Lee Acres Landfill Site Background

The Lee Acres Landfill was operated by San Juan County from May 1962 through April 1986 on land leased from the BLM. Unlined liquid waste lagoons at the landfill, which were present beginning in approximately 1979 (Roy F. Weston, Inc., 1995), accepted a variety of wastes, including chlorinated solvents, produced water from oil and gas fields, spent acid, and septic waste (NMED, 1986). The landfill's county operators applied potassium permanganate to the liquid waste lagoons to control hydrogen sulfide emissions generated by septic wastes (Norman Norvelle, El Paso Natural Gas Company, written comm., 2022). While performing a broader investigation into groundwater contamination related to landfills in New Mexico, NMED (1986) detected chlorinated solvents in liquid waste lagoons at the Lee Acres Landfill in 1985. Also, in 1985, a berm at the landfill’s northernmost liquid waste lagoon was breached, resulting in the release of liquid waste to the adjacent arroyo and the release of hydrogen sulfide gas that adversely affected individuals present at the time of the breach. As an emergency response to the release of hydrogen sulfide gas, NMED applied ferric chloride to the liquid wastes in the breached lagoon (NMED, 1986; U.S. Environmental Protection Agency [EPA], 2004). Shortly after the lagoon breach in 1985, the landfill discontinued the acceptance of liquid waste. The liquid waste lagoons had mostly dried through evaporation and infiltration by January 1986 (96–97 percent), and the former lagoons were covered with fresh soil later that year (BLM, 2009). The Lee Acres Landfill was placed on the National Priorities List in 1990 by the EPA because of the presence of heavy metals and volatile organic compounds related to the landfill in groundwater and soils surrounding the site (EPA, 2004).

Alluvial aquifer background concentration ranges were published in the landfill’s remedial investigation report (Roy F. Weston, Inc., 1995), and cleanup levels for seven constituents of concern were set by the EPA in the site’s record of decision (ROD; EPA, 2004). The alluvial aquifer background concentration ranges for several parameters and cleanup levels for the constituents of concern at the landfill are listed in table 1. Dissolved manganese is the only constituent of concern that is persistent in monitoring wells directly downgradient from the landfill in concentrations greater than the cleanup levels described in the ROD (Gray and Ferguson, 2022). The following remedies to address groundwater contamination were selected in the ROD:

The USGS—in cooperation with the BLM, which oversees the Lee Acres Landfill site—regularly monitors groundwater in wells at the site to satisfy the monitored natural attenuation requirement. The landfill’s location adjacent to an unnamed arroyo and the locations of site features, including liquid waste lagoons, monitoring wells, and the outlet of a culvert where surface-water runoff from County Road 350 is directed, are shown in Figure 2.

Giant Bloomfield Refinery Site Background

The GBR is a former crude oil refinery operated by Giant Industries from 1973 to 1982. Following the discontinuation of refinery operations in 1982, the site was used as a truck maintenance and dispatching headquarters for Giant Industries. In 1986, Giant Industries began investigating the impact of several historical releases of diesel fuel, crude oil, and gasoline in soils and groundwater at the refinery (Geoscience Consultants, Ltd., 1988). Shortly thereafter, a groundwater monitoring plan was established, and a treatment system was installed that targeted groundwater and soil-water contamination related to a large (10,000 to 15,000 gallon) diesel spill, regular small-scale spills at a truck fueling area, and regular releases of crude oil and gas in a firefighting drill area (fig. 2). During regular monitoring at the site between 2010 and 2020, dissolved manganese and chloride were detected in GBR monitoring wells upgradient from the refinery’s primary oil and fuel release areas (the diesel spill area, truck fueling area, and firefighting drill area) and downgradient from the landfill (fig. 2) at concentrations exceeding alluvial aquifer background concentration ranges established in the landfill’s remedial investigation (table 1; Roy F. Weston, Inc., 1995; LT Environmental, Inc., 2020).

Hydrogeologic Setting

The Lee Acres Landfill receives less than 10 inches of rainfall a year and is therefore considered arid. From 1991 to 2020, the mean annual precipitation in the nearby City of Farmington was 7.76 inches, and the daily mean temperature ranged from 31.1 degrees Fahrenheit in January to 77.1 degrees Fahrenheit in July (National Oceanic and Atmospheric Administration, 2023).

Lee Acres Landfill is within a broad physiographic region known as the San Juan Basin. The basin is filled mainly with sedimentary rocks that are Triassic to Quaternary in age (Craigg, 2001). The landfill is on San Juan River Quaternary floodplain alluvium, which consists of pebbly, gravely sand with clay lenses and boulders, and associated soil. Much of the fine-grain sediment in the area is of eolian origin, but fine-grained sediment at the site may come from fluvial deposits of the nearby unnamed arroyo. Larger clasts are likely derived from glacial outwash deposits from the Pleistocene era. These larger clasts are typically metamorphic or igneous, well-rounded, and lithologically similar to glacial outwash deposits along the Animas River (Brown and Stone, 1979). Also present at the site are soils of various development stages (Ward, 1990). Based on geologic logs from the landfill’s remedial investigation (Roy F. Weston, Inc., 1995), the combined alluvium and soils generally range from 13 to 62 feet thick and lie directly atop the Paleocene Nacimiento Formation. Alluvium is generally thickest within a buried channel incised in the underlying bedrock Nacimiento Formation (Peter and others, 1987). Groundwater in the alluvial aquifer is typically present where the alluvium is sufficiently deep; the water table is typically between 25 to 40 feet below land surface in the wells surrounding the landfill (Gray and Ferguson, 2022).

The Nacimiento Formation comprises shales, mudstones, and sandstones deposited in lake beds or stream channels (Craigg, 2001). The sandstones are typically fine- to coarse-grained, consisting of igneous-derived grains, and may be cemented by clay, hematite, silica, or calcite (Pederson and Dehler, 2004). The shales in this formation tend to display a popcorn-like texture on the surface that is associated with the presence of swelling clays, and the shales often display piping where the unit is exposed (Stone and others, 1983; Pederson and Dehler, 2004). The lower part of the formation contains black, carbonaceous mudstones interbedded with white sandstones, whereas the upper part consists of more neutral-colored sandstones and mudstones (Stone and others, 1983). The thickness of the formation varies but is generally thickest (as much as 1,300 feet) in the central part of the San Juan Basin, where the landfill is located (fig. 2; Stone and others, 1983; Craigg, 2001).

The two aquifers of interest in the study area are the alluvial aquifer and the underlying Nacimiento Formation aquifer, also referred to as the bedrock aquifer in this report. Groundwater flow in the alluvial aquifer at the site is generally from north to south, and groundwater flow in the bedrock aquifer is generally from northeast to southwest (Roy F. Weston, Inc., 1995). According to the landfill’s remedial investigation (Roy F. Weston, Inc., 1995), the alluvial aquifer at the site is heterogeneous and composed of bedded clay, silt, fine to coarse sand, and fine gravel. Although the dominant clast size is coarse sand, borehole and cone penetrometer data collected during the landfill’s remedial investigation (Roy F. Weston, Inc., 1995) indicate the presence of thick (more than 10 feet thick) clay layers in the alluvium. However, such clay deposits are not present at the surface, and their deposition cannot be explained by current stratigraphy. The bedrock aquifer at the site comprises shales and sandstones of various degrees of cementation and is discontinuously confined by a shale unit (Roy F. Weston, Inc., 1995). Previous work in the area has shown that the alluvial aquifer in the San Juan Basin is generally fresh and is a sodium-calcium-sulfate-bicarbonate type of water (Kelley and others, 2014; Newton and others, 2017). The bedrock aquifer is generally described as having a sodium-sulfate water type with high total dissolved solids relative to alluvial water (Brown and Stone, 1979; Kelley and others, 2014; Newton and others, 2017). Newton and others (2017) found water in the Nacimiento Formation aquifer to be pre-modern (at least several thousand years old) and water in the alluvial aquifer to be recharged within the last 10 years. Recharge to the Nacimiento Formation aquifer happens north of the study area near the San Juan Mountains in Colorado (Newton and others, 2017). This distance from the recharge source to the withdrawal site could account for this aquifer's generally older and more geochemically evolved water.

In a study focused on the neighboring Animas River alluvial aquifer within the San Juan Basin—which is assumed to be similar to the alluvial aquifer in the study area—Newton and others (2017) found that the primary source of recharge to that alluvial aquifer is from infiltration of irrigation water. Irrigation water originates as river water and infiltrates the shallow alluvium during conveyance and after application to crops and pastures. They also found that leakage from the regional bedrock aquifer and infiltration of focused precipitation along ephemeral arroyos were also possible sources of recharge to the Animas River alluvial aquifer, though relatively minor contributors—leakage from the regional bedrock aquifer accounted for as much as 20 percent but generally less than 5 percent of the alluvial aquifer water and the contribution of precipitation recharge to the aquifer was not quantified. A dense network of canals, crops, and pastures immediately surrounding the San Juan River near the study area (fig. 3) indicates that the recharge of irrigation water is likely a recharge mechanism to the alluvial aquifer in that area. However, upgradient from and directly surrounding the landfill, no evidence of irrigation is present (fig. 3). This suggests that infiltration of precipitation through the adjacent unnamed arroyo channel and leakage of the Nacimiento Formation aquifer could be the primary sources of recharge to the alluvial aquifer at the landfill. Although annual potential evapotranspiration exceeds annual precipitation in the study area, recharge from precipitation may happen periodically when precipitation exceeds runoff and evapotranspiration (Kernodle, 1996). The landfill’s remedial investigation found that a shale unit near the landfill discontinuously confines the Nacimiento Formation aquifer and that upward gradients are present, particularly on the east side of the arroyo where the landfill is located (Roy F. Weston, Inc., 1995). Upward leakage from the Nacimiento Formation aquifer is a likely source of recharge to the alluvial aquifer in areas with upward gradients but where the confining layer is not present, thin, or fractured.

