Well selection for this study was a collaborative process between the USGS, the coordinator of the ENWRA group, geologists from the University of Nebraska Conservation and Survey Division, and staff from the participating NRDs. Wells selected for sampling were chosen from the Nebraska Department of Natural Resources registered well database (Nebraska Department of Natural Resources, 2020). A detailed, reliable driller’s log was required to ensure that the chosen well was screened within the targeted aquifer (Nebraska Department of Natural Resources, 2020). Monitoring wells were preferred over other well types because the depth and length of the screened interval is usually based on lithologic descriptions and geophysical logs collected by a professional geologist (Conservation and Survey Division, University of Nebraska-Lincoln, 2020a). In many instances, a monitoring well was not available and domestic or stock wells were chosen for sampling. Wells with relatively short screen intervals (10 ft or less) were preferred to ensure a representative sample from a specific interval within the aquifer. The type of well, aquifer sampled, and screen interval are listed in table 2. One well, LC–6S domestic, was not completed as a screened well, rather a casing was installed through 40 ft of unconsolidated deposits until the borehole reached competent bedrock, then the borehole was left open to the formation from 40 to 95 ft below land surface (table 2).

This report uses three different methods of well identification. The first method uses the USGS site number, which is a 15-digit number (table 2). The 15-digit USGS site number is particularly useful for accessing available water-quality or water-level data for a given well in the USGS National Water Information System (NWIS; U.S. Geological Survey, 2020b). The next identification method is the USGS station name. The station name uses the legal description of the well location and a local well name, which is the well name typically used by participating NRDs (hereinafter referred to as the well’s “field name”), such as 19N 8E 20AD (LE–1). The last well identification method is the field name, such as LE–1 (table 2).

For this report, a total of 31 wells were sampled from the three secondary bedrock aquifers. Of these 31 wells, 22 wells were sampled from wells completed in the Dakota aquifer, 3 were sampled from wells completed in the Niobrara aquifer, and 6 were sampled from wells completed in Paleozoic aquifers (table 2, fig. 2). Well depths for sampled wells completed in the Dakota aquifer ranged from 95 to 750 ft below land surface. Well depths for sampled wells completed in the Niobrara aquifer ranged from 120 to 180 ft below land surface. Well depths for sampled wells completed in the Paleozoic aquifers ranged from 79 to 172 ft below land surface (table 2). Of the 22 sampled wells completed in the Dakota aquifer, 15 were monitoring wells and 7 were domestic wells. Two monitoring wells and one domestic well were sampled for the Niobrara aquifer. One stock and five domestic wells were sampled for the Paleozoic aquifers (table 2).

Groundwater samples were collected from August 9, 2016, to June 20, 2018 (table 2). Sample collection procedures followed the guidelines and protocols set forth in the USGS “National Field Manual for the Collection of Water-Quality Data” (U.S. Geological Survey, variously dated), the USGS Menlo Park Tritium Laboratory (U.S. Geological Survey, 2020c), and the USGS Reston Stable Isotopes Laboratory (U.S. Geological Survey, 2020d). The sampling protocols described in this section ensured that a representative sample of groundwater was collected at each well and that the samples were collected and handled in a way that minimized contamination.

Prior to sampling monitoring wells, the water level was measured to the nearest 0.01 ft using a graduated electric tape following guidelines and protocols described in Cunningham and Schalk (2011). For domestic and stock wells, water level and well depth were not measured because the wells generally were in operation prior to sampling. Groundwater samples were collected from monitoring wells using a stainless-steel submersible pump. Groundwater samples from domestic or stock wells were collected at a hydrant or spigot closest to the well so that water was sampled prior to reaching a pressure tank or any other water treatment such as water softeners. Groundwater was sampled through Tygon tubing connected to the submersible groundwater pump or the chosen hydrant or spigot. Samples were collected in-line so that groundwater samples were not exposed to the atmosphere prior to collection.

Groundwater from all sampled wells was pumped into a flow-through chamber to facilitate the monitoring of the physical properties of the water to ensure that enough water was purged from the well casing prior to the collection of a representative sample. The physical properties—specific conductance, pH, water temperature, and dissolved oxygen—were measured in the field at 3-minute intervals by a multiparameter sonde (YSI Incorporated, 2020). Representative samples were collected after three well volumes had been purged from the well and after sequential readings of the physical parameters stabilized within limits described by USGS “National Field Manual for the Collection of Water-Quality Data” (U.S. Geological Survey, variously dated).

