Purpose and Scope

This report builds on initial observations made by Penn and Everett (2019), who examined groundwater-level elevations measured in Elbert County from 2015 through 2018. Some wells analyzed by Penn and Everett (2019) were not evaluated in this report because they were discontinued. This report uses data collected from a network of 36 groundwater wells where discrete measurements of groundwater levels were measured bimonthly from 2015 through 2023. There are at least two wells in each of the bedrock aquifers. In 7 of the 36 wells, a vented pressure transducer with an internal data logger records hourly groundwater-level measurements (also referred to as continuous groundwater levels). The purpose of this report is to summarize groundwater levels measured during the study period and present results from statistical analyses of changes in groundwater-level elevations through time (trends) in the Denver Basin aquifer system in Elbert County, from April 2015 through November 2023.

Previous Studies

One of the first published studies of the Denver Basin aquifer system described the artesian groundwater conditions (Cross and others, 1884). Since then, numerous studies have documented the geology, physiography, climate, stratigraphy, and hydrologic conditions of the Denver Basin aquifer system. By 1989, a bibliography of geology and groundwater geology for the Denver Basin (Wireman and Romero, 1989) contained more than 160 references (Everett, 2014). Paschke (2011) cited more than 190 references in a detailed description of previous work. Throughout the 1970s and 1980s, the USGS and the Colorado Division of Water Resources (CDWR) mapped and characterized the Denver Basin aquifer system (Romero and Hampton, 1972; Romero, 1976; Robson and Romero, 1981a, 1981b; Robson, Romero, and Zawistowski, 1981; Robson, Wacinski, and others, 1981; and Robson, 1983), which helped lead to the development of a groundwater-flow model (Robson, 1987) and a three-dimensional MODFLOW-2000 groundwater-flow model (Paschke, 2011) of the Denver Basin aquifer system.

Groundwater-level measurements in the Denver Basin aquifer system began in the 1890s (Emmons and others, 1896). From 1956 to 1963, the USGS conducted the first basinwide assessment of groundwater levels (McConaghy and others, 1964), followed by a comprehensive set of groundwater-level data for the bedrock and alluvial aquifers measured from 1956 to 1981 (Major and others, 1983). During the 1980s, the CDWR established a groundwater-level monitoring network with data published in annual reports (Pottorff and Horn, 2013). From 2007 to 2017, CDWR conducted long-term groundwater monitoring in the Denver Basin aquifer system, which indicated decreasing groundwater-level elevations in all bedrock aquifers; however, changes in groundwater levels from 2016 to 2017 and 2012 to 2017 were inconsistent (Flor, 2017). The USGS published a study of groundwater-level elevations in the Denver Basin aquifer system within Elbert County (Penn and Everett, 2019), which presented results showing all aquifers except the upper Dawson aquifer had more wells with statistically significant trends indicating decreasing groundwater-level elevations than increasing groundwater-level elevations. In Douglas and El Paso Counties (fig. 1), groundwater-level monitoring networks, similar to the Elbert County network presented in this report, has been operating since 2011 (Everett, 2014; Malenda and Penn, 2020) and 2021 (Kisfalusi and others, 2025), respectively.

Description of Study Area

Elbert County is 1,851 square miles of mostly rural land in eastern Colorado, southeast of Denver, and northeast of Colorado Springs. Elbert County is bordered by Arapahoe County to the north, Lincoln County to the east and southeast, El Paso County to the south and west, and Douglas County to the west (fig. 1). Elbert County receives approximately 12 to 18 inches of precipitation per year, accounting for both rain and snowfall, compared to the average 16 inches for Colorado (Elbert County Planning Commission, 2018). Drought-tolerant plants such as Bouteloua gracilis (blue grama), Bouteloua dactyloides (buffalograss), Pascopyrum (wheatgrass), and Fescuta, sp. (fescue) dominate the plains, whereas Pinus ponderosa (ponderosa pine) lines the western region of Elbert County. Populus deltoides (cottonwood) and Salix, sp. (willow) populate the riparian corridors (Elbert County Planning Commission, 2018). Numerous intermittent tributaries of the South Platte River originate in Elbert County, including Kiowa Creek, East, Middle, and West Bijou Creeks. Big Sandy Creek, an intermittent stream, flows into the Arkansas River (fig. 1).

