Long-Term Permitted Groundwater Use

A total of 28 of the 125 long-term groundwater permits active in 2022 for the Antlers aquifer were prior-right permits. The OWRB defines a prior-right groundwater permit as the right to use groundwater established by compliance with laws in effect prior to July 1, 1973 (OWRB, 2023c). An additional 29 long-term groundwater permits, 5 of which were prior rights, became inactive in or before 2022 and are also included in a statistical analysis. Mean annual reported groundwater use associated with long-term groundwater permits for the Antlers aquifer was 3,483 acre-ft/yr during the study period (table 5); median annual groundwater use was 2,927 acre-ft/yr during this same period. Public supply accounted for 44 percent of the mean annual reported groundwater use during 1967–2022, followed by irrigation (37 percent), industrial (17 percent), mining (1 percent), with the rest of the use categories together accounting for the remaining 1 percent (fig. 8A).

Groundwater-use data were also analyzed for four time periods: 1967–80, 1981–2010, 1980–2022, and 2011–22. The reported irrigation groundwater use was generally higher during 1967–80 than during the other time periods, whereas overall groundwater use, especially industrial groundwater use, was generally higher in 1980–2022 and 2011–22 than it was in 1967–80 and 1981–2010 (fig. 8B). The higher reported amounts of irrigation groundwater use pre-1980 may stem from a change in estimation methodology for irrigation amounts by the OWRB; the large increase in industrial groundwater use in 2012 is from a single permit in McCurtain County (table 4), which first reported groundwater use for that year. Mean annual groundwater use was 2,982 acre-ft/yr during 1967–80, 2,706 acre-ft/yr during 1981–2010, 5,429 acre-ft/yr during 2011–22, 3,358 acre-ft/yr during 1967–2022, and 3,483 acre-ft/yr during 1980–2022 (table 3). The median annual groundwater use during 1967–80 was 2,807 acre-ft/yr, which was similar to the median groundwater use during 1981–2010 (2,643 acre-ft/yr), 1967–2022 (2,840 acre-ft/yr), and 1980–2022 (2,927 acre-ft/yr), but differed appreciably from the median groundwater use during 2011–22 (5,202 acre-ft/yr) (table 5). The minimum and maximum annual reported groundwater use totals for 1967–2022 were 945 acre-feet (acre/ft) in 1968 and 8,884 acre-ft in 2016 (table 5; OWRB, 2023b).

For the 125 active long-term groundwater permits issued for the Antlers aquifer, 48,222 acre-ft of groundwater withdrawals were allocated, and 6,675 acre-ft of groundwater use were reported for 2022. The large discrepancy between allocated groundwater withdrawals and reported groundwater withdrawals resulted from permit holders that did not submit a groundwater use report for 2022 and permit holders that reported a groundwater use that was appreciably less than their allocation. There were 14 active permit holders with a right to use more than 1,000 acre-ft/yr of groundwater in 2022, and 6 of these permit holders did not submit groundwater use reports for that year. The remaining eight permit holders reported groundwater use totals that were less than their allocated amount; only one of those permit holders used more than 15 percent of their total allocation. Of all active permit holders with rights to withdraw water from the Antlers aquifer, 59 percent did not report groundwater use in 2022 and 65 percent of those who submitted reports used less than half of their allocation.

Provisional-Temporary Groundwater Use Permits

The OWRB issues provisional-temporary groundwater permits that expire 90 days after issuance (OWRB, 2023c). These permits are used to provide a short-term water supply or supplement the water supply of existing permit holders. Unlike for long-term permits, groundwater-use reports are not currently (2024) required for provisional-temporary permits with volumes assumed not to exceed the authorized amount.

Although OWRB permit records extend back to 1992, the first recorded provisional-temporary permit for the Antlers aquifer was issued in 1996. A total of 38 provisional-temporary permits were issued between 1996 and 2022, with a mean authorized amount of 75 acre-ft per permit (fig. 9A); the median authorized amount was only 4 acre-ft per permit. A single temporary permit for 1,452 acre-ft in 2011 skews the distribution of the provision-temporary amounts; excluding the 2011 provisional-temporary permit for industrial groundwater use reduces the mean authorized provisional-temporary amount to 38 acre-ft per permit. By volume, industrial groundwater use accounts for the majority of provisional-temporary groundwater use at 53 percent, or 95 acre-ft/yr, followed by public supply at 22 percent; irrigation at 17 percent; oil, gas and mining at 7 percent; and agriculture at 1 percent for the 1996–2022 period (fig. 9B). There were 20 unique provisional-temporary permits issued for oil, gas, and mining groundwater use—5 of which were issued for agriculture and public-supply groundwater use, and 4 of which were issued for both irrigation and industrial groundwater use.

Aquifer Depths

In the unconfined part of the Antlers aquifer, the top of the aquifer was defined as the land-surface altitude obtained from a 10-m (horizontal resolution) digital elevation model (DEM; USGS, 2015). The altitudes of the top of the aquifer in the confined part and the base of the aquifer in the unconfined and confined parts of the Antlers aquifer were digitized and interpolated from contour maps by Morton (1992) and Hart and Davis (1981), which were derived from interpretation and correlation of 230 geophysical logs from oil and gas wells.

