Purpose and Scope

This USGS Scientific Investigations Report (SIR) is comparable to other USGS SIRs written for the GAMA-PBP study units and is the second in a series of reports presenting the water-quality data collected in the NSV-SA (Shelton and others, 2020). Reports addressing the status, understanding, and trends of the water-quality assessments done as part of the GAMA-PBP are available from the USGS (https://webapps.usgs.gov/gama/products/products.html) and the SWRCB (https://www.waterboards.ca.gov/water_issues/programs/gama/). This report provides (1) a description of the hydrogeologic setting of the NSV-SA, (2) a status assessment of the groundwater quality in the resources used for domestic wells of the NSV-SA in 2018–19, (3) an understanding assessment to identify natural and anthropogenic factors that could be affecting groundwater quality, and (4) a comparison between the quality of groundwater in the groundwater resources used by domestic wells and the quality of groundwater resources used for public drinking water. Temporal trends in groundwater quality in the domestic and public-supply aquifer systems are not discussed in this report.

The status assessment was designed to provide a statistically representative characterization of groundwater resources used for domestic drinking water at the study-area scale for December 2018 to April 2019. This report describes methods used to design the sampling network for the status assessment and estimate aquifer-scale proportions for constituents (Belitz and others, 2010). Aquifer-scale proportion is defined as the areal proportion of the groundwater resource having water of defined quality (Belitz and others, 2010). Water-quality data from 50 wells sampled by USGS-GAMA for the NSV-SA were used for the status assessment (fig. 2; Shelton and others, 2020). Aquifer-scale proportions for constituents and classes of constituents were computed for the NSV-SA as a whole and separately for the two study areas in the study unit (Redding and Red Bluff) by using a stratified-random sampling design (the USGS grid method) based on a 50-cell, approximately 50-square-kilometer (km²) equal-area grid covering the study unit (fig. 2; Belitz and others, 2010; Shelton and others, 202074).

To provide context, the water-quality data discussed in this report were compared to California and Federal regulatory and non-regulatory benchmarks for treated drinking water delivered by public water systems (Toccalino and others 2014; California State Water Resources Control Board, 2015a; U.S. Environmental Protection Agency, 2016). The State of California does not regulate the quality of drinking water provided by domestic wells. The status of water quality reported herein is for ambient, untreated groundwater resources in domestic wells of the study unit.

The evaluation of natural and human factors affecting groundwater quality in the study unit is based primarily on relations between groundwater quality and potential explanatory variables. The following potential explanatory variables were evaluated: aquifer lithology, land use, hydrologic conditions, depth, elevation, aridity, groundwater age, geochemical conditions, underground storage tank (UST) density, and septic tank density. The relations were examined with statistical tests and graphical analyses. The results of the analyses are discussed in the context of the hydrogeologic and geographic setting of the study unit to support the observed associations. Results also are discussed in relation to other water-quality assessments in the region and differences observed between the studies are addressed (California State Water Resources Control Board, 2009; Central Valley Regional Water Quality Control Board, 2016).

To investigate the differences between groundwater resources used for domestic drinking water and those used for public drinking-water supply, statistical comparisons were made between groundwater-quality results from the NSV-SA (domestic drinking-water sources) samples and results obtained by the GAMA-PBP sampling of the north Sacramento Valley public-supply aquifer study unit (NSV-PA; public drinking-water sources) in 2007 (Bennett and others, 2009, 2011). Direct comparisons between the two study units were made after evaluating and adjusting for subtle differences in the designs of each study unit (discussed in the "Comparison of Domestic and Public-Supply Aquifer Systems" section). Variances in numerous study-unit characteristics, including explanatory variables and the results of water-quality analyses in each study unit, were identified and discussed at study-unit and study-area scales.

Status Assessment

The status assessment was designed to quantitatively summarize groundwater quality of the groundwater resources used for domestic drinking-water wells of the Redding–Red Bluff study unit (NSV-SA) and Redding and Red Bluff study areas. This section describes the methods used for (1) defining groundwater quality, (2) assembling the data used for the assessment, (3) selecting constituents for evaluation, and (4) calculating aquifer-scale proportions.

Understanding Assessment

In the understanding assessment, the physical and chemical processes that could be affecting water quality were identified by examining relations between water quality and potential explanatory variables. The GAMA-PBP uses statistical tests of associations between potential explanatory variables and water-quality conditions to identify potential relations between water quality and various natural and human-induced processes in a study unit. Methods used for the understanding assessment included (1) selecting constituents for additional evaluation and (2) applying statistical tests of relations between potential explanatory variables and selected groundwater-quality constituents.

Comparison Between the Redding–Red Bluff Shallow Aquifer Study Unit and North Sacramento Valley Public-Supply Study Unit

Groundwater quality results were compared between the two study units to identify differences in water quality of groundwater resources used for domestic and public supply. Variations in the explanatory variables between the NSV-SA and NSV-PA were explored using contingency table comparisons to identify potential drivers of water-quality differences between the groundwater resources used for domestic and public drinking water supplies. Multiple nonparametric comparison tests, Kruskal-Wallis rank-sum test, and Dunn’s tests from the R statistical program (RStudio Team, 2021) were used to evaluate relations among categorical and continuous explanatory variables, including well depth, groundwater age and redox class, primary aquifer lithology, and land use or land cover. Geochemical differences between the study units were further evaluated using Piper diagrams (Piper, 1944) of the major ion proportions (R function transform_piper_dater; https://rpubs.com/gavan_mcgrath/PiperTrilinear), which identified the major geochemical processes that influence the chemical composition of the groundwater resources.

Aquifer-scale proportions of groundwater resources with high, moderate, and low RCs and categorical explanatory variables (groundwater age, redox class, aquifer lithology, land use/land cover) for the study units and study areas were tested using contingency tables and Fisher exact tests (2 by 2; R function fisher.test; RStudio Team, 2021). The significance level (p-value) used when testing these differences was based on a threshold value (α) of 10 percent (α=0.1). As already noted, if the test statistic p was less than α, the relation was considered statistically significant. If the test statistic p was greater than α, the relation was not considered statistically significant. A threshold value for significance of 0.1 was used for the study unit comparison to allow discussion of differences that would have been defined as not significant relative to the threshold value of 0.05 used for the rest of this study. The contingency tables were constructed to compare to frequency of sites with high or moderate RCs against those with low RCs and to compare sites with high RCs against those with low or moderate RCs. These tests were run for the constituent classes and for individual constituents.

Using the framework established in the status and understanding assessments of the NSV-PA and NSV-SA, constituents selected for additional evaluation were the focus of the comparison between the study units with respect to groundwater quality. Individual constituent aquifer-scale proportions and proportions of constituent groups by class presented for the NSV-SA were recalculated for the comparison analysis with the NSV-PA. Adjustment was necessary to prevent sites outside of the area of comparison from being included in the comparison of the two study units. Less than half of the wells in the NSV-SA study unit extent overlapped with the NSV-PA extent, but all the NSV-PA wells were within the NSV-SA study unit extent. A more even set of wells for statistical comparison between the two study units was achieved when sites from the NSV-SA within a 3-km distance of the NSV-PA extent were included in the analysis. The distance of 3 km was chosen because this was the buffer distance around public-supply wells used to create the NSV-PA study unit extent (Bennett and others, 2009).