Yields for wells completed in the alluvial aquifers throughout the San Juan Basin are between 1 and 1,100 gallons per minute, and those wells have a median output of about 15 gallons per minute (Kernodle, 1996). During the landfill’s remedial investigation (Roy F. Weston, Inc., 1995), slug testing was performed to measure the hydraulic conductivity in the alluvial and bedrock aquifers underlying the site. The mean hydraulic conductivity measured in the alluvial aquifer was 3.13 feet per day (ft/d). A reasonable upper boundary of hydraulic conductivity in the alluvial aquifer was estimated to be 10.0 ft/d, two standard deviations above the mean. Wells completed in the Nacimiento Formation aquifer throughout the San Juan Basin generally have yields between 2 and 90 gallons per minute and a median yield of about 7.5 gallons per minute (Kernodle, 1996). The mean hydraulic conductivity measured among bedrock wells at the site during the remedial investigation was 0.33 ft/d, an order of magnitude lower than hydraulic conductivity in the alluvial aquifer (Roy F. Weston, Inc., 1995).

The shallow alluvial aquifer can also be considered distinct from the deeper parts of the alluvial aquifer because of differing geochemical processes at various depths. For instance, oxidizing conditions are more likely to be found near the water table, where oxygen from the vadose zone is available. However, reducing conditions may become prevalent in deeper portions of the saturated alluvium. As in the landfill’s remedial investigation (Roy F. Weston, Inc., 1995), “shallow alluvial aquifer wells” are defined in this report as wells in which the screened interval intersects the water table in the alluvial aquifer, as opposed to being completely submerged within the alluvial aquifer. The landfill’s remedial investigation also defines “bedrock wells” as those having screens that encompass the top of the uppermost saturated bedrock zone (Roy F. Weston, Inc., 1995), and this report will use that definition. The aquifer classifications of the wells used in this study are listed in table 2. These classifications represent usual conditions while acknowledging that, for short periods, water level fluctuations may have fully submerged the screen of a shallow alluvial well or exposed the screen of a well classified as alluvial. Wells can be reclassified if conditions within that well change. For example, the aquifer classification of well BLM-39 has changed since the initial well installation because of rising groundwater levels relative to the screened interval in that well. In areas where the alluvial aquifer is thin, some wells have screens that intersect the water table in the alluvium and also extend into the bedrock. These are classified as shallow alluvial and bedrock aquifer wells in this report and in table 2. Instances where the groundwater elevation was lower than the bottom of the screen could indicate inaccuracy in the well construction information.

Conceptual Site Model

Liquid waste stored in the landfill’s lagoons likely leached into the subsurface during the period that the lagoons were wetted, from approximately 1979 to 1985. During infiltration, leachate could have encountered low-hydraulic-conductivity silt or clay lenses present in the unsaturated alluvium, causing it to spread laterally and seek alternate vertical infiltration routes. In addition, some portion of the leachate was probably retained in the vadose zone (the undersaturated portion of the subsurface above the groundwater table), particularly in fine-grained layers. Once the leachate reached the water table, it could have mixed with groundwater and flowed advectively (with the bulk of the fluid) downgradient. Though the primary alluvial groundwater gradient at the site is directed toward the south, steeper gradients caused by local mounding of the leachate could have also resulted in the lateral spread of leachate in all directions around the liquid waste lagoons. Seasonal and annual variation in groundwater levels likely resulted in vertical smearing of leachate mixed with groundwater. During initial infiltration and subsequent groundwater-level fluctuations (the latter of which is ongoing), landfill leachate could be retained in the vadose zone through the following mechanisms: incorporation in dead-end pore spaces, sorption onto sediment surfaces driven by surface charges, changes in the oxidation-reduction (redox) state of the leachate leading to precipitation, or evaporation of pore water leading to precipitation.

These processes may have led to the development of a dispersed vadose-zone source of leachate and leachate-related chemistry. Despite installing the capillary barrier and surface-water runoff controls over the landfill in 2005, a vadose-zone source of leachate could be mobilized into groundwater where it is encountered by rising groundwater levels. Furthermore, though the surface-water runoff controls direct surface water away from infiltrating within the landfill’s boundary, the location of the runoff culvert outlet near the landfill’s southern boundary (fig. 2) indicates that focused runoff could potentially mobilize landfill leachate stored in the vadose zone into groundwater. This study’s assessment of groundwater elevations, groundwater flow directions, and the timing of changes in groundwater chemistry could help distinguish among the following leachate-transport pathways: (1) a vadose-zone source of leachate mobilized by a fluctuating water table, (2) a vadose-zone source of leachate mobilized to groundwater through the infiltration of precipitation, possibly by focused storm-water runoff, and (3) leachate mixed with groundwater flowing downgradient.

Landfill leachate frequently contains organic chemicals and other compounds that can be electron donors in redox reactions. Redox reactions are chemical reactions in which electrons are transferred from an electron donor (oxidized compound) to an electron acceptor (reduced compound). Landfill groundwater plumes, therefore, commonly contain elevated concentrations of reduced manganese and other reduced species, but electron acceptors like nitrate and manganese oxides become depleted. This process is frequently associated with distinct redox zones within an aquifer; the distribution of these zones is controlled by the extent of leachate transport, the availability of electron acceptors, and the rate of electron donor degradation (Lyngkilde and Christensen, 1992a, 1992b). Therefore, in addition to investigating contaminants directly associated with the landfill, this study’s assessment of the distribution of redox conditions can also indicate aquifer zones affected by landfill leachate.

Purpose and Scope

The purpose of this report is to use the site investigation (Roy F. Weston, Inc., 1995) and long-term monitoring data in the LAGBRD to describe groundwater conditions within the study area and their corresponding implications regarding hydrogeochemical processes affecting groundwater chemistry. First, the water types among various aquifers underlying the landfill are classified. This classification is used to identify differences among groundwater upgradient, adjacent to, or downgradient from the landfill, providing insight into the impacts of landfill activities on groundwater chemistry. Next, temporal groundwater-quality patterns are assessed with the aim of understanding the impacts on groundwater chemistry of the 2005 installation of a capillary barrier and surface-water runoff controls over the landfill’s former liquid waste lagoons. Finally, the relationship of groundwater chemistry at the site to groundwater elevations and groundwater flow directions is assessed with the goal of better understanding contaminant transport processes.

Water Chemistry—Major Ions and Bromide

Major ion chemistry for groundwater samples collected at the regularly monitored BLM and GBR wells is plotted on a Piper diagram (fig. 4), along with data for the bedrock aquifer, precipitation, and Lee Acres Landfill liquid waste lagoon water. Groundwater from the bedrock aquifer is a sodium-sulfate type water, precipitation is a calcium-sulfate type water, and liquid waste lagoon water is a sodium-chloride type water. In a San Juan Basin-wide study by Kelley and others (2014), produced water (for oil and gas operations) from Cretaceous units in the San Juan Basin, which was accepted as waste at the lagoons, was found primarily to be sodium-chloride type water. This could explain the ionic composition of the liquid waste lagoons.

Recent (2019–20) groundwater data from the site falls into two groups on the central (anion-cation) portion of that diagram, both of which are between the precipitation end member and the bedrock and lagoon end members (fig. 4A). The first group, comprised of BLM-39, BLM-45, BLM-60, BLM-69, BLM-75, BLM-77, GBR-17, and GBR-50, is calcium-sulfate type water and plots closer to precipitation in figure 4A. The wells in this group are mostly shallow alluvial and alluvial wells (except for GBR-50, which has screens that extend from the shallow alluvial aquifer to the bedrock aquifer) and are in all areas surrounding the landfill: upgradient, adjacent, and downgradient from the landfill. The second group, comprised of BLM-62, BLM-68, BLM-80, GBR-32, GBR-48, and GBR-49, is a calcium-sodium-sulfate type water and plots closer to bedrock aquifer water in the central and cation plots in figure 4A. This group of wells is a mixture of shallow alluvial wells and wells with screens that extend from the shallow alluvial aquifer to the bedrock aquifer. The wells in this group are exclusively downgradient from the landfill. The distribution of all three end members in a straight line and the major ion composition of the recent groundwater monitoring data compared to that of the end members, particularly on the central and cation plots, makes it difficult to determine whether alluvial groundwater is mixing with water from the bedrock aquifer or the liquid waste lagoons. However, some wells from the second group (BLM-62, GBR-32, and GBR-48) do not plot between the bedrock aquifer and precipitation on the anion plot and are slightly shifted off the precipitation compositions toward the liquid waste lagoons, indicating that groundwater in these wells could potentially be mixing with a minor fraction of leachate from the lagoons in this recent time period.