The top of the Antlers aquifer dips south-southwest at a rate of 35–90 feet per mile (ft/mi), with a mean dip of approximately 60 ft/mi (Morton, 1992). The depth of the top of the Antlers aquifer exceeds 1,500 ft in the far southeastern corner of McCurtain County, Okla. The base of the Antlers aquifer also dips south-southeast at a steeper rate of 35–105 ft/mi with a mean dip of approximately 75 ft/mi (Morton, 1992). The base of the aquifer dips greater to the southeast than does the top, resulting in a thickening of the aquifer to the south-southeast. The deepest depth to the base of the Antlers aquifer from land surface is greater than 2,500 ft. The thickness of the Antlers aquifer varies greatly from north to south (Hart and Davis, 1981). The Antlers aquifer saturated thickness ranges from 0 ft in the northernmost unconfined part—increasing to the southeast to almost 1,200 ft in southeastern Oklahoma—to more than 2,000 ft south of the Red River (Morton, 1992).

Aquifer Base

For the purposes of this report, the base of the Antlers aquifer is considered to be the base of fresh groundwater (OWRB, 2023c). Data from Robinson and Deeds (2019) that were collected as part of the Texas Water Development Board’s (TWDB) Brackish Resources Aquifer Characterization System, were used to determine the base of fresh groundwater for the Antlers aquifer.

Robinson and Deeds (2019) spatially estimated the depths of TDS concentrations in the five geologic units (Hosston Formation, Pearsall Formation, Hensell Sand, Glen Rose Formation, and Paluxy Formation; fig. 12) in the northern part of the Trinity aquifer at 1,000, 3,000, and 10,000 mg/L for each of the geologic units using a combination of groundwater samples and interpretation of geophysical logs from wells completed in the Trinity aquifer. The altitude of the top surface of each geologic unit was correlated to the 1,000-, 3,000-, and 10,000-mg/L TDS contours for the respective surface. For the purposes of this report and to align with water-quality limits set by the State of Oklahoma, a 5,000-mg/L TDS contour (fig. 16A) was interpolated as the boundary between fresh and saline groundwater for each of the geologic units by using the 3,000- and 10,000-mg/L TDS extents determined by Robinson and Deeds (2019). Altitude values were assigned to points along the 5,000-mg/L contour lines that corresponded to the altitude of the top of each geologic unit (Robinson and Deeds, 2019). An altitude surface was interpolated between the 5,000-mg/L line of the lowermost geologic unit (Hosston Formation) and the uppermost geologic unit (Paluxy Formation). The altitude surface created from this interpolation was merged with altitude surfaces for the top and base of the Antlers aquifer from Morton (1992) as well as the land surface altitude (for the unconfined part). This resulted in a three-dimensional boundary of altitudes for the base of usable groundwater (freshwater zone) for the Antlers aquifer (fig. 16). Beyond where the 5,000-mg/L TDS altitude surface intersected the altitude of the top of the Antlers aquifer, groundwater was considered to have TDS concentrations that exceeded 5,000 mg/L (saline zone) and was therefore considered to be unusable as freshwater for the purposes of the State of Oklahoma (OWRB, 2023c).

Hydraulic Properties Estimated From a Multiple-Well Aquifer Test

A multiple-well aquifer test involving preexisting wells GW-04, GW-07, and GW-08 (fig. 1) was completed as part of this investigation in November 2022 in the eastern part of the Antlers aquifer to determine transmissivity, hydraulic conductivity, and storage coefficient values for the aquifer. Aquifer transmissivity is a measure of the amount of water that can be transmitted horizontally through a unit width by the full saturated thickness of the aquifer (Fetter, 2001). Hydraulic conductivity is a measure of the capacity of a porous medium to transmit water (Driscoll, 1986). The storage coefficient of an aquifer is the volume of water that a permeable unit will absorb or expel from storage per unit surface area per change in head (Fetter, 2001). Data from this aquifer test are included in the accompanying data release (Fetkovich and others, 2025).

