Groundwater Forensics to Evaluate Molybdenum Concentrations Near a CCR Landfill

Transcription

1 15 World of Coal Ash (WOCA) Conference in Nasvhille, TN - May 5-7, 15 Groundwater Forensics to Evaluate Molybdenum Concentrations Near a CCR Landfill Bruce R. Hensel 1, Thomas J. Jansen 2, Timothy C. Muehlfeld 2 1 Natural Resource Technology, Inc., 234 W Florida St. Milwaukee, WI 534 * ; 2 We Energies, 333 W. Everett, Milwaukee, WI 533 CONFERENCE: 15 World of Coal Ash ( KEYWORDS: molybdenum, isotopes, groundwater, coal combustion residuals, CCR BACKGROUND Identification of molybdenum concentrations at levels higher than state standards near a coal fired power plant triggered a series of investigations to determine whether or not operations at the power plant were the source of molybdenum. The molybdenum concentrations of concern were initially detected in off-site water supply wells that draw groundwater from a dolomite bedrock aquifer. Molybdenum is a common constituent in coal ash leachate, and is present in leachate samples from leachate head wells or the leachate collection system of the coal combustion residual (CCR) landfills of the subject property. One of these landfills (CAL) is lined with a leachate collection system, and the other (OCS) does not have an engineered liner, but is underlain by more than 30-feet of clay with very low hydraulic conductivity. Conversely, published nation-wide data show molybdenum typically has low naturally occurring concentrations in groundwater (Table 1). These initial observations caused the state to explore whether or not the fly ash landfills were a potential source. * Currently with the Electric Power Research Institute, 34 Hillview Ave., Palo Alto, CA 94304

2 Table 1. Molybdenum Concentrations in US Aquifers[1] Nationwide Carbonate Aquifers Glacial Aquifers Range of values (10 th to 90 th Percentile) Percent of samples with concentration higher than ug/l 0.13 to 8.0 <0.01 to to % 0.0% 2.0% INVESTIGATION METHODS In response to these findings and in light of a revised state groundwater standard for molybdenum, We Energies and the Wisconsin Department of Natural Resources (WDNR) performed independent investigations to characterize the molybdenum in groundwater. The We Energies investigations focused on groundwater quality and groundwater flow. This investigation included samples from monitoring wells and off-site water supply wells within approximately one mile of the CCR landfills. WDNR collected groundwater samples from a wider area (up to 5 miles from the CCR landfills) for tritium age-dating and isotope analysis. The WDNR samples were obtained from We Energies monitoring wells, monitoring wells at other landfills in the area, and off-site water supply wells. WDNR tritium and isotope sample collection and analysis are documented in [2]. Combined, these data enabled a three-phased approach to groundwater forensics that could be used to determine whether or not the CCR landfills were the source of molybdenum. The three phased approach to groundwater forensics consisted of: Phase I Hydrogeology Groundwater flow Molybdenum distribution Phase II Major ion chemistry Distribution of other CCR indicators Phase III Tritium age-dating Boron isotopes Strontium isotopes Molybdenum isotopes In many cases, the first or the first two phases are sufficient to evaluate potential sources of groundwater contamination. The tritium and isotope data of Phase III can be

3 used as additional evidence in the event of any ambiguities in the results, and are included here because they were available and applicable. PHASE I RESULTS Geology in the area consists of more than 150 feet of glacial deposits overlying a dolomite bedrock aquifer. The glacial deposits are largely composed of silt- and clayrich till. In some areas, there are discontinuous horizontal sand lenses within the glacial deposits. Most of these sand lenses are small; although one of these sand lenses, referred to as the intermediate sand, occurs beneath the northern portion of the power plant property and extends both east (downgradient) and west (upgradient) of the property. The underlying dolomite is regional in extent, and feet thick in the area. Groundwater flow in the dolomite is through fractures, which provide sufficient water for domestic water supplies. Vertical hydraulic gradients within the glacial deposits are downward toward the bedrock aquifer. Groundwater flow in both the intermediate sand and dolomite aquifer is to the east (Figures 1 and 2). Figure 1. Lines of groundwater equipotential in the intermediate sand indicating eastward flow.

