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

The San Antonio Pool spans that portion of the Edwards (Balcones Fault Zone) Aquifer from Knippa Gap on the west (eastern Uvalde County, Texas) to the groundwater divide at Onion Creek in Hays County, Texas, on the east (Fig. 1). The aquifer is characterized by three distinct hydrogeological zones: contributing zone, recharge zone, and artesian zone. Surface-water and groundwater flow in watersheds in the contributing zone is mostly north-to-south in the western portion of the San Antonio Pool, shifting to northwest-to-southeast in the eastern portion of the San Antonio Pool. Flow within each watershed is separate in the contributing zone but commingles once the aquifer transitions from unconfined to confined conditions and the direction of flow changes to eastward and eventually northeastward near the points of major discharge at Comal and San Marcos Springs.

The major watersheds from west to east in the San Antonio Pool include the Dry Frio, Frio, Sabinal, Medina, Cibolo, Guadalupe, and Blanco Rivers (Fig. 1). There are additional interwatershed areas in the recharge and artesian zones that do not extend north into the contributing zone. Due to the diversity in recharge from the watersheds, the San Antonio Pool is described by subareas from west to east in the following sections. Note that much of this discussion is from Fratesi et al. (2015).

The western section of the San Antonio Pool is recharged by the Dry Frio, Frio, and Sabinal River watersheds in addition to recharge from the Uvalde Pool via the Knippa Gap. Early assessments of the geology, hydrology, and water resources of the western San Antonio Pool by the U.S. Geological Survey (i.e., Holt, 1956; Maclay and Small, 1984; Maclay and Land, 1988; Maclay, 1995) have been augmented with more focused recent studies on geologic structure (Small and Clark, 2000; Blome et al., 2005, 2007; Green et al., 2006; Clark et al., 2006, 2009; Pantea et al., 2008), groundwater flow paths (Clark and Journey, 2006; Green et al., 2006, 2009a, 2009b, 2012b), and water balance (Lambert et al., 2000; Slattery and Miller, 2004; Okerman, 2005; Pedraza and Okerman, 2012; Green et al., 2012b; Fratesi et al., 2015) to develop a conceptual model of the western San Antonio Pool (Fig. 2).

The Dry Frio River recharges the Edwards Aquifer in close proximity with, and downgradient to, the Knippa Gap. There is no evidence that recharge from the Dry Frio River watershed enters the Uvalde Pool. Mapped concentrations of specific conductance provide a strong case that the low-specific-conductance water from groundwater sampling in the Dry Frio River watershed recharges the Edwards Aquifer downgradient from the Uvalde Pool (Fig. 3; Green et al., 2006). The Dry Frio and Frio Rivers seldom flow in the lower reaches of the recharge zone, a consequence of the fact that long reaches of both the Dry Frio and Frio Rivers pass over the recharge zone. This results in all base flow from these two rivers infiltrating in the contributing and recharge zones of the San Antonio Pool.

Holt (1956) first characterized large faults in the San Antonio segment as groundwater flow barriers and characterized large faults in northern Medina County as flow-path–controlling features. Holt stated that groundwater flows in dissolution channels along fractures generally parallel to the fault pattern. Holt also noted that faults with sufficiently large displacements to offset the aquifer from itself can form barriers to groundwater flow (Fig. 4). Water entering the Edwards Formation limestone when Medina Lake levels are high and water from Diversion Lake (located on the Edwards Formation limestone) flow downdip to the south, where movement is diverted to the southwest, along Haby Crossing fault, where the throw is less than the thickness of the aquifer. From there, groundwater flows across the fault and into the downthrown block to the south.