Not all inorganic or organic constituents were sampled in both the NSV-SA and NSV-PA, so some constituents at high or moderate RCs in one study unit were not included in the comparison. Relative concentrations and aquifer-scale proportions presented in Bennett and others (2011) for the NSV-PA were recalculated to ensure that the same water-quality benchmarks were applied for both study units. Benchmark threshold levels for some constituents and the relative hierarchy among benchmarks changed since the NSV-PA analysis in 2007. For example, in the NSV-SA, the boron concentrations were compared to the US-HAL of 6,000 µg/L (Toccalino and others, 2014). In the NSV-PA, concentrations of boron were compared to the CA-NL of 1,000 µg/L (California State Water Resources Control Board, 2016). For a consistent comparison of RCs, benchmark thresholds and the benchmark hierarchy used for the 2018–19 NSV-SA sampling results were used for study-unit comparisons. Lastly, reporting levels for organic and special-interest constituents were re-evaluated if a constituent’s reporting level changed between 2007 and 2019; constituent non-detections were re-censored at the higher of the two reporting levels.

Aquifer Lithology

To examine broad relations between aquifer lithology and groundwater quality, the various geologic units of the NSV-SA, as represented on the state geologic map (Saucedo and others, 2000), were grouped into three general rock types: (1) sediments of the Tehama Formation, (2) volcanoclastic sediments of the Tuscan Formation, and (3) Quaternary alluvium in the floodplains of the Sacramento River (fig. 3; appendix 1; Harkness, 2022).

The lithology of the Redding and Red Bluff study areas is composed primarily of sedimentary deposits of the Tuscan and Tehama Formations or Quaternary alluvial sediments that host the freshwater aquifers, with limited Quaternary pyroclastic volcanic rocks that crop out in the east (fig. 3). Three of the grid sites were in the Tuscan Formation, but all other sites were in the sedimentary deposits of the Sacramento Valley, so the statistical comparison was between the two predominant sedimentary lithologies: marine and continental sediments of the Tehama Formation and the Quaternary alluvium along the Sacramento River (fig. 3). Wells in the Tehama Formation sediments were significantly deeper than wells in the Quaternary alluvium (p=0.004; table 4).

Land Use and Land Cover

Land use and land cover was generalized into three categories: (1) natural (unused), (2) urban, and (3) agricultural (fig. 4; appendix 1; Harkness, 2022). Percentages of the three categories were calculated for the study unit, study areas, and for a circular area, with a radius of 500 m around each USGS grid site (Johnson and Belitz, 2009). As of 2012, land use in the NSV-SA was 67-percent natural, 14-percent urban, and 19-percent agricultural (fig. 5A; Falcone, 2015). Land use in the Redding study area was classified as 64-percent natural, 22-percent urban, and 14-percent agricultural, whereas land use in the Red Bluff study area was 71-percent natural, 5-percent urban, and 24-percent agricultural (fig. 5B).

Like the overall study-unit composition, land use in the buffered areas surrounding the USGS grid sites was primarily classified as natural (fig. 4). On average, in the study unit, urban and agricultural land use in the buffered areas (18 and 27 percent, respectively) was greater than in the overall study unit and natural land use was less (55 percent on average), indicating that more drinking-water wells were located near population centers. Around individual grid sites, land use ranged from 0 to 98-percent urban and 0 to 90-percent agriculture, with 20 of 50 sites (12 in the Red Bluff study area) surrounded by greater than 20-percent agricultural land use and 12 of 50 sites (8 in the Redding study area) surrounded by greater than 20-percent urban land use (fig. 5B). The primary crops grown in Shasta County are alfalfa, hay, rice, walnuts, and wheat, whereas agriculture in Tehama County is dominated by tree nuts (almonds, pecans, pistachios, walnuts), olives, and prunes (California Department of Pesticide Regulation, 2017).

Septic tanks and USTs are markers of urban areas or areas with higher populations, as shown by the strong positive correlation of septic tanks and UST densities in the NSV-SA with urban (ρ=0.68, p<0.0001, and ρ=0.41, p=0.003, respectively) and agricultural land use (ρ=0.31, p<0.027, septic tank density only) and negative correlation with natural land use (ρ= −0.67, p<0.0001 and ρ= −0.50, p<0.001; table 3). The average densities of USTs and septic tanks near a groundwater site can be indicators of potential sources of anthropogenic contaminants near the land surface. The density of USTs around grid sites ranged from 0 to about 2.5 tanks per square kilometer (tanks/km²), and the median density was 0.02 tanks/km². A description of how the Thiessen polygon method (Thiessen, 1911) was used to calculate UST density is included in appendix 1 and Harkness (2022). The density of septic tanks around grid sites ranged from 0 to nearly 72 tanks/km², and the median density was 5.0 tanks/km².

Differences in land use were significant between the sediments of the Tehama Formation that dominates the region and the Quaternary alluvium. The Quaternary alluvium aquifer lithology had a higher average percentage of agricultural land use and a lower average percentage of natural land use and the differences between aquifer lithologies were significant (p=<0.001 and p<0.001, respectively; table 4). This relation indicates that agricultural land uses are preferentially located near the Sacramento River and smaller creeks because low-relief landforms favor such development, and the alluvial soil is typically more suitable for agriculture. This relation is further demonstrated by a higher density of septic tanks in areas with Quaternary alluvium lithology because development is denser along the river and the difference between aquifer lithologies is statistically significant (p=0.004; table 4).

Hydrologic Conditions

Hydrologic conditions were represented by elevation and aridity index at each groundwater site (appendix 1; Harkness, 2022; Shelton and others, 2020). Land-surface elevations referenced to the North American Vertical Datum of 1988 (NAVD 88) in the NSV-SA ranged from about 200 ft at the southernmost extent of the Sacramento River to more than 1,000 ft in the west, with the Redding study area at higher elevation than the Red Bluff study area (Shelton and others, 2020). Sampled groundwater sites were at elevations ranging from 233 to 914 ft above NAVD 88 (Shelton and others, 2020).

The climate in the study unit is typical of the Central Valley in California, with warm, dry summers and cold, wet winters (U.S. Department of Commerce National Climatic Data Center, 2020). The National Climatic Data Center station in Red Bluff, California, reported an average annual temperature of 17.0 degrees Celsius (°C) and an average annual precipitation of 62.2 cm in 2018. In contrast, Redding, California, which is in the north end of the study unit, had an average annual temperature of 16.8 degrees °C and average annual precipitation of 87.9 cm in 2018. This general decrease in precipitation from north to south is due to the rain-shadow effect of the mountain ranges in the study unit and prevailing winter weather patterns (U.S. Department of Commerce National Climatic Data Center, 2020).

Aridity index was used as an indicator of climate. Aridity index is defined as average annual precipitation divided by average annual potential evapotranspiration and is identical to the United Nations Educational, Scientific and Cultural Organization Aridity Index (United Nations Educational, Scientific and Cultural Organization, 1977; United Nations Environment Programme, 1997). Typically, conditions are wetter at higher elevations because of orographic effects, and the aridity index was positively correlated to elevation in the NSV-SA (ρ = 0.61, p-value <0.0001; table 3). Aridity-index values at sampled USGS grid sites ranged from 0.49 to 1.00 (Harkness, 2022). Of the 50 sites sampled in the NSV-SA, 21 sites (42 percent) had an aridity index in the humid or wet category (aridity index>0.65), as defined by the United Nations Environment Programme (1997), with only two sites having aridity-index values less than 0.50 (semiarid; Harkness, 2022).

All 21 of the sites with an aridity index in the humid or wet category were in the Redding study area, whereas the Red Bluff study area is drier, and encompasses the two semiarid groundwater sites. Sites in the sediments of the Tehama Formation had significantly greater aridity-index values and elevations than sites in the Quaternary alluvium aquifer lithology (p=0.002 and p<0.001, respectively; table 4); however, this is an example of correlation between two explanatory variables that is unrelated to direct causation. That is, aridity-index values are directly correlated to elevation (table 3), and the Quaternary alluvium found along the Sacramento River is at a relatively lower elevation. The aridity index also is negatively correlated to agricultural land use because agriculture in the region is concentrated along the low-lying Sacramento River (table 3).