End-member data, as well as all available monitoring data from the GBR wells, which spans from 1988 to 2019, are shown in Figure 4B. These data show that historical groundwater chemistry at the GBR wells differs from more recent (2019–20) conditions. Data from GBR-32, GBR-48, and GBR-49 display ion compositions that imply a greater degree of mixing with the sodium-chloride type water of the liquid waste lagoons than is indicated by more recent data. Bicarbonate data are not available for most historical BLM samples, so those data cannot be plotted on a Piper diagram. Other methods were used to ascertain whether BLM wells also historically displayed mixing with a sodium-chloride type water. Furthermore, other methods can help distinguish between natural causes, such as evaporite dissolution, and anthropogenic causes, such as leaching from the landfill, of the sodium and chloride enrichment seen in the historical GBR well data.

The relation between chloride and sodium (fig. 5) provides information on natural geochemical processes and anthropogenic impacts that are likely affecting groundwater in the alluvial aquifer. Alluvial aquifer data plotting between precipitation and bedrock aquifer end members in figure 5 could be the result of mixing between these two sources or could be the result of chemical evolution of the precipitation end member. During or following recharge to the alluvial aquifer, precipitation could acquire sodium through silicate mineral weathering (Meybeck, 1987) or cation exchange in sodium-rich clays (Cerling and others, 1989). The latter process is particularly associated with shales such as those found in the Nacimiento Formation (Cerling and others, 1989) and could explain the sodium content in the bedrock aquifer. Few natural processes cause chloride to increase independently of sodium, and groundwater plotting above the one-to-one chloride-to-sodium line in figure 5 is likely the result of mixing with the liquid waste lagoon end member. The high chloride content in the lagoons relative to groundwater and other end members could be the result of biotic and abiotic transformation of chlorinated solvents stored at the landfill (Witt and others, 2002; He and others, 2015). Furthermore, this process could potentially cause increasing chloride concentrations along flow paths as parent compounds in leachate are depleted. Some oil and gas produced waters from Cretaceous units in the San Juan Basin (produced water is a component of the liquid waste lagoons) also exhibit a chloride excess relative to sodium, which is indicative of the presence of connate seawater (Kelley and others, 2014).

All the wells identified as producing calcium-sulfate type water in figure 4A (BLM-39, BLM-45, BLM-60, BLM-69, BLM-75, BLM-77, GBR-17, and GBR-50) plot between precipitation and the bedrock aquifer in figure 5A. Those wells are from all portions of the alluvial aquifer underlying the site (shallow alluvial, alluvial, and screened across the shallow alluvial and bedrock aquifers) and are in all areas relative to the landfill (upgradient from, adjacent to, and downgradient from the landfill). As discussed, the relation between chloride and sodium in these data is indicative of mixing between the precipitation and bedrock aquifer end members or geochemical evolution of precipitation in the shallow alluvium. However, some data from BLM-60, GBR-17, and GBR-50 have exhibited higher chloride concentrations than can be explained by available data from those two end members (chloride concentrations more than 4 millimoles per liter [mmol/L]). This is particularly the case at GBR-17, where a sample from 1988 plotted above the one-to-one chloride-to-sodium line (chloride molar concentration of 18 mmol/L). Although BLM-80, which is screened across the shallow alluvial aquifer and the bedrock aquifer, plots in the same area as bedrock aquifer data, some samples exhibit higher chloride concentrations (as much as 5.6 mmol/L in figure 5A) than can be explained by the available bedrock aquifer data. The available bedrock aquifer data, which is limited to samples collected between 1987 and 1990 for the landfill’s remedial investigation (Roy F. Weston, Inc., 1995), might not capture the full variability of chloride concentration in the bedrock aquifer. Also, groundwater data with high chloride concentrations might be the result of mixing with a more concentrated source, such as leachate from the liquid waste lagoons.

Data from shallow alluvial and bedrock aquifer wells GBR-32 and GBR-48 plot between groundwater located between precipitation and bedrock end members ( BLM-39, BLM-45, BLM-60, BLM-69, BLM-75, BLM-77, GBR-17, and GBR-50) and the liquid waste lagoon end member (fig. 5), suggesting that those wells have been impacted by leachate originating from the lagoons. Based on chloride concentrations, shallow alluvial groundwater at BLM-68 and GBR-49 also appears to be a mixture of alluvial aquifer water and the liquid waste lagoon end member, but to a lesser degree. Data from BLM-68 and GBR-49 generally overlap with the lower chloride and sodium data range of GBR-32 and GBR-48 in figure 5. Data from BLM-62, a shallow alluvial aquifer well, are plotted in several locations in figure 5. Some BLM-62 data plots near the calcium-sulfate type groundwater that has more dilute concentrations of sodium and chloride (chloride less than 5.4 mmol/L and sodium between 8.7 mmol/L and 16.1 mmol/L). Another group of data from BLM-62 overlaps with upgradient bedrock aquifer data. Finally, some BLM-62 data overlap with groundwater from BLM-68 and GBR-49, and with the less-concentrated data from GBR-32 and GBR-48, or plots along a mixing line between that area and the bedrock-aquifer end member. Groundwater at BLM-62 has likely mixed with landfill leachate to varying degrees over the monitoring period.

From the analysis of the relation of chloride and sodium, it is apparent that a source outside of the bedrock alluvial aquifer and precipitation end members has contributed chloride to groundwater at BLM-62, BLM-68, BLM-80, GBR-32, GBR-48, and GBR-49. In particular, the groundwater at GBR-32 and GBR-48 appears to have likely mixed, to varying degrees, with leachate from the landfill’s liquid waste lagoons. Chloride-to-bromide ratios can be used to evaluate the possibility of chloride inputs from other sources, as this ratio is distinct among various natural and anthropogenic sources (Davis and others, 1998; Katz and others, 2011). Analysis of this ratio could be useful at Lee Acres, where road salts and general runoff from County Road 350 might also account for the high sodium and chloride concentrations observed in figure 5. The relation between chloride-to-bromide molar ratios and chloride molar concentrations in the bedrock aquifer water and in alluvial groundwater at the site is shown in figure 6. Bromide data were not available for the precipitation or the liquid waste lagoon end members, so literature values (table 4) for precipitation and various anthropogenic sources that could potentially impact groundwater at the site have been included.

Bedrock aquifer and precipitation data have lower chloride molar concentrations and generally lower chloride-to-bromide ratios than other groundwater data from the site (fig. 6), suggesting that another source contributes chloride to groundwater at the site. Data from BLM-39, BLM-45, BLM-69, GBR-17, and GBR-50 plot together in figure 6, generally between background bedrock aquifer water from well BLM-40 and septic tank effluent (wastewater discharge). The chloride-to-bromide molar ratios range between 176 and 761 and chloride concentrations range between 1.2 mmol/L and 2.5 mmol/L. Notably, these wells are all included in the group of calcium-sulfate type groundwater that appear to be a mixture of precipitation and bedrock aquifer water in figure 4A and figure 5A. Data from BLM-62, BLM-68, GBR-32, GBR-48, and GBR-49 plot together in figure 6 with similar to slightly higher chloride-to-bromide molar ratios (between 329 and 959) and higher chloride concentrations (between 3.4 mmol/L and 33 mmol/L) relative to wells in the first grouping. In figure 6, data from this group are generally located between the septic tank effluent and landfill leachate data. Some data points from BLM-62, BLM-68, and GBR-49 spread farther toward the more dilute groundwater of the first group. The second grouping of wells also showed evidence of mixing with liquid waste lagoon water in figure 5B. Other groundwater data at the site (data from BLM-60, BLM-75, BLM-77, and BLM-80) mainly plot between the first two groups, implying a mixture of water from those other groups.

Bromide tends to sorb (take up and hold by either adsorption or absorption) to organic materials, causing the depletion of bromide in solution (Davis and others, 1998). This observation is consistent with literature values (table 4) that show higher chloride-to-bromide ratios in septic tank effluent (where organic materials are present) as compared to landfill leachate and produced water from oil and gas operations in the San Juan Basin. Waste stored in the liquid waste lagoons at Lee Acres included septic wastes, produced water from oil and gas operations, and industrial wastes, and likely had a chloride-to-bromide molar ratio within the range of literature values for those sources, which appears reasonable based on figure 6 and the range in composition of produced water indicated in table 4. The high chloride-to-bromide ratios in the second group of wells in figure 6 (those with data plotting closer to a landfill end member) are likely explained by mixing with leachate from the Lee Acres Landfill. The compositions of groundwater from these wells appear to be following along a mixing line with landfill leachate as the higher-chloride end member rather than toward road-salt and highway-runoff compositions, which have higher chloride-to-bromide ratios than the landfill leachate. This pattern indicates that landfill leachate is more likely than road salt and runoff from County Road 350 to be a source of elevated chloride concentrations in groundwater at the site.