Maclay and Small (1984), Maclay and Land (1988), Maclay (1995), and Lindgren et al. (2004) retained the concept that Balcones fault zone faults in northern Medina County act as barriers to flow and that these barriers divert groundwater flow from northeast to southwest in Medina County. In this conceptualization, a fault was considered a barrier if offset equated to 50% or greater of the aquifer thickness (Fig. 4). Clark and Journey (2006) evaluated geological structural data, hydraulic correlations, and water chemistry to support the conceptualization of faults acting as barriers to flow in Medina County, although the hydraulic and water chemistry data were mostly from the eastern half of the county. Additional analysis of geologic structure and water chemistry by Green et al. (2012b) confirmed this conceptualization (Fig. 5). However, in northern Bexar County, tracer testing results indicated rapid movement of groundwater across a fault with more than 50% displacement (Johnson et al., 2010). The fault plane at that location was suspected of being a pathway between upper and lower units within the Edwards Formation limestone. Faults as preferential flow paths have not been tested by tracing in northern Medina County.

The water budget of the western San Antonio Pool, prior to the addition of water diverted from northeast Medina County, can be expressed

where Q_DIS = discharge to eastern San Antonio Pool; Q_KG = recharge from the Knippa Gap; Q_DFR = recharge from Dry Frio River Basin; Q_FR = recharge from Frio River Basin; Q_SR = recharge from Sabinal River Basin; Q_if = interformational flow; ΔS = change in storage; and Q_pumping = discharge by pumping.

For this evaluation, steady-state conditions were assumed; thus, there will be no change in storage, ΔS. The points of natural discharge from the western San Antonio Pool are several springs on the Frio River (Fig. 6). These springs appear as waterholes with discharge sufficiently low that these springs do not significantly affect the San Antonio Pool water budget. The first opportunity for natural discharge from the San Antonio Pool is San Antonio and San Pedro Springs in Bexar County (Fig. 1). There are very few data quantifying the contribution of interformational flow to the Edwards Aquifer; therefore, it is considered negligible.

Total average annual recharge to the San Antonio segment of the Edwards Aquifer for the period 1934–2006 has been estimated using gain-loss river flow measurements to be 887.8 Mm³, of which ~25% is attributed to the three principal watersheds in Medina County, namely, the Dry Frio River, Frio River, and the Sabinal River watersheds (Tremallo et al., 2014). Recharge estimates published in Tremallo et al. (2014) assumed that river-gauge measurements accurately reflect the amount of water that enters and exits the Edwards Aquifer along reaches of the creeks and rivers that cross the recharge zone. The accuracy of this recharge calculation also is predicated on the assumption that little or no subsurface flow occurs in the floodplains where the creeks and rivers enter and exit the recharge zone.

The combined area of recharge of the Dry Frio, Frio, and Sabinal River watersheds above the confined zone is 2405 km² (LBG-Guyton Associates and Aqua Terra Consultants, 2005). Based on river discharge measurements, the U.S. Geological Survey calculated median and mean recharge by the combined Dry Frio, Frio, and Sabinal River watersheds for the period 2001–2011 to be 1177.5 Mm³/yr and 250.5 Mm³/yr, respectively (Tremallo et al., 2014). A mean recharge of 177.5 Mm³/yr over 2405 km² equates to 73.9 mm/yr. This value is somewhat less than the 84.8 mm/yr recharge estimate for the Frio River Basin derived using base-flow separation (Fratesi et al., 2015). Given that the U.S. Geological Survey calculation does not account for recharge that occurs in the contributing zone upstream of the river gauges, values of recharge for the Dry Frio, Frio, and Sabinal River watersheds in excess of the U.S. Geological Survey calculations appear to be justified.

The watershed located between the Sabinal River and Medina River watersheds includes the floodplains of Seco, Parker, Live Oak, Hondo, Verde, Elm, and Quihi Creeks. These watersheds coalesce to form a contiguous expanse of Leona Formation sediments in an outwash plain (Figs. 7 and 8). The combined watersheds are referred here as the Seco Creek watershed for convenience. Water that is conveyed south in the Seco Creek watershed in central Medina County (below the recharge zone) as either surface water or shallow groundwater is not in hydraulic connection with the Edwards Aquifer. Based on this conceptualization, the only recharge that occurs in this region would have to be in the upland headwaters of the watershed where the Edwards Formation limestone is subaerially exposed. Water that flows south is soon hydraulically perched above the Edwards Aquifer when it enters the Austin Chalk Aquifer and related hydraulic formations.