Depth and Groundwater Age Characteristics

Well-depth information was available for all 50 wells sampled (appendix 1; Shelton and others, 2020). When compared, no differences were significant between well depths and depths to top of open or screened interval for USGS grid wells in the Redding study area and the Red Bluff study area (p=0.884; table 4; fig. 6). Depths of USGS grid wells ranged from 64 to 528 ft below land-surface datum; the median well depth was 157 ft (fig. 6; Shelton and others, 2020). Depths to the top of the screened or open interval ranged from 49 to 528 ft, with a median of 148 ft. The screened- or open-interval length ranged from 0 to 57 ft, with a median of 0 ft, which indicates that many (52 percent) of the wells in the study unit are open at the bottom of the well rather than screened over an interval (Shelton and others, 2020).

Groundwater age refers to the length of time that the water has resided in the aquifer system, which is the amount of time elapsed since the water was last in contact with the atmosphere at recharge. Groundwater samples were assigned to groundwater-age classes based on the tritium (³H) and carbon-14 (¹⁴C) activities in the samples (see “Groundwater Age” section in appendix 1 and Harkness [2022]). Samples from 13 sites were classified as modern water (recharged after 1952; Schlosser and others, 1988; Lindsey and others, 2019), samples from 20 sites were classified as mixed (having a component of modern and premodern water), and samples from 17 sites were classified as premodern water (recharged prior to 1952).

Groundwater age typically increases with well depth and with depth to the top of the open or screened interval. Modern groundwater came from wells with significantly shallower depths and depths to the top of the open or screened interval than premodern groundwater (p<0.001 and p<0.001, respectively; fig. 7; table 4). Wells were deeper and groundwater ages older in the sediments of the Tehama Formation, and these wells had a greater proportion of mixed and premodern groundwater compared to the shallower wells with younger aged groundwater completed in the Quaternary alluvium along the river (Kruskal-Wallace test p=0.004 and contingency table test p=0.004; tables 4 and 5). Agricultural land use and septic-tank density were higher for wells with modern groundwater due to higher development along the Sacramento River where the wells are shallower and the differences between modern and premodern groundwater were statistically significant (p=0.013 and p=0.003, respectively; table 4). Percent natural land use and elevation were higher for wells with premodern groundwater in the more rural northern and western parts of the study unit and the differences between premodern and modern groundwater were statistically significant (p=0.003 and p=0.023, respectively; table 4).

Inorganic Constituents

Most inorganic constituents are naturally present in groundwater, although their concentrations can be influenced by human activities (Hem, 1985). Of the 50 inorganic constituents analyzed by the USGS-GAMA, 46 were detected in groundwater in the NSV-SA (Shelton and others, 2020). Of these 46 constituents, 26 had regulatory or non-regulatory health-based benchmarks, 8 had non-regulatory, aesthetic-based benchmarks, and 12 did not have established benchmarks (table 2). The constituents without benchmarks are major or minor ions that are present in nearly all groundwater (table 2).

Of the 26 inorganic constituents that had regulatory and non-regulatory health-based benchmarks, 8 were detected at moderate or high RCs in the grid sites. These constituents included the trace elements arsenic, barium, boron, chromium, and strontium; the nutrient nitrate; the radionuclide gross alpha (a measure of radioactivity in water); and the special-interest constituent hexavalent chromium (table 2; figs. 8, 9A–C). The four inorganic constituents with aesthetic-based secondary maximum contaminant level (SMCL) benchmarks detected at moderate or high RCs in the grid sites were chloride, iron, manganese, and total dissolved solids (TDS; table 2; figs. 8, 9D). Previous investigations of domestic well-water quality in Tehama County (California State Water Resources Control Board, 2009) and the Shasta-Tehama sub-watershed (Central Valley Regional Water Quality Control Board, 2016) also observed moderate and high concentrations of nitrate, TDS, arsenic, boron, chromium, iron, and manganese. Five constituents—arsenic, chromium, hexavalent chromium, manganese, and nitrate—were selected for further evaluation in the understanding assessment because they were present at high or moderate RCs in greater than 2 percent of the groundwater resources used by domestic drinking-water wells (table 6).

Inorganic constituents with human-health benchmarks (nutrients, radionuclides, and trace elements) had high RCs in 6 percent of the groundwater resources used by domestic wells in the NSV-SA and moderate RCs in 18 percent of the NSV-SA (table 7). Inorganic constituents having aesthetic-based benchmarks had high and moderate RCs in 4 and 2 percent, respectively, of the groundwater resources used by domestic wells in the NS-SA (table 7).

Groundwater can be grouped in a few distinct water types based on the relative proportions of major ions measured in the samples. Samples with the same water types have likely undergone similar processes that affected the major ion profile and the presence of constituents that are concerning for human health. For example, in the NSV-SA, two major water types are observed. In modern and mixed-aged groundwater, primarily in the Red Bluff study area, calcium and magnesium are the dominant cations and sulfate and chloride are the dominant anions (fig. 10). This type of water is typical of younger groundwater that has interacted with marine rocks during recharge. Younger, typically fresh water (less than 1000 mg/L dissolved solids concentration) with a calcium-magnesium water type are recently recharged and can have higher concentrations of anthropogenic compounds, including nutrients, perchlorate, and organics that are present at the surface and brought into the aquifer during recharge.

The second water type is a sodium-bicarbonate water that is predominant in older, premodern groundwater that has undergone extensive ion exchange due to interactions with aquifer materials (fig. 10). The increased contact time allows for exchange of calcium for sodium on the surface of the aquifer matrix that results in sodium-dominated, premodern groundwater (fig. 10). The longer interactions with aquifer materials could allow for the accumulation of naturally occurring constituents, and higher salinities and concentrations of trace elements are found in sodium-bicarbonate water types.

A third water type was observed in a single sample, RB-01, in the southwesternmost part of the Red Bluff study area (fig. 2). The groundwater in this well has a sodium-chloride water type that is found in groundwater derived from seawater. This water type has been found in old, deep saline groundwater throughout California and can migrate into shallower, fresh-water aquifers used for drinking water though natural pathways such as faults and fractures in the subsurface (Kang and Jackson, 2016).

Trace Elements

Trace elements with health-based benchmarks had high and moderate RCs (for one or more constituents) in 6 and 14 percent, respectively, of the groundwater resources used by domestic wells (table 7). The proportion of the groundwater resources used by domestic wells in the Red Bluff study area with high RCs of trace elements (12 percent) was greater than the corresponding proportion of the Redding study area (0 percent). Barium, hexavalent chromium, and strontium were present at high RCs in samples from the Red Bluff study area (fig. 9A). Moderate RCs of arsenic, boron, chromium, and hexavalent chromium were reported in the Red Bluff study area, and moderate RCs of chromium and hexavalent chromium were reported in the Redding study area (table 6; fig. 9A).

Arsenic

Arsenic is a semi-metallic element that has been associated with increased cancer risk and non-cancerous effects, including skin damage and circulatory problems (U.S. Environmental Protection Agency, 2001). Natural sources of arsenic in groundwater include dissolution of arsenic-bearing minerals, desorption of arsenic from mineral surfaces, and mixing with hydrothermal fluids (Welch and others, 2000; Smedley and Kinniburgh, 2002). Potential anthropogenic sources of arsenic can include copper ore smelting, coal combustion, arsenical pesticides, wood preservatives, and dissolution of natural arsenic-bearing minerals from mining activities (Smedley and Kinniburgh, 2002; Welch and Stollenwerk, 2003). In an aqueous environment, arsenic occurs as an oxyanion in either the oxidized form, arsenate (As(VI)), or the reduced form, arsenite (As(III)). The reduced form, As(III), is more mobile under ambient conditions and potentially more toxic. The redox conditions and pH determine the dominance aqueous species in each well (Smedley and Kinniburgh, 2002).