Water samples from all sites were analyzed for major ions, trace elements, nutrients including nitrate, and stable isotopes. Age tracers (tritium [³H] and carbon-14 [¹⁴C]) were analyzed for a subset of sampled wells. The constituent, analyzing laboratory, references to methods used, and field preservation procedures are given in table 3. Sampling results from all analyses are available online in the USGS NWIS (U.S. Geological Survey, 2020b). For the convenience of the reader, selected results are presented in this report to support the interpretations and conclusions.

For this report, conventional nomenclature was used to describe analyses of water samples for stable isotopes. The composition of stable isotopes of low-mass (light) of oxygen (O) and hydrogen (H) commonly are reported as “δ” (delta) values, which indicate parts per thousand or per mil. The reported value is compared to the isotopic ratio of the Vienna Standard Mean Ocean Water (Clark and Fritz, 1997). The general expression for the δ value is calculated by the following equation (Clark and Fritz, 1997): δ  in per mil = Rx Rs −1 *1,000 (1)where

A negative δ value indicates that the sample is depleted of the rare isotope relative to the standard; that is, the sample is isotopically “light.” A positive δ value indicates that the sample is enriched in the rare isotope relative to the standard; that is, the sample is isotopically “heavy.”

Understanding and interpreting the stable isotopic composition of groundwater samples is aided by understanding hydrologic processes that can affect isotopic ratios. Isotopic fractionation occurs where the isotopic composition is altered by chemical, biological, or physical processes that may result in the preferential enrichment or depletion of one isotope over another. This fractionation process partitions isotopes as a function of the differences in the masses of the isotopes. Because the heavier isotope has a stronger molecular bond (oxygen-18 [¹⁸O] has a stronger molecular bond than oxygen-16 [¹⁶O] and deuterium [²H] has a stronger molecular bond than hydrogen-1 [¹H]), the liquid phase of water generally is isotopically “heavier” than the gaseous phase (Kendall and Caldwell, 1998). As a result, evaporation is a major fractionation process in which the lighter isotope is concentrated in water vapor, whereas the heavier isotope enriches the aqueous phase. Additional information on isotopes and their presence in the environment can be found in references such as Clark and Fritz (1997).

Data-analysis procedures included general statistical analyses, creation of Piper diagrams (Piper and Garrett, 1953), scatter plots, and summary tables. This report only contains selected plots and tables necessary to support the interpretations and conclusions presented herein. Analyses of stable isotope data and age tracer results, including radiocarbon dating procedures, are also described in this section of the report.

Statistical analyses were completed to determine if differences in the concentrations of selected constituent were significant between aquifers. Two nonparametric tests were performed for this report: Kruskal-Wallis and Wilcoxon Rank Sum (Helsel and others, 2020). The Kruskal-Wallis test was used to determine differences between distribution of concentrations of selected constituents in samples by aquifer. For this report, this test was performed on samples from wells screened in the Dakota, Niobrara, or Paleozoic aquifers. The null hypothesis for the test was that the distributions for the groups are identical, and the alternative hypothesis was that at least one of the group medians is different from the others (Helsel and others, 2020). The Wilcoxon Rank Sum test (Helsel and others, 2020) was used to determine if differences between distributions of the two subsets of samples collected from the Dakota aquifer were significant. Water samples collected from the Dakota aquifer were divided into two groups based on the dominant water type, which is discussed in further detail in the “Major Ions and Trace Elements” subsection later in this report. The null hypothesis for the test was that the distributions for the groups are identical, and the alternative hypothesis was that at least one of the distributions is different from the others (Helsel and others, 2020). For both types of nonparametric statistical tests, the null hypothesis was rejected when the calculated p-value was less than an alpha value (α) of 0.05, which indicates a 95-percent confidence level that the probability value from the statistical test generated an accurate representation of the populations being tested (Helsel and others, 2020). A p-value is defined as the measure of the probability that an observed difference occurred by random chance (Helsel and others, 2020). If a calculated p-value is less than 0.05, then it can be inferred there is a significant difference between the medians of the sample groups with 95-percent confidence.

Nonparametric statistical methods rely on the conversion of data to ranked values and do not require that the data be distributed normally (Helsel and others, 2020). Censored data, which have values less than the reporting limit of the analytical procedure, were kept for analyses. In this report, for the nonparametric statistical analyses, each censored value was assigned a value equal to one-half the reporting limit. A value of one-half the reporting limit ensures that the censored data are smaller than the smallest number above the reporting limit. Because the tests performed are nonparametric and compare independent populations by ranking the data, assigning a minimum value to the censored data does not affect the hypothesis-test results (Helsel and others, 2020).