Ancestral headwater locations of the paleostream channels in Medina County are interpreted to be located to the south of the Austin Chalk outcrop. This interpretation is based on the coincident alignment of the downdip boundary of the Austin Chalk and the most updip occurrence of wells located in the paleostream channels (Fig. 4). If this interpretation is valid, the paleostream channels are recharged directly from the Austin Chalk and not the Edwards Aquifer. Aerial exposure of the ancestral springs likely ceased in the late Pleistocene, after which subsequent alluviation buried the springs and paleostream channels with 12–15 m of sediments, resulting in the current state of the floodplain (Doyle, 2003).

Elevations of the ancestral springs that discharge to the paleostream channels in Seco, Parker, Live Oak, Hondo, Verde, Elm, and Quihi Creeks were calculated by subtracting 20 m, i.e., the maximum depth of the base of the Leona Formation (Cary Spurgeon, well driller, 2005, personal commun.), from the surface elevation of the creek beds south of the Austin Chalk outcrop. These estimated elevations vary from 254 to 271 m relative to mean sea level (msl) and are believed to be conservatively low. Actual elevations of the base of the paleostream channels could be higher if the depth of the Leona Formation is less than 20 m (Green et al., 2012b). Given that the potentiometric surface of the Edwards Aquifer in this area is no greater than 250 m msl, it is likely the water discharged from the Austin Chalk to these paleostream channels is perched above the Edwards Aquifer and conveyed south to the Seco Creek watershed and is lost from the Edwards Aquifer water budget.

The role of the Medina River watershed in recharging the Edwards Aquifer has been a significant source of uncertainty when calculating the water budget of the San Antonio Pool. Past conceptualizations attributed water loss from Medina and Diversion Lakes to recharge of the Edwards Aquifer (Lambert et al., 2000; Slattery and Miller, 2004; Clear Creek Solutions, 2009). Field studies by Green et al. (2012b) provided the basis for a conceptual model that refines this hydrologic relationship. A significant paleostream channel composed of the Leona Formation gravel (Fisher, 1983) was detected in the Medina River floodplain downgradient from the Haby Crossing fault during an electrical resistivity survey conducted in 2012 (Fig. 9; Green et al., 2012b). Given the absence of alluvial deposits upstream from Haby Crossing fault, the Leona Formation gravel paleostream channel appears to originate at Haby Crossing fault (Fig. 8).

It is hypothesized that a spring was once present at the origin of the gravel deposits at Haby Crossing fault. The spring and the Pleistocene-age, paleostream channel deposits were later overlain by the fluvial terrace deposits; however, the significant transmissive capacity of the paleostream channel deposits remained. The Leona Formation gravel paleostream channel allows for significant groundwater flow out of Medina and Diversion Lakes and discharge into the Leona Formation gravel paleostream channel present in the Medina River floodplain.

The average annual discharge in Medina River at the Bandera U.S. Geological Survey gauging station is assumed to be representative of Medina River flow upgradient from the hydrologic effect of the dam at Medina River. Base flow was separated from flow measurements taken at the U.S. Geological Survey gauging station at Bandera (Fig. 7; Arnold et al., 1995; Arnold and Allen, 1999). The long-term average flow for the Medina River watershed upstream of this station equates to 171.9 mm/yr when averaged for the watershed area. The base-flow fraction of 0.68 indicates that 116.8 mm/yr of average flow is attributed to recharge (Fratesi et al., 2015). The contributing and recharge zones of the Medina River watershed have an area of 1836 km² (LBG-Guyton Associates and Aqua Terra Consultants, 2005). This suggests that 214.5 Mm³/yr is recharged to the Edwards Aquifer in the Medina River watershed. This estimate for recharge, which is considered high when compared with recharge estimates for adjoining basins, suggests the groundwater basin for Medina River may extend beyond the boundary of the surface watershed. Insufficient information is available to surmise the groundwater basin boundary that extends beyond the surface watershed boundary. Alternatively, it is possible that a recharge rate of 116.8 mm/yr is representative of actual recharge, given the localized high in precipitation (i.e., 939 mm/yr) due to an orographic effect associated with local highlands at the headwaters of the Medina River watershed.