Temporal Changes in Groundwater Chemistry

Patterns in major ion and bromide chemistry, as discussed in the “Water Chemistry—Major Ions and Bromide” section, indicate that groundwater at the site, particularly at BLM-62, BLM-68, GBR-32, GBR-48, and GBR-49, has likely mixed with leachate from Lee Acres Landfill liquid waste lagoons as these wells are downgradient from the landfill. The data also establish that elevated chloride is an indicator of likely mixing with the landfill leachate. Piper diagrams (fig. 4) and the relation of chloride and sodium (fig. 5) show substantial changes in groundwater chemistry at some wells. This section assesses temporal changes in the concentrations of chloride (an indicator of contamination from the landfill’s liquid waste lagoons), manganese (a constituent of concern at the Lee Acres Landfill), and nitrate (an indicator of the redox state of groundwater for which available data are in the LAGBRD).

Chloride concentrations over time from the regularly monitored wells at Lee Acres are shown in figure 7. In the “Water Chemistry—Major Ions and Bromide” section, molar concentrations were a useful tool in relating the chloride and sodium content of groundwater, then relating those inferences derived about recharge sources from those concentrations to chloride-to-bromide ratios. Moving forward, chloride concentrations and other groundwater constituents will also be described in mass concentrations, which are more commonly used, particularly regarding regulatory limits. Two vertical axes display chloride in mass and molar concentrations in figure 7. Chloride concentrations in groundwater from most wells in this study (fig. 7) are generally much lower than the EPA’s secondary drinking water standard of 250 mg/L (EPA, 2023); chloride concentration did not exceed or only rarely exceeded 250 mg/L at over half of the 14 wells. Secondary standards are non-mandatory water-quality standards often based on aesthetic effects; in the case of chloride, the 250 mg/L secondary standard is based on salty taste (EPA, 2023). At BLM-39, BLM-45, BLM-60, BLM-69, BLM-75, BLM-77, BLM-80, GBR-17, and GBR-50, chloride concentrations generally did not exceed 140 mg/L except during specific time periods at the following subset of these wells downgradient from the landfill: BLM-60, BLM-75, BLM-77, BLM-80, GBR-17, and GBR-50. The first period with an exceedance of 140 mg/L at one of these wells is between 1986 and 1989, when the chloride concentration of groundwater from GBR-17 reached as high as 1,105 mg/L and from GBR-50 reached 141.85 mg/L (the number of significant figures is as reported). There are data gaps between 1991 and 1997, particularly at BLM wells, but during that period, GBR-17 and GBR-50 returned to lower chloride concentrations (generally less than 100 mg/L). The next period of high chloride concentrations among these generally low-chloride wells was between December 1997 and November 1999, when chloride concentrations exceeded 140 mg/L at BLM-60, BLM-75, and BLM-77, and again at GBR-50. Chloride concentrations did not exceed 140 mg/L again at these wells until 2016 when chloride reached 220 mg/L in September at BLM-60 and 220 mg/L in December at BLM-80. Chloride also increased at BLM-75 and BLM-77 in 2016 (as high as 130 mg/L in September at BLM-75 and 130 mg/L in March at BLM-77). Chloride concentrations have remained consistently low (less than 91 mg/L) throughout the assessed period at BLM-39, BLM-45, and BLM-69.

Compared to the low chloride concentrations (usually less than 140 mg/L) at BLM-39, BLM-45, BLM-60, BLM-69, BLM-75, BLM-77, BLM-80, GBR-17, and GBR-50, chloride concentrations have been consistently high (higher than 140 mg/L, and at times as high as 2,110 mg/L) at BLM-62, BLM-68, GBR-32, GBR-48, and GBR-49. Chloride concentrations were particularly high (higher than 800 mg/L) in some early samples (from 1991 and earlier) from GRB-17, GBR-32, GBR-48, and GBR-49 (fig. 7). Given that the landfill’s liquid waste lagoons had dried and were covered with fresh soil in 1986, the substantial decline of chloride concentrations after the early monitoring period is inferred evidence of the slowing of active leaching from the lagoons because of their drying. Chloride concentrations exceeded 800 mg/L again at GBR-32 between 1996 and 1999 and at GBR-48 between 1998 and 2004.

A Wilcoxon rank-sum test with continuity correction was performed to test for a step change in chloride concentrations following the installation of the capillary barrier and surface-water runoff controls. The test rejected the null hypothesis that the chloride concentrations for GBR-32 and GBR-48 were similar before and after installation in April 2005 (p-values of 5.9×10⁻³ and 4.3×10⁻⁶, respectively), indicating that chloride concentrations were significantly different between the two time periods. The change is particularly pronounced at GBR-48, where the chloride concentration decreased from 890 mg/L in December 2004 to 420 mg/L in December 2005 and has not exceeded 560 mg/L since the barrier’s installation. This pattern in concentration could indicate that the capillary barrier did reduce the leaching of liquid waste lagoon water to groundwater around GBR-32 and GBR-48. Even though areal recharge (recharge across a broad area) is generally minimal compared to focused recharge in arid regions, several berms within the landfill boundary and covered by the barrier in 2005 may have caused runoff to pool and resulted in focused recharge prior to 2005. These berms were identified in the site’s remedial investigation (Roy F. Weston, Inc., 1995). Chloride concentrations at BLM-68 and GBR-49 are generally moderate, having median values of 240 mg/L and 310 mg/L, respectively. No significant difference in chloride concentrations appears before and after capillary barrier installation at BLM-68 (Wilcoxon rank-sum p-value of 0.97), but there is a significant decrease in concentrations (p-value of 6.1×10^-3) at GBR-49. No significant difference in chloride concentrations at BLM-68, despite a significant decrease at nearby GBR-49, might be because of the gap in monitoring data at BLM-68 from 1991 to 1998, when increased chloride concentrations were observed at GBR-49. The hydraulic conductivity at BLM-68 was measured as 8.70 ft/d, higher than the mean hydraulic conductivity in the alluvial aquifer of 3.13 ft/d (Roy F. Weston, Inc., 1995), indicating that the lack of a significant decrease in chloride at that well is probably not a result of increased time for the arrival and dispersion of landfill leachate. This is confirmed by the apparent arrival of landfill leachate at GBR-17, farther from the landfill’s liquid waste lagoons than BLM-68, as early as 1986. Interestingly, chloride concentrations at BLM-62 rose significantly (p-value of 7.7×10⁻⁶) after the capillary barrier was installed. The median chloride concentration at BLM-62 was 36 mg/L prior to capillary barrier installation and 150 mg/L after the barrier was installed. Chloride reached its highest value at BLM-62 of 470 mg/L in July 2014, then subsequently decreased to 47 mg/L in March 2016 before climbing again to the most recent value in this dataset of 350 mg/L in August 2020. These changes suggest that the capillary barrier installation could have altered groundwater flow paths, causing the groundwater at BLM-62 to subsequently mix to a greater degree with water influenced by the liquid waste lagoons. A more compelling hypothesis may be that increasing chloride concentrations at BLM-62 could be the result of chloride being flushed from the vadose zone: surface-water runoff controls also installed in 2005 directed runoff from County Road 350 to a culvert with an outlet about 175 feet north of BLM-62 (fig. 2). Concentrated runoff may have recharged through the vadose zone, mobilizing chloride into groundwater. The source of this flushed chloride could be connected to the landfill or could be natural, as chloride and other salts are known to accumulate in the vadose zone in arid regions, particularly in settings where recharge is negligible (Scanlon and others, 2005); however, the timing of high manganese concentrations at this well, discussed later, suggests that groundwater at this well has mixed with landfill leachate.

Dissolved manganese concentrations in the regularly monitored wells at Lee Acres are shown in figure 8. The range of dissolved manganese at BLM-39, which is in the arroyo upgradient from the landfill and is described as the background well in the landfill’s ROD (EPA, 2004), is between 30 µg/L and 1,200 µg/L, and has a median concentration of 472 µg/L. Analysis of data from BLM-39 indicates that background manganese concentrations in the alluvial aquifer, which are unaffected by the landfill, have frequently exceeded the regulatory cleanup level of 346 µg/L. Most dissolved manganese concentrations in the regularly monitored wells do not exceed the maximum concentration of 1,200 µg/L observed in BLM-39 (fig. 8.) Therefore, in this report, manganese results for other wells will be compared to the maximum dissolved manganese concentration at BLM-39, instead of the regulatory cleanup level, as this is likely to be more representative of the maximum concentration for background aquifer conditions. For comparison, the range of dissolved manganese concentrations in the bedrock aquifer is between 149 µg/L and 504 µg/L, and the range of manganese concentrations (of unknown fraction, total or dissolved) at the liquid waste lagoons is between 800 µg/L and 2,100 µg/L.