Quality control samples were collected for this study and included field blanks and field replicates. Quality control samples were collected to evaluate and determine if samples had been contaminated or if the data were biased by the collection, processing, storage, or analysis of the samples. For this study, six field blanks and five field replicate samples were collected. Selected statistics for the field blank results and the relative percent difference between replicate samples are shown in table 4. The relative percent difference is calculated using equation 3 given below. Table 4

Table 4. Selected statistics of field blank samples and relative percent difference between replicate samples, eastern Nebraska, 2016–18.

[mg/L, milligram per liter; --, not applicable or not sampled; <, less than; NH₃, ammonia; NH₄⁺, ammonium; N, nitrogen; µg/L, microgram per liter; ¹⁴C, carbon-14; %, percent; pCi/L, picocurie per liter; δ¹³C, carbon isotope ratio in per mil Vienna Peedee belemnite; δ²H, hydrogen isotope ratio in per mil Vienna Standard Mean Ocean Water; δ¹⁸O, oxygen isotope ratio in per mil Vienna Standard Mean Ocean Water]

Field blanks are collected to determine the occurrence and magnitude of sample contamination from collection, processing, storage, and analysis. Field blanks are collected using water that has been certified to be free of inorganic constituents and were analyzed for major ions, trace elements, and nutrients. Generally speaking, few detections, or “hits,” were analyzed from the field blanks (table 4). Most detections were within an order of magnitude of the lower detection limit for the specific constituent; however, two blank samples, D–SWSarpy1 and LC–4S, had concentrations of cobalt and zinc, respectively, that exceed an order of magnitude of the detection limit. Concentrations of copper were detected in blank samples that exceeded an order of magnitude of the detection limit for wells D–SWSarpy1 and D–Sp2 (table 4).

Five field replicate samples were collected to assess the variability in analytical results. The relative percent difference in the concentrations of the replicate samples for major ions all were 5 percent or less of each other, with the exception of the concentration of magnesium, sodium, and fluoride in one set of replicate samples, which were 5.1, 5.9, and 7.7 percent, respectively (table 4). It should be noted that when concentrations for replicate samples are low, small differences in concentration can result in higher relative percent differences. For nutrients, the relative percent differences in the concentrations of the replicate samples for most nutrients were less than 5 percent or within an order of magnitude of the detection limit. Notable exceptions included the concentration of nitrate plus nitrite in one set of replicate samples that had a 9.8-percent (0.39 milligram per liter [mg/L]) difference and nitrite in one set of samples which had a 32.3-percent (0.005 mg/L) difference (table 4). For trace elements, the six trace elements with the greatest maximum relative percent differences were copper (16.2 percent), iron (28.1 percent), cadmium (11.8 percent), manganese (13.9 percent), nickel (12.8 percent), and vanadium (11.0 percent); however, the associated differences in concentration were within one order of magnitude of the lower detection limit (table 4). It is important to note that because the concentrations of these constituents are low, the calculated relative percent difference values can be anomalously high. For the trace element cobalt, the maximum relative percent difference was calculated at 9.5 percent, which was an order of magnitude greater than the detection limit. Exceeding the order of magnitude threshold indicates the possibility that some samples may have a positive bias. The median relative percent difference for all replicate samples was 0.3 percent. The median relative percent difference for iron was 7.8 percent but the difference in concentrations was within an order of magnitude of the lower detection limit (table 4). Replicates were only collected for two ³H samples (D–A1 and D–SWSarpy1). The relative percent difference was not calculated for D–A1 because the ³H concentrations were below the detection limit. The relative percent difference calculated for ³H in the sample from well D–SWSarpy1 was 24, which was due to the fact that ³H concentrations were fairly low and differences between samples are close to the lower detection limit.