Medina Lake and Diversion Lake have been evaluated to determine the role that they play in the water budget of the Edwards Aquifer and the amount of water they recharge to the Edwards Aquifer (Green et al., 2012b). Lambert et al. (2000) evaluated the hydrogeology of the Medina Lake and Medina Lake diversion and used the following water budget from Lee and Swancar (1997) and prior work by Espey, Huston & Associates, Inc. (1989) to calculate the outflow from Medina and Diversion Lakes:

where ΔS = change in lake storage; P = precipitation on the lake; E = evaporation from the lake surface; SW_in = surface-water inflow to the lake; SW_out = surface-water outflow from the lake; GW_in = subsurface inflow into the watershed; GW_out = subsurface outflow from the lake (losses from lakes/recharge to the groundwater system); and e_i = uncertainty or error in each term.

The water-budget expression can be rearranged in terms of discharge as groundwater:

Detection of the extensive paleostream channel in the Medina River floodplain provides the basis to estimate the water budget of the Medina River watershed. The paleostream channel in the Medina River floodplain is estimated to be 1830 m wide and 2.1 m thick, for a cross-sectional area of 3901 m² (Green et al., 2012b). The gradient of the paleostream channel is assumed to be consistent with the gradient of the current Medina River, or 0.0025 m/m. The hydraulic properties of the Leona Formation in the Medina River floodplain have not been measured with an aquifer test and were estimated using documented values for coarse gravel (i.e., hydraulic conductivity of 0.070–0.70 m/s; Bear, 1979; Freeze and Cherry, 1979). This equates to ~0.68–6.8 m³/s or 21.4–214 Mm³/yr of flow through the Medina River floodplain paleostream channel deposits. If an average hydraulic conductivity of 0.35 m/s is assumed for the Leona Formation deposits, underflow would be 3.4 m³/s or 108 Mm³/yr in the Medina River floodplain.

Average annual Medina River surface discharge downstream from Haby Crossing fault is 5.72 m³/s (or 180 Mm³/yr) for the period 1981–2011 (U.S. Geological Survey website measured at station near Macdona, Texas, https://waterdata.usgs.gov/tx/nwis/uv/?site_no=08180700&PARAmeter_cd=00065,00060; Fig. 7). If average hydraulic values for the paleostream channel are assumed, total flow via the Medina River floodplain is ~9.12 m3/s (288 Mm3/yr). Underflow in the paleostream channel in the Medina River floodplain accounts for ~38% of the total average discharge lost to the Edwards Aquifer via the Medina River floodplain.

The groundwater outflow term in Equation 3 can be defined as:

where groundwater outflow has two components, recharge to the Edwards Aquifer (i.e., from Medina Lake) and outflow to the Leona Formation gravel in the Medina River floodplain. The GW_Leona component is underflow in the Medina River floodplain. If the conceptualization of significant outflow through the Medina River paleostream channel deposits is correct, then recharge from the Medina Lake diversion system to the Edwards Aquifer is negligible or possibly nonexistent.

In summary, the average annual discharge in Medina River at Bandera is representative of Medina River flow upgradient from the hydrologic effect of the dam at Medina River. The base-flow fraction of flow indicates that recharge equates to 117 mm/yr. This suggests that the 1836 km² area of contributing and recharge zones of the Medina River watershed could provide significant recharge (i.e., 215 Mm³/yr) to the Edwards Aquifer, even when accounting for discharge via the paleostream channel in the Medina River channel.