The multiple-well aquifer test involved withdrawing groundwater from well GW-07 (hereinafter referred to as the “pumping well”) at a constant rate of approximately 160 gallons per minute for approximately 48 hours until groundwater levels in nearby observation wells GW-04 and GW-08 stabilized (figs. 1, 19). The withdrawal of groundwater induced a maximum drawdown of 32.87 ft in the pumping well, 2.38 ft in well GW-04 (approximately 978 ft from the pumping well), and 1.20 ft in well GW-08 (approximately 1,530 ft from the pumping well) after approximately 45 hours. After the pump was turned off, groundwater levels continued to be monitored to observe the recovery in each well until groundwater levels returned to their pretest static levels observed prior to the approximately 48-hour period of groundwater withdrawals. Groundwater levels in wells GW-04 and GW-07 recovered to 100 percent and 98 percent, respectively, of their pretest levels. Data collection from well GW-08 was stopped prematurely, which resulted in a recovery of only 46 percent of the original groundwater level in that well. Maximum recovery of groundwater levels was achieved after about 7.5 days post-pumping, as observed in wells GW-04 and GW-07. Water-level observations varied between the three wells used for this aquifer test, so each well was analyzed independently using both the drawdown and recovery data, where data were sufficient. Owing to the recorder being removed prematurely in GW-08 resulting in insufficient recovery, recovery data at this well were not viable for analysis. The pumping well (GW-07) was not constructed in a way that allowed for installation of a recorder, and although water-level data were collected intermittently from this well during the test by using a manual water-level tape, these data were not sufficient for analysis. Data that were analyzed included the drawdown and recovery data from well GW-04 and the drawdown data from well GW-08. The aquifer test data were analyzed using the AQTESOLV software package (fig. 19A–C; Hydrosolve, Inc., 2011). The depths of the test and observation wells with casing information were input to the AQTESOLV program to correct for partial penetration of the aquifer. Groundwater levels from the pumping and recovery periods were matched to the Hantush (1960) method, which accounts for partially penetrating wells and estimates anisotropy in a leaky confined aquifer (fig. 19A–C). The test-well location and analytical-model solution indicated a leaky-confined aquifer (Lohman, 1972). One analysis was performed for each well by using the drawdown and recovery observations and the Hantush (1960) method in either standard time or Agarwal equivalent time (Duffield, 2025). Agarwal equivalent time is an adjustment of the time to an equivalent time that only includes time of recovery (Duffield, 2025). The values from each of the analyzed datasets were compiled into ranges for transmissivity, hydraulic conductivity, and the storage coefficient.

Transmissivities determined by using the Hantush (1960) method ranged from 3,598 to 5,647 square feet per day (ft²/d). Using the screened interval of 61 ft for the test well, the hydraulic conductivities ranged from 6.79 to 10.65 feet per day (ft/d). The storage coefficient from the Hantush (1960) method ranged from 0.00056 to 0.00080. The transmissivity, hydraulic conductivity, and storage coefficient derived from the multiple-well aquifer test are likely to be more accurate than hydraulic values derived using other methods, such as from grain size or laboratory tests of aquifer material, because flow is induced across a larger volume of the aquifer during the aquifer test than during other tests (Lohman, 1972). However, these hydraulic property values represent local conditions and are not necessarily indicative of the mean or range of regional hydraulic property values.

Hydraulic Properties Estimated From Slug Tests

Multiple slug tests were performed at four wells (GW-03, GW-04, GW-05, and GW-09; fig. 1) to assess repeatability and well integrity. One selected well that was not included in this report did not appear to be hydraulically connected with the aquifer and was not included in the analyses. Both mechanical and poured slugs were utilized depending on the construction of each well.

For each poured slug test, a volume of water, either 5, 10, or 15 gallons was rapidly poured into the well casing. Water-level changes were recorded with an electronic recording pressure transducer set at a 0.25-second measurement frequency. The near instantaneous rise and subsequent fall of the water level in response to the poured slugs were recorded and analyzed as a falling head test using the Bouwer and Rice method, as described by Halford and Kuniansky (2002).

In a mechanical slug test, the existing groundwater in the well is displaced instead of adding a “slug” of actual water to the well. The groundwater is displaced by rapidly lowering a 4-ft-long section of polyvinyl chloride (PVC) pipe filled with sand and capped on each end below the groundwater level in the well; the section of PVC pipe is then rapidly raised above the groundwater level in the well. The change in groundwater level and time for the well to return to the original pretest groundwater level is recorded. The test is performed multiple times. Each time, the change in water level and time for the well to return to the original groundwater level are recorded. The slug test responses (changes in groundwater levels) were analyzed by using the AQTESOLV software package (Hydrosolve, Inc, 2011) and were matched to an analytical solution dependent on the construction of the well and the observed response of each test.

Methods explained in Butler (1998) were used in combination with well construction information, where available, to aid in determining the most appropriate analytical solution for each well used in the slug tests. Selected wells were completed in either the unconfined or confined part of the Antlers aquifer (fig. 1). Analytical solutions used in the analysis of the slug tests included the Hvorslev (1951) solution for wells completed in the confined part of the aquifer and the Bouwer and Rice (1976) and Kansas Geological Survey (Hyder and others, 1994) solutions for wells completed in the unconfined part of the aquifer. One analytical solution was selected for each well, depending on available well construction information and following guidelines from Butler (1998). Further details for the analyses of the slug tests are included in the accompanying data release (Fetkovich and others, 2025).