Major ion concentrations for each sample are plotted as a Piper diagram (Piper and Garrett, 1953) shown in figure 3. Analysis of the major ion concentrations indicates samples collected from the Dakota aquifer show no single primary water type; however, the two most common water types within the Dakota aquifer are calcium bicarbonate (10 samples) and calcium sulfate (7 samples; figs. 3 and 4, table 5). Samples that were characterized as calcium bicarbonate generally had lower specific conductance and concentrations of total dissolved solids (table 5A). This is consistent with previous studies, which determined that calcium bicarbonate was the dominant water type where the Dakota aquifer was unconfined and groundwater was recharged from local precipitation (Gosselin and others, 2001). Sampled wells that are not characterized as calcium bicarbonate water type typically had higher concentrations of total dissolved solids and lower dissolved oxygen concentrations (table 5A). Calcium sulfate water types often are present where Cretaceous-age shales overly the Dakota aquifer, such as in Cedar County in the northern part of the study area (fig. 4). In Dakota County, wells D–J1 and D–J2 are characterized as calcium sulfate waters (fig. 4). Both wells are screened beneath sequences of clay, claystone, siltstone, and sand deposits, indicating the Dakota aquifer is at least partially confined in this location and meteoric recharge would be limited.

Sodium was the dominant cation for one sample (LPN–1; fig. 4, table 5A) and a subdominant cation for three samples (LE–3, D–Va3, and D–A1), all from the Dakota aquifer. With the exception of D–Va3, these sodium dominant and subdominant samples display higher specific conductance and concentrations of the total dissolved solids concentration compared to most samples collected in the study area (table 5A). The Dakota aquifer in these locations is confined by interbedded mudstone or shales within the Dakota or by overlying Late Cretaceous-age shales (Conservation and Survey Division, University of Nebraska-Lincoln, 2020a; Katie Cameron, Eastern Nebraska Water Resources Assessment, written commun., 2019), which can limited meteoric recharge (fig. 4).

In some locations, the dominant water type of the Dakota aquifer varied by depth. Domestic wells LE–4 and LE–3 are located in northwest Dodge County (fig. 1) and are approximately 1 mile apart. Well LE–4 is 239 ft deep and screened in the upper part of the Dakota aquifer, and well LE–3 is 526 ft and is completed in the lower part of the Dakota aquifer. Sampling data indicate that groundwater sampled from well LE–4 is calcium bicarbonate water type and likely recharged by meteoric water (fig. 4). Well LE–3 is a calcium sodium sulfate water type and underlies interbedded shale units within the Dakota aquifer (Conservation and Survey Division, University of Nebraska-Lincoln, 2020a; Katie Cameron, Eastern Nebraska Water Resources Assessment, written commun., 2019) where meteoric recharge is likely limited. The total dissolved solids concentration of well LE–3 was more than four times greater than that of well LE–4 (table 5A).

Samples collected from wells completed in the Niobrara aquifer and Paleozoic aquifers were characterized as either calcium bicarbonate or calcium sulfate water types (figs. 3 and 4, table 5A). Of the three samples collected from the Niobrara aquifer, one sample (LC–4S) was characterized as calcium bicarbonate and two samples (LE–5 and LE–6) were characterized as calcium sulfate. The sample characterized as calcium bicarbonate water type (LC–4S) had the lowest concentration of total dissolved solids compared to the other two samples (table 5A). The sample collected at well LE–5 indicated higher concentration of total dissolved solids and lower concentrations of dissolved oxygen when compared to well LE–6. Well LE–5 had a well depth of 157 ft and was screened below a thick sequence of clay; this clay layer may be restricting meteoric recharge to the aquifer at this location. A clay layer was not encountered when drilling well LE–6, which has a well depth of 180 ft and is about 0.75 mile northwest of well LE–5 (Nebraska Department of Natural Resources, 2020).

Of the six samples collected from wells screened in Paleozoic aquifers, all but one sample were characterized as calcium bicarbonate water type (well N–3; fig. 3, table 5A). Well N–3 was characterized as calcium sulfate water type and had much higher total dissolved solids compared to other Paleozoic aquifers wells. The calcium-sulfate-rich water may be explained by the dissolution of gypsum, which was noted in the driller’s log (Nebraska Department of Natural Resources, 2020). Except for well N–5, all Paleozoic aquifers water samples indicated oxic conditions (dissolved oxygen greater than 0.5 mg/L).

The Kruskal-Wallis test was performed on selected constituents to determine if a group of samples collected from each of the three aquifers (Dakota aquifer, Niobrara aquifer, or Paleozoic aquifers) had significantly different distributions from the others. The calculated p-values from the tests are presented in table 6. For this report, all values less than 0.05 are considered significant with 95-percent confidence. Results indicated that the distributions of one of the three aquifer sample groups were significantly different for specific conductance; dissolved oxygen; iron; strontium; uranium; and nutrients, including nitrate plus nitrite as N, ammonia (NH₃+NH₄⁺), and orthophosphate as phosphorus (table 6). Because the Kruskal-Wallis test is unable to discern which of the three groups’ distributions were significantly different, there is some ambiguity in the results; however, some useful insights can be interpreted. Examining the water-quality data presented in tables 5A and 5B, it can be determined that most of the samples collected from the Paleozoic aquifers were calcium bicarbonate water type and typically had high dissolved oxygen concentrations, low specific conductance, and low concentrations of trace elements including iron and uranium.

