Article pubs.acs.org/est
Naturally Occurring Contamination in the Mancos Shale Stan J. Morrison,*,† Craig S. Goodknight,† Aaron D. Tigar,† Richard P. Bush,§ and April Gil§ †
Environmental Sciences Laboratory, 2597 Legacy Way, Grand Junction, Colorado 81503, United States, Operated by SM Stoller for the U.S. Department of Energy Office of Legacy Management § U.S. Department of Energy Office of Legacy Management, 2597 Legacy Way, Grand Junction, Colorado 81503, United States S Supporting Information *
ABSTRACT: Some uranium mill tailings disposal cells were constructed on dark-gray shale of the Upper Cretaceous Mancos Shale. Shale of this formation contains contaminants similar to those in mill tailings. To establish the contributions derived from the Mancos, we sampled 51 locations in Colorado, New Mexico, and Utah. Many of the groundwater samples were saline with nitrate, selenium, and uranium concentrations commonly exceeding 250 000, 1000, and 200 μg/L, respectively. Higher concentrations were limited to groundwater associated with shale beds, but were not correlated with geographic area, stratigraphic position, or source of water. The elevated concentrations suggest that naturally occurring contamination should be considered when evaluating groundwater cleanup levels. At several locations, seep water was yellow or red, caused in part by dissolved organic carbon concentrations up to 280 mg/L. Most seeps had 234U to 238U activity ratios greater than 2, indicating preferential leaching of 234U. Seeps were slightly enriched in 18 O relative to the meteoric water line, indicating limited evaporation. Conceptually, major ion chemical reactions are dominated by calcite dissolution following proton release from pyrite oxidation and subsequent exchange by calcium for sodium residing on clay mineral exchange sites. Contaminants are likely released from organic matter and mineral surfaces during weathering.
■
INTRODUCTION Chemicals leached from the Upper Cretaceous Mancos Shale cause contamination to groundwater and surface water. The potential for release of contaminants to groundwater is large because Mancos Shale is exposed over an area of about 3200 km2 in the semiarid to arid western United States (Figure 1). Contaminants of major interest that are released from the Mancos include nitrate, selenium, sulfate, and uranium. Contamination of surface drainages by salts and selenium released from the Mancos is well established;1−20 however, little is known about concentrations in groundwater and few data are available for uranium.21 Because Mancos Shale contains constituents similar to those in uranium mill tailings, differentiating the natural (Mancos-derived) component from the anthropogenic (mill-derived) component22 in proximal groundwater is problematic. It is important to understand natural contamination to ensure application of realistic cleanup standards and to properly evaluate the progress of groundwater remediation. Mancos Shale was deposited during Late Cretaceous time in offshore and open-marine environments of the epicontinental Western Interior Seaway. The exposed formation is up to 1265 m thick in Colorado and Utah.23 Mancos Shale consists of gray © 2012 American Chemical Society
siltstone and shale with minor limestone, marlstone, bentonite, and sandstone beds.24 Weathering causes oxidation of pyrite and loss of organic matter,25 resulting in a change in color from dark gray to yellowish gray and release of contaminants. The main goal of this study was to determine if the elevated contaminant concentrations we had measured in a few Mancos groundwater samples prior to this study were localized or were a regional feature of Mancos groundwater. The variability and distribution of natural contamination in Mancos Shale groundwater was determined by measuring chemical concentrations in groundwater over much of its outcrop area. Uranium was found to be a natural contaminant and we measured uranium isotopic activity ratios to help understand its origin. Another study goal was to formulate a basin-wide conceptual model of potential release mechanisms and apply a reaction progress model. While this work provides a regional framework of natural groundwater contamination in the Mancos Shale, Received: Revised: Accepted: Published: 1379
September 13, 2011 January 4, 2012 January 6, 2012 January 6, 2012 dx.doi.org/10.1021/es203211z | Environ. Sci. Technol. 2012, 46, 1379−1387
Environmental Science & Technology
Article
Figure 1. Sampling regions (rectangles), areas (circles), and locations (dots). Mancos Shale outcrop is shaded green.
further work is needed to better define the exact nature of the chemical release mechanisms.
■
beds or from a veneer of gravel alluvium underlain by bedrock. Seeps issued from shale at all other locations. Groundwater was also sampled from 15 deep (∼27 m) Mancos wells at DOE’s uranium mill tailings disposal sites at Crescent Junction and Shiprock (Figure 1). Data reported by Golder Associates28 from a deep well in the Delta Reservoir area were included in our analyses of deep Mancos groundwater. Drill core from 10 borings at the Crescent Junction site provided mineralogic data on unweathered Mancos.