In 1990, the earliest date with available dissolved manganese concentrations for groundwater from the regularly monitored wells at this site, the dissolved manganese concentration exceeded the maximum 1,200 µg/L background aquifer concentration and the maximum 2,100 µg/L liquid waste lagoon concentration at BLM-45, BLM-60, GBR-32, GBR-48, and GBR-50 (these wells are adjacent to and downgradient from the landfill). The only other wells with manganese data for 1990 are BLM-39, BLM-62, and GBR-49. After 1990, the next available dissolved manganese data are from 1998. In 1998, dissolved manganese had returned to levels below 1,200 µg/L at all wells with previously higher concentrations except GBR-50, where concentrations decreased below 1,200 µg/L by June 1999 and have not exceeded 1,200 µg/L since that time. As with early high chloride concentrations, early high dissolved manganese concentrations are evidence of direct leaching from the landfill’s liquid waste lagoons to wells adjacent to and downgradient from the landfill; subsequent decreases in manganese concentrations indicate that this leaching slowed over a period of several years after the lagoons had dried around 1986.

Since 1990, dissolved manganese exceeded the liquid waste lagoon maximum of 2,100 µg/L at BLM-62, BLM-77, and GBR-49. Dissolved manganese exceeded 2,100 µg/L at BLM-77 and GBR-49 before and after the capillary barrier and surface-water runoff controls were installed in April 2005. However, dissolved manganese exceeded 2,100 µg/L at BLM-62 only after 2005, coinciding with the period discussed above, when chloride was also high at BLM-62. Chloride and dissolved manganese are highly correlated at BLM-62, having a Spearman’s correlation coefficient of 0.73 and a p-value of 1.0×10⁻⁸. High dissolved manganese concentrations found at BLM-62 after 2005 provide an indication that altered flow paths or recharge locations induced by the capillary barrier or surface-water runoff controls have increased mixing of groundwater at that well with leachate from the landfill’s liquid waste lagoons. This finding is corroborated by the high positive correlation between dissolved manganese and chloride at BLM-62.

Dissolved manganese concentrations reached maximums of 9,200 µg/L at BLM-62 and 29,000 µg/L at BLM-77 in December 2013. Particularly at BLM-77, where the maximum detected dissolved manganese concentration is an order of magnitude above the concentrations detected in the landfill’s liquid waste lagoons, manganese originating from the lagoons and leaching directly into groundwater is probably not responsible for these high concentrations observed in groundwater. Rather, the redox state of the groundwater, which can determine the mobility of dissolved constituents—including manganese (Chapelle and others, 1995; Paschke and others, 2007; McMahon and Chapelle, 2008), could help explain the high dissolved manganese observed in some groundwater at the site. Microbially mediated reduction of manganese oxides, resulting in increased manganese mobility, can be stimulated by the presence of organic carbon (Chapelle and others, 1995; Christensen and others, 2001; McMahon and Chapelle, 2008; McMahon and others, 2019). Wastes stored in the landfill’s liquid waste lagoons, including septic wastes and chlorinated solvents, could act as a source of organic carbon for microbially-mediated manganese reduction at the site.

Data in the LAGBRD, though limited, could be used to infer the redox state of the monitored groundwater; for instance, some nitrate data are available in the LAGBRD. Redox frameworks developed for studies on diverse aquifers suggest that manganese reduction can happen at nitrate concentrations of less than 0.5 mg/L as nitrogen (Chapelle and others, 1995; Paschke and others, 2007; McMahon and Chapelle, 2008). The maximum nitrate value observed at BLM-62 was 0.52 mg/L as nitrogen (this value was below the reporting level and qualified as estimated). At BLM-77, nitrate is moderately negatively correlated to dissolved manganese and has a Spearman’s correlation coefficient of −0.63 and a p-value of 1.9×10⁻⁵ (nitrate values at BLM-62 were all below the censoring level, and so the ranked correlation between nitrate and manganese could not be calculated). Furthermore, nitrate was not detected (at a reporting level of 0.5 mg/L as nitrogen) in the BLM-77 sample with a dissolved manganese concentration of 29,000 µg/L.

For comparison, the range of nitrate concentrations in precipitation is 0.42 to 1.0 mg/L as nitrogen, and the median concentration is 0.93 mg/L as nitrogen (NADP, 2022), and the range of nitrate in the background well BLM-39 is 0.57 to 13 mg/L as nitrogen, and the median concentration is 4.4 mg/L as nitrogen. Nitrate was not detected (the reporting level was unknown) in the single liquid waste lagoon sample with available nitrate data. Nitrate concentrations over time from the regularly monitored wells at Lee Acres are shown in figure 9. Nitrate concentrations vary at BLM-77 and frequently exceed the 0.5 mg/L as nitrogen threshold for manganese reduction (fig. 9), suggesting that conditions at this well are not always sufficiently reducing to the point of manganese reduction and mobilization. The negative correlation between nitrate and dissolved manganese at BLM-77 suggests that the redox state of groundwater is a control on manganese concentrations at that well.

Dissolved manganese data are available at the GBR wells until the installation of the capillary barrier and surface-water runoff controls in April 2005, except for one sample date in January 2009 (there is no dissolved manganese result for GBR-17 on this date). Among the available data, dissolved manganese concentrations at GBR wells have generally been below the 1,200 µg/L maximum at background well BLM-39 (except when noted, above). This is true even at GBR-32 and GBR-48, where chloride concentrations and chloride-to-bromide ratios indicate mixing with landfill leachate, particularly (in the case of GBR-48) prior to capillary barrier and surface-water runoff control installation, when dissolved manganese concentration data are available. A possible explanation for the elevated chloride concentrations concurrent with lower (less than 1,200 µg/L) manganese concentrations observed at wells like GBR-32 and GBR-48 is that oxidizing conditions could be limiting manganese mobility. All nitrate concentrations, except for one GBR-32 result from December 1991, were above the nitrate threshold of 0.5 mg/L as nitrogen for manganese reduction (fig. 9). Nitrate concentrations are particularly high at GBR-48, reaching a maximum value of 23.1 mg/L as nitrogen in June 1997. Though nitrate was not detected in the one liquid waste lagoon sample analyzed for nitrate, the lagoons, which accepted septic waste, might be a source of nitrate that may have spread to the vadose zone and to groundwater. This would be feasible under oxidizing conditions, which may have happened periodically throughout the lifespan of the lagoons. The source of nitrate in groundwater at the site might be natural, as the vadose zone can act as a reservoir where natural nitrate can accumulate in arid regions where recharge is negligible (Scanlon and others, 2005; Linhoff and Lunzer, 2021). Nitrate from natural and anthropogenic sources can be mobilized from the vadose zone and released to groundwater in arroyos by migration of the stream channel (Linhoff and Lunzer, 2021; Linhoff, 2022). Land use changes have also been linked to the mobilization of vadose-zone nitrate into groundwater (Scanlon and others, 2005; Linhoff and Lunzer, 2021). Beyond natural arroyo channel migration, vadose-zone nitrate at Lee Acres may have been mobilized by infiltration of leachate from the landfill’s liquid waste lagoons, which had mostly dried by 1986, or by recharge of concentrated surface-water runoff at a culvert outlet (fig. 2) installed in 2005. A full investigation of the sources of nitrate in groundwater at the site is outside of the scope of this investigation, though further research could enhance understanding of how and if nitrate controls manganese concentrations in groundwater at the site.

Temporal Changes in Groundwater Elevation

As discussed in the “Hydrogeologic Setting” section, the alluvial aquifer at the landfill is likely recharged through leakage from the Nacimiento Formation aquifer and infiltration of precipitation focused in the adjacent arroyo channel. In this arid region, the timing of precipitation and potential evapotranspiration determine whether precipitation is sufficient to recharge groundwater (Kernodle, 1996). To determine the amount of precipitation that could potentially reach the aquifer in the study area, the annual water surplus was calculated. Annual precipitation and the annual water surplus are shown in figure 10, overlain by groundwater elevations at regularly monitored BLM and GBR wells at the site. The date of the 2005 installation of the capillary barrier and surface-water runoff controls is also shown in figure 10.

Comparison of temporal changes in groundwater elevation to the annual water surplus (fig. 10) reveals several periods when focused recharge of precipitation in the arroyo channel probably happened, as indicated by increasing groundwater elevations concurrent with years of positive annual water surplus. For example, groundwater elevations increased at most wells from 1997 to 1999 and from 2005 to 2008—periods when the annual water surplus was positive. Groundwater decline in response to annual water deficit is also apparent, as during the 2000–2004 period when these four successive years of annual water deficit were concurrent with gradual groundwater decline. These observations confirm that precipitation is a component of groundwater recharge in this area. However, wells near and downgradient from the landfill’s southern boundary (within 500 feet of the southern landfill boundary) exhibited a substantial decline in groundwater elevation between about 2010–13 despite substantial annual water surplus in 2010. This was apparent at BLM-60, BLM-62, BLM-75, BLM-77, and BLM-80. GBR-32, GBR-48, GBR-49, and GBR-50 did not have sufficient groundwater-elevation data in the LAGBRD during this period to determine whether a similar decline happened at those wells. Notably, no concurrent decline in groundwater elevation was observed at wells BLM-15 and BLM-39 upgradient from the landfill, at BLM-45 adjacent to the landfill, or at GBR-17 farthest downgradient from the landfill, though during other periods groundwater elevation variability was generally consistent across all wells in the study area. The location of the decreased water levels suggests that this is the effect of the landfill’s capillary barrier retarding local infiltration, though given the more immediate response of groundwater elevations to climatic forcing in this area, one would expect a groundwater decline related to the barrier to happen more immediately after barrier installation. Another explanation of the observed groundwater decline is downwards infiltration of water from the alluvial aquifer to the bedrock aquifer, which may not be fully confined at the location of these wells (Roy F. Weston, Inc., 1995), and for which no groundwater elevation data are in the LAGBRD during this 2010–13 period. Additionally, the wells that exhibited groundwater decline are near the surface-water runoff control outfall; changes in recharge dynamics related to the capillary barrier and surface-water runoff controls may have caused the decline.

