al., 2014). Similar high levels of recharge occurred in 1990 and 2007. The median groundwater recharge to the Edwards Aquifer from the Nueces–West Nueces River watersheds calculated by the U.S. Geological Survey was 1.23 × 10⁸ m³/yr for the period 1934–2013 and 9.00 × 10⁷ m³/yr for the period 2003–2013. The median annual discharge by pumping and spring discharge for Uvalde County was calculated to be 9.24 × 10⁷ m³/yr for the period 1934–2013 and 9.66 × 10⁷ m³/yr for the period of 2003–2013. Discharge in Kinney County by spring flow and pumping is included in this estimate for Uvalde County (Tremallo et al., 2014).

U.S. Geological Survey recharge estimates of the Uvalde Pool (Tremallo et al., 2014) are predicated on several important assumptions. Both recharge and discharge are regional estimates (i.e., for the combined West Nueces River–Nueces River watersheds and the combined Dry Frio River–Frio River watersheds). An additional limitation in these calculations is that recharge that occurs upstream from the river gauges used to calculate recharge is not captured in the calculations. Since river gauges are placed close to the recharge zone upstream boundary, recharge that does occur in the contributing zone is not included in either the U.S. Geological Survey or the Hydrologic Simulation Program FORTRAN (HSPF) calculations of recharge to the Uvalde Pool of the Edwards Aquifer (Clear Creek Solutions, Inc., 2009, 2012, 2013; Tremallo et al., 2014). Interformational flow studies by the Edwards Aquifer Authority are actively exploring this enduring issue (https://www.edwardsaquifer.org/science-and-maps/research-and-scientific-reports/interformational-flow-study).

Discharge from the Uvalde Pool has been quantified using results from recent investigations. Field-study results performed on the Leona River floodplain at a location 6 km south of Uvalde, Texas, were evaluated to determine the quantity of water discharged from the Edwards Aquifer through the Leona River floodplain (Green, 2003; Green et al., 2008). Principal components to the field studies were electrical resistivity and magnetic surveys to determine the lateral and vertical extents of sand and gravel deposits and an aquifer test to determine the hydraulic properties of the Leona Formation Aquifer. These studies were augmented by well logs to corroborate the geophysical survey interpretation.

Results from the geophysical surveys were used to delineate a paleo–stream channel embedded in the fluvial sediments of the Leona River floodplain (Fig. 9). Flow through the Leona Formation gravels that compose the paleo–stream channel was calculated to be ~9.13 × 10⁷ m³/yr (Green et al., 2008). When combined with the estimate of 9.00 × 10⁶ m³/yr of surface-water flow in the Leona River, the total quantity of water discharged from the Edwards Aquifer through the Leona River floodplain may be as great as 1.00 × 10⁸ m³/yr (Green et al., 2008). During periods of severe drought, discharge through the Leona River floodplain could be less than 2.31 × 10⁷ m³/yr, with no surface flow and reduced underflow (Green et al., 2008). Even this lowflow calculation exceeds the 1.38 × 10⁷ m³/yr target value for discharge from the Leona Springs specified by Lindgren et al. (2004) for steady-state calibration of their groundwater model of the Edwards Aquifer.

Discharge via Knippa Gap can be estimated using the water budget of the Uvalde Pool. The Uvalde Pool water budget can be posed as:

where:

• $QNR−GWin$ = recharge from the Nueces River as groundwater,

• $QNR−SWin$= recharge from the Nueces River as surface water,

• $QKG$= discharge out via Knippa gap,

• $QLR$= discharge out via Leona River,

• $QNRout$= discharge out via Nueces River springs,

• $Qpumping$=.discharge out via pumping, and

• ΔS = change in storage in the Uvalde Pool.