The US-MCL was lowered from 50 to 10 µg/L in 2002 because of studies linking chronic exposure to arsenic concentrations between 10 µg/L and 50 µg/L in drinking water to increased incidence of cancer and other health risks (U.S. Environmental Protection Agency, 2001). An estimated 5 percent of groundwater resources used for domestic drinking water in the United States have high RCs of arsenic (greater than 10 µg/L; Ayotte and others, 2017, Focazio and others, 2000; Welch and others, 2000), and high concentrations of arsenic in groundwater resources used for drinking water are a worldwide concern (Smedley and Kinniburgh, 2002; Welch and others, 2006).

Arsenic was present at moderate RCs in 8 percent (table 6) of the groundwater resources used by domestic wells in the Red Bluff study area but was not found in the groundwater resources used by domestic drinking water wells in the Redding study area (table 6; fig. 9). As part of the SWRCB GAMA domestic well project, 223 wells were sampled in 2005 for water quality in Tehama County, which overlaps with parts of the Redding and Red Bluff study areas (California State Water Resources Control Board, 2009). During the SWRCB Tehama County study, private wells were sampled on a first-come, first-serve basis and the results from the program were collated at the county level, rather than at the basin level. The resulting well distribution in the SWRCB study is spatially biased and includes higher well densities near population centers. Of the wells sampled for the SWRCB Tehama County study, 157 wells were located within the current (2022) Red Bluff study area and 17 percent of those wells contained high arsenic concentrations (California State Water Resources Control Board, 2009). When reassessing the water-quality data from the SWRCB study using the areal proportion calculations from the GAMA-PBP domestic well assessment, 7 percent of the groundwater resources in the Red Bluff study area had high arsenic concentrations (Belitz and others, 2003; Belitz and others, 2010; Belitz and others, 2015). The differences between the well and area proportions reflect the clustering of wells sampled by the SWRCB study. The two wells with moderate arsenic values measured in the GAMA-PBP domestic well study (RB-16 and RB-22; fig. 2) were in areas where the SWRCB sampled 59–77 wells within a 5-km radius and those clusters included wells with high, moderate, and low arsenic—and were the only places the SWRCB sampling found high arsenic. The comparison of the two studies highlights that the GAMA-PBP grid-based method does not capture water-quality variations on a more local scale.

Previous studies of elevated arsenic in groundwater have identified two primary mechanisms for arsenic mobilization related to the observed conditions at sampled sites: (1) desorption from, or inhibition of sorption to, aquifer materials at elevated pH levels and (2) release of arsenic from dissolution of iron or manganese oxyhydroxides under iron- or manganese-reducing conditions (Smedley and Kinniburgh, 2002; Welch and others, 2006; Bennett and others, 20115). The first mechanism described requires pH values to be greater than 7.8, the level above which the primary arsenate species is relatively soluble (Smedley and Kinniburgh, 2002). The second mechanism requires low oxygen, anoxic conditions in which arsenic is released from oxyhydroxides as arsenite (As(III)). Moreover, arsenite release is enhanced by long contact and reaction times between groundwater and aquifer material (Smedley and Kinniburgh, 2002).

Elevated arsenic concentrations in anoxic groundwater from fine-grained sediment lithologies along the Sacramento River were observed in previous studies of the Sacramento Valley (Hull, 1984; Dawson, 2001; Bennett and others, 2011). Although the wells in the alluvium and volcanic lithologies along the river generally are shallower and contain younger groundwater, DO concentrations are relatively low indicating anoxic conditions that typically favor mobilization of arsenic in groundwater. However, low DO values do not appear to be the cause of arsenic along the river as arsenic was not correlated to DO values (table 8), and differences in arsenic concentrations between oxic and anoxic groundwaters were not significant. Additionally, arsenic was not correlated to iron or manganese, which would indicate that reductive dissolution of iron and manganese-oxyhydroxides is not the primary source of arsenic in the domestic wells of the NSV-SA (table 9).

Table 8.    

Summary of Spearman’s rho (ρ) rank correlation test results for selected potential explanatory variables and water-quality constituents, Redding–Red Bluff shallow aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, 2018–19.

[Tabled values of rho are shown for tests in which the variables were determined to be statistically significant (p-value <0.05). Abbreviations: ns, statistical test indicates no significant correlation between variables; positive values, significant positive correlation; negative values, significant negative correlation; USTs, underground storage tanks]

Table 8.    Summary of Spearman’s rho (ρ) rank correlation test results for selected potential explanatory variables and water-quality constituents, Redding–Red Bluff shallow aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, 2018–19.

Water-quality constituents	Land use			Distance to nearest geothermal site	Density of septic tanks	Density of USTs	Aridity index	Elevation	Well depth	Depth to top of open interval	pH	Dissolved oxygen concentration	Tritium	Carbon-14
	Agricultural	Natural	Urban
Nutrient and trace elements with health-based benchmarks
Arsenic	ns	ns	ns	−0.30	ns	ns	ns	ns	ns	ns	0.36	ns	−0.31	−0.43
Barium	ns	ns	ns	ns	−0.29	ns	−0.33	ns	0.45	0.48	0.46	ns	ns	−0.39
Boron	ns	ns	ns	ns	ns	ns	−0.33	−0.37	ns	ns	ns	−0.51	ns	ns
Chromium	−0.32	0.48	−0.35	ns	−0.39	−0.35	ns	ns	0.54	0.52	0.54	0.29	−0.45	−0.41
Hexavalent Chromium	−0.35	0.48	−0.33	ns	−0.31	−0.30	ns	ns	0.50	0.47	0.48	0.32	−0.40	−0.36
Nitrate	0.56	−0.58	0.31	ns	0.44	0.40	−0.38	−0.62	−0.35	ns	−0.32	ns	0.57	0.50
Strontium	ns	ns	ns	ns	−0.48	−0.52	−0.48	ns	0.46	0.50	0.53	ns	ns	−0.38
Inorganic constituents with aesthetic-based benchmarks
Chloride	ns	ns	ns	ns	ns	ns	−0.38	ns	ns	ns	ns	−0.29	ns	ns
Iron	ns	ns	ns	−0.30	ns	ns	ns	ns	ns	ns	ns	−0.28	ns	ns
Manganese	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	−0.41	ns	ns
Total dissolved solids (TDS)	ns	ns	ns	ns	−0.34	ns	−0.52	ns	ns	ns	ns	ns	ns	ns
Radionuclides
Adjusted gross-alpha particle radioactivity	ns	ns	ns	ns	−0.36	0.30	−0.37	ns	ns	ns	ns	ns	ns	ns
Organic and special-interest constituents
Chloroform	ns	−0.34	0.35	ns	ns	0.30	ns	ns	ns	ns	ns	ns	ns	ns
Perchlorate	ns	ns	ns	ns	ns	ns	ns	ns	0.33	0.32	ns	ns	ns	ns

Table 9.    

Summary of Spearman’s rho (ρ) rank correlation test results for selected water-quality constituents, Redding–Red Bluff shallow aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, 2018–19.