Transmissivities determined from the analytical solutions ranged from 399 to 6,416 ft²/d. Hydraulic conductivity values determined from the analytical solutions ranged from 0.84 to 12.15 ft/d. The storage coefficient of an aquifer is equal to the specific storage multiplied by the saturated thickness (Fetter, 2001). The storage coefficient determined from the analytical solutions for the confined part of the aquifer was 0.004. The storage coefficient of the confined part of the aquifer was determined from analysis of GW-09, which was the only slug test performed in a confined well. Because specific storage is generally very small, the storage coefficient in unconfined aquifers is generally considered to be equivalent to specific yield (Driscoll, 1986). Specific storage was estimated to range from 1.1 × 10⁻⁶ to 5.0 × 10⁻⁵ foot⁻¹ using a combination of values estimated from slug test analyses and values taken from Batu (1998). Because the storage coefficient is dependent upon specific yield, a specific yield value of 0.10 was used, estimated for the Trinity aquifer by Jigmond and others (2014). Using this specific yield value in slug test analyses from unconfined wells resulted in storage coefficient values that ranged from 0.10 to 0.11. These hydraulic values represent local conditions and are not necessarily indicative of the hydraulic property values of larger areas (regions).

Horizontal Hydraulic Conductivity Estimated From Lithologic Logs

Horizontal hydraulic conductivity distribution across the Antlers aquifer was estimated by using information obtained from lithologic logs (OWRB, 2023d) that were reported to the OWRB by drillers. Reported lithologic logs were filtered to include only sections of logs that were located between the top and base of the Antlers aquifer. Textural descriptions and terms provided in the lithologic logs were categorized and converted to percent-coarse-material values by using methods described in Mashburn and others (2014). Textural descriptions and terms varied between drillers. To simplify and standardize the lithologic logs, lithologic descriptions from wells completed within the Antlers aquifer and the lithologic descriptions for the surrounding geologic units were reclassified into 12 categories that were each assumed to include a specific percentage of coarse material based on known grain sizes of materials listed in the lithologic logs. This reclassification is similar to a modified version of the methods described in Mashburn and others (2014) in which granite (lower confining unit of the Antlers aquifer), shale, clay, silt, very fine sand, dolomite, limestone, fine sand, coal, medium sand, coarse sand, and gravel each contain 0, 10, 10, 10, 20, 30, 30, 30, 40, 50, 70, and 90 percent-coarse material, respectively. The respective percent-coarse-material value was then assigned to each lithologic depth interval. Lithologic depth intervals assigned as granite were removed, as granite is not considered part of the Antlers aquifer. The percent-coarse-material value for each lithologic log was computed as the thickness-weighted mean of percent-coarse-material values assigned to the lithologic categories found within the log. The theoretical maximum percent-coarse-material value for any lithologic log was 90 percent (all gravel), and the theoretical minimum percent-coarse-material value for any lithologic log was 10 percent (all clay, shale, or silt). A total of 1,520 usable lithologic logs (OWRB, 2023d) were included in the percent-coarse-material analysis. Logs with obvious errors were corrected to extract as much useful information as possible, whereas logs with inscrutable errors were discarded.

The horizontal hydraulic conductivity estimated from lithologic logs ranged from 0.87 to 10.65 ft/d, using the maximum and minimum estimated horizontal hydraulic conductivity between the aquifer test, slug tests, and previously published values (Hart and Davis, 1981). Hart and Davis (1981) reported a horizontal hydraulic conductivity range of 0.87–3.75 ft/d. Although the slug test analyses determined a maximum horizontal hydraulic conductivity of 12.15 ft/d, this value was determined from a slug test on the same well that was used for the multiple-well aquifer test, which resulted in a value of 10.65 ft/d. Multiple-well aquifer tests are considered to be more accurate than slug tests because the former are affected by a wider area extending beyond the influence of a slug test; therefore, the horizontal hydraulic conductivity range of 0.87–10.65 ft/d was used. Using this range and methods from Ellis and others (2017), the following equation was developed by correlating the minimum and maximum horizontal hydraulic conductivity values to 10 and 90 percent-coarse material, respectively, and performing a linear regression to characterize the relation between horizontal hydraulic conductivity and the percentage-coarse material value for the Antlers aquifer:

Equation 1 is used in this report to estimate horizontal hydraulic conductivity values for lithologic logs from wells completed in the Antlers aquifer. The lithologic-log estimated horizontal hydraulic conductivity values ranged from 0.87 to 9.02 ft/d with a mean of 3.31 ft/d (fig. 20).

Aquifer Storage Properties and Estimated Groundwater Storage

Total groundwater storage, in acre-feet, for the Antlers aquifer was estimated by the following formula, modified from Fetter (2001):

For this report, the value for specific yield (Sy) was set to 0.1 in accordance with Jigmond and others (2014). The value for specific storage (Ss) was estimated to be in the range of 1.1 × 10⁻⁶ to 5 × 10⁻⁵ foot⁻¹ based on values from slug tests and ranges in Batu (1998). The base of the Antlers aquifer and the 2022 potentiometric surface were used to create surface rasters for each set of data. The value for the mean potentiometric saturated thickness (ST_p) was determined by subtracting the altitude surface raster of the base of the Antlers aquifer from the 2022 potentiometric altitude surface raster and calculating the mean value for the aquifer in Oklahoma. The value for the mean saturated thickness (ST_a) was determined by subtracting the altitude of the base of the Antlers aquifer from either the altitude of the 2022 potentiometric surface or the altitude of the top of the aquifer (Morton, 1992) and calculating the mean for the aquifer in Oklahoma. Using the values for Sy and Ss along with the values of 434 ft for the mean saturated thickness (ST_a) and 621 ft for the mean potentiometric saturated thickness (ST_p) in equation 1 with the aquifer area (2,746,648 acres), the total groundwater storage for the Antlers aquifer is estimated to range from about 120,000,000 to 156,000,000 acre-ft.