The interbasin area between the Medina River and Cibolo Creek basins includes San Geronimo, Helotes, Leon, and San Pedro Creeks and the San Antonio River. Recharge in the San Geronimo Creek and Helotes Creek watersheds was evaluated by assessing each water budget using river-flow measurements. U.S. Geological Survey river gauges on Helotes Creek at Helotes and on Laurel Canyon Creek (Government Canyon State Natural Area) near Helotes were used in the evaluation (Fig. 7; U.S. Geological Survey National Water Information System, https://waterdata.usgs.gov/nwis/uv?site_no=08181400). The average annual discharge in Laurel Canyon Creek (located in the San Geronimo Creek watershed) for the period 2003–2009 was 0.002 m3/s, which equates to 39.1 mm/yr over the 1.55 km2 watershed. The average annual discharge in Helotes Creek for the period 1969–2013 was 0.02 m3/s. This equates to 12.7 mm/yr over the 39 km2 watershed.

Recharge values for both of these creeks are unrealistically low. This is explained by the fact that most flow in Helotes Creek recharges the Upper Glen Rose Formation upstream from the river gauge, which is located at the boundary of the Edwards Aquifer recharge zone. Similarly, flow in San Geronimo Creek only reaches the Edwards Aquifer recharge zone during periods of significant surface runoff and flow (Fratesi et al., 2015). Farther east, Leon Creek is believed to exhibit similar characteristics. This observation is consistent with the interpretation that the Upper Glen Rose Formation has sufficiently high permeability to accept the base flow in Helotes, San Geronimo, and Leon Creeks (Gary et al., 2011).

Cibolo Creek exhibits perennial flow only in its upper reaches and is ephemeral elsewhere (Slade et al., 2002). The absence of flow in Cibolo Creek, except after large precipitation events, is evidence of the large capacity of the bedrock along Cibolo Creek to accept recharge. Water loss in Cibolo Creek is recognized as Edwards Aquifer recharge by the designation of the Upper Glen Rose Formation limestone exposed in the creek bed as part of the Edwards Aquifer recharge zone. The Texas Commission on Environmental Quality has acknowledged this by designating that area within the 100 yr floodplain of Cibolo Creek, from where it begins at Herff Falls in Kendall County, as recharge zone (Texas Commission on Environmental Quality, 2009; see Fig. 10).

River-flow data for three gauging stations on Cibolo Creek were analyzed. Two gauges are located near its headwaters, where the creek has mostly perennial flow (U.S. Geological Survey gauge near Boerne and U.S. Geological Survey gauge at the Cibolo Nature Center near Boerne). The third gauge is located upstream of Selma, near where Cibolo Creek enters the Edwards Aquifer recharge zone, flow is ephemeral, and the creek is dry 90% of the time (Fig. 10).

The average annual flow at the Cibolo Nature Center gauge for the period 2007–2011 was 0.71 m³/s (23.6 Mm³/yr). The watershed area above this gauging station is 146 km², which equates to 137 mm of recharge and runoff per year. The average annual flow at the Boerne gauge for the period 1963–1995 was 0.83 m³/s (26.1 Mm³/yr). The watershed above this gauging station is 177 km², which equates to 147 mm of recharge and runoff per year. The average annual flow at the Selma gauge for the period 1978–2011 was 0.78 m³/s (24.6 Mm³/yr), which is less than flow at the Boerne gauge. Thus, the reach between the gauge at the Cibolo Nature Center in Boerne and the gauge at Selma is losing. This observation is consistent with a compendium of gain-loss studies by the U.S. Geological Survey (Fig. 11; Slade et al., 2002). Water recharged along this reach is characterized as recharge to the Edwards Aquifer via unidentified subsurface flow paths.