Description of the Denver Basin Aquifer System

The Denver Basin aquifer system, which covers an approximate 7,000 square miles, is bound on the western edge by the base of the Colorado Front Range Mountains and extends into the eastern plains of Colorado (Bauch and others, 2014; fig. 1 of this report). The northern extent ends near Greeley, Colorado, and the southern extent reaches into El Paso County. The structure of the basin is synclinal (bowl shaped) and is composed of Late Cretaceous to Tertiary sandstone bedrock aquifers separated by claystone confining units (Fenneman, 1931; Robson, 1987; Paschke, 2011). The four principal bedrock aquifers, from youngest (shallowest) to oldest (deepest), are the Dawson aquifer in the Late Cretaceous to Eocene Dawson Formation, Denver aquifer in the Late Cretaceous to Paleocene Denver Formation, Arapahoe aquifer in the Late Cretaceous Arapahoe Formation, and Laramie-Fox Hills aquifer in the Late Cretaceous Laramie Formation and Fox Hills Sandstone. The principal bedrock aquifers are underlain by the confining Cretaceous Pierre Shale (fig. 3). The Dawson and Arapahoe aquifers are divided into lower and upper units in parts of the basin by discontinuous confining units. In Elbert County, the Arapahoe aquifer is undivided, and the Dawson aquifer is divided, thus totaling five distinct bedrock aquifers. Outcrops of each aquifer can be found along the outer edge of their extents and are generally considered unconfined, whereas confined conditions exist towards the interior of the basin in each aquifer where it is overlain by a younger confining unit (Paschke, 2011). The physical characteristics of the bedrock aquifers in Elbert County are summarized in table 2. Studies from which the information in table 2 was acquired include Romero (1976), Kirkham and Ladwig (1979), Schneider (1980), Robson and others (1981a), Robson and others (1981b), Robson (1987), Crifasi (1992), Raynolds and others (2001), Raynolds (2002, 2004), and Paschke (2011).

Groundwater-Level Measurements and Groundwater-Level Elevations

This section presents the methods used for making and processing the discrete and continuous groundwater-level measurements analyzed in this study. Groundwater levels are presented as depth to groundwater in ft below land surface datum (LSD). Calculated groundwater-level elevations are presented in ft above the North American Vertical Datum of 1988 (NAVD 88). Well common names are a combination of aquifer of completion abbreviation and an assigned number: upper Dawson aquifer well, UDAW; lower Dawson aquifer well, LDAW; upper or lower Dawson aquifer well, DAWMAS; Denver aquifer well, DENV and DENMAS; Arapahoe aquifer well, ARAP and ARAPMAS; Laramie-Fox Hills aquifer well, LARA (table 1).

Groundwater-Level Summary, Groundwater-Level Elevation Trend Analysis, and Mapping

After measurement of, and corrections to, discrete and continuous groundwater levels, all groundwater levels used in this report went through an internal review and an independent approval process. Groundwater levels, both discrete and continuous, were converted to groundwater-level elevations prior to trend analysis and graphical representation. The hydrographs showing groundwater-level elevations of the discrete and continuous data through time are available in appendix 2. The presence of temporal trends in both the discrete and continuous groundwater-level elevation data was evaluated using nonparametric statistical methods. This study used a similar approach to Malenda and Penn (2020) using a seasonal Mann-Kendall (sM–K) trend test (Helsel and others, 2020). These trend analyses were completed using the R statistical software (R Core Team, 2018) and the “EnvStats” package (Millard, 2013) as described in appendix 3. The calculated trend estimate, in ft per year, was quantified using the Theil-Sen slope estimate (Sen, 1968; Hirsch and others, 1982) with the “EnvStats” package (Millard, 2013). This approach performs a nonparametric test for a monotonic trend within each season and summarizes the trend as the median of all within season slopes (Hirsch and others, 1982; Helsel and others, 2020). Refer to appendix 3 for a more detailed description of the statistical methods and the respective equations.