[Tabled values of rho are shown for tests in which the variables were determined to be statistically significant (p-value <0.05). Abbreviations: —, not applicable; ns, Spearman’s test indicates no significant correlation between variables; positive values, significant positive correlation; negative values, significant negative correlation]

Table 9.    Summary of Spearman’s rho (ρ) rank correlation test results for selected water-quality constituents, Redding–Red Bluff shallow aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, 2018–19.

ρ	Arsenic	Barium	Boron	Chloride	Chromium	Hexavalent chromium	Iron	Manganese	Nitrate	Strontium	Total dissolved solids (TDS)	Perchlorate	Adjusted gross-alpha radioactivity
Arsenic	—	ns	ns	ns	0.35	0.40	ns	ns	ns	ns	ns	ns	ns
Barium		—	ns	0.29	ns	ns	ns	ns	ns	0.68	0.35	ns	ns
Boron			—	0.61	ns	ns	ns	ns	ns	0.32	0.42	ns	ns
Chloride				—	ns	ns	ns	ns	ns	0.35	0.54	ns	ns
Chromium					—	0.97	ns	−0.49	ns	0.39	ns	0.32	ns
Hexavalent chromium						—	ns	−0.49	ns	0.32	ns	0.33	ns
Iron							—	0.56	−0.32	ns	ns	ns	−0.48
Manganese								—	ns	ns	ns	ns	ns
Nitrate									—	ns	ns	0.32	ns
Strontium										—	0.60	0.36	ns
Total dissolved solids (TDS)											—	ns	ns
Perchlorate												—	−0.35
Adjusted gross-alpha radioactivity													—

Arsenic concentrations had a positive correlation with pH (ρ=0.36, p=0.01) and a negative correlation to percent modern carbon (pmC; ρ= −0.43, p<0.0001) and tritium (ρ= −0.31, p<0.0001; table 8). Higher arsenic concentrations in older groundwater with higher pH are consistent with arsenic accumulation due to increased aquifer contact times and groundwater evolution. Older, premodern groundwater in the NSV-SA is associated with a sodium-bicarbonate water type that is indicative of exchange of ions over long residence times (fig. 10). In addition to allowing increased time for groundwater to interact with arsenic-rich materials, dissolved oxygen is consumed, and pH increases with groundwater age, creating conditions that promote arsenic mobilization.

One of the two sites with a moderate arsenic RC value, S9-RB-16, had a pH value greater than 7.8, was premodern, anoxic, and from the volcanic sediments of the Tuscan Formation in the eastern part of the study area (fig. 11A). The second site, S9-RB-22, was slightly acidic (pH=6.6) and had modern and oxic groundwater from the Quaternary alluvium. Groundwater samples with moderate RC values for arsenic were close to the Sacramento River. The presence of arsenic in alkaline, anoxic water and more acidic, oxic water indicates that more than one mechanism could be contributing to mobilization of arsenic in wells in the Sacramento River floodplain.

Hydrothermal waters can be a potential source of arsenic in some areas of the NSV-SA. Mineral solubility tends to increase with temperature, and thermal waters are often characterized as having elevated concentrations of trace elements like arsenic, boron, and fluoride (Hem, 1985). In the NSV-SA, arsenic concentrations were negatively correlated with the distance to the nearest geothermal site and the correlation was statistically significant, indicating that thermal water could be contributing arsenic to groundwater (table 8; fig. 11A).

Chromium

Chromium is an oxyanion-forming trace metal that exists in the aquatic environment in two forms: (1) as the reduced species trivalent chromium, and (2) as the oxidized form hexavalent chromium. Hexavalent chromium is more stable under oxic groundwater conditions and has higher solubility and mobility in natural waters than trivalent chromium (California State Water Resources Control Board, 2017). Natural sources of chromium in groundwater include dissolution of chromium bearing minerals associated particularly with ultramafic rocks, such as serpentinites found in ophiolite complexes (Oze and others, 2004, 2007; Izbicki and others, 2015; Hausladen and others, 2018). Hexavalent chromium is a common industrial contaminant, and the anti-corrosive properties of chromium make it a common chemical for use in metal plating, stainless-steel manufacturing, leather tanning, anticorrosion of industrial cooling waters, and wood preservative treatment.

The US-MCL for total chromium is 100 µg/L, and the CA-MCL for total chromium is 50 µg/L. Trivalent chromium is an essential nutrient and is considered to have very low health risks; however, hexavalent chromium is considered a health risk and carcinogen and can cause nausea, gastrointestinal distress, stomach ulcers, skin ulcers, allergic reactions, kidney and liver damage, reproductive problems, and lung and nasal cancer (U.S. Environmental Protection Agency, 2001; California State Water Resources Control Board, 2017). Hexavalent chromium detected in groundwater was widely presumed to be associated primarily with anthropogenic contamination, until more recent reports of the prevalence of naturally occurring hexavalent chromium (Hem, 1985, Izbicki and others, 2015; Hausladen and others, 2018). Hexavalent chromium is highly soluble and mobile in alkaline, oxic groundwater, and is often the dominant chromium species in more natural waters (Oze and others, 2004, 2007; Izbicki and others, 2015 Hausladen and others, 2018). A national MCL for hexavalent chromium has not been developed, but the USGS has an HBSL of 20 µg/L (Toccalino and others, 2014). California established an MCL of 10 µg/L for hexavalent chromium in 2014, but it was withdrawn in 2017 (California State Water Resources Control Board, 2017).

The Coast Ranges and Sierra Nevada contain known chromium-rich ultramafic units (Morrison and others, 2009, Manning and others, 2015, Izbicki and others, 2015 Hausladen and others, 2018). Samples of soil derived from the ultramafic rocks in the region contain up to 10,000 milligrams per kilogram (mg/kg) chromium, and sedimentary soils in the Sacramento Valley can contain up to 1,400 mg/kg chromium, particularly from the west side of the valley where there is a substantial contribution of chromium from ultramafic rocks in the Coast Ranges (Morrison and others, 2009). Valley sediments on the east side are primarily from the lower Cascade Range, which does not contain substantial ultramafic units compared to the Coast Ranges and parts of the northern Sierra Nevada (Manning and others, 2015).

Total chromium was present at moderate RCs in 4 percent of the groundwater resources used by domestic wells in the NSV-SA (table 6; fig. 9), and hexavalent chromium was present at high RCs in 4 percent and moderate RCs in 8 percent of the groundwater resources used by domestic wells (table 6; fig. 9). High RCs for hexavalent chromium were only found in the Red Bluff study area, but moderate RCs for total chromium and hexavalent chromium were found in 8 percent of groundwater resources in both study areas (table 6; fig. 9). Elevated total chromium and hexavalent chromium concentrations were measured in samples from wells in the valley west of the Sacramento River, where the chromium-rich mafic sediments from the Coast Ranges are incorporated into the sediments of the Tehama Formation (fig. 11B). Total chromium and hexavalent chromium were highly correlated (ρ=0.97, p<0.05; table 9) and hexavalent chromium is the dominant species (>75 percent) in most samples. About 80 percent of samples contained more than 90 percent of the total chromium as hexavalent chromium, which is consistent with previous studies that indicate hexavalent chromium is the dominant species of chromium in oxic groundwater (Oze and others, 2004, 2007; Izbicki and others, 2015 Hausladen and others, 2018). Hexavalent chromium was not reported in groundwater in the northern Sacramento Valley in previous domestic well studies, but total chromium was measured at elevated concentrations above the EPA-MCL in Tehama County (California State Water Resources Control Board, 2009). Similar to the GAMA-PBP Red Bluff study area, the moderate and high total chromium RC values reported for Tehama County occurred only in wells west of the Sacramento River and Red Bluff (California State Water Resources Control Board, 2009).