The highest level of annual water surplus was in 2015 (fig. 10) and was followed by a distinct increase in groundwater elevations in all wells. Despite the apparent recharge response at all wells, including those near the landfill, infiltrating precipitation associated with this surplus probably did not recharge directly through the surface of the landfill. According to the most recent five-year site review (BLM, 2019), the capillary barrier has been functioning optimally since installation, as indicated by lysimeters within the barrier. Focused recharge through the sandy bottom of the adjacent arroyo probably causes groundwater mounding that can be observed as increased groundwater elevation in alluvial wells outside of the main arroyo channel. Focused runoff from the surface-water runoff control outfall probably supplements recharge at the wells downgradient from the landfill.

As discussed, years of substantial water surplus are generally followed by increased groundwater elevations at all wells in figure 10, indicating that focused recharge of precipitation is the cause of these increases. Groundwater mounding would be indicated by local increases in groundwater elevation that propagate away from the landfill over time. No evidence of groundwater mounding that may be associated with landfill leachate infiltration from the time that the liquid waste lagoons were active is seen in figure 10. The groundwater-elevation behavior at BLM-15 provides evidence that mounding is not observed in figure 10. The location of the BLM-15 well suggests that a groundwater mound would probably not be evident at this well at the scale or measurement frequency shown in figure 10; the well is over 1,000 feet upgradient from the landfill, and the groundwater elevation at that well is about 13 feet higher than at BLM-39, near the landfill’s northern boundary. If a propagating groundwater mound were observable at this well, it would be attenuated (would have a smaller relative increase and would happen later) relative to wells like BLM-39, BLM-45, and BLM-60, that are closer to the landfill. However, the timing and magnitude of groundwater-elevation changes observed at wells near and downgradient from the landfill are also reflected at BLM-15, suggesting that a groundwater mound spreading from the landfill’s liquid waste lagoons is probably not influencing the groundwater-elevation changes observed in figure 10.

Groundwater Elevation and Groundwater Chemistry

Assessing temporal variation in groundwater elevation and groundwater chemistry can provide insight into the mechanisms that control groundwater mixing with landfill leachate. In addition to assessing the timing of groundwater-elevation and groundwater-chemistry changes, well location in relation to the landfill and to the surface-water runoff culvert can be incorporated to consider the effects of the 2005 installation of the capillary barrier and surface-water runoff controls. In this section, the following two mechanisms of mixing with leachate outlined in the “Conceptual Site Model” section will be assessed: (1) a vadose-zone source of leachate captured by a fluctuating water table, and (2) a vadose-zone source of leachate mobilized to groundwater through the infiltration of precipitation, possibly by focused storm-water runoff. A third mechanism described in the “Conceptual Site Model” section, leachate mixed with groundwater flowing downgradient, is discussed in the “Groundwater Gradients and Transport” section. This section focuses on comparing the concentrations of manganese and chloride, two elements that are associated with Lee Acres Landfill leachate, with groundwater elevations over time in selected regularly monitored wells at the Lee Acres Landfill site. Groundwater elevations, manganese concentrations (dissolved, in µg/L), and chloride concentrations (dissolved and total, in mg/L) for BLM-39, BLM-62, BLM-77, BLM-80, GBR-32, GBR-48, and GBR-50 are shown in Figure 11. These wells were selected for figure 11 because of their locations or chemical characteristics: BLM-39 is upgradient from the landfill and was established as the background well in the remedial investigation (Roy F. Weston, Inc., 1995); BLM-62 and BLM-77 have historically had the highest dissolved manganese concentrations; BLM-80 is adjacent to alluvial wells BLM-62 and BLM-77 but is screened in the alluvial and bedrock aquifers and displays different behaviors with respect to manganese and chloride concentrations; and GBR-32, GBR-48, and GBR-50 are screened in the bedrock and alluvial aquifers near BLM-62 and BLM-77, but show different behaviors with respect to their manganese and chloride concentrations.

Several groundwater-elevation changes that are relevant to the interpretation of the chemical data and are common across multiple wells are shown in figure 11. A 3–4-ft decrease in groundwater elevations around 2011 is apparent in figure 11 at BLM-62, BLM-77, and BLM-80 (wells directly downgradient from the landfill) but not at BLM-39 (upgradient from the landfill). As previously discussed, this could be the result of recharge from the alluvial aquifer to the bedrock aquifer at the location of the downgradient wells or to changes in recharge dynamics related to the 2005 installation of the landfill capillary barrier and surface-water runoff controls. The same decrease is not observed at GBR wells in figure 11 in a similar area, though groundwater-elevation data are not available at the GBR wells in 2011. A general rapid increase is shown in all wells in figure 11 between around 2013 (2012 at GBR wells) and 2015. Increases range from around 4 to 6 ft. This is followed by a more gradual decline in water levels (ranging from 2 to 3 ft at most wells) at all wells except GBR-48 until the end of the record around 2020. Several groundwater-elevation measurements at GBR-48 between 2015 and 2019 are higher relative to elevations observed at nearby wells, but these observations do not affect the interpretations of this section and are not discussed in additional detail in this report.

The highest dissolved manganese concentration at BLM-39 (1,200 µg/L) was in May 1998 and did not coincide with a substantial change in groundwater elevation (fig. 11A). However, dissolved manganese concentrations were consistently lower after 2013, corresponding with the period of highest groundwater elevations at the well. At this well, the increases in groundwater elevation in the 2013–15 period are associated with decreases in dissolved manganese concentrations and increases in chloride concentrations. Groundwater elevations are moderately negatively correlated with dissolved manganese concentrations at this well (Spearman’s correlation coefficient of −0.67 and a p-value of 3.2×10⁻⁷) and moderately positively correlated with chloride concentrations (Spearman’s correlation coefficient of 0.70 and a p-value of 9.9×10⁻⁸). Chloride concentrations and dissolved manganese concentrations are strongly negatively correlated (Spearman’s correlation coefficient of −0.73 and a p-value of 9.4×10⁻¹⁰). A plausible explanation for increases in chloride concentrations as groundwater elevations increase at BLM-39 is that mixing with landfill leachate is happening as rising groundwater encounters leachate stored in the vadose zone. This could be feasible even at this location upgradient from the landfill because leachate could have both spread laterally and been retained in the vadose zone during infiltration upon interception with low conductivity units such as the thick clay layers known to be present in the alluvium at the site (Roy F. Weston, Inc., 1995). Redox conditions likely favor oxidation and precipitation of manganese at these groundwater elevations, as indicated by generally high nitrate concentrations at BLM-39 during this period (fig. 9) and low manganese concentrations. Installation of the landfill capillary barrier and surface-water runoff controls in 2005 does not appear to have affected groundwater elevations or groundwater chemistry at this well.

Groundwater elevations and dissolved manganese concentrations are strongly positively correlated at BLM-62 (Spearman’s correlation coefficient of 0.73 and a p-value of 5.0×10⁻⁸), though it appears that dissolved manganese concentrations did not change substantially with changes in groundwater elevation until after the 2005 installation of the landfill’s capillary barrier and surface-water runoff controls (fig. 11B). After 2005, dissolved manganese concentrations increased overall but increases were more rapid during periods of groundwater-level rise during two periods, first in 2007 and then in 2013. During both periods, the groundwater level at BLM-62 was around 5,382 ft above datum, suggesting that the increase in dissolved manganese concentrations could be caused by a horizon in the aquifer at this elevation that contains reducing constituents or soluble manganese, possibly related to landfill leachate. Alternatively, or possibly in addition to, increased manganese concentrations after 2005 could be related to focused runoff being directed to the land surface near this well after surface-water runoff control installation. This focused runoff may have mobilized leachate containing manganese or reducing compounds stored in the vadose zone (as discussed in the “Conceptual Site Model” section) into groundwater or into new areas of the vadose zone.

Manganese and chloride concentrations are strongly positively correlated at BLM-62 (Spearman’s correlation coefficient of 0.73 and a p-value of 1.0×10⁻⁸), suggesting that landfill leachate is the likely source of manganese or reducing constituents that mobilize manganese into groundwater at these horizons in the aquifer or during these precipitation recharge conditions. Increases in chloride concentration that began after capillary barrier and surface-water runoff control installation in 2005 at this well (discussed in the “Temporal Changes in Groundwater Chemistry” section) and the correlation between manganese and groundwater elevation that only began after capillary barrier installation suggest that the capillary barrier or surface-water runoff control installation altered flow paths or recharge locations and caused increased mixing with landfill leachate at BLM-62.