4 Figure 2. Lines of groundwater equipotential in the dolomite aquifer indicating eastward flow. Based on site-specific hydrogeologic observations, two Site Conceptual Models (SCM) were developed (Figure 3). Simplified summaries for these models are: The first SCM is for areas where the intermediate sand is present, beneath OCS landfill. Water at the land surface slowly percolates downward through the upper glacial till to the intermediate sand. Once water reaches the intermediate sand it flows eastward. Vertical hydraulic gradient information indicate that vertical percolation is also possible from the intermediate sand to the dolomite bedrock aquifer, where it will then flow eastward. The second SCM is for areas where the intermediate sand is not present, beneath CAL landfill. Water at the land surface slowly percolates downward to the dolomite bedrock aquifer. Once water reaches the bedrock, it flows eastward.

5 Figure 3. Site Conceptual Models of groundwater flow in the vicinity of the power plant. Groundwater quality is monitored near the landfills on the power plant property. Monitoring occurs at multiple depths: within the glacial till, in the intermediate sand where present, and in the dolomite bedrock. Molybdenum concentrations in the monitoring wells screened in the glacial deposits are mostly lower than concentrations observed in the underlying dolomite (Figure 4). This pattern of increasing molybdenum concentration with depth is observed at the two monitoring well nests with wells screened in shallow till, the intermediate sand, and the dolomite aquifer (Figure 5) Molybdenum Concentrations (mg/l) Bedrock Till Figure 4. Box-whisker plot showing that molybdenum concentrations in the underlying dolomite bedrock are mostly higher than in the till surrounding the power plant landfills. The black diamonds represent statistical outliers.

6 700 W39A W37A Elevation (ft-msl) W37C W39B Till Bedrock W37D Till Bedrock W39C Figure 5. Molybdenum concentration profiles showing increasing concentration as a function of depth. The W37 well nest is immediately west of the OCS landfill (between the landfill and the offsite water supply wells where molybdenum concentrations were greater than ug/l). The W39 nest is east (downgradient) of the landfill. PHASE II RESULTS Pieper diagrams were prepared to compare the general water chemistry between 1) coal ash leachate, 2) background groundwater, and 3) downgradient groundwater for the intermediate sand and dolomite aquifer. These formations were plotted because the SCM identified them as the strata where lateral migration would occur. Major ion chemistry for the intermediate sand shows three distinct groupings, one for background groundwater, one for downgradient groundwater, and one for leachate (Figure 6). If the downgradient grouping plotted between background and leachate, then it would provide evidence of a release and impacts to groundwater, but in this case it does not lie between the end members and simply indicates that groundwater quality in the intermediate sand is variable by nature. The Pieper diagram for monitoring wells in the bedrock shows the background and downgradient water quality overlapping, and does not indicate mixing with coal ash leachate (Figure 7).

7 Piper Diagram Bedrock Groundwater Unit-background Bedrock Groundwater Unit CAL Leachate (median) OCS Leachate (median) Sulfate (SO4) + Chloride (Cl) Calcium (Ca) + Magnesium (Mg) X OCS Leachate (samples) Magnesium (Mg) Mg SO 4 Sodium (Na) + Potassium (K) Carbonate (CO3) + Bicarbonate (HCO3) Sulfate (SO4) Ca Na+K HCO 3+CO 3 Cl Calcium (Ca) Chloride (Cl) C A T I O N S %meq/l A N I O N S Figure 6. Major ion chemistry for the intermediate sand. Piper Diagram Bedrock Groundwater Unit-background Bedrock Groundwater Unit CAL Leachate (median) OCS Leachate (median) Sulfate (SO4) + Chloride (Cl) Calcium (Ca) + Magnesium (Mg) X OCS Leachate (samples) Magnesium (Mg) Mg SO 4 Sodium (Na) + Potassium (K) Carbonate (CO3) + Bicarbonate (HCO3) Sulfate (SO4) Ca Na+K HCO 3+CO 3 Cl Calcium (Ca) Chloride (Cl) C A T I O N S %meq/l A N I O N S Figure 7. Major ion chemistry for the dolomite aquifer.

