Predicting Post-mining Water Quality of Surface Mines by Acid-base Accounting

Acid-Base Accounting (ABA) is an analytical procedure that provides values to help assess the acidor alkaline-producing potential of overburden rocks prior to coal mining. This procedure was developed at West Virginia University in the late 1960s. After the passage of laws requiring an assessment of surface mining on water quality, ABA became the preferred method to predict post-mining water quality, and permitting decisions for surface mines are based on the values determined by ABA. In order to predict the post-mining water quality, the amount of acid-producing rock is compared to the amount of alkalineproducing rock, and a prediction of the water quality at the site (whether acid or alkaline) is obtained. We gathered geologic and geographic data for 56 mined sites in West Virginia, which allowed us to estimate total overburden amounts, and values were determined for maximum potential acidity (MPA), neutralization potential (NP), net neutralization potential (NNP), and NP/MPA ratios for each site based on ABA. These values were correlated to post-mining water quality from springs or seeps on the mined property. Overburden mass was determined by three methods, with the method used by Pennsylvania researchers showing the most accurate results to actual overburden mass. A poor relationship existed between MPA and post-mining water quality, NP was intermediate, and NNP and the NP/MPA ratio showed the best prediction accuracy. In this study, NNP and the NP/MPA ratio gave identical water quality prediction results. So using NP/MPA ratios, values were separated into categories: <1 should produce acid drainage, between 1 and 2 can produce either acid or alkaline water conditions, and >2 should produce alkaline water. On our 56 sites, NP/MPA ratios varied from 0.1 to 31, and six sites (11%) did not fit the expected pattern using this category approach. Two sites with ratios <1 did not produce acid drainage as predicted (the drainage was neutral), and four sites with a ratio >2 produced acid drainage when they should not have. These latter four sites were either mined very slowly, had inaccurate ABA data, received water from an adjacent underground mine, or included some mining practice that degraded the water. In general, an NP/MPA ratio of <1 produced mostly acid drainage sites, between 1-2 produced mostly alkaline drainage sites, while NP/MPA ratios of >2 produced alkaline drainage with a few exceptions. Using the NP/MPA ratio, ABA is a good tool to assess overburden quality before mining and to predict post-mining drainage quality after mining, and was correct in 50 out of 56 cases (89%). Introduction Most coal mining state regulatory agencies began requiring the prediction of acidand alkalineproducing materials in the overburden of surface mine operations in the early 1970s. For example, the West Virginia legislature passed a law in 1971 requiring mine operators to show in the permit “the presence of any acid-producing materials, which, when present, may cause minesoils with a pH of less than 3.5 and prevent effective revegetation” (West Virginia Surface Mining and Reclamation Law, 1971). Dr. Richard M. Smith and his associates at West Virginia University, in conjunction with the existing West Virginia coal regulatory agency, began working on a procedure to identify acid-producing materials in 1965 (Perry, 1998). Throughout the late 1960s and 1970s, the procedure was refined and termed “Acid-Base Accounting” (ABA) (Smith et al., 1976). ABA was originally designed to distinguish layers in the overburden that could be used as topsoil substitutes or as hard durable rock for valley fills. But since the method identified acidand alkaline-producing materials in the overburden, this method was the first technology available to predict the quantity of acid-producing materials prior to mining (Skousen et al., 1990). Since the late 1970s and the passage of the Surface Mining Control and Reclamation Act (1977), ABA has become widely adopted as a method of overburden characterization and prediction of post-mining drainage quality (Sobek et al., 2000). ABA, as originally developed, is designed to measure neutralization potential (NP) and sulfur content of individual overburden strata. From these measurements, maximum potential acidity (MPA) and net neutralization potential (NNP) are calculated for each geologic layer from the surface of the land down to, including, and immediately underlying the coal seam. The NP is a measure of the amount of neutralizing compounds (mostly carbonates, exchangeable alkali and alkali earth cations) present in the coal and overburden. The NP is calculated from the amount of acid neutralized by the sample and is expressed in metric tons/1000 metric tons (mt/1000 mt, or parts/1000 parts) of overburden (Kania, 1998). Refinements have been made recently on the NP method to discount artifacts from slower-reacting, non-alkalineproducing carbonates, such as siderite (Skousen et al., 1997). The MPA is the maximum amount of sulfuric acid that can be produced from the oxidation of sulfur minerals in the rock material. Although acid production is associated with pyritic sulfur, the ABA procedure typically measures total sulfur because it is easier to measure than pyritic sulfur. Total sulfur is currently the most common value used to calculate MPA (Kania, 1998). The simplest and most frequently used method of total S determination is high-temperature furnace combustion (Skousen, 2000). This test results in a percent of sulfur present in the rock and is then multiplied by a constant to determine the MPA in mt/1000 mt. For the complete oxidation of pyrite and neutralization of all generated acidity, FeS2 + 2CaCO3 + 3.75 O2 + 1.5 H2O Fe(OH)3 + 2 SO4 + 2 Ca + 2 CO2 (1) 1 mol of FeS2 (64 g of sulfur) is neutralized by 2 mols of CaCO3 (200 g of CaCO3). Therefore, the constant is 31.25 and it takes 31.25 mt of CaCO3 to neutralize 1000 mt of rock containing 1% pyritic sulfur. Cravotta et al. (1990) suggested that in a closed system (such as that of a surface mine backfill) CO2 would not be driven off in the reaction, but would instead react with water to form carbonic acid as in the following reaction: FeS2 + 4CaCO3 + 3.75O2 + 3.5H2O Fe(OH)3 + 2SO4 + 4Ca + 4HCO3 (2) In this reaction, 1 mol of FeS2 is neutralized by 4 mols of CaCO3. So using this equation, 1000 mt of rock containing 1% pyritic sulfur requires 62.5 mt of CaCO3 for neutralization. Studies analyzing ABA have used both methods. The results have been mixed (Brady and Cravotta, 1992; Brady et al., 1994). In general, the 31.25 factor for overburden MPA calculation is most commonly used. After the NP and MPA are calculated using the above methods, total NNP is determined for each stratigraphic layer by subtracting the MPA from the NP. Conceptually, a positive number indicates more potentially alkaline-producing strata in the overburden and a negative number indicates more potentially acid-producing strata in the overburden. The NP/MPA ratio is simply computed by dividing the NP by the MPA. A ratio of 1 is equivalent to an NNP of 0. These values for each rock unit can then be used to identify potentially toxic materials in the overburden and can assist in planning overburden handling and placement (Skousen et al., 1987). The use of ABA data to predict post-mining water quality involves numerous assumptions: 1) all sulfur in a sample will react to form acid, 2) all material in the sample which consumes acid in the laboratory will generate alkalinity in the field, 3) pyrite oxidation rate is less than or equal to the rate of neutralization, 4) NP and %S below certain thresholds do not influence water quality. Because of these assumptions, many researchers have questioned the ability of ABA to accurately predict post-mining drainage quality (diPretoro and Rauch, 1988; Erickson and Hedin, 1988). In the initial usage of ABA as a predictor of post-mining water quality, overburden calculations were made only according to layer thickness (thickness-weighted), giving equal weight to layers at the top and bottom of the column. However, most surface mines of Appalachia have hilly to mountainous topography with horizontal strata. So, a 2-meter-thick rock stratum high in the overburden contains much less volume of material than a 2-meter-thick rock stratum low in the overburden. Therefore, this method of interpretation overestimated the amount of alkaline material high in the overburden column, and created a situation where insufficient alkaline material was available for neutralization of acidity in high sulfur rocks near the coal seam. Many sites were mined under this scenario and some subsequently generated severe, post-mining acid mine drainage due to inaccurate overburden interpretation. In the early 1980s, computer spreadsheets came into use for volume adjustment of ABA data. Once volumes of specific rock types were determined, total mass of acid, neutral, or alkaline rocks could be calculated for the entire site (mass-weighted), thereby providing one value for MPA and NP for each site (Smith and Brady, 1990). In 1988 two studies by diPretoro and Rauch (1988) and Erickson and Hedin (1988) compared massweighted ABA data with post-mining water quality. In both studies, mass-weighted values were calculated from volumes assuming a right triangle-shaped area to be mined. Although this method can only be considered an approximate technique for volume adjustment, it yielded more realistic volumes than the simple thickness-weighted approach of the past. Researchers at the Pennsylvania Department of Environmental Protection (PADEP) developed a method of volume adjustment using actual measurements of the areas to be mined (Brady et al., 1994). Due to the difficulty of measuring the surface area of each individual stratum, only the surface area of the upper and lower strata was measured. A computer spreadsheet then interpolated the area of each layer between the two, assuming a constant slope. Volumes were calculated for each interval using the surface area and the measured thickness (Smith and Brady, 1990). No volume estimation technique has ever been compared to actual amounts of overburden moved on a surface mine. Even though mass-weighting using volume adjustments gave more realistic ABA information, the question remained as to which of the ABA parameters (MPA or NP) or a combination of parameters (NNP or NP/MPA ratio) best predicted post-mining water quality. In studies by diPretoro and Rauch (1988), Erickson and Hedin (1988) and Brady et al. (1994), NP and NNP were found to be the best indicators of post-mining drainage quality (Table 1). diPretoro and Rauch (1988) showed that sites with NP >40 mt/1000 mt and NNP >30 mt/1000 mt produced net alkalinity in post-mining drainage. Erickson and Hedin (1988) found that an NNP >80 produced alkalinity and an NNP <20 typically produced acidity, but this study had no sites with NNPs between 20 and 80 mt/1000 mt. A mass-weighted study by Brady et al. (1994) resulted in NP and NNP values much lower than these earlier studies. They showed that sites with NP >21 mt/1000 mt and NNP >12 mt/1000 mt produced alkaline drainage. In addition to NP and NNP, the ratio of NP to MPA has also been used as an indicator of post-mining water quality (Perry, 1998). diPretoro and Rauch (1988) found that sites with a NP/MPA ratios of <2.4 generally resulted in acid mine drainage and sites with >2.4 usually produced alkaline drainage. Although all three studies