Dissolved manganese concentrations and groundwater elevations are not significantly correlated at BLM-77 (Spearman’s correlation coefficient of 0.10 and a p-value of 0.57). Prior to the installation of the capillary barrier and surface-water runoff controls in 2005, the highest dissolved manganese concentration was 8,700 µg/L in November 2001. The first sample collected following the capillary barrier and runoff control installation (June 2005) had a dissolved manganese concentration of 2,900 µg/L. After that, dissolved manganese concentrations were below the 1,200 µg/L background alluvial aquifer threshold until the highest concentration (29,000 µg/L) was detected in December 2013, during the 2013–15 period of rapid groundwater-elevation rise. Subsequent dissolved manganese concentrations were much lower than 29,000 µg/L, even though groundwater elevations continued to rise. In fact, beginning with the sample collected in November 2015, all dissolved manganese concentrations at this well were below the background alluvial aquifer threshold concentration of 1,200 µg/L. This pattern of an abrupt, but short-lived, increase in dissolved manganese concentrations following a substantial increase in groundwater elevations suggests that soluble manganese or reducing constituents were either flushed from the vadose zone by infiltrating precipitation or that the water table had risen into an area of the vadose zone where such constituents were stored. The temporary decline in nitrate concentrations (fig. 9) at this well during the period of increased manganese indicates that reducing conditions are responsible for the observed mobilization of manganese. Reducing conditions could have been stimulated by acids or organic constituents leached from the landfill. The short-term nature of the increased manganese concentrations shows that oxidizing conditions prevailed as groundwater elevations continued to rise or that the compounds stimulating reducing conditions were rapidly flushed.

Chloride concentrations at BLM-77 have a statistically significant but weak positive correlation with groundwater elevations (Spearman’s correlation coefficient of 0.41, p-value of 8.8×10⁻³). Prior to 2005, the highest concentration of chloride was 170 mg/L in June 1999, around the same time that several other wells with generally low chloride concentrations (less than 140 mg/L) showed increased chloride, as discussed in a previous section. Chloride concentrations fell rapidly between June 1999 and May 2002, then remained stable between 44 and 60 mg/L until they begin to rise again in December 2013, coinciding with the sharp rise in groundwater elevations discussed above. The highest concentration of chloride after this date was 130 mg/L in March 2016, right after groundwater elevations reached their highest level observed for BLM-77 and about 2 years after the highest dissolved manganese observation at this well. Around the same time, chloride concentrations at BLM-62 reached their lowest values since 2007 (fig. 11B) and were within a similar range of concurrent BLM-77 concentrations (chloride concentrations at BLM-62 ranged between 47 and 130 mg/L between December 2015 and September 2016). Groundwater elevations were similarly high at BLM-77 and BLM-62 during this 2016 period (5,386.63 and 5,385.97 ft above datum, respectively, in March 2016). The presence of similar chloride concentrations at BLM-62 and BLM-77 around a groundwater elevation of 5,386 ft could indicate that a conductive layer exists around this elevation, and that groundwater at these nearby wells was mixed as it encountered this layer. The presence of a conductive layer in this area at an elevation of 5,386 ft is confirmed by Roy F. Weston, Inc.’s (1995) geologic logs from BLM-62 and BLM-77. These logs show a sandy layer at this elevation, and furthermore show a thick, about 8-ft layer of sandy clay at lower elevations at BLM-62 that is not present at BLM-77 and that likely inhibits groundwater mixing between these wells at lower groundwater elevations (Roy F. Weston, Inc., 1995). After 2016, chloride concentrations fell rapidly at BLM-77 (they did not exceed 70 mg/L after March 2017), chloride concentrations increased at BLM-62 (reaching 310 mg/L by March 2019), and groundwater elevations gradually decreased at both wells. This evidence indicates that at a groundwater elevation of 5,386 ft, landfill leachate from the source impacting BLM-62 is mixing with groundwater at BLM-77, resulting in the dilution of the chloride leachate signal at BLM-62 and simultaneous increased chloride concentrations at BLM-77. Below this groundwater elevation, low conductivity aquifer materials at BLM-62 may inhibit landfill leachate from mixing with groundwater at BLM-77.

Because BLM-80 was not installed until late 2005, no data exist prior to the installation of the capillary barrier or surface-water runoff controls. Dissolved manganese concentrations are not significantly correlated with groundwater elevations at this well (Spearman’s correlation coefficient of 0.01, p-value of 0.98). All dissolved manganese concentrations for this well are below the 1,200 µg/L background alluvial aquifer threshold value established at BLM-39. For most of the record, manganese concentrations remain fairly steady, ranging from 340 to 500 µg/L. Only four samples have higher manganese concentrations, and these were during the 2013–15 period of groundwater elevation rise. Dissolved manganese reached as high as 990 µg/L during that period. As at BLM-77, the timing of low nitrate concentrations (fig. 9) during periods of high manganese concentrations suggests that reducing conditions are responsible for the mobilization of manganese at BLM-80. Reducing conditions may have been captured by rising groundwater levels or reducing constituents may have been flushed from the vadose zone during periods of high recharge. The availability of more energetically favorable electron acceptors than manganese oxides may explain the variability of manganese chemistry at BLM-80 during this period.

Chloride concentrations at BLM-80 are moderately correlated with groundwater elevations (Spearman’s correlation coefficient of 0.65, p-value of 4.8×10⁻⁴). Chloride concentrations prior to July 2014 were low and fairly steady, ranging between 19 and 27 mg/L. Beginning in July 2014, the concentrations of chloride increased until reaching their historical maximum of 220 mg/L in December 2016. This concentration increase corresponds with generally increasing groundwater elevations during that time. The groundwater elevation of 5,386 ft above datum is associated with apparent groundwater mixing at BLM-62 and BLM-77 (groundwater elevation ranged from 5,385.12 ft above datum in July 2014 to 5,387.43 ft above datum in March 2017 at BLM-80). After 2016, chloride concentrations and groundwater elevations decreased at BLM-80, though chloride did not return to the lower concentrations seen previously. The rise in chloride concentrations combined with increased groundwater elevations could be indicative of a flushing event wherein landfill leachate stored in the vadose zone (discussed in the “Conceptual Site Model” section) was mobilized into groundwater through recharge of precipitation or was encountered by rising groundwater. Alternatively, and similar to BLM-77, mixing with leachate at this location could be controlled by low-conductivity clays around BLM-62 that inhibit leachate mixing at BLM-80 until groundwater at BLM-62 encounters more conductive aquifer materials around the elevation of 5,386 ft.

Dissolved manganese concentrations and groundwater elevations at GBR wells do not have enough overlapping data (n=1) to provide correlation statistics. This is because of the 30-year gap in groundwater-elevation data and because most of the manganese data collected after 2009 were total concentrations, as opposed to the dissolved concentrations reported previously. The total concentrations are not shown in figure 11 because interpretation of changes in the concentration of metals could be affected by examining different fractions (total or dissolved; as discussed in the “Methods” section).

The highest concentrations of dissolved manganese at GBR-32 (2,330 µg/L) and at GBR-48 (2,120 µg/L) were measured in 1990. At GBR-32, the high dissolved manganese sample occurred shortly after an increase in groundwater elevations. During this period, not enough groundwater data are available to establish a similar relation at GBR-48. The high dissolved manganese at GBR-32 and GBR-48 in 1990 also occurred shortly after the period of early elevated chloride concentrations at these wells that is likely associated with infiltration of leachate from landfill liquid waste lagoons before they had dried. Most remaining manganese samples at these wells were collected between 1998 and 2005, prior to the installation of the capillary barrier and surface-water runoff controls, but all remaining dissolved manganese results are below the background alluvial aquifer dissolved manganese concentration of 1,200 µg/L established at BLM-39. As previously discussed, redox conditions (as indicated by nitrate concentrations; fig. 9) likely control manganese concentrations at these wells; early high manganese could have been caused by mixing of reducing constituents in landfill leachate with groundwater but the increased availability of nitrate that started around 1998, a more energetically favorable electron acceptor than manganese oxides, likely prevented the reductive dissolution of manganese oxides.

Chloride concentrations and groundwater elevations are not significantly correlated at either GBR-32 or GBR-48 (Spearman’s correlation coefficient of −0.71 with a p-value of 7.1×10⁻² at GBR-32 and Spearman’s correlation coefficient of 0.40, p-value of 0.60 at GBR-48). Even though the landfill’s capillary barrier installation apparently reduced mixing with the landfill end member at these wells (discussed in a previous section), chloride concentrations remained high relative to alluvial chloride concentrations upgradient from the landfill at BLM-39, which never exceeded 65 mg/L; after capillary barrier installation, chloride ranged from 190 to 530 mg/L at GBR-32 and from 140 to 560 mg/L at GBR-48. The lingering high chloride at GBR-32 and GBR-48 indicates that some mixing with leachate has persisted in groundwater at these nearby wells. Chloride concentrations did not increase at these wells (or GBR-50, discussed below) following periods of increased groundwater elevation that are associated with mixing with landfill leachate at nearby BLM-77 and BLM-80. However, this lack of chloride increase at GBR-32 and GBR-48 does not preclude the possibility that a similar source of leachate was encountered as water levels increased in these wells; rather, the pre-existing, elevated chloride concentrations at these wells may have obscured the signal of mixing with a relatively less-concentrated leachate source. Heterogeneity in the vadose zone and the aquifer leading to preferential leachate flow paths could explain the uneven distribution of chloride among these nearby wells. Chloride concentrations and dissolved manganese concentrations at GBR-32 and GBR-48 are not significantly correlated (Spearman’s correlation coefficient of −0.20 and a p-value of 0.49 at GBR-32 and Spearman’s correlation coefficient of 0.20, p-value of 0.49 at GBR-48), confirming that the availability of energetically favorable electron acceptors like nitrate likely inhibits the mobilization of manganese into groundwater at these wells despite apparent mixing with leachate.