Recharge

For this report, recharge is defined as the amount of precipitation that infiltrates at the land surface and reaches the saturated zone of the aquifer. Areal recharge from precipitation is the predominant inflow to the Antlers aquifer. Factors such as precipitation rates, evapotranspiration rates, soil and sediment permeability, vegetation cover types, and the gradient of the land surface affect rates of recharge (Rogers and others, 2023). Recharge rates are difficult to quantify because of spatial and temporal variability; however, recharge rates can be estimated from other measurements and climatological data. Two methods were used to estimate recharge rates for this report: (1) the Soil-Water-Balance (SWB) code (Westenbroek and others, 2010) was used to estimate spatially distributed recharge rates for the 1980–2022 study period; and (2) the water-table-fluctuation (WTF) method (Healy and Cook, 2002) was used to estimate localized recharge rates during 2013–22.

Soil-Water-Balance Code

For the 1980–2022 study period, the amount and spatial distribution of daily groundwater recharge was used to estimate the mean annual recharge to the Antlers aquifer for the study period. Recharge was estimated by using the SWB code (SWB version 1; Westenbroek and others, 2010). The SWB code uses a modified Thornthwaite-Mather (Thornthwaite and Mather, 1957) SWB method on a gridded data structure to compute the daily amount of infiltration, accounting for losses, that exceeds the storage capacity of the plant root zone. Input data required to estimate recharge using the SWB code include precipitation, air temperature, soil-water storage capacity, hydrologic soil group, surface-water flow direction, and land-cover type (Westenbroek and others, 2010). The input data files and output recharge data files were included in the data release (Fetkovich and others, 2025) that accompanies this report. The Soil-Water-Balance code uses the following equation (modified from Westenbroek and others, 2010):

$R=\left(P+S+R_i\right)-\left(Int+R_0+P_{et}\right)-\Delta Sm$

(3)

where

R

is recharge, in inches per day;

P

is precipitation, in inches per day;

S

is snowmelt, in inches per day;

R_i

is surface runoff, in inches per day;

Int

is plant interception, in inches per day;

R₀

is surface runoff outflow, in inches per day;

P_et

is potential evapotranspiration, in inches per day; and

ΔSm

is the change in soil moisture, in inches per day.

Input data needed for the SWB code were assigned to a user-specified grid that consisted of 917 columns by 372 rows of cells that were each 1,000 ft by 1,000 ft. Climate data inputs (daily precipitation, minimum temperature, and maximum temperature grids for 1980–2022) were obtained from the Daymet database (version 4; Thornton and others, 2020). Soil properties (soil-water storage capacity and hydrologic soil group) were obtained from the Gridded Soil Survey Geographic database (USDA, 2021). Land-cover types were obtained from the National Land Cover Database (fig. 2; Multi-Resolution Land Characteristics Consortium, 2023) and resampled to the SWB grid resolution by using the most common land-cover type as a percentage of the total coverage within each cell. Flow direction was derived by calculating the land-surface gradient by using the D8 method (Greenlee, 1987) from a 10-m DEM (USGS, 2015); any depressions were filled by using the ArcGIS Fill tool (Esri, 2023) after the DEM was resampled to the SWB grid. Filling depressions in the DEM ensures correct routing of surface runoff and eliminates isolated areas that could result in unrealistically high amounts of recharge.

Potential evapotranspiration was calculated by using the Hargreaves and Samani (1985) method for a reference latitude range of 34.5–35.4 degrees. Land-cover types (fig. 2; Multi-Resolution Land Characteristics Consortium, 2023) were used in conjunction with hydrologic soil groups to partition daily precipitation into plant interception (Int) and surface runoff (R_i and R₀) components and assign plant root-zone depths. The root-zone depths for grass/pasture and forest/shrubland (the dominant land-cover types for land overlying the aquifer; fig. 2) varied with soil texture but ranged from about 0.7 to 1.5 ft after being scaled to 40 percent of the values used by Westenbroek and others (2010), which were in permeable glacial deposits in Wisconsin. The root-zone depths were scaled by 40 percent to account for the difference in root zones between the study area and Wisconsin, where Westenbroek and others (2010) estimated the initial root zone depths. The maximum volume of water available in the root zone is calculated by multiplying the soil-water storage capacity by the root-zone depth. Changes in soil moisture (ΔSm) exceeding the soil-water storage capacity were assumed to be recharge (R) to the saturated zone. Smaller root-zone depths resulted in increased recharge and decreased evapotranspiration of water from the root zone, and larger root-zone depths resulted in decreased recharge and increased evapotranspiration of water from the root zone. Recharge from irrigation was not simulated by SWB but was assumed to be negligible given the relatively small amount of irrigation groundwater use in the study area.