Samples collected from wells completed in the Dakota aquifer were subdivided into two groups based on water type: calcium bicarbonate and all other water types. The Wilcoxon Rank Sum test was completed to determine if the distributions of selected constituent were statistically significant between the two Dakota aquifer subgroups. The p-values are presented in table 6 and indicate that differences between the distributions for all selected field properties and constituents are significant. Given these results, the samples from the Dakota aquifer were subdivided into two groups when discussed in the remainder of this report: (1) Dakota aquifer (noncalcium bicarbonate water type) and (2) Dakota aquifer (calcium bicarbonate water type).

A Wilcoxon Rank Sum test was completed to determine if the distributions of selected constituents were statistically significant between two samples groups for the Niobrara and Paleozoic aquifers: (1) calcium bicarbonate dominant samples and (2) calcium sulfate dominant samples (table 6). The p-values presented in table 6 indicate that differences between the distributions for all selected field properties and constituents are insignificant, with the exception of specific conductance. Given these results, samples dominated by calcium sulfate had higher specific conductance, and by inference total dissolved solids, compared to samples dominated by calcium bicarbonate.

Sampling results for nutrients, which include nitrate plus nitrite as nitrogen (hereinafter referred to as “nitrate”) and orthophosphate as phosphorous (hereinafter referred to as “orthophosphate”), are presented in table 5B. Of the 31 samples collected from the three bedrock aquifers, 18 wells had concentrations of nitrate that were below the detection limit of 0.040 mg/L. In general, nitrate concentrations for the Dakota aquifer were the lowest of the three aquifer groups and all samples were below the EPA maximum contaminant level of 10 mg/L. One of the three samples collected for the Niobrara aquifer and two of the six samples collected from the Paleozoic aquifers exceeded the 10 mg/L EPA maximum contaminant level for nitrate (table 5).

Nitrate in groundwater is often the primary contaminant of concern for local water managers and users because land use in eastern Nebraska is dominated by row-crop agriculture. For this report, the background nitrate concentration is assumed to be less than 2 mg/L (Mueller and Helsel, 1996). Denitrification may play a role in reducing nitrate concentrations in secondary bedrock aquifers in eastern Nebraska. Denitrification is an important microbially mediated natural attenuation process where anaerobic microorganisms, in the absence of dissolved oxygen, reduce nitrate to generate energy and nitrogen gas (Green and Bekins, 2010). Dissolved oxygen measurements can be used to assess the prevalence of denitrification and to predict the extent to which natural attenuation may help control nitrate concentrations in groundwater.

Dissolved oxygen and nitrate concentrations are plotted against well depth for water samples from both Dakota aquifer subgroups, the Niobrara aquifer, and the Paleozoic aquifers (figs. 5A and 5B, table 5A and 5B). Of the 31 wells sampled, 12 wells had dissolved oxygen concentrations of 0.5 mg/L or greater and 7 sampled wells had nitrate concentrations that exceeded 2.0 mg/L (figs. 5A and 5B). For this report, groundwater with dissolved oxygen concentrations less than 0.5 mg/L is considered anoxic. Samples in table 5 were coded as “M” when the instrument used measured dissolved oxygen concentrations of less than 0 mg/L. The less than 0 mg/L measured values for dissolved oxygen occurred because the 1-point calibration of the sensor was completed at 100-percent saturation (U.S. Geological Survey, 2020a) and, therefore, may not predict the dissolved oxygen concentration as accurately when values approach 0 mg/L. Samples coded as “M” were assumed to be 0 mg/L for statistical analyses and in figure 5A. Generally, groundwater samples collected from wells screened deeper than 220 ft had dissolved oxygen concentration that were less than or approximately equal to 0.5 mg/L, and all wells screened deeper than 270 ft had dissolved oxygen concentrations less than 0.5 mg/L. Of the 10 Dakota aquifer (calcium bicarbonate) wells, 4 have dissolved oxygen concentrations greater than 0.5 mg/L, compared to only 2 of the 12 Dakota aquifer (noncalcium bicarbonate) wells. The highest dissolved oxygen concentration for Dakota aquifer (calcium bicarbonate) wells was 9.1 mg/L and for the Dakota aquifer (noncalcium bicarbonate) wells was 0.6 mg/L (table 5A).