The maximum rate at which J-27 drops during times of drought provides an indication of the minimum flow rate via Knippa Gap. The minimum discharge is assumed to have occurred in the early 1950s. The water level at J-27 dropped by ~0.15 m/mo. This equates to ~4.74 × 10⁷ m³/yr. The water level at J-27 during the drought of the 1950s was below 258 m msl, and so there was minimal discharge via the Leona Formation gravel in the Leona River floodplain and no flow from the springs on the Nueces River. Discharge via the Leona Formation gravel, Q_LR, during periods of drought is estimated at 2.31 × 10⁷ m³/yr (Green et al., 2008).

Pumping in Uvalde County has been highly variable over the past 80 yr (Fig. 10; Green et al., 2009). Approximately 56% of pumping of Edwards Aquifer wells in Uvalde County occurred in the Uvalde Pool (i.e., 6.24 × 10⁷ m³/yr for the period of 1998–2007), with the remainder pumped from the San Antonio Pool. This estimate for pumping does not include amounts pumped by wells in the Buda Limestone, Austin Chalk, and Leona Gravel Aquifers. Pumping, Q_pumping, in the early 1950s was ~1.85 × 10⁷ m³/yr for Uvalde County (Fig. 10), of which 56%, or 1.04 × 10⁷ m³, is estimated to have been from the Uvalde Pool. The U.S. Geological Survey measured recharge to the Uvalde Pool (⁠$QNR−SWin$⁠) at an average of 2.40 × 10⁷ m³/yr for the period 1950–1954 (Tremallo et al., 2014). As previously discussed, recharge to the Uvalde Pool from the Nueces River as groundwater, $QNR−GWin$⁠, is not included in this measurement. Discharge during this period via the Knippa Gap is estimated at 5.60 × 10⁷ m³/yr plus recharge from Nueces River as groundwater, which is unknown.

Volumetric flow through the Knippa Gap was also estimated using Darcy’s law. The use of Darcy’s law to determine flow velocity in a karst aquifer, such as the Edwards Aquifer, and particularly in the Knippa Gap, where well-developed conduit flow is likely, may not be appropriate; however, groundwater flow out of the Knippa Gap using Darcy’s law and estimates of the dimensions and hydraulic properties of Knippa Gap may be justified when strictly limited to a volumetric calculation. The Knippa Gap near the intersection of U.S. Highway 83 and Ranch Road 2960 was selected for analysis of flow. The width of the cross section is estimated at 4.0 km (Fig. 9). For drought conditions when groundwater elevation at J-27 is 253 m msl, and if the base of the Edwards Aquifer in the Knippa Gap is estimated to be 162 m msl, then the saturated thickness of the aquifer in Knippa Gap is 91 m. For a hydraulic gradient if 0.001 m/m and hydraulic conductivity of 450 m/d (Hovorka et al., 1996, 1998; Mace and Hovorka, 2000; Painter et al., 2002, 2007), flow through the Knippa Gap is calculated to be 6.17 × 10⁷ m³/yr for drought conditions. There was no discharge from the Uvalde Pool via the Nueces River and minimal discharge from the Leona Formation gravels in the Leona River channel during drought conditions. Most discharge from the Uvalde Pool is via the Knippa Gap and by pumping during periods of drought.

Evaluation of discharge via the Knippa Gap, Q_KG, at high stage in the Uvalde Pool was evaluated for J-27 at 270 m msl. Compared with a period of drought, the saturated thickness of the Edwards Aquifer in the Knippa Gap is increased to 108 m. If the hydraulic conductivity and width of the Knippa Gap are assumed constant and the hydraulic gradient is increased to 0.002 m/m, discharge from the Uvalde Pool via the Knippa Gap increases to 1.45 × 10⁸ m³/yr. This is approximately one quarter of the estimated discharge in 2004 when groundwater at J-27 was 270 m msl and recharge to the Uvalde Pool by the Nueces River watershed as measured by the U.S. Geological Survey was 5.94 × 10⁸ m³/yr (Tremallo et al., 2014).