The sM–K test (Mann, 1945; Kendall, 1975; Helsel and others, 2020) was applied to static discrete and continuous groundwater-level elevations measured between April 2015 and November 2023. The sM–K test accounts for temporal correlation caused by seasonality by comparing data from a user-defined season only to data from the same season (appendix 3). For the sM–K test on discrete data, seasons were defined as months when measurements were made to minimize effects of temporal correlation among months (totaling six seasons for bimonthly site visits). For the sM–K test on continuous data, the test was completed on the maximum hourly groundwater-level elevation measured each day, and seasons were defined as individual months to minimize effects of temporal correlation among months (totaling 12 seasons). The null hypothesis of no monotonic trend and an alpha (α), or probability value (p-value), of 0.05 was used to evaluate trend significance. Therefore, when the p-value was less than or equal to 0.05, the null hypothesis was rejected, and a trend in groundwater-level elevations was considered statistically significant (Helsel and others, 2020). Previous reports by Malenda and Penn (2020) and Penn and Everett (2019) used a p-value of 0.1. A more conservative confidence level was used in this report to reduce the likelihood of incorrectly rejecting the null hypothesis (Helsel and others, 2020). A negative trend indicates generally declining groundwater-level elevations in the well through time, and a positive trend indicates generally rising groundwater-level elevations in the well through time.

In addition to the sM–K test, the sensitivity of wells to their environment and conditions was assessed by calculating the departure of static discrete groundwater-level elevations from the 2015 median groundwater-level elevation for each well, in each aquifer. For this study, the 2015 median groundwater-level elevation was selected because it is the earliest available USGS groundwater-level data for Elbert County. Additionally, 2015 had the highest groundwater-elevation for the entire study period for the most wells (USGS, 2023). Ideally, the median groundwater level would be selected from predevelopment or early development in order to represent stable aquifer conditions; however, these data are not available from the wells in the study site. As such, the earliest (2015) groundwater-level data were selected. If a well did not have data in 2015, it was not included in this analysis. The departure from the 2015 median was compared with the average 30-year normal precipitation for Elbert County (PRISM Climate Group, 2024) where dry periods are less than, and wet periods are more than the 30-year normal. The 30-year normal of precipitation, provided by Parameter-elevation Regression on Independent Slopes Model (PRISM) climatological datasets, for Elbert County between the years 1991 and 2020 is 17.2 inches (PRISM Climate Group, 2024). Potentiometric-surface maps show the hydraulic head distribution of an area with contour lines of equal hydraulic head. Groundwater flow is from areas of high hydraulic head to areas of low hydraulic head; flow direction is perpendicular to the contours (Winter and others, 1998). Groundwater-level elevations from April 2023 observations were interpolated spatially using the Python programming language (Van Rossum and Drake, 2009) Numpy package (Harris and others, 2020) to derive the hydraulic head distribution and contour lines of equal hydraulic head using the static discrete values of groundwater-level elevation above NAVD 88 in ft in the upper and lower Dawson aquifers. Potentiometric-surface maps in Penn and Everett (2019) were compiled from April 2018 data. April 2023 data were used in this report for consistency between both reports and to provide a comparison between April 2018 and April 2023 data.

Discrete Groundwater-Level Elevation Summary and Trends

Groundwater levels in each of the bedrock aquifers varied both temporally and spatially. In general, groundwater-level elevations were lowest during summer and fall (June through October) and recovered to higher elevations in winter and spring (December through April) (appendix 2). Some wells exhibited strong seasonal fluctuations of about 10 ft (for example, well UDAW 13, fig. 2.8), whereas others show minimal seasonality with fluctuations less than 1 ft between measurements (for example, well DENV 16, fig. 2.20). Seasonal variations are caused by natural processes, including precipitation and evapotranspiration in aquifer zones connected to the land surface and timing of aquifer recharge in confined aquifer zones (Paschke, 2011). Human activities, such as increased irrigation for agriculture (during the growing season) and domestic pumping (for lawns), can also affect seasonal variations.