SAMPLING LOCATIONS
Samples were collected from 51 locations identified using the U.S. Geological Survey (USGS) National Water Information Service,26 Google Earth, DOE’s National Uranium Resource Evaluation program,27 and field reconnaissance.21 Coordinates and descriptions of sampling locations are provided in the Supporting Information. Locations were grouped into six sampling regions named after the nearby towns of Grand Junction, Green River, Hanksville, Montrose, Price, and Shiprock (Figure 1). Three areas with well-defined hydrologic systems are discussed in detail: Daly Reservoir, Delta Reservoir, and Loutsenhizer Arroyo (Figure 1). Seeps were sampled at 45 locations and surface water at six. Surface locations were water bodies that provided the source of water for nearby groundwater seeps. Many seeps were conspicuous by the presence of white salt coatings (efflorescence) on the ground surface. At nine locations, referred to as sandstone locations, groundwater issued either from sandstone
■
MATERIALS AND METHODS Sampling. A single round of sampling was conducted during November and December 2010.21 This period is considered base flow conditions, a time when irrigation and runoff are at a minimum.20 At all groundwater sampling locations, an effort was made to locate the farthest upgradient point where seepage emerges from bedrock. Flow from most seeps was less than 1 L/min. Groundwater was collected either through a 0.3−1 m long, by 5 cm diameter, slotted casing (sampling pipe) or a hand-augured hole. The water was pumped from the “well” with a peristaltic pump. Specific conductivity, temperature, dissolved oxygen, oxidation−reduc1380
dx.doi.org/10.1021/es203211z | Environ. Sci. Technol. 2012, 46, 1379−1387
Environmental Science & Technology
Article
Figure 2. Concentrations of selected chemicals in groundwater samples. Only those locations that had analyses of all five constituents are shown. Values that exceed the graph maxima are indicated. Red bars are samples of groundwater from sandstone, all others are from shale. Arrows at top indicate primary drinking water standards for uranium, selenium, and nitrate;33 and secondary drinking water standard for sulfate; DOC has no standard. For each region, the locations (bars) are in order of increasing specific conductivity.
potassium were analyzed by flame atomic absorption spectroscopy. Uranium was analyzed by laser-induced kinetic phosphorescence analysis (KPA). Dissolved organic carbon (DOC) was analyzed by colorimetry on filtered samples, after acidifying to remove inorganic carbon and heating to 105 °C with persulfate. Samples were analyzed for 234U and 238U by alpha spectrometry; and for arsenic, selenium, and vanadium by inductively coupled plasma mass spectrometry. Uranium concentrations determined from alpha spectrometry were consistent with those determined by KPA. δ2H was determined by hydrogen equilibration29 and concentrations were estimated from activity values.30 δ18O was determined by carbon dioxide equilibration31 and activity values were converted to concentrations using salinity corrections.32
tion potential, and pH were measured either in a flow-through cell or directly in the well. Care was taken to ensure that groundwater was flowing by pumping until a steady flow rate was achieved. The proximity of some seeps to the ground surface suggested that some upward capillary flow and evaporation was possible. Surface water samples were collected by submerging a container about 0.3 m into the water. Samples for chemical analyses were field-filtered into Nalgene bottles through 0.45 μm in-line filters using a peristaltic pump. Samples for anions, δ18O, and δ2H analyses were placed on ice. All other samples were preserved with concentrated nitric acid to maintain pH less than 2. Analysis. Alkalinity was determined in the field on filtered samples by titration with sulfuric acid. Color analyses were performed on filtered samples by light absorbance at 465 nm and normalized to a platinum−cobalt standard. Sulfate, chloride, and nitrate were analyzed by ion chromatography (IC). Because some IC concentrations were surprisingly high, samples were analyzed at least twice at different dilutions and often in triplicate. Standard additions, run on at least every fifth sample, indicated minimal matrix interference. To minimize biodegradation, samples were analyzed for nitrate within 48 h of collection. Calcium, magnesium, sodium, iron, and
■
RESULTS Concentrations of nitrate, selenium, and uranium in Mancos groundwater collected from shale seeps commonly exceeded drinking water standards by a factor of 10 or more (Figure 2). In general, concentrations of nitrate, selenium, sulfate, and uranium are correlated suggesting a common source. Salinity was lower in groundwater seeping from Mancos sandstone as indicated by geometric means of specific conductivity of 2362 1381
dx.doi.org/10.1021/es203211z | Environ. Sci. Technol. 2012, 46, 1379−1387
Environmental Science & Technology
Article
and 11 970 μS/cm in sandstone and shale, respectively. Most seeps had pH values between 6.1 and 7.5. Groundwater from most seeps had a sodium sulfate composition (Figure 3).
exceeded 2000 mg/L in seeps in the Grand Junction, Montrose, and Shiprock regions (Figure 2). Uranium concentrations in seeps ranged from 0.20 to 822 μg/L (Figure 2). Seeps located throughout the sampling domain had uranium concentrations more than 100 μg/L. Uranium concentrations in shale groundwater were higher than those in sandstones, as indicated by geometric means of 83.4 and 7.3 μg/L, respectively. Uranium activity ratios (ARs) of 234U to 238U ranged from 1.22 to 4.08 (Figure 4). ARs in seeps from shale and sandstone
Figure 3. Piper diagram of Mancos groundwater and surface water (black dots = seeps, blue stars = surface water, red dots = deep Mancos wells).