Nitrate concentrations plotted against well depth indicate that concentrations for wells deeper than 200 ft are below the background concentration of 2 mg/L (Mueller and Helsel, 1996) and are often below the detection limit of 0.040 mg/L. Of the 12 Dakota aquifer (noncalcium bicarbonate) wells sampled, only two wells had nitrate concentrations above the detection limit, and all were below the background concentration of 2 mg/L (Mueller and Helsel, 1996; fig. 5B). Three of the 10 Dakota aquifer (calcium bicarbonate) wells had nitrate concentrations above the background concentration of 2 mg/L (fig. 5B, tables 5A and 5B). The only Niobrara aquifer sample with measurable dissolved oxygen concentrations had a nitrate concentration of 13.6 mg/L, which exceeds the EPA maximum contaminant level of 10 mg/L (fig. 5A and 5B, tables 5A and 5B). Sampled wells completed in Paleozoic aquifers, which have well depths that range from 79 to 172 ft, had some of the highest concentrations of nitrate compared to the other aquifer groups, with nitrate in three of the six wells exceeding the background nitrate concentration of 2 mg/L, and nitrate concentrations in two of the six wells exceeding the EPA maximum contaminant level of 10 mg/L (fig. 5B, tables 5A and 5B).

Stable isotope data from samples collected from wells completed in the Dakota aquifer (including both subgroups), Niobrara aquifer, and Paleozoic aquifers tend to cluster together along the LMWL from Mead, Nebr. (fig. 6; Harvey, 2001). Generally, samples collected in the northern part of the study area are isotopically depleted, and samples collected in the southern part of the study area are isotopically enriched (table 5B). The latitudinal effect described in Clark and Fritz (1997) states that for every degree of increase in latitude, δ¹⁸O decreases by approximately 0.6 per mil. The latitude effect explains some, but not all, of the variability in δ¹⁸O values in figure 6.

The seasonality of recharged groundwater likely plays a role because the northern part of the study area receives more snow fall compared to the southern part of the study area. Examining the data from both Dakota aquifer subgroups, the samples from the Dakota aquifer (noncalcium bicarbonate) subgroup are generally more isotopically depleted compared to the samples from the Dakota aquifer (calcium bicarbonate) subgroup (fig. 6). Some of the most isotopically depleted water samples were collected in Lewis and Clark NRD in the northern part of the study area (fig. 1), which is consistent with samples collected by Gosselin and others (2001) in that area. Values of δ¹⁸O are known to be temperature dependent; globally colder regions have more isotopically depleted δ¹⁸O compared to more temperate regions (Clark and Fritz, 1997). The values of the most isotopically depleted samples from the Dakota aquifer (noncalcium bicarbonate) subgroup indicate the climate at the time of recharge was approximately 15 °C cooler than present day (2020) climate. Samples from the Dakota aquifer (calcium bicarbonate) subgroup are isotopically enriched and are similar to modern precipitation (fig. 6). Samples collected from Paleozoic units were the most isotopically enriched of any aquifer group, indicating groundwater is recharged by more summertime precipitation compared to other aquifer groups (table 5B).

Groundwater age tracer results indicated that samples from the Dakota aquifer (noncalcium bicarbonate) subgroup generally had lower values of percent Modern Carbon (fig. 7) and had older radiocarbon ages (table 7, fig. 8) compared to samples from the Dakota aquifer (calcium bicarbonate) subgroup. Of the 11 samples from the Dakota aquifer (noncalcium bicarbonate) subgroup where ¹⁴C samples were collected, 3 were considered modern (post-1950). The oldest radiocarbon ages estimated in this report are from samples of four Dakota aquifer (noncalcium bicarbonate) subgroup wells located in Cedar County (figs. 1 and 8, table 7) and dated the time of recharge to the Pleistocene when continental ice sheets covered much of the study area. The oldest radiocarbon ages coincided with the most depleted stable isotope samples, indicating groundwater from the Dakota aquifer in Cedar County was recharged from Pleistocene-age water, which is consistent with the findings by Gosselin and others (2001). One Dakota aquifer (noncalcium bicarbonate) subgroup well, LPN–1, which is within the Todd Valley (fig. 2), had a radiocarbon age of 30,500 years before present (table 7) with stable isotope values that were very isotopically depleted (table 5B). Groundwater sampled from LPN–1 likely had the same recharge history as wells in Cedar County farther to the north. Other samples from Dakota aquifer (noncalcium bicarbonate) subgroup wells, including D–A1 and LE–3 (fig. 1) with radiocarbon ages of 20,500 and 17,500 years before present (table 7), respectively, and isotopically depleted values (table 5B), also indicate Pleistocene-age recharge. Of the samples from the Dakota aquifer (noncalcium bicarbonate) subgroup, the three ³H samples collected indicated that groundwater is considered premodern and likely to have recharged at least several hundred years ago (figs. 9 and 10, table 7).