Recharge was assumed to only occur in the unconfined part of the Antlers aquifer; therefore, the output data from the SWB model were summarized within the unconfined part of the Antlers aquifer. Recharge over large lakes was considered to be zero because open bodies of water do not contribute to recharge in the SWB calculation. The spatially distributed mean annual SWB-estimated recharge rate (SWBR) for the 1980–2022 study period was 8.58 in/yr, or about 19 percent of the mean annual precipitation of 45.2 in/yr. The annual SWBR ranged from 3.3 in. for 2005 to 18.8 in. for 2015, which respectively were years of much lower and higher precipitation compared to the mean annual precipitation (figs. 4, 22A). The ratio of mean monthly recharge to mean monthly precipitation is the recharge efficiency. During the study period, recharge efficiency was greatest during November–March when evapotranspiration is generally at a minimum; recharge during these months exceeded 20 percent of precipitation (fig. 22B). Recharge efficiency was lowest in July and August when evapotranspiration is generally at a maximum; recharge during these months was less than 9 percent of precipitation (fig. 22B). Spatially, SWBR increases from west to east across the unconfined part of the aquifer (fig. 23), which follows the general increase in precipitation from west to east (Thornton and others, 2020; Oklahoma Mesonet, 2023), with the highest SWBR rates generally being in the eastern half of the aquifer (fig. 23).

The modeled SWBR rates were compared to published estimates of mean annual recharge rates for the Antlers aquifer. Morton (1992) estimated that recharge was between 0.32 and 0.96 in/yr. However, Morton (1992) estimated recharge as the net groundwater gain to the aquifer from stream seepage, which was calculated as gains or losses between streamflow measurement locations along streams. Morton (1992) did not account for the wider spatial distribution of recharge that occurs away from streams. This study used stream gain or loss measurements to calculate stream seepage similar to Morton (1992), but recharge was also estimated for spatial distribution of recharge across the Antlers aquifer by using the SWB method.

SWBR was used as the primary recharge estimate for the conceptual model because it is a model that accounts for multiple inputs over the entire aquifer, as opposed to local estimates. SWBR for the Antlers aquifer is estimated as an inflow of 799,252 acre-ft/yr and is the only net inflow to the aquifer (table 6).

Streambed Seepage

Streambed seepage is commonly estimated by making streamflow measurements during periods without precipitation runoff. Total streamflow is the sum of base flow and runoff into the stream from precipitation that falls on the drainage area. Base flow is the component of streamflow that is supplied by the discharge of groundwater to streams, whereas runoff is attributed to other factors such as precipitation and overland flow. Base flow can be measured directly in streams during periods when there is essentially no runoff component to streamflow (Garner and Bills, 2012). To directly measure base flow, a series of streamflow measurements were made for this study using the methods of Rantz (1982) during a period when the runoff component of streamflow was at or near 0 cubic feet per second, referred to hereinafter as “seepage run measurements.” The seepage-run measurements were collected at 23 sites within the spatial extent of the Antlers aquifer during January 17–18, 2023 (fig. 24). In addition to quantifying base flow at individual measurement sites, comparing measurements along a stream reach made it possible to delineate tributary inflow and outflow and to estimate streamflow gains and losses across the aquifer in January 2023.

Seepage calculations along streams are done by subtracting inflow, including tributary inflow where applicable, from outflow and dividing this value by the total stream distance between measurement points. Seepage runs and estimates can help determine which stream reaches over the aquifer are gaining (that is, show a downstream increase in flow across the aquifer) or losing (that is, show a downstream decrease in flow across the aquifer). The stream lengths used in these calculations were obtained from the National Hydrography Dataset (Gary and others, 2009; Horizon Systems Corporation, 2015).

Tributaries with unmeasured base flows that were visually observed as having minimal or no flows were considered to contribute no base flow. Streamflow could not be measured in several tributaries that might have provided meaningful inflow data. Daily mean streamflow values measured during the same time period as the seepage run measurements from gaged locations on tributaries were used to supplement streamflow measurements obtained during seepage runs for the purpose of computed seepage gains and losses. Where gaged streamflow data or seepage-run measurements were not available, a seepage estimate was made by using a “zero-flow headwaters method.” The zero-flow headwaters method was done by estimating the length of the stream section from the furthest upstream measurement point up to the uppermost extent of the headwaters and assuming a gradient of increasing flow over the stream section assuming there was no flow at the uppermost extent of the headwaters. Thus, the seepage estimate was equal to the streamflow at the measurement point divided by the total upstream length of the stream from that point. This method was used to estimate streamflow for Walnut Bayou (figs. 1, 24; table 8). Another alternative method involved the use of USGS StreamStats (Smith and Esralew, 2010; Ries and others, 2017), which provided a watershed streamflow estimate to a selected point on the designated stream. This method was used to estimate streamflow for tributary inflow to the Little River in Arkansas (figs. 1, 24; table 8), as the zero-flow headwaters method was not viable owing to the large number of tributaries to this stream.