Johnson and Schindel (2008) provided insight on the relationship between recharge from Cibolo Creek and discharge to the major springs (Fig. 12). The Selma gauge on Cibolo Creek is located in the Artesian fault block. There is a positive correlation between flow on Cibolo Creek and discharge at Hueco Springs. This correlation varies with Cibolo Creek flow rate. There is one correlation when Cibolo Creek flow at Boerne (which is west of the area mapped in Fig. 12) is between 0.85 m³/s and 1.98 m³/s and a different correlation at higher flow rates. When there is no flow in the Cibolo River at Selma, Hueco Springs discharge averages 1.53 m³/s. When Cibolo Creek is flowing at Selma, Hueco Springs discharge averages 2.83 m³/s. This suggests that flow in Cibolo Creek recharges the Hueco Springs fault block only during high-river-flow periods.

Cibolo Creek gains in the Bat Cave fault block and loses in the Hueco Springs fault block. There is no correlation between Cibolo Creek flow at Boerne and groundwater elevation in wells in the Bat Cave fault block. There is no correlation between Cibolo Creek flow at Boerne and groundwater elevation in wells in the Artesian fault block. Although Johnson and Schindel (2008) did not examine the correlation between Cibolo Creek flow at Boerne and discharge at Comal Springs, it is expected that this correlation would be minimal because Comal Springs is recharged by the Artesian fault block.

Johnson and Schindel (2008) noted that there is little evidence to suggest that Cibolo Creek is a major contributor to the San Marcos Springs system. Based on limited data, Cibolo Creek contributes a relatively small fraction of San Marcos Springs discharge below 4.4 m³/s and a slightly larger fraction above 4.4 m³/s. They noted that George et al. (1952) held the same conceptualization and that they estimated that Cibolo Creek contributes one quarter to one third of the flow at Comal Springs.

The Cibolo Creek watershed obviously provides significant recharge to the Edwards Aquifer. Examination of the geologic structure near Cibolo Creek indicates that fault offset in the Balcones fault zone south of Cibolo Creek is less significant than north of Cibolo Creek. This local reduction in fault offset is consistent with the increased width of the Edwards Aquifer recharge zone to the south of Cibolo Creek.

Recharge of the Edwards Aquifer from the Cibolo Creek watershed can be estimated using upstream river-gauge measurements and an empirical relationship between annual average precipitation and recharge. A correlation between precipitation and recharge was developed for the western Edwards Plateau region using recharge calculations based on stream base-flow separation analyses (Eq. 1; Green and Bertetti, 2010a; Green et al., 2012a). If this correlation is appropriate for the Cibolo Creek watershed, then the average annual precipitation of 857 mm in Boerne should provide ~66 mm of recharge when averaged over the watershed. For the 710 km² watershed upstream from Selma, recharge from Cibolo Creek watershed should average ~46.6 Mm³/yr, if this evaluation is valid.

The Guadalupe River is the only river in the Edwards Aquifer contributing zone that is perennial from its headwaters through the recharge zone. Other rivers in the contributing zone may be perennial near their headwaters but are ephemeral at downstream locations, particularly at reaches close to the recharge zone where the Upper Glen Rose Formation is exposed in riverbeds. In comparison, the Guadalupe River retains perennial flow even where it overlies the Upper Glen Rose Formation. The Guadalupe River crosses the Comal Springs fault block and Bat Cave fault block at locations where the faults that separate the major fault blocks are believed to act as barriers to flow. Johnson and Schindel (2008) noted that although the Guadalupe River carries large volumes of water, discharge measurements above and below the recharge zone indicate it is a minor contributor to the Edwards Aquifer.

Long-term average flow for eight U.S. Geological Survey gauges on the Guadalupe River upstream of New Braunfels, Texas, is included in Table 1 (U.S. Geological Survey National Water Information System, https://waterdata.usgs.gov/nwis/rt). The generally increasing flow measurements are consistent with gain-loss measurements summarized by Slade et al. (2002) and in Figure 11, which suggest that the Guadalupe River is essentially a gaining river with limited flow lost to recharge. Exceptions to this generalization are reaches of the Guadalupe River upstream from Kerrville, at Center Point, and immediately upstream from the Edwards Aquifer recharge zone. Data at the Center Point gauge are limited to less than 5 yr (2008–2012).