Age tracer samples collected from the Dakota aquifer (calcium bicarbonate) subgroup indicate that nearly all samples have higher values for percent Modern Carbon compared to the Dakota aquifer (noncalcium bicarbonate) subgroup samples (table 7, fig. 7). Of the six ¹⁴C samples from the Dakota aquifer (calcium bicarbonate) subgroup, all were considered modern (table 7). Of the five ³H samples from the Dakota aquifer (calcium bicarbonate) subgroup, four were determined to be premodern and likely several hundred years old. The fifth sample is a mixture of modern and premodern waters (table 7). One sample (PM D–1) was considered modern based on ¹⁴C sampling and premodern based on ³H sampling; however, the difference in ³H concentration between the measured ³H concentration and premodern/mixed threshold was within the reported error of the analysis (U.S. Geological Survey, 2020c). Three samples (D–SWSarpy1, D–Sp1, D–Sp2) from the Dakota aquifer (calcium bicarbonate) subgroup had nitrate concentrations above the background concentration of 2 mg/L (table 5B), indicating some post–1952 water in each sample. These three samples were characterized as mixed, modern, and premodern, respectively (table 7).

Two of the three samples (LE–5 and LE–6) collected from wells completed in the Niobrara aquifer are considered modern based on ¹⁴C and (or) ³H sampling. Unlike samples collected from the Dakota aquifer subgroups, these younger samples were characterized as being calcium sulfate water type, which is most likely the result of weathering of pyrite and gypsum in the overlying Pierre Shale (Divine and Sibray, 2017). One Niobrara aquifer sample, LC–4S, had a radiocarbon age of 3,400 years before present (fig. 8, table 7) and is characterized as having a calcium bicarbonate water type. LC–4S is within a monitoring well cluster with a Dakota aquifer (noncalcium bicarbonate) subgroup well (LC–4M), which had a radiocarbon age of 31,000 years before present (fig. 8, table 7). Well LC–4S was more isotopically depleted and had the lowest concentration of total dissolved solids compared to the other two samples from wells completed in the Niobrara aquifer (tables 5A and 5B). It is possible that the sample collected from LC–4S is a mixture of older Dakota aquifer water and recent recharge, but it is difficult to make that interpretation with the data available.

Four of the six samples collected from wells completed in Paleozoic aquifers were characterized as modern based on ¹⁴C results (table 7). These four modern samples had dissolved oxygen concentrations that ranged from 3 to 10.8 mg/L and nitrate concentrations that ranged from 1.89 to 16.8 mg/L (tables 5A and 5B, figs. 5A and 5B), indicating that the Paleozoic aquifers have been affected by agricultural activities in these locations. Stable isotope results for samples from wells completed in Paleozoic aquifers were the most isotopically enriched samples collected for this report and plot close to the LMWL (fig. 6), which indicates that recharge to Paleozoic aquifers is from local precipitation and appears to occur during warmer months. The radiocarbon ages for samples LPS–2 and N–5 were 13,000 and 7,000 years before present, respectively (table 7; fig. 8). Both of these samples had nitrate concentrations that were below the background concentration of 2 mg/L (table 5B); however, their stable isotope values plotted very close to the LMWL and to other samples determined to be modern, indicating that these samples are likely a mixture of modern recharge and older water (table 5B, fig. 6). Three ³H samples collected from the Paleozoic aquifers indicated that two samples (N–3 and LPS–4) were determined to be modern with elevated nitrate concentrations and one sample (LPS–1) was mixed with nitrate concentration below 2 mg/L (tables 5B and 7). Based on the six samples collected from the Paleozoic aquifers, water type did not appear to be diagnostic for determining the age and recharge history of groundwater in these aquifers.