Total and hexavalent chromium concentrations were positively correlated to natural land-use percentage (table 8), likely due to chromium-rich minerals in the sediments of the Tehama Formation in the western side of the valley (Morrison and others, 2009; Manning and others, 2015; fig. 11B) where there is less urban and agricultural land use and wells have higher percentage of natural land cover (fig. 11B). Total chromium and hexavalent chromium concentrations were positively correlated to pH, DO, and well depth and negatively correlated to tritium and pmC (table 8). Total and hexavalent chromium were higher in oxic, premodern groundwater (p=0.007 and p=0.021, respectively; table 4; fig. 12). Studies in the Sacramento Valley and other parts of California have indicated that long residence times of oxygenated, alkaline (pH>8) groundwater promote mobilization of hexavalent chromium in groundwater (Izbicki and others, 2015; Manning and others, 2015). As groundwater age increases, weathering of primary silicate minerals increases groundwater pH, and hexavalent chromium desorption will occur in the more alkaline environment under persistent oxic conditions (Izbicki and others, 2015). This mechanism is consistent with the relations between pH (ρ=0.48, p<0.001), DO (ρ=0.32, p=0.024), tritium (ρ= −0.40, p=0.004), carbon-14 (ρ= −0.36, p=0.01) and hexavalent chromium concentrations observed in the NSV-SA (table 8).

In previous studies of hexavalent chromium occurrence in alluvial aquifers along the western side of the Central Valley, researchers hypothesized that irrigation return flow through a thick undersaturated zone can mobilize chromium in alluvial aquifers (Mills and others, 2011; Izbicki and others, 2015). Correlations between nitrate and hexavalent chromium concentrations in groundwater wells from the Sacramento Valley were used to support this hypothesis. However, in groundwater wells in the NSV-SA study unit, we did not observe a significant relation between total or hexavalent chromium and nitrate as observed in the other studies. Mills and others (2011) and Izbicki and others (2015) looked at the relation between chromium and nitrate in groundwater wells in the southwestern Sacramento Valley and in public-supply wells across California, respectively. In the NSV-SA, we sampled shallow, domestic wells only in the northern part of the Sacramento Valley, where hexavalent chromium concentrations were correlated with high natural land use and low agricultural land use areas. The results of the understanding assessment in the NSV-SA for hexavalent chromium supports long aquifer contact times with chromium-rich sediments as the primary mechanism driving the occurrence of high RCs in this part of the Sacramento Valley rather than irrigation return flow mobilizing chromium that has been observed in other parts of the Central Valley.

Constituents with Secondary Maximum Contaminant Level Benchmarks

Constituents with aesthetic-based benchmarks that provide secondary maximum contaminant levels (SMCLs) had high RCs (for one or more inorganic constituents) in 4 percent of the groundwater resources used by domestic wells (table 7) and moderate RCs in 2 percent. Chloride, iron, manganese, and TDS were the constituents with high RCs in the study unit (table 6). Saline-water indicators such as chloride, sulfate, and TDS have recommended and upper CA-SMCL benchmarks. In this report, the upper CA-SMCL values are used for constituents with aesthetic-based benchmarks.

Manganese

Manganese is naturally present in most rock types and can be abundant in mafic and ultramafic igneous and metamorphic rocks (Homoncik and others, 2010). As a result of weathering processes, manganese can accumulate in aquifer sediments and is naturally present in most groundwater because of water-aquifer interactions (Homoncik and others, 2010). Manganese is redox sensitive and is present most commonly as hexavalent manganese in oxic conditions and divalent manganese in reducing conditions. The relative solubility of manganese is strongly influenced by oxidation-reduction (redox) conditions in the aquifer, with reduced divalent manganese more mobile in near-neutral pH groundwater (Groschen and others, 2008; Homoncik and others, 2010; DeSimone and others, 2014). In oxic conditions, hexavalent manganese and iron oxides precipitate and coat aquifer materials. The primary mechanism of manganese release to groundwater is the reductive dissolution of iron and manganese oxyhydroxides in anoxic conditions (Chapelle and others, 1995; McMahon and Chapelle, 2008; McMahon and others, 2019).

The comparison benchmark for manganese used in this study was the CA-SMCL of 50 µg/L (table 2). The HBSL for manganese is equal to the EPA lifetime health advisory level (US-HAL) of 300 µg/L and the CA-NL is 500 µg/L. The CA-SMCL is established at the concentration at which manganese can affect the aesthetic properties of water but has negligible adverse health effects. Chronic exposure of manganese at elevated concentrations has been associated with neurological effects in children, among other more minor health effects (U.S. Environmental Protection Agency, 2004; Wasserman and others, 2006).

Manganese was detected at high RCs in 4 percent of the groundwater resources used by domestic wells and at moderate RCs in 2 percent (table 6; fig. 11D). Manganese concentrations were detected at high RCs in two groundwater samples, one from each study area (figs. 9D, 11D). High manganese concentrations in the Sacramento Valley have been reported in previous studies, and manganese was detected at high concentrations in 8 percent of domestic wells sampled in the domestic well study in Tehama County, which is comparable to the GAMA-PBP study, considering the difference in sample methods (California State Water Resources Control Board, 2009). Manganese concentrations were significantly negatively correlated to DO concentrations (table 8); moreover, correlations were statistically significant for higher manganese concentrations in groundwater classified as anoxic than in groundwater classified as oxic (p<0.001; table 4; fig. 13). Anoxic conditions in groundwater persist when DO concentrations are less than 1 mg/L, per the criteria established in McMahon and Chapelle (2008; appendix 1; Harkness, 2022).

The negative correlation with DO concentration is consistent with the redox processes controlling manganese solubility (fig. 13; Groschen and others, 2008; Homoncik and others, 2010; DeSimone and others, 2014). Manganese and iron concentrations are positively correlated, further supporting the release of manganese during reductive dissolution of manganese and iron oxides in the groundwater resources used by domestic wells in the study area (table 9). Arsenic also can be released to groundwater by the dissolution of iron and manganese oxides, but manganese is not correlated to arsenic (table 9). However, manganese is negatively correlated to chromium and hexavalent chromium. Manganese oxides have been shown to effectively oxidize chromium, resulting in the transformation of insoluble trivalent chromium to soluble hexavalent chromium (Eary and Rai, 1987; Oze and others, 2007; Hausladen and others, 2018). Manganese oxides from mafic-derived sediments could be oxidizing chromium in the oxic groundwater west of the Sacramento River; however, reducing conditions in the alluvial sediments along the river would mobilize manganese to the groundwater and reduce hexavalent chromium.

Saline-water Indicators

The high RCs of chloride, iron, and TDS, as well as the high and moderate RCs of the trace elements barium, boron, and strontium, were all measured in one sample collected from a grid site in the southwestern corner of the study unit (fig. 11E). The chemistry in this sample is consistent with mixing of saline water (TDS>1000 mg/L) from deeper formations into the shallow freshwater aquifer (figs. 10, 11E; Hull 1984; Kang and Jackson, 2016). This sample is from well RB-01 (fig. 2), which is located near the western fault systems that intersect in the southwestern most part of the study area (fig. 11E; Harwood and Helley, 1987); these faults could be conduits for deeper saline water to migrate upward into the freshwater aquifers the area. Although there is evidence this process influences the water quality in the study unit, it only affects a small (<2 percent) proportion of the groundwater resources used for the domestic well supply (table 7).