Streamflow measurements across the unconfined part of the Antlers aquifer indicated that streams are net gaining across the study area (except for Blue River [fig. 24; tables 6, 8]), including the following: Red River, Walnut Bayou, Hickory Creek, part of the Washita River, Muddy Boggy Creek, and Little River. The longest measured base-flow net gaining stream reach was a 50.3-mi reach along the Red River overlying the far western part of the Antlers aquifer (fig. 24). The greatest base-flow gain (+2.27 cubic feet per second per mile) occurred over a 32.17-mi reach of the Little River (fig. 24).

Net streambed seepage for the conceptual model was estimated from mean annual base flows computed by using the BFI code (Wahl and Wahl, 1995) in the USGS Groundwater Toolbox (Barlow and others, 2015) on data from selected streamgages in the study area (figs. 1, 24; tables 2, 8). Base-flow values for streamgages immediately upstream from the measurement site and any other tributary inflows were subtracted from the base-flow of the downstream streamgage to calculate the total base flow fed to that stream reach and lost by the aquifer. The total gain or loss between the stream measurements was then multiplied by an approximate percentage of the stream that flowed over the unconfined part of the aquifer to account for net streambed seepage to the underlying aquifer (fig. 24; tables 6, 8). This estimation specifies the net base-flow gain from or loss to the aquifer. These estimates and calculations are summarized in tables 2 and 8. Net streambed seepage in the unconfined part of the Antlers aquifer for the period 1980–2022 is a net outflow from the Antlers aquifer of 660,253 acre-ft/yr or 82.6 percent of total outflow from the Antlers aquifer (tables 6, 8). Streambed seepage is estimated to be the largest contributor to outflow from the Antlers aquifer.

Lateral Groundwater Flows

No data were available to estimate lateral groundwater flows across the defined boundaries of the Antlers aquifer in Oklahoma. Lateral groundwater flows are assumed to be from flow across the aquifer from northwest to southeast as groundwater flows downgradient from Oklahoma into Texas and is pushed by the incoming recharge against the stagnant saline zone and out of the study area. Net lateral groundwater flow was calculated as the difference between the aquifer inflows (recharge) and the aquifer outflows (saturated-zone evapotranspiration, streambed seepage, and well withdrawals) and was used to balance the water budget. Net lateral groundwater flows for the Antlers aquifer in Oklahoma were estimated to be a net outflow of 128,355 acre-ft/yr or 16.1 percent of the overall water budget (table 6).

Saturated-Zone Evapotranspiration

Evapotranspiration is the process by which water is transferred to the atmosphere directly through evaporation and indirectly through plant transpiration. Most evapotranspiration occurs at or near the land surface where precipitation pools as surface water or where it infiltrates the soil unsaturated zone and becomes available to plant root zones; most precipitation does not reach the saturated zone of the aquifer (Lubczynski, 2009). The land-surface and unsaturated-zone components of evapotranspiration were not a part of the conceptual model for the Antlers aquifer because they occur before infiltrating precipitation has reached the saturated zone to become groundwater recharge. Saturated-zone evapotranspiration occurs in areas of the aquifer where the saturated zone intersects the plant-root zone, most commonly in lower lying or wetland areas along streams (Lubczynski, 2009). Saturated-zone evapotranspiration was an important part of the conceptual-model water budget.

Saturated-zone evapotranspiration is difficult to estimate over a large area such as the study area. However, saturated-zone evapotranspiration rates were assumed to be proportional to (1) the area where the saturated zone intersects the plant root zone, (2) the mean depth to groundwater in that area during the growing season, and (3) the mean rate of transpiration associated with the assemblage of plants in that area (Smith and others, 2021). Because saturated-zone evapotranspiration generally occurs in wetland areas, wetland data from the National Wetlands Inventory (U.S. Fish and Wildlife Service, 2014) were used to determine the total area of wetlands overlying the unconfined part of the Antlers aquifer. Land areas with frequently saturated or flooded soils are classified as wetlands (Cowardin and others, 1979). An estimated 42,965 acres of wetland area (about 4 percent of the total unconfined area) overlie the unconfined part of the Antlers aquifer. The saturated-zone component of evapotranspiration was assumed to be active during the growing season (April–October [National Agricultural Statistics Service, 2023; Oklahoma Climatological Survey, 2023]). On an annual and monthly basis, the saturated-zone component of evapotranspiration is greatest in wet and hot years, and greatest in early summer months when precipitation and temperature are above their mean annual values (figs. 4, 5, 22; Scholl and others, 2005). In an earlier groundwater flow model for the Antlers aquifer, Morton (1992) concluded that evapotranspiration from the saturated zone was a trivial component of the overall water budget because the depth of the water table (about 50 ft below land surface) was deep enough to eliminate any concerns about saturated-zone evapotranspiration contributing to groundwater losses from the Antlers aquifer. Modeling tools have improved since 1992, and for this assessment, evapotranspiration from the shallower parts of the water table along streams and wetlands overlying the unconfined part of the aquifer was considered in the computation of the water budget.

