4.1.1 Discharge

It has long been recognized that natural aquifer discharge is primarily due to the large springs that characterize the Edwards Aquifer, with well pumpage as the second major discharge. In some cases, springs discharge directly into alluvial systems (e.g., Leona Springs) or as base flow to major streams or rivers (e.g., Medina River in the San Antonio pool (Green et al., 2019a). Some discharge from the Cold Springs pool to the Colorado River (Hauwert et al., 2004; Hunt et al., 2019). There is also limited cross-formational flow to other aquifers, which has dropped significantly as well yields increased and potentiometric surfaces lowered. There are fewer flowing artesian wells than in the time of Hill (1890) and upward cross-formational flow is a minor source of the discharge because of the effectiveness of the Del Rio Clay as a confining unit.

There is (or was because of lower heads due to well pumping) the potential for upward cross-formational flow. Many wells in the Edwards Aquifer confined zone are (or were) free flowing, particularly during periods of greater rainfall. An extreme example of this is the Catfish Farm well (Figure 24), which was one of the world’s largest free-flowing wells capable of flowing on the order of 2.5 m³/sec (40,000 gallons per minute). This well is now closed, but this video demonstrates both the great productivity resulting from the high permeability and some of the political issues that arise when dealing with the Edwards Aquifer.

Figure 24 – Catfish farm well (Photo by Greg Eckhardt, 1995).

On a regional scale, cross-formational flow as a process of natural discharge has not been quantified, but it is small. This is supported quantitatively by numerical model simulations.

4.1.2 Recharge

Recharge to the aquifer can be estimated in two ways: indirect and direct. The indirect (lumped system) method is to assume steady-state flow for years so that recharge from all sources must equal the measured or estimated long-term discharge. This method depends on the accuracy of discharge data and does not address spatial or temporal variations or analyze different sources of recharge that are important for understanding hydrogeological processes and for aquifer management.

Direct estimates generally show that the largest source of recharge is from losing streams that flow over the contributing and recharge zones. The second-largest source of direct recharge is from precipitation on the recharge zone as diffuse recharge (e.g., Hauwert and Sharp, 2014). Cross-formational flow from The Edwards-Trinity Aquifer is another form of recharge. Although there are no direct measurements of cross-formational flow, the current model of the San Antonio Segment of the Edwards Aquifer (Liu et al., 2017) uses a value of 74,000 acre-feet/year (91.2 GL/year) and the Texas Water Development Board’s Trinity model (Jones et al., 2009) uses a value 110,000 acre-feet/year (136 GL/year). That would account for 10 to 20 percent of the average input to the Edwards Aquifer. Smaller amounts come from anthropogenic sources (urbanization, artificial recharge, and irrigation return flow (Garcia-Fresca, 2004; Passarello et al., 2012 and 2014).

Estimates of recharge from losing streams are based on stream gaging and estimates of precipitation in contributing zone watersheds. Puente (1978) established the basic method of estimating recharge from losing streams by stream gaging above and below the recharge zone. The difference was the calculated recharge. Puente’s estimates are shown in Table 2.

Table 2 – Estimated recharge from losing streams for the San Antonio and Kinney segments of the Edwards Aquifer by basin (after Puente, 1978). 1 GL is one-million cubic meters.

Puente’s method forms the basis for direct stream-loss recharge calculation, but it has had to be modified because of several factors. First, the great variability in rainfall at local scales along the Balcones Escarpment (e.g., Nielsen-Gammon, 2013; Nielsen et al., 2015) makes prediction of stream flow and subsequent recharge from losing streams difficult when using broad-scale precipitation measurements. Second, geological mapping and tracer tests have led to revised delineations of the recharge zone. Tracer tests in Bexar County (Johnson et al., 2019) show that the Upper Glen Rose Formation is more hydraulically similar to the Edwards Aquifer than the rest of the Trinity Aquifer. Accordingly, the upper Glen Rose Formation acts as part of the Edwards Aquifer recharge zone (Figure 17). Supporting this conceptualization are multiport well data in the Barton Springs segment that show that the Edwards Aquifer and the Upper Glen Rose Formation function as a single aquifer (Smith and Hunt, 2019). The long-accepted basic aquifer framework that negligible recharge to the Edwards Aquifer occurs in the contributing zone can be expected to be modified to reflect this finding and thus result in refined conceptualizations in the future.