Of the 36 wells monitored, 35 were analyzed (UDAW 20 [fig. 2.9] was excluded because of insufficient data) in the study period, with 29 exhibiting statistically significant trends in discrete groundwater-level elevations, based on the sM–K test (table 3, appendix 3). Potential reasons for wells not resulting in statistically significant trends could be inadequate static measurements, such as DENV 17 (fig. 2.19) with approximately fewer than 35 percent static measurements. Continued monitoring and more static measurements at the wells where trends were not identified may increase the likelihood of statistically significant trend occurrence at those wells in the future. Of the remaining five wells that did not have statistically significant trends, three wells (ARAP 7, DAWMAS22, and ARAPMAS27) have sufficient data (more than 75 percent of measurements with a static status), but based on the sM–K results and hydrographs are considered in apparent steady state (table 3, figs. 2.26, 2.3, and 2.27, respectively). The two remaining wells (LDAW 15 and ARAP 6) had more than 57 percent of measurements with a static status but did not meet the criteria of a statistically significant p-value of less than or equal to 0.05 (table 3, figs. 2.12 and 2.28, respectively). Based on the criteria used by Penn and Everett (2019) however, the trends would be considered statistically significant with p-values less than or equal to 0.10.

In wells with statistically significant trends, groundwater-level elevations were declining in all aquifers except the lower Dawson aquifer. The lower Dawson aquifer had two wells (DAWMAS27 and DAWMAS16) with increasing groundwater-level elevations of 0.040 and 0.63 ft/yr, respectively, near the border of Douglas and El Paso Counties. In this study, the groundwater-level elevation trend medians in the upper Dawson, lower Dawson, Denver, Arapahoe, and Laramie-Fox Hills aquifers were −0.23, −0.66, −0.64, −0.39, and −0.63 ft/yr, respectively (table 3). The groundwater-level elevation trend medians in this study, compared to the overall trends in Penn and Everett (2019), both indicated decreasing groundwater-level elevations except in the upper Dawson aquifer, where the trend direction was opposite, a positive trend (increasing groundwater elevations) from 2015 to 2018 and a negative trend (decreasing groundwater elevations) from 2015 to 2023. Trends from the previous study (Penn and Everett, 2019) were determined by a linear regression analysis, whereas trends in this study were determined using the sM–K test. Additionally, trends from Penn and Everett (2019) used a shorter period of analysis; therefore, trend magnitudes (amount groundwater-level elevation changed, in ft/yr) may be affected by both the period and method applied. Statistically significant trends throughout Elbert County are relatively consistent for grouping and distribution, where the largest trends are near the western border of Elbert County and along Colorado State Highway 86 (fig. 4). Trend magnitude and direction were generally in agreement for wells near and in the same aquifer (for example, UDAW 13 and UDAW 15; LDAW 12 and DAWMAS21; DENV 14 and DENV 15; ARAP 5 and ARAPMAS28; LARA 5 and LARA 6; fig. 4).