8 Water quality data for boron, a primary indicator constituent for coal ash leachate[3], was also reviewed to determine if there had been a release to either of the intermediate sand or dolomite aquifer units. Boron concentrations in site monitoring wells plotted within the ranges of background concentrations, providing no evidence of a release (Figure 8). Background Compliance Background Compliance Background Figure 8. Boron concentrations for compliance monitoring wells screened in the intermediate sand (top) and bedrock (bottom), compared to concentrations in background monitoring wells

9 PHASE III RESULTS The Wisconsin Department of Natural Resources performed a separate investigation in which groundwater and leachate samples were collected from the coal ash landfills, and groundwater samples were collected from off-site water supply wells and monitoring wells at other facilities. Samples were collected as far as 5 miles from the power plant. The samples were analyzed for tritium, and for isotopes of boron, strontium, and molybdenum. Sampling, analysis methods, and results are discussed in [2]. The discussion that follows uses only the data from [2]. Tritium provides a semi-qualitative measure of the period when water entered the subsurface as recharge, because atomic testing of nuclear warheads beginning in the early 1950s introduced tritium to the atmosphere. As a result, detectable levels of tritium in groundwater indicate that it recharged since the 1950s. Conversely, if tritium is not detected, then it indicates that groundwater either recharged before the early 1950s or that the water is a mixture with a large volume recharging before the early 1950s and a small volume recharging after the 1950s. The power plant began operation in the mid- 1950s, so any impact on groundwater quality caused by power plant operations would be reflected in relatively recent groundwater with detectable tritium concentrations. Figure 9 illustrates the relationship between molybdenum concentrations and tritium. The highest molybdenum concentrations are in wells with tritium concentrations that are below detection limits, indicating recharge prior to power plant operation or mixing with much larger volumes of groundwater. 1.0 Tritium vs Molybdenum Molybdenum (ug/l) Bedrock > 1 mile Bedrock Glacial Tritium (TU) Figure 9. Comparison of molybdenum concentrations to tritium showing that the highest molybdenum concentrations occur in groundwater with low to non-detectable tritium levels. Data are from [2].

10 Boron and strontium isotopes have been used in CCR applications by the United States Geological Survey and others [e.g., 3, 4, 5]. Boron isotopes are particularly useful for investigations at CCR facilities. Boron has two naturally-occurring stable isotopes, 11 B with a natural abundance of.1%, and 10 B with a natural abundance of 19.9%. Fly ash leachate tends to have a lower percentage of the 11 B isotope than the NIST standard, while sea water and many groundwaters tend to be isotopically heavy relative to the NIST standard. As a result, if a CCR leachate that is isotopically light mixes with groundwater that is isotopically heavy, then the resulting mixture will have an isotopic signature that falls between background groundwater and leachate both in terms of boron concentration and the relative percentage of boron isotopes. Figure 10 compares boron concentration to boron isotope ratio (as δ11b). Boron concentrations in coal ash leachate collected at the CCR landfills has high concentration and is isotopically light. Therefore, if leachate was mixing with groundwater, then the data would include a subset of results where the water becomes isotopically lighter (decreasing δ11b) as concentration increases (i.e., an inverse trend). The data from [2] indicate that monitoring wells and water supply wells exhibit a trend where the water becomes isotopically heavier (increasing δ11b) as boron concentration increases (i.e., a direct trend) Boron (ug/l) Leachate Mix Glacial Bedrock Bedrock (> 1 mile) δ 11 B (per mil) Figure 10. Comparison of boron concentration to isotope ratios. Data are from [2]. NIST (National Institute of Standards and Technology) 951

11 Figure 11 presents strontium concentrations versus isotope ratios. The CCR landfill leachate has relatively high concentration and is isotopically heavy with respect to strontium, and a direct trend would be anticipated to provide evidence that water was affected by coal ash leachate. Conversely, the data from [2] show an inverse trend where water becomes isotopically heavier as strontium concentration decreases Strontium (ug/l) Leachate Mix Glacial Bedrock Bedrock (> 1 mile) Sr/ 86 Sr Figure 11. Comparison of strontium concentration to isotope ratios. Data are from [2]. Molybdenum isotopes have not been vetted in the literature for CCR applications as have boron and strontium isotopes. Molybdenum concentration is plotted versus isotope ratios in Figure 12. This plot shows a generally flat trend in isotope ratio as concentration increases for the monitoring and water supply wells. Molybdenum concentrations in CCR landfill leachate are higher than in groundwater, while the leachate isotope ratio is not distinct and falls within the range of groundwater ratios. Two water supply wells (in bedrock) have molybdenum concentrations greater than 100 ug/l and molybdenum isotope ratios similar to CCR leachate; however, one of these wells is five miles upgradient from the power plant, and since there is no reasonable mechanism for leachate migration over that distance, it indicates that molybdenum can have a concentration greater than 100 ug/l in groundwater without being mixed with CCR leachate.