Organic and Special-Interest Constituents

The organic constituent classes assessed in this study were volatile organic compounds (VOCs) and pesticides (including degradates). Volatile organic compounds can be naturally present in groundwater associated with hydrocarbon (natural gas and oil) deposits, but their presence in groundwater in most areas outside oil and gas fields is related to anthropogenic sources (Gilliom and others, 2006; Zogorski and others, 2006; Rowe and others, 2007). Volatile organic compounds are characterized by their tendency to evaporate, and they generally persist longer in groundwater than in surface water because groundwater is more isolated from the atmosphere (Zogorski and others, 2006; Rowe and others, 2007). Volatile organic compounds can be present in paints, solvents, fuels, refrigerants, and fumigants or can be formed as byproducts of water disinfection (Zogorski and others, 2006; Rowe and others, 2007). Pesticides are used to control weeds, fungi, or insects in agricultural and urban settings and are entirely derived from anthropogenic applications at the land surface.

One or more organic constituents were detected in 8 of the 50 grid sites (16 percent) sampled in the study unit (Shelton and others, 2020). Of the 167 organic constituents analyzed, 8 were detected in at least 1 site in the NSV-SA (table 2). Of these eight constituents, six have regulatory health-based benchmarks (table 2). Organic constituents were not present at high or moderate RCs in groundwater resources used by domestic wells in the NSV-SA (table 6). The most common organic compounds detected were VOCs, primarily disinfection-byproducts, and the only organic constituent detected in greater than 10 percent of the grid-site samples was chloroform (table 2; fig. 14). Other VOCs detected include trichloroethylene (TCE), chlorodifluoromethane, bromodichloromethane, and methyl tert-butyl ether (MTBE). The detected pesticides were triazine herbicides atrazine and simazine and their degradation byproduct 2-chloro-4-isopropylamino-6-amino-s-triazine (table 2; fig. 14). Finding low levels of organic compounds, including VOCs and pesticides, in groundwater is consistent with other studies in the Redding–Red Bluff region that found no detections or only low concentrations of organic compounds in domestic wells (California State Water Resources Control Board, 2009; Central Valley Regional Water Quality Control Board, 2016). During the study in Tehama County, a different set of VOCs were detected, including acetone, the fumigant 1,3-dichloropropane, and the solvents 1,1,2-trichloroethane and 1,1,3-trichloro-1,2,2-trifluoroethane, but all detections that were at low concentrations were in less than 1 percent of sampled wells (California State Water Resources Control Board, 2009).

The trihalomethane (THM) chloroform was the most commonly detected VOC in the NSV-SA, with a detection frequency of 12 percent (fig. 14), and detections were all at low RCs (fig. 15A). Chloroform is among the most detected VOCs in groundwater nationally (Zogorski and others, 2006). Chloroform is formed during wastewater treatment when chlorine is used for disinfection but can also form in groundwater when chlorine interacts with natural organic matter (Zogorski and others, 2006). Water used for household consumption from domestic wells is commonly disinfected with solutions that contain chlorine, called shock chlorination. Shock chlorination is a recommended procedure for the treatment of bacterial contamination and odor problems in domestic wells (Centers for Disease Control and Prevention and U.S. Department of Housing and Urban Development, 2006). In addition to disinfecting the water, the chlorine can react with organic matter to produce THMs and other chlorinated or brominated disinfection byproducts that can persist in the well.

Chloroform concentrations were higher in samples from wells with anoxic conditions and wells in the Quaternary alluvium and the differences between redox class and aquifer lithology were statistically significant (table 4). Concentrations of chloroform in the NSV-SA also were positively correlated to percent urban land use and UST density and negatively correlated to percent natural land use (table 8). The correlation between chloroform concentrations and urban land use or UST density could reflect a higher frequency of chlorination of domestic wells in densely populated areas. Domestic wells are more susceptible to contamination in urban areas and, therefore, more likely to be treated with chlorine than domestic wells in less populated areas. Municipal or public-supply wells also could leak disinfected water containing chloroform and other disinfection by-products to groundwater resources in the NSV-SA study unit (Ivahnenko and Zogorski, 2006). The higher chloroform concentrations in the alluvial aquifer are related to higher proportions of urban centers and development along the Sacramento River.

Microbial Indicators

Microbes are ubiquitous and naturally occurring in the environment. Most of the microbes found in groundwater are harmless to humans; however, some microbes found in human and animal waste can be harmful when consumed in drinking water. The concentrations of these microbes typically are higher in surface waters, and microbes do not travel very far from the source through aquifer material. However, microbes may be introduced to wells through a “short-circuit pathway” such as improperly sealed casing or compromised well construction. The microbial indicators reported in this assessment were used to evaluate the potential for fecal contamination of water sources. Total coliforms, including Escherichia coli (E. coli), and Enterococcus spp. occur naturally in the digestive tracts of animals. In the NSV-SA, total coliforms were present in 8 percent of the groundwater resources used for domestic wells, and Enterococcus spp. were present in 4 percent of the groundwater resources used for domestic wells with total coliforms present (table 7). Escherichia coli bacteria are enumerated on samples that tested positive for total coliform bacteria but were not detected in any samples.

Enterococcus spp. were detected in 4 percent of the groundwater resources used for domestic wells in the Red Bluff and Redding study areas (table 7). Total coliforms were detected only in the Red Bluff study area, in about 16 percent of the groundwater resources used for domestic wells. Microbial indicators also were present in domestic wells sampled by the SWRCB in Tehama County, which includes parts of the Redding and Red Bluff study areas (California State Water Resources Control Board, 2009). Total coliforms were detected in 25 percent of domestic wells, and fecal coliforms were observed in 1 percent of wells in the county, mostly in domestic wells around Red Bluff, in the same locations where total coliforms were detected in the GAMA-PBP Red Bluff study area (California State Water Resources Control Board, 2009). Although the detection frequency of total coliforms was higher in the Tehama County study, it is likely related to higher densities of wells that were sampled in urban centers compared to the GAMA-PBP study because total coliforms are more likely to be present in domestic wells in densely populated areas with septic systems or agriculture.

The presence of microbial indicators provides information about surface contamination of groundwater wells. Wells with total coliform and Enterococcus spp. detections were in regions with agriculture and urban land use, and six of the eight wells with detections (including the two wells with detections of Enterococcus spp.) were shallow (<100 feet), indicating a relation between human activity and microbial contamination of wells (fig. 11F). However, relations between the microbial indicator detections and any of the inorganic or organic constituents associated with potential anthropogenic sources (nitrate, chloroform, or pesticides) were not statistically significant.

Comparison of Study-Unit Characteristics

The NSV-PA and NSV-SA largely coincide areally (fig. 16); however, they represent different parts of the aquifer system vertically (fig. 17). The NSV-PA represents deeper groundwater primarily used for public supply, and the NSV-SA represents shallower groundwater primarily used for domestic supply. The differences in well depth between the NSV-PA and NSV-SA are statistically significant (p<0.0001; fig. 17). Additionally, the NSV-SA included the more rural western side of the valley that lacks public-supply wells, although wells that did not overlap with the extent of the NSV-PA were not included in the comparison. The differences between proportions of different land use and covers and aquifer lithologies around the public-supply and domestic wells used in the comparison were not statistically significant (p>0.1; figs. 18, 19).

Groundwater samples from the public-supply and domestic wells were geochemically similar at the study unit and study-area scales (figs. 20–22). The differences in dissolved oxygen concentrations and pH between the two study units were not statistically different (p=0.11 and p=0.35, respectively), and the difference in the proportion of wells in oxic or anoxic groundwater between the two study units was not statistically significant (fig. 20). The difference in proportion of wells with modern, mixed, and premodern aged groundwater and the distribution of tritium concentrations between the study units were not statistically difference (contingency table p=0.85 and Kruskal-Wallace rank sum test p=0.09, respectively; fig. 21). The same three water types were observed in the NSV-SA and NSV-PA: (1) a younger calcium-magnesium-sulfate-chloride water type found more commonly in the Red Bluff area, (2) an older or mixed sodium-bicarbonate type found in both areas, and (3) a sodium-chloride water type from mixing with deeper saline water (fig. 22). The overall water geochemistry in both study units was similar, indicating that water chemistry varies minimally with depth in the aquifer system.