The departure from the 2015 median groundwater-level elevation for each well in each aquifer (fig. 6A−E) indicated departures from the median in each aquifer were negative, meaning groundwater level was less than the median, more frequently (at least 60 percent) than positive (groundwater level greater than the median), supporting the overall pattern of negative trends in groundwater-level elevation in Elbert County. The maximum negative departures from the 2015 median groundwater-level elevations were greater than the maximum positive departure for all wells in each aquifer (maximum negative and positive departures, respectively, were as follows: −17.06 ft and 3.81 ft in the upper Dawson aquifer [fig. 6A]; −59.55 ft and 20.10 ft in the lower Dawson aquifer [fig. 6B]; −17.96 ft and 1.83 ft in the Denver aquifer [fig. 6C]; −13.26 ft and 2.48 ft in the Arapahoe aquifer [fig. 6D]; and −12.28 ft and 1.86 ft in the Laramie-Fox Hills aquifer[fig. 6E]). Wells nearest to Elizabeth, Colorado, and the border between Elbert and Douglas Counties had the greatest negative departures from the 2015 median in the upper and lower Dawson aquifers (UDAW 13, UDAW 15, LDAW 12, LDAW 13, LDAW 14, and DAWMAS 21, figs. 5, 6A−B). Wells in the Denver aquifer (DENV 14, DENV 15, and DENMAS05), Arapahoe aquifer (ARAP 8 and ARAP 3), and Laramie-Fox Hills aquifer (LARA 5, LARA 6, and LARA 7) that showed the greatest negative departures from the 2015 median also indicated the largest decreasing groundwater-level elevations (figs. 4, 6C−E). Departures from the 2015 median groundwater-level elevations were greatest during the dry period between the years 2020 and 2023. Notably, departures from the 2015 median were largest in LDAW 12 well (approximately −60 ft; fig. 6B), which has proximity to a multiresidential development beginning between 2017 and 2019 based on imagery of the area (Google, 2023); these years coincide with the largest drawdowns during the study period at this site (fig. 6B; fig. 2.14). Based on CDWR well records, the recent residential development included the addition of two commercial wells intersecting the Denver and Arapahoe aquifers (CDWR, 2023). Although LDAW 12 had the largest departures from the 2015 median, seasonal recharge is shown, which returns the departure from the 2015 median to a near zero difference annually. Conversely, most wells with the greatest departures in the Denver, Arapahoe, and Laramie-Fox Hills aquifers do not have departures returning to zero annually and instead, some continue to decline (fig. 6C−E). These aquifers are predominately under confined conditions (Ruybal and others, 2019) and are generally dominated by thousands of years old groundwater which has not been affected by mixing with young recharge groundwater (Musgrove and others, 2014). Substantial drawdown from increased pumping in the bedrock aquifers resulted in lowered potentiometric surfaces and increased areas of unconfined conditions, making them more susceptible to varying recharge (Paschke, 2011; Ruybal and others, 2019). The departure from the 2015 median groundwater-level elevation can be used as a tool to observe potential effects from increased groundwater withdrawals, especially in confined aquifers in the Denver Basin bedrock aquifer system.

Continuous Groundwater-Level Elevation Summary and Trends

Hydrographs of continuous groundwater-level elevations from the wells containing pressure transducers were generally in agreement with discrete groundwater levels measured at the same well (figs. 2.4, 2.5, 2.14, 2.18, 2.24, 2.36). The pressure transducer at site LDAW 15 (fig. 2.12) was installed in December 2022 and was not included in the trend analysis because of insufficient data. Groundwater-level elevations were generally highest during winter and spring (December through April) and lowest during summer and fall (June through October), except for DENV 12 (fig. 2.24), which does not have a strong seasonal pattern. Based on the results from the sM–K trend test conducted on the continuous daily maximum groundwater-level elevations, the trend direction (increasing or decreasing groundwater-level elevations) and magnitude (change in groundwater-level elevation, in ft/yr) generally agreed with the trend analysis results calculated from discrete groundwater-level elevations (table 3). The hydrograph for LARA 3 (fig. 2.36) clarifies the additional information that continuous measurements can provide about groundwater-level elevation changes between discrete measurements.

Potentiometric-Surface Maps

Potentiometric-surface maps were compiled from static groundwater-level elevations made during April 2023 in the upper and lower Dawson aquifers. Only wells with static measurements were used to derive the groundwater-level elevations using the Python programming language (Van Rossum and Drake, 2009) Numpy package (Harris and others, 2020). The potentiometric surface as 40-ft interval contours for April 2023 in the upper and lower Dawson aquifers are shown in figures 7 and 8, respectively. Groundwater flow in the upper and lower Dawson aquifers in western Elbert County, based on the derived potentiometric-surface maps, is generally from south to north. The potentiometric surface maps are generalized representations of a complex groundwater system in the upper and lower Dawson aquifers. Because of the small spatial dataset size for each aquifer, the potentiometric-surface maps are limited to a generalized interpretation of the groundwater-flow direction (Anderson and Lundgren, 2024). Despite generalizations of the interpretation, potentiometric surface maps can be used to assist water resource managers to identify areas for future monitoring and gain a general understanding of groundwater flow (Anderson and Lundgren, 2024).