12 Molybdenum (ug/l) Leachate Glacial Bedrock Bedrock (>1 mi) δ 98/95 Mo (per mil) Figure 12. Comparison of molybdenum concentration to isotope ratios. Data are from [2]. CONCLUSIONS Groundwater forensics, performed in three phases as outlined for this example case, provides a weight of evidence approach for evaluating whether or not a CCR landfill is the source of groundwater quality issues. A phased approach is recommended because in many cases it may be possible to make necessary and appropriate conclusions without performing all three phases. In this case, Phase I established a Site Conceptual Model (SCM) and was based on standard hydrogeologic and groundwater quality data, Phase II was based on the major ion chemistry of the migration zones identified by the Site Conceptual Model, and on CCR indicator constituent concentrations within the migration zones, and Phase III used tritium age dating and analysis of boron, strontium, and molybdenum isotopes. Results indicated: Phase I o The CCR landfills are underlain by a thick sequence of glacial till with low hydraulic conductivity, and are not hydraulically connected to the dolomite aquifer. While slow vertical percolation through the till is possible, the

13 volume of water percolating through the till will be small compared to the volume flowing through the dolomite aquifer. o Groundwater flow direction in both horizontal migration zones identified by the SCM are eastward, and not toward the water supply wells with high molybdenum concentrations. o Molybdenum concentrations near the CCR landfills increases with depth, whereas a decreasing vertical profile would be anticipated if the CCR landfills were the source. Phase II o Major ion concentrations provided no evidence of groundwater mixing with CCR leachate o Concentrations of boron, a primary indicator constituent for CCR leachate, was within background levels within both horizontal migration zones identified by the SCM. Phase III o The highest molybdenum concentrations occurred in water samples with non-detectable to very low tritium levels, indicating that the samples either recharged prior to power plant operation or represent small volumes of recent recharge with much larger volumes of groundwater that recharged prior to the 1950s. o Boron and strontium isotope analyses indicated that groundwater samples with the highest molybdenum concentrations were isotopically distinct from CCR leachate, and provided no evidence of mixing. o Molybdenum isotope analyses were inclusive. Combined, this weight of evidence approach indicates that the CCR landfills were not the source of molybdenum observed in water supply wells near the power plant. If another source of molybdenum had been identified in the region, then this investigation might have stopped at Phase I or Phase II. However, no anthropogenic source has been identified, even after Phase III, suggesting by elimination that the molybdenum is naturally-occurring, although additional research would be needed to prove that hypothesis. REFERENCES [1] United States Geological Survey (USGS), Trace Elements and Radon in Groundwater Across the United States, , Scientific Investigations Report , 11, 131 p. [2] Wisconsin Department of Natural Resources (WDNR), Caledonia Groundwater Molybdenum Investigation, Southeast Wisconsin, PUB-WA 1625, 13, 101 p. [3] Electric Power Research Institute (EPRI), Groundwater Quality Signatures for Assessing Potential Impacts from Coal Combustion Product Leachate, , 12, 162 p.

14 [4] United States Geological Survey (USGS), Evaluation of Ground-Water and Boron Sources by Use of Boron Stable-Isotope Ratios, Tritium, And Selected Water-Chemistry Constituents Near Beverly Shores, Northwestern Indiana, 04, Scientific Investigations Report Series , 07, 46 p. [5] Davidson, G.R. and R.L. Bassett, Applications of Boron Isotopes for Identifying Contaminants such as Fly Ash Leachate in Groundwater, Environmental Science and Technology, 27, 1993, pp