The results of this study indicate that the dominant water type determined from samples collected from the Dakota aquifer are diagnostic for assessing the quality, recharge source, and age of groundwater. Groundwater samples collected from the Dakota aquifer, where calcium bicarbonate was the dominant water type, show very different geochemical characteristics from samples with other water types. Calcium bicarbonate dominant samples from the Dakota aquifer were characterized as modern or mixed using radiocarbon results, indicating that in these areas, groundwater is unconfined and is recharged by precipitation and (or) surface water (table 7, figs. 7 and 8). Stable isotope ratios of calcium bicarbonate dominant samples from the Dakota aquifer were more isotopically enriched compared to other water types and were similar to modern precipitation values (tables 5A and 5B, fig. 6). If groundwater extraction rates exceed recharge rates, total dissolved solid concentrations may increase as a result of upwelling of groundwater from deeper units, which can adversely affect groundwater quality. Concentrations of nitrate for calcium bicarbonate dominant samples from the Dakota aquifer were generally near or below background, but some wells indicated effects from agricultural activities (fig. 5B). It should be noted that seven samples were considered anoxic (dissolved oxygen less than 0.5 mg/L) and could indicate that denitrification is reducing nitrate concentrations in groundwater. Sampling results presented in this report indicate water quality in the Dakota aquifer with a calcium bicarbonate water type is good, but that the groundwater may be vulnerable to surface contamination. Future groundwater sampling plans could focus on understanding how groundwater quality in the Dakota aquifer with a calcium bicarbonate water type changes in the future.

In contrast, groundwater sampled from the Dakota aquifer, having a dominant water type other than calcium bicarbonate, generally has low dissolved oxygen and nitrate concentrations, and higher concentrations of total dissolved solids and trace elements, including iron and strontium. The geochemical characteristics of these samples indicated confining conditions and limited groundwater recharge from local precipitation. Of the 12 wells sampled from the Dakota aquifer where calcium bicarbonate was not the dominant water type, only 2 were characterized as modern. Similar to what Gosselin and others (2001) reported for Cedar County, apparent groundwater age estimated from ¹⁴C sampling dates groundwater recharge to the Dakota aquifer to Pleistocene time. Depleted stable isotope results from these samples indicate recharge occurred during a colder climate. Other samples collected in wells completed in sandstone below layers of shale, mudstone, or clay within the Dakota aquifer (Dakota, Washington, Saunders, and Douglas Counties for this report) show similar geochemical and age characteristics. Groundwater in these areas is not easily recharged from precipitation or surface water. Future groundwater–level monitoring could provide information to evaluate whether groundwater supplies remain sufficient to meet future domestic and irrigation needs. Sampling results from the Dakota aquifer, having a dominant water type other than calcium bicarbonate, indicate that groundwater is not vulnerable to surface contamination; however, groundwater-quality issues include elevated concentrations of total dissolved solids and trace elements.

For the Niobrara aquifer and Paleozoic aquifers, the dominant water type was not a diagnostic indicator of recharge source, age, and groundwater quality as it was with the Dakota aquifer. Most likely this is because the host aquifer was dominated by calcium-carbonate-rich rocks, and prolonged rock-water interaction would result in higher concentrations of these major ions; however, relatively few samples were collected from these aquifers to be able to confirm this interpretation. Samples collected from wells completed in the Niobrara aquifer (2 wells) and Paleozoic aquifers (1 well) and characterized as calcium sulfate water type have statistically significantly higher concentrations of total dissolved solids compared to other samples from the Niobrara aquifer and Paleozoic aquifers characterized as calcium bicarbonate. Given that six of the nine of samples collected from the Niobrara and Paleozoic aquifers indicated modern recharge, these secondary bedrock aquifers are reliant on recharge from precipitation to sustain groundwater levels and may be vulnerable to a multiyear drought. Divine and Sibray (2017) state that yields to wells completed in the Niobrara are dependent on the presence of bedding planes, and Tanner and Steele (1991) noted that well yields in Paleozoic aquifers are dependent on fracturing and secondary porosity. It can be inferred that these units offer little storage and that the Niobrara and Paleozoic aquifers are susceptible to overpumping or a multiyear drought. Samples collected from wells completed in Paleozoic aquifers were the most isotopically enriched, isotopically similar to modern precipitation, and had the highest concentrations of nitrate. Sampling indicated that groundwater in Paleozoic and Niobrara aquifers is affected by agricultural activities and future groundwater sampling would be needed to characterize groundwater–quality changes with time.