Assumptions about the base elevation of the Edwards Aquifer in the Knippa Gap and constant hydraulic conductivity in the Edwards Aquifer may not be valid. The Upper Salmon Peak Formation may be more permeable with greater karst development than the lower portion of the Edwards Aquifer in the Knippa Gap (Maclay 1995; Clark and Small, 1997). If valid, the 108 m thickness of saturated aquifer that would be added during high stage could convey more water per unit thickness. In addition, more water could be conveyed if the base of the Knippa Gap were deeper. Regardless, this added capacity is unlikely to be able to account for the tremendous increase in recharge to the Uvalde Pool witnessed in 2004 and other wet years. Other avenues of discharge from the Uvalde Pool are available to accommodate this increased quantity, namely, spring discharge on the Nueces River and discharge via the paleo–stream channel deposits in the Leona River channel. These two features act as overflow outlets.

Discharge via the Leona River channel has been calculated to be as high as 1.00 × 10⁸ m³/yr during wet years (Green et al., 2008). Pumping in Uvalde County was 1.13 × 10⁸ m³ in 2004 (Tremallo et al., 2014), of which 6.30 × 10⁷ m³/yr was from the Uvalde Pool. Discharge via the Nueces River springs is difficult to calculate. Flow measurements downstream from the springs include storm surge. Base-flow separation of flow measurements at the Nueces River gauge suggests that storm flow accounts for approximately half of all Nueces River flow.

Based on the conceptualization described in this document, actual recharge of the Uvalde Pool from the Nueces River is determined to be greater than either the U.S. Geological Survey or the HSPF calculations indicate (Tremallo et al., 2014; Clear Creek Solutions, Inc., 2009, 2012, 2013). The increased recharge is due to Nueces River underflow and interformational flow from the Trinity Aquifer to the Edwards Aquifer, neither of which is accommodated in those earlier recharge calculations. Last, the Knippa Gap appears to be incapable of accommodating the large quantities of recharge to the Uvalde Pool experienced during 1990, 2004, and 2007. This suggests that either the Nueces River springs or Leona River channel underflow areas have greater capacity for discharge than previously thought.

The Uvalde Pool forms the western portion of the San Antonio segment of the Edwards (Balcones Fault Zone) Aquifer. The northern extent of the Uvalde Pool is defined by the recharge zone, and the southern extent is defined as the transition from fresh to saline water. The Uvalde Pool extends from eastern Kinney County on the west to the Knippa Gap on the east. The Knippa Gap is a structurally controlled constriction to groundwater flow in the Edwards Aquifer. This constriction results in significantly higher groundwater elevations in the Uvalde Pool compared with the San Antonio Pool. For this reason and others, the Uvalde Pool is governed by different drought management triggers (i.e., groundwater elevations) at index well J-27 in the Uvalde Pool than at index well J-17 in the San Antonio Pool (Texas Senate Bill 1477, 1993).

Recharge to the Uvalde Pool is from surface-water flow and underflow associated with the Nueces River. The Uvalde Pool discharges to springs in the Nueces River channel, paleo–stream channel gravels in the Leona River channel, and the Knippa Gap. Within the Uvalde Pool, the Edwards Aquifer is in hydraulic communication with two secondary aquifers, the Buda Limestone and the Austin Chalk. Although the Edwards Aquifer is confined in the Uvalde Pool, it acts as an unconfined aquifer due to its hydraulic communication with aquifers exposed at the ground surface.

Recharge to and discharge from the Uvalde Pool can be highly variable. When recharge is low, discharge via the springs on the Nueces River can cease, and discharge via the paleo–stream gravels in the Leona River floodplain is greatly diminished. During these times, discharge from the Uvalde Pool is mostly via the Knippa Gap. During times of increased recharge, discharge via the Nueces River and the Leona River channel acts as an overflow feature, supplementing increased levels of discharge via the Knippa Gap. This added capacity for discharge via all three avenues of discharge moderates groundwater elevations in the Uvalde Pool during periods of high recharge and limits the maximum groundwater elevation attainable in the Uvalde Pool.