Finally, stream-gaging data have shown that infiltration along the losing stream reaches is not uniform spatially (e.g., Slade et al., 2002) or temporally. Zahm (1998) found that infiltration in the losing reaches along Barton Creek was not uniform but concentrated in a limited number of locations. Stream-gaging studies on the Nueces River (Kromann, 2015: Hackett, 2019) and the Blanco River (Hunt et al., 2017) show that the location of losing and gaining reaches in rivers that cross the recharge zones are not fixed but may vary with time. Consequently, the precise location of river gages and use of only a few gages can bias estimates of recharge.

Direct recharge from precipitation occurs on the recharge zone as autogenic recharge. This has also been estimated indirectly by subtracting the estimated recharge from losing streams from total discharge (spring flows and wells). This should give a general estimate, but in some early cases before the pools in the segments were clearly defined, this calculation method caused significant underestimation of direct recharge (Hauwert and Sharp, 2014). Direct aquifer recharge is best estimated from water-balance analysis using precipitation, surface runoff, and evapotranspiration data. Hauwert and Sharp (2014) compiled existing data from central Texas flux towers (Figure 25) to show that water available for recharge is a function of total annualized precipitation. In dry years, most precipitation infiltrating the land surface is lost to evapotranspiration.

Figure 25 – Central Texas evapotranspiration flux tower data and annual precipitation (from Hauwert and Sharp, 2014).

In wetter years, direct recharge is greater. Hauwert’s (2009) data for the Barton Springs segment recharge zone on two control plots with internal drainage (i.e., no surface runoff to streams) indicated that, during the period of study (2005 to 2006), about 30 percent of the precipitation was available for aquifer recharge, which is in the typical range for karst aquifer outcrop areas in sub-humid environments (Hauwert, 2009, Table 3.1 therein).

There are other minor sources of recharge such as the effects of urbanization (Passarello et al., 2012, 2014; Sharp, 2019), including water losses from water and sewage lines, irrigation return flow, and artificial recharge. Artificial recharge is where groundwater is recharged by redirecting water across the land surface through canals, infiltration basins, or ponds; adding irrigation furrows or sprinkler systems; maintaining streamflow in losing streams; or injecting water directly into the subsurface through injection wells, injection galleries, or sinkholes. Artificial recharge has been considered as a strategy for the Edwards Aquifer to maintain critical spring flows to preserve endangered aquatic species, specifically by injection wells or galleries near Comal and San Marcos springs (McKinney and Sharp, 1995; Uliana and Sharp, 1996), but has never been implemented. Consideration has also been given to retention dams in the contributing zone that would preserve high flood flows and then release flows gradually to the recharge zone, but, these have only been implemented to a limited degree. The Edwards Aquifer Authority (EAA) has, however, developed the Seco Creek Sinkhole (Figure 12), see the video here, to receive high stream flows via a diversion ditch from Seco Creek.

The Barton Springs Edwards Aquifer Conservation District (BSEACD) has constructed a concrete structure over a cave in the bed of Onion Creek in south Austin, which directs water in the creek into the Edwards Aquifer (Smith et al., 2011). This system was designed to automatically close a valve during periods of stormwater flow so that the high-turbidity stormwater doesn’t enter the aquifer. When the water quality in Onion Creek improves, the valve opens allowing the better-quality water to recharge the aquifer. This decreases the amounts of contaminants entering the aquifer during flood events and increases recharge to the aquifer by preventing the cave from getting clogged with debris.

In areas where there is irrigation on the recharge zone, the possibility exists that there is recharge from irrigation return flow to the Edwards Aquifer. However, this has been considered minimal because many irrigated lands are found in the confined zone where direct recharge to the Edwards Aquifer is negligible. However, recharge due to urbanization (leaky water and sewer systems as well as over-irrigation of parks and lawns) may be significant in large metropolitan areas and areas of expansion such as along the I-35 corridor (Garcia-Fresca, 2004; Passarello et al., 2012 and 2014). This is especially true during periods of very low rainfall. Passarello and others (2012) show that, in the Barton Springs segment, during times of drought urban recharge may be the largest component. Average annual recharge estimates for all four segments are given in Table 3.

Table 3 – Average annual recharge.