Flow data and analyses on the Guadalupe River watershed provide insight into the hydraulic relationship of the river with the subsurface. Data include flow measurements and base-flow separation analyses (Table 2). The base-flow fractions at four of the Guadalupe River gauges have been calculated. The four values of 0.816, 0.695, 0.865, and 0.635 are relatively similar, with an average value of 0.76. If this base-flow fraction is representative of the Guadalupe River, then recharge for the Guadalupe River watershed would be 100.5 mm/yr when using average annual flow measured at the Guadalupe River above Comal River.

Average annual flow of the Comal River (270 Mm³/yr) when added with the Guadalupe River above the confluence with the Comal River (515 Mm³/yr) sums to 786 Mm³/yr; however the average flow at downstream gauge is only 685 Mm³/yr, a difference of 101 Mm³/yr (Fig. 13). Average annual flow in the Guadalupe River at the Seguin gauge, which is farther downstream, is 726 Mm³/yr, i.e., still a difference of almost 59.2 Mm³/yr less than the sum of the Comal River and the upstream Guadalupe River flows. All river gauges illustrated in Figure 12 are within the Edwards Aquifer confined zone. None is within the recharge zone. Thus, loss of flow between the confluence of the Comal and Guadalupe Rivers cannot be directly attributed to recharge of the Edwards Aquifer, although this prospect cannot be completely dismissed.

George et al. (1952) posited that the Guadalupe River contributes little water to Comal Springs. LBG-Guyton Associates (2004) concluded that San Marcos Springs receive water from local and regional sources and that the Guadalupe River watershed may contribute to Comal Springs as a local source. Woodruff and Abbott (1986) reported that the Guadalupe River contributes little or no water to the San Marcos hydrologic system. Johnson and Schindel (2008) used well data near the Guadalupe River to estimate that the maximum recharge from the Guadalupe River to the Edwards Aquifer was 10% of river flow when the flow is below 14.1 m³/s, for a maximum recharge rate of 1.41 m³/s, although they acknowledged that this estimate may be high. Musgrove and Crow (2012) examined the chemistry of water discharged from San Marcos Springs and found little evidence of dilution by local recharge. These analyses suggest that the Guadalupe River provides a maximum of 44.4 Mm³/yr of recharge to the Edwards Aquifer, i.e., less than 9% of the annual average river flow of 515 Mm³/yr measured at the gauge located above the confluence with the Comal River.

The consensus of these evaluations is that most recharge from the Guadalupe River watershed is conveyed out of the contributing zone and past the recharge zone without recharging the Edwards Aquifer. This attribute sets the Guadalupe River watershed apart from the other major watersheds in the Edwards Aquifer contributing zone. Nonetheless, it appears that a minor component of recharge from the Guadalupe River does recharge the Edwards Aquifer.

Blanco River flow at Wimberley ranged from 0.11 to 16.9 m³/s during the period 1956–2006, with an average of ~0.42 m³/s (Johnson and Schindel, 2008). A comparison of flow at U.S. Geological Survey gauges at Wimberley (upstream) and at Kyle (downstream) suggests that the Blanco River is losing where it crosses the Bat Cave fault block and the Hueco Springs fault block (Fig. 12). When there is no flow at the Kyle gauge, all flow that flows past the Wimberley gauge recharges the Edwards Aquifer and is available for discharge at San Marcos Springs.