Groundwater from deep Mancos wells at DOE’s uranium mill tailings disposal sites at Crescent Junction and Shiprock, and from a deep Mancos well in the Delta Reservoir area28 had sodium chloride compositions (Figure 3). 18 O and 2H values were measured to help determine evaporation effects (see Supporting Information). Most of the groundwater samples plot to the right of the local meteoric water line (LMWL) for the Colorado Plateau,34 indicating that evaporation had occurred in some samples. Most samples plotted close to the LMWL, but samples known to have been partially evaporated plotted farthest to the right. From these data, we surmise that most of the evaporation observed in carefully collected (those that avoided recent surface exposure) groundwater samples occurred during recharge; therefore, the samples are representative of in situ Mancos groundwater conditions. Only one sample of seep water exceeded the arsenic drinking water standard and all except one sample had vanadium concentrations less than 3.8 μg/L. Because concentrations of arsenic and vanadium in Mancos seeps were low, they are not considered natural contaminants. DOC concentrations in Mancos seeps ranged from 2.9 to 280 mg/L (Figure 2). DOC values exceeded 100 mg/L in the Grand Junction, Montrose, and Shiprock regions; and exceeded 50 mg/L in the Price region. Mancos seepage was commonly yellow or red. Sample color, as measured by light absorbance, correlated positively with DOC concentration suggesting DOC as a cause of the coloration. Selenium concentrations in seep samples ranged from 0.14 to 4700 μg/L with a geometric mean of 52 μg/L (Figure 2). Locations in the Grand Junction, Green River, Montrose, and Shiprock regions had selenium concentrations more than 1000 μg/L. Geometric means for selenium were 3.0 and 120 μg/L for sandstone and shale, respectively. Nitrate concentrations
Figure 4. AR versus uranium concentration. Black dots and open squares represent groundwater and surface water samples, respectively.
were similar, with geometric means of 2.06 and 1.95, respectively. Daly Reservoir, Delta Reservoir, and Loutsenhizer Arroyo. These three areas were selected for further discussion because their groundwater systems are well understood.21 Daly Reservoir water infiltrated into Mancos bedrock and emerged at two seeps located 70 and 110 m downgradient. The specific conductivity of water in Daly Reservoir was 1720 μS/cm but increased to 9187 and 15 380 μS/cm at the seeps. Concentrations of nitrate, selenium, sulfate, and uranium also increased owing to interaction with the Mancos (Table 1). At Delta Reservoir, pristine surface water (specific conductivity of 114 μS/cm) infiltrated Mancos Shale and emerged at a line of seeps about 420 m downgradient. The seeps have specific conductivities more than 22 000 μS/cm and nitrate, selenium, sulfate, and uranium values as high as 534, 0.95, 18 500, and 0.137 mg/L, respectively (Table 1). At Loutsenhizer Arroyo, water infiltrated the Mancos at the upland Bostwick Canal and emerged as saline seeps with specific conductivities up to 22130 μS/cm and concentrations of nitrate, selenium, sulfate, and uranium up to 3360, 4.7, 14900, and 0.135 mg/L, respectively (Table 1). These three areas illustrate the potential for groundwater to acquire elevated concentrations of natural contaminants during relatively short travel paths through Mancos Shale. Data from the Delta Reservoir area were used to estimate a flow rate through shale beds of the Mancos. Water was conveyed through a 1.6 km long pipe from Delta Reservoir to 1382
dx.doi.org/10.1021/es203211z | Environ. Sci. Technol. 2012, 46, 1379−1387
Environmental Science & Technology
Article
Table 1. Concentrations of Constituents at Daly Reservoir, Delta Reservoir, and Loutsenhizer Arroyo Areas
a
area
location typea
specific cond. (μS/cm)
NO3 (mg/L)
Se (μg/L)
SO4 (mg/L)
U (μg/L)
Daly Reservoir Daly Reservoir Daly Reservoir Delta Reservoir Delta Reservoir Delta Reservoir Loutsenhizer Arroyo Loutsenhizer Arroyo Loutsenhizer Arroyo Loutsenhizer Arroyo Loutsenhizer Arroyo
surface seep (70 m) seep (110 m) surface seep (420 m) seep (420 m) surface seep (1,500 m) seep (900 m) seep (250 m) seep (400 m)
1720 9187 15 380 114 27 250 22 840 197 15 330 22 130 8844 5190
3.7 150 449