Comparison of Water Quality

Constituents selected for additional evaluation (table 6) were the focus of the groundwater-quality comparison between the study units. Manganese was the only inorganic constituent found at high RCs in both NSV-PA and NSV-SA study units. Hexavalent chromium was selected for further evaluation in the NSV-SA due to high RCs in more than 2 percent of the groundwater resource but was not included in the comparison analysis because it was not analyzed in the NSV-PA dataset. However, hexavalent chromium is typically the dominant chromium species in natural waters because it is the more mobile species, and total chromium can be used to estimate the concentration of hexavalent chromium in groundwater (Manning and others, 2015). Chloroform was the only organic constituent that was detected at a frequency greater than 10 percent in either study unit (Bennett and others 2009; Shelton and others, 2020). Neither chloroform nor any other organic constituents were found at high RCs in either study unit.

Inorganic Constituents

When comparing the NSV-PA and NSV-SA, higher proportions of high and moderate RCs for inorganic constituents (fig. 23) and trace metals (fig. 24) were observed in the groundwater resources used by domestic wells. These higher proportions also were demonstrated in differences between the individual study areas, except for total chromium, which was found at moderate concentrations only in the NSV-SA Redding study area. However, the difference in proportions of wells between the study units and study areas were not statistically significant for inorganic constituents (fig. 23) or trace metals (fig. 24), Although the difference is not significant, the higher occurrence of elevated trace metal concentrations in the NSV-SA likely is related to interactions of groundwater with sediments derived from the North Coast Ranges (fig. 1) west of the Sacramento River where there are more domestic wells and fewer public-supply wells. This observation is supported by the higher proportion of wells in the Tehama Formation and in areas with natural land use in the NSV-SA (figs. 18, 19).

Manganese was the only trace element present at both high and moderate concentrations in the NSV-SA and NSV-PA; however, the difference in proportions of high or moderate RCs between the two study units was not statistically significant (figs. 25A–B). Total chromium was only found at moderate concentrations in 5 percent of the NSV-SA Redding study area, (figs. 25C–D). High and moderate RCs of total and hexavalent chromium were found in wells in the west side of the valley that did not overlap with the extent of the NSV-PA. Moderate levels of total chromium in the NSV-SA are not necessarily related to differences in depth of the groundwater resources used by the domestic and public-supply wells and could be related to locations of public-supply wells. Public-supply wells are primarily on the east and south side of the valley where there is less chromium-rich minerals in the aquifer sediments. However, the difference in number of grid wells located in each aquifer lithology between the study units was not statistically significant (fig. 19B).

Organic Constituents and Perchlorate

Wells in the NSV-PA are significantly deeper than wells in the NSV-SA (fig. 17). The groundwater used by public wells in the NSV-PA is slightly older than groundwater resources used by domestic wells in the NSV-SA (fig. 21), and the older groundwater would be less likely to contain constituents from anthropogenic sources if flow paths and residence times are relatively long. However, differences in the proportions of wells in each groundwater age category between the two study units were not statistically significant (contingency table test p>0.1; fig. 21). The organic constituents were present in higher proportions in the NSV-PA than the NSV-SA (fig. 26). Organic constituents were present at concentrations greater than reporting levels in about 18 percent of the groundwater resources in the NSV-SA and 30 percent in the NSV-PA (fig. 26A, average no detections of about 82 and 70 percent, respectively). This pattern also was observed for differences between the individual study areas, with higher proportions of organic constituent detections in the NSV-PA study areas compared to the NSV-SA study areas; however, the difference in the proportions between the study units was not statistically significant (contingency table test p>0.1). The differences in proportions of the aquifer with detections of organic constituents between the Redding and Red Bluff study areas for either the NSV-SA or the NSV-PA study units was not significant (contingency table text p>0.1).

Disinfection by-products were detected in 15 percent of the area in the NSV-SA, but detection was lower at 9 percent of the area in the NSV-PA (fig. 27A); however, the difference was not significant (fig. 27B). Chloroform was the most common disinfection by-product detected in both study units and was found in all four study areas (Bennett and others, 2009; Shelton and others, 2020). The disinfection by-product nitrosodimethylamine (NDMA) was detected at moderate RCs in the NSV-PA Red Bluff study area but was not analyzed as part of the NSV-SA assessment; therefore, NDMA was not included in the comparison (Bennett and others, 2009). The presence of similar occurrences of trihalomethanes (including chloroform) and pesticides in both study units is consistent with the results of the geochemical and groundwater age comparisons showing that the groundwater resources used by the domestic and public-supply wells likely are not distinct. However, the slightly higher presence of chloroform detection in the Redding study area of the NSV-SA could be related to more wells in modern or mixed groundwater in the NSV-SA compared to the NSV-PA (fig. 21).

The difference in the proportion of pesticide detections between the study units was not statistically significant (fig. 27B). The herbicides, simazine and atrazine, were the most detected pesticides in both study units, and simazine was observed in all study areas (Bennett and others, 2009; Shelton and others, 2020). The herbicides were the most detected organic constituent class in the NSV-PA and were found in 16 percent of wells from the NSV-PA and 5 percent of the NSV-SA (fig. 27A). The difference in agricultural land use around grid wells between the two study units was not statistically significant, which may explain why proportions of pesticides are similar in the two study units (fig. 18). Simazine is among the Nation’s most detected pesticide compounds in groundwater (Gilliom and others, 2006). However, use of atrazine and simazine as herbicides has declined in California (California Department of Pesticide Regulation, 2017). Atrazine is more commonly used in Shasta County, and use of atrazine peaked in 1999, whereas simazine is used in Tehama County, and use of simazine peaked in 2006. Uses of simazine and atrazine have declined to less than 250 pounds per year (peak values were more than 10,000 pounds per year) in the last decade, so the higher proportions of these pesticides in the NSV-PA may have resulted from legacy contaminations through migration to deeper wells (California Department of Pesticide Regulation, 2017).

Perchlorate

Moderate RCs of perchlorate were detected in the NSV-PA Red Bluff study area but not the NSV-PA Redding study area or in either of the study areas in the NSV-SA (fig. 28). The application of imported fertilizers containing minor perchlorate impurities until the mid-1900s can leave residue in the unsaturated zone (Interstate Technology and Regulatory Council, 2005; U.S. Environmental Protection Agency Federal Facilities Restoration and Reuse Office, 2005; Dasgupta and others, 2005; Böhlke and others, 2009), and natural perchlorate salts also can accumulate in the unsaturated zone over time (U.S. Environmental Protection Agency Federal Facilities Restoration and Reuse Office, 2005; Rao and others, 2007; Fram and Belitz, 2011). These natural or fertilizer-based perchlorate salts can be remobilized by infiltrating irrigation water. The return-flow waters remobilizing perchlorate can contain detectable levels of pesticides and bring both perchlorate and pesticides into the groundwater. Pesticide detections were higher in the NSV-PA, with the higher proportions of wells with pesticide detections in the NSV-PA Red Bluff study area, which is consistent with influence from agricultural return flow and legacy pesticide use in the groundwater resources used for public supply. The co-occurrence of moderate perchlorate RCs with pesticide detections in the NSV-PA Red Bluff study area may be a result of agricultural activities and return-flow infiltration.