A factor of importance is the amount of water the Blanco River contributes to discharge at San Marcos Springs under varying hydrological conditions. The northern orifices of San Marcos Springs—Cabomba, Hotel, Johnny (aka Weissmuller), and Divergent (aka Diversion)—are recharged by the Blanco River, Sink Creek, and areas downgradient to the north. These recharge areas are south of the groundwater divide near Onion Creek that separates the San Marcos Springs recharge basin from the Barton Springs recharge basin (Johnson and Schindel, 2008). The Blanco River also carries large volumes of water through the area believed to be within the San Marcos Springs capture area. Discharge of both the Blanco River at Wimberley and San Marcos Springs is flashy, responding quickly to storm events. Statistically, however, correlation of Blanco River flow and discharge at San Marcos Springs is relatively low (~60%; Johnson and Schindel, 2008). Guyton (1979) contended that regional groundwater comprises 55%–60% of San Marcos Springs discharge. Consistent with this assessment, McKinney and Sharp (1995) found that local sources such as Purgatory and York Creeks and the Blanco River contribute up to 35% of San Marcos Springs discharge. LBG-Guyton Associates (2004) concurred that San Marcos Springs receive water from local and regional sources. They identified Blanco River and other parts of the Blanco River watershed and possibly the Guadalupe River watershed as local sources.

Previous estimates of the amount of recharge that the Blanco River provides to San Marcos Springs are mixed. Klemt et al. (1979) attributed 47% of the Blanco River flow directly to San Marcos Springs discharge. In contrast, Texas Board of Water Engineers (1960), Watson (1985), and Slade et al. (2002) measured channel losses in the Blanco River over the recharge zone at less than 0.56 m³/s, meaning there is limited infiltration in the Blanco River available for discharge at San Marcos Springs. Streamflow measurements by the Texas Department of Water Resources (1960) indicated 0.42 m³/s as the maximum loss rate for the Blanco River. However, Watson (1985) concluded that rate was appropriate at river stages up to 0.6 m. He increased his estimate to 0.84 m³/s because additional water would flow into karst features along the Blanco River at higher river stages. When San Marcos Springs discharge is less than 2.83 m³/s, the Blanco River contributes only a small percentage of water to San Marcos Springs discharge. If all flow from the Blanco River recharged San Marcos Springs during periods of no flow at Kyle, its contribution would range from ~8% to 25% of spring discharge (Johnson and Schindel, 2008). Recharge of the Edwards Aquifer from the Blanco River Basin remains uncertain; however, the consensus of these assessments suggests that the Blanco River does provide limited, but meaningful recharge to the Edwards Aquifer and that this recharge is of the order of 0.42 m³/s (or 13.4 Mm³/yr), although this amount could vary significantly. Tracer testing also shows that the Blanco River contributes water to the Barton Springs watershed during low-flow periods when the groundwater divide at Onion Creek flattens and disappears. However, this contribution is estimated at less than 0.14 m³/s.

The San Antonio Pool forms the eastern portion of the San Antonio segment of the Edwards (Balcones Fault Zone) Aquifer. The Edwards Aquifer in the San Antonio Pool has three principal zones, the contributing zone, the recharge zone, and the artesian zone. The hydraulic relationship among these zones is a function of the Balcones fault zone, a structural feature that defines the hydraulic relationship between the Edwards Aquifer and the underlying Trinity Aquifer. The northern extent of the San Antonio Pool is defined by the boundary of the contributing zone, and the southern extent is defined as the transition from freshwater to saline water. The San Antonio Pool extends from the Knippa Gap on the west to a groundwater divide near Kyle in the east.

Discharge from the Uvalde Pool via the Knippa Gap provides recharge to the San Antonio Pool. The San Antonio Pool is also recharged from the contributing zone to the north via surface flow and underflow associated with the Dry Frio, Frio, Sabinal, Medina, Cibolo, Guadalupe, and Blanco Rivers. The San Antonio Pool discharges to two major springs in the east, Comal Springs and San Marcos Springs, in addition to two minor springs in Bexar County, San Antonio Springs and San Pedro Springs, in addition to a few spring-fed waterholes on the Frio River. Significant water is discharged from the Medina Lake and Diversion Lake system via a paleostream channel in the Medina River floodplain. This discharge had previously been assumed to recharge the Edwards Aquifer.

We are grateful for reviews by Van Brahana, Bill Hutchinson, Brian Smith, and Jack Sharp, whose comments greatly improved this manuscript.