The highest dissolved manganese concentration at GBR-50, 2,420 µg/L, was in April 1990. Dissolved manganese again exceeded the 1,200 µg/L background alluvial aquifer threshold value in 1998, reaching 1,500 µg/L in May 1998 before decreasing to concentrations below 65 µg/L beginning in November 2002. This decrease in dissolved manganese concentrations was during the gap in groundwater-elevation data, so it is not clear if the concentrations were affected by changes in groundwater elevations. Only one sample for dissolved manganese was collected after the installation of the landfill capillary barrier and surface-water runoff controls. This January 2009 sample had a dissolved manganese concentration of 50 µg/L.

Chloride concentrations and groundwater elevations at GBR-50 are not significantly correlated (Spearman’s correlation coefficient of 0.50 and a p-value of 0.39), though groundwater-elevation data are not consistently available during two periods when chloride concentrations were higher than the typical range at this well (between 39 and 69 mg/L). Those periods of elevated chloride concentration are concurrent with the two periods of high manganese concentration discussed above, and chloride and dissolved manganese concentrations are strongly positively correlated at this well (Spearman’s correlation coefficient of 0.83, p-value of 2.5×10⁻⁴). The correlation between chloride and dissolved manganese concentrations at GBR-50 indicates that, even though less prevalent than at BLM-62, BLM-77, GBR-32, and GBR-48 (which have exhibited more extreme chloride and dissolved manganese concentrations than those seen at GBR-50), historical mixing with landfill leachate is evident at this well. Following the 2005 installation of the capillary barrier and surface-water runoff controls, chloride concentrations remained within the more typical range for this well (39 to 69 mg/L), indicating that mixing with landfill leachate has not been substantial at this well since the barrier’s installation. The heterogeneity in the alluvium surrounding the landfill suggests that preferential flow paths could contribute to the complex nature of chloride distribution in groundwater. The location of GBR-50, farther north and west of BLM-62, BLM-77, BLM-80, GBR-32, and GBR-48 (which are very near one another), could place this well outside of flow paths connected to landfill leachate sources that affect other wells. Furthermore, the location of GBR-50 slightly upgradient from the surface-water runoff culvert (fig. 2) suggests that this well might not be affected by recharge of concentrated surface-water runoff that is likely mobilizing landfill leachate from the vadose zone into groundwater observed at the downgradient wells.

Groundwater Gradients and Transport

Water-table and potentiometric contours and the extent of saturated alluvium during three time periods (1987, 1993, and winter 2013–14) are shown in figure 12. The 1987 water-table contours are from Peter and others (1987) and represent an early monitoring period shortly after the landfill’s liquid waste lagoons had dried. In 1987, water was still present in the fire water storage pond, which stored water from the San Juan River that was used for fire training that took place at the refinery. This pond was emptied shortly after 1988 (Geoscience Consultants, Ltd., 1988). A limited number of wells were available to create the 1987 contours; many monitoring wells were installed after 1987 for the landfill’s remedial investigation (Roy F. Weston, Inc., 1995). The sparsity of data is reflected in the number of contours that are approximately located (fig. 12). The 1993 water-table and bedrock aquifer potentiometric contours are from the landfill’s remedial investigation (Roy F. Weston, Inc., 1995). Bedrock aquifer potentiometric contours are only available in 1993. The 1993 contours are from several years after the landfill’s liquid waste lagoons had dried, at a time when chloride concentrations had decreased from initial elevated values at GBR-17, GBR-32, GBR-48, and GBR-49. Chloride data are not available at BLM wells in 1993. The 1993 contours benefited from the most data because a large number of wells installed for the remedial investigation were available (many were abandoned prior to the winter 2013–14 contour period). This relatively rich data availability in 1993 can be seen in the definition in the saturated alluvium boundary. Winter 2013–14 water-table contours were created using alluvial groundwater elevation data from table 5, which were calculated using the groundwater-monitoring data in the LAGBRD and updated land-surface elevation data in table 1. Groundwater elevation from wells screened across part of the bedrock aquifer was not used for creating the alluvial contours but is labeled in figure 12. The 2013–14 period was selected for contouring because during this time groundwater elevations were rising at all monitoring wells at the site, and dissolved manganese concentrations spiked at BLM-62, BLM-77, and BLM-80. Wells used to create these contours are shown in figure 12. No alluvial wells were dry during that period, so the extent of saturated alluvium could not be estimated.

Though data scarcity around the fire water storage pond led to many of the 1987 contours being approximated, these contours suggest that the extent of mounding related to the pond reached GBR wells directly downgradient from the landfill (GBR-32, GBR-48, and GBR-49). These wells exhibited high chloride prior to 1991, suggesting that some early chloride concentrations in this area may, in fact, have been diluted by this relatively fresh source (dissolved chloride concentrations between 1961 and 2021 range from 0.8 to 6.0 mg/L at the nearby USGS site 09357000, San Juan River at Bloomfield, New Mexico; USGS, 2023). Groundwater mounding in relation to the landfill’s liquid waste lagoons is not evident in any of the periods of groundwater contours shown in figure 12, though the monitoring data density was likely sufficient to discern such mounds only in the 1993 period.

To assess advective leachate transport in groundwater relative to other mechanisms of leachate transport at the site (noted in the “Conceptual Site Model” section), groundwater transport times from the landfill’s liquid waste lagoons to BLM-62 and GBR-17 have been calculated. BLM-62 was selected as a well near the landfill that exhibited mixing with landfill leachate. Furthermore, BLM-62 is in close proximity to other wells near the landfill’s southern boundary (particularly BLM-77, BLM-80, GBR-32, GBR-48, and GBR-49), and these wells may, therefore, share similar groundwater transport times to the lagoons. Because GBR-17 is farthest from the landfill, it was selected as the monitoring well. The range of average linear groundwater velocity at the site, calculated using the average hydraulic gradient (see “Methods” section) from the three alluvial surfaces in figure 12, is 77 to 230 ft per year. Groundwater traveling from the northernmost lagoon would take 6–19 years to arrive at BLM-62 and 10–30 years to arrive at GBR-17. Groundwater traveling from the southernmost lagoon would take 2–7 years to arrive at BLM-62 and 6–17 years to arrive at GBR-17.

Despite these wide ranges of groundwater transport time estimates, it is possible to infer that groundwater transporting landfill leachate could have reached BLM-62 and GBR-17 by the start of groundwater monitoring at GBR-17 in 1986 (liquid waste lagoons were present at the landfill starting in approximately 1979; Roy F. Weston, Inc., 1995). This could explain the high chloride concentrations at that well prior to 1990. By around 1991, about 5 years after the liquid waste lagoons had dried, chloride at this well had decreased substantially, suggesting dilution with relatively fresh groundwater after the leachate supply diminished. This indicates that groundwater transport between the liquid waste lagoons and GBR-17 is relatively rapid because the dilution of leachate happens around the low end of the transport time range. No chloride data are available for BLM-62 until 1990 (and chloride concentrations at this well were low until after the capillary barrier and surface-water runoff controls were installed in 2005), though groundwater at nearby GBR-32, GBR-48, and GBR-49 exhibited a similar pattern of high initial chloride concentrations that tapered by 1991. However, unlike at GBR-17, groundwater at GBR-32, GBR-48, and GBR-49 continued to remain higher than background concentrations. Chloride concentrations did not stabilize at lower values until after the capillary barrier’s installation. Differences in the timing of chloride concentrations between the near and far locations could be attributed to local low conductivity units retarding groundwater flow and inhibiting dilution with relatively fresh water at wells near the landfill. The relatively high chloride concentrations at wells on the landfill’s southern boundary prior to roughly 2005 may also be the result of leachate stored in the vadose zone being mobilized into groundwater, a process that is apparently inhibited by the landfill’s capillary barrier.

Advective groundwater transport of landfill leachate does not explain the presence of leachate observed at BLM-62, BLM-77, or BLM-80 after the landfill’s capillary barrier and surface-water runoff control installation in 2005. At the upper range of groundwater transport times, a plume of leachate traveling from the furthest, northernmost liquid waste lagoon would have reached these locations around 2004. Leachate from nearer lagoons would have arrived earlier. Evidence of groundwater mixing with leachate at BLM-62 persists until 2020 in the latest records available in this dataset, which show generally increasing chloride concentrations between 2016 and 2020. Leachate transported advectively from the landfill’s liquid waste lagoons would have likely arrived earlier at this well. Groundwater mixing with leachate near BLM-62, BLM-77, or BLM-80 is more likely to be controlled by local conductivity, and mobilization of vadose-zone leachate into groundwater is likely enhanced by surface-water runoff controls in this area.