White (1932) estimated annual saturated-zone evapotranspiration rates of 0.75–1.9 ft/yr for undisturbed Distichlis spicata (saltgrass) cover in southwestern Utah with a mean depth to water of 1–2 ft. Relative humidity and dewpoint temperature are comparatively low in southwestern Utah compared to southeastern Oklahoma (National Climatic Data Center, 2023); however, precipitation is much higher in southeastern Oklahoma compared to southwestern Utah, which likely results in an overall higher annual rate of evapotranspiration for the study area compared to what White (1932) reported for his study area. Smith and others (2021) and Rogers and others (2023) estimated saturated-zone evapotranspiration for the Salt Fork of the Red River alluvial aquifer and reaches 3 and 4 of the Washita River alluvial aquifer to be 1 ft/yr and 1.33 ft/yr, respectively. These two aquifers are in southwestern Oklahoma, where mean precipitation rates are lower than they are in southeastern Oklahoma. Although the methods described in White (1932) account for the WTF method assumptions described in this section of the report, annual data that covered the entire study period (1980–2022) were not available from any of the shallow wells. Thus, the White (1932) methods were not used in this report to estimate groundwater outflows attributable to saturated-zone evapotranspiration from daily WTF data at wells with shallow depths to water. Wells GW-01, GW-04, and GW-06 (fig. 1; table 1) each show some level of daily fluctuation during the daylight hours in the summer months (fig. 7). These daily fluctuations are indicative of daily declines in groundwater level during daylight hours in summer conditions. The observation of these daily fluctuations in groundwater levels indicated that saturated-zone evapotranspiration was an active process in the Antlers aquifer. Because the unconfined part of the Antlers aquifer receives greater amounts of precipitation (Oklahoma Mesonet, 2023) and is generally warmer than the areas overlying the Salt Fork of the Red River alluvial aquifer and reaches 3 and 4 of the Washita River alluvial aquifer, a saturated-zone evapotranspiration rate of 2.0 ft/yr was assumed for the study area. If the estimated 42,965 acres of wetland area that overlie the unconfined part of the Antlers aquifer are assumed to have a constant depth to water and saturated-zone evapotranspiration rate, the saturated-zone evapotranspiration rate would correspond to an annual saturated-zone evapotranspiration outflow of 7,161 acre-ft/yr or 0.9 percent of outflow from the Antlers aquifer (fig. 21; table 6).

Well Withdrawals

Well withdrawals were assumed to be equivalent to the mean annual reported groundwater use during 1980–2022, or 3,483 acre-ft/yr (fig. 21; tables 3, 5, 6). These withdrawals were generally greatest during dry and hot years because more groundwater was required in those years for crops to make up for the lack of precipitation and increased evapotranspiration. The altitude of the water table generally declines during dry and hot years (especially during extended droughts) and rises during wet and cool years (figs. 4, 7). The degree to which the water table fluctuates annually at a given location is related in part to the volume of nearby well withdrawals and the distribution (or concentration) of recharge near that location. The well withdrawal amount of 3,483 acre-ft/yr for the study period accounts for 0.4 percent of the conceptual water budget (fig. 21; table 6).

Vertical Leakage

Groundwater exchange is probably substantial between the Antlers aquifer and the Goodland-Walnut confining unit (Hart and Davis, 1981) but probably not substantial between the Antlers aquifer and the lower confining units. Groundwater exchange likely involves groundwater inflow to the Antlers aquifer from the overlying Goodland-Walnut confining unit in the shallow parts of the confined Antlers aquifer; the direction of groundwater flow eventually reverses from inflow to outflow as the altitude of the top of the Antlers aquifer decreases. The hydrograph from continuous water-level recorder well GW-02 (fig. 7C; USGS, 2023), which is completed in the confined part of the Antlers aquifer, shows rapid responses to precipitation, indicating that the Antlers aquifer is vertically hydraulically connected to the overlying units and that vertical leakage is occurring. The observed response to precipitation could also be an effect of site-specific issues with the given well, such as compromised well casings or surface seals designed to keep surface runoff from entering the well. Because of the change from groundwater inflow to outflow with depth, net vertical leakage is considered to be net zero and a negligible part of the water budget (fig. 21; table 6).

Groundwater Storage

Annual water-level measurements for the 1980–2022 study period were not available from wells completed in the Antlers aquifer, so estimating net storage change in the aquifer was not possible. The net storage change of the Antlers aquifer was assumed to be a negligible component of the conceptual-model water budget (fig. 21; table 6), which corresponds with the minimal change between the potentiometric maps in Hart and Davis (1981) and Morton (1992) and in this report.