identified the importance of NP, NNP, and NP/MPA ratio in predicting post-mining water quality, all three gave different values of NP and NNP as thresholds. Also, these values only represented trends in the relationship of NP and NNP to water quality prediction, and all sites did not fit the predictions. There are also questions as to the most reliable and accurate method of volume adjustment to obtain mass of acid and alkaline overburden materials for use in prediction. Therefore, we collected data for 56 sites, used three methods of volume adjustment to compute overburden mass for determining MPA, NP, NNP, and NP/MPA ratio values, and then compared these values with post-mining water quality. Methods Site Selection and Data Study sites were chosen to represent surface mining operations throughout the state of West Virginia (Figure 1). No special consideration was given to the surface mining area, coal seam, mining method, special handling plans, or post-mining water quality. For all sites, special handling plans were not considered and estimates were based on overburden properties from drill cores only and ABA calculations. If alkaline amendments were added, the amount of material was added to the overall NP of the site. Data for each site were obtained from mine operator files or from surface mine permits on file with the WV Department of Environmental Protection. In order to be included in the data set, each site had to meet the following four criteria. First, a detailed topographic map was necessary to show the boundaries of mining, location of core holes, coal outcrops, depth of overburden, and post-mining water sampling points. Second, a complete ABA data set was required from overburden cores drilled above or near the highest point of mining (ABA data must include depth, thickness, rock type, %S and NP for all layers down to and including the coal pavement). Third, amounts of alkaline material imported to the site were needed, if practiced. Fourth, post-mining drainage data were necessary including flow, pH, alkalinity, and acidity from seeps or springs discharging from the site. The first three criteria were relatively easy to meet, but the postmining water quality data were harder to obtain, and many additional sites were eliminated from the data set due to insufficient water data. Fifty-three of the 56 sites had water data for at least five separate sampling times and values were averaged to determine the post-mining discharge quality for each site. The other three sites had water samples taken in the fall of 2000 only, but other records and regulatory personnel were contacted to assure that these samples represented the historical water quality on these sites. Overburden Volume and Mass Calculations Volumes of each layer in the overburden were determined in three ways. The first way (Ziem) was a method modified from diPretoro and Rauch (1988) by Paul Ziemkiewicz. The cross-sectional area perpendicular to the highwall was calculated from depth and pit width, assuming the area was a right triangle. The cross-sectional area of each layer was calculated by interpolation using layer depths and the average slope. Layer volumes were calculated by multiplying layer cross-sectional area by total pit length. A second method, Simmons and Skousen (S&S), is a modification of the Ziem method in that it separated each mining site into unique topographic shapes (concave vs convex slopes, narrow vs bowlshaped valleys, steep vs gentle slopes, etc.). Widths, lengths, depths, and measured slopes of unique shapes were measured, then the volumes for all the unique shapes were summed together to obtain total overburden amounts for the site. This method incorporated varying slopes and land surfaces into the volume calculation. The third method, the PADEP method, used a planimeter to determine the surface area of the uppermost and lower-most strata within the mined area boundary. A spreadsheet then interpolated the upper and lower area of each layer, assuming a constant slope, and then multiplied the average layer area by its thickness to obtain volume (Smith and Brady, 1990). The Ziem and PADEP methods were far less timeconsuming in estimating volumes than the S&S method. Volumes for each method were then converted to mass by multiplying the volume of each stratigraphic unit by the unit weight of the rock type present (Table 2). The metric tons of overburden present in each layer were used to convert %S into metric tons of MPA and metric tons of NP. Spreadsheets for each overburden volume/mass calculation method are available from the authors. On five different-sized coal mine operations used in this study, we were able to determine actual mass of overburden materials hauled during the operation by collecting data from computers mounted on overburden haulage trucks. These actual masses were compared to calculated masses from overburden volume estimates by the three methods. Results and Discussion Thirteen of the 56 sites were found in southern West Virginia (Figure 1). The remaining 43 northern West Virginia sites included 17 sites in Preston, 12 sites in Monongalia, and a few sites in each of nine other counties (Table 3). The reason for the majority of sites to be found in northern counties is because of the general geology of West Virginia and its impact on water quality. The coal geology of West Virginia is divided into the northern and southern coalfields, both of which were formed during the Pennsylvanian Period (Barlow, 1974). The southern coalfield contains coal seams found in the Pottsville Group (Pocahontas, New River, and Kanawha Formations), which generally have higher overall rank and heating value, and lower sulfur and ash contents than northern coals. The northern coalfield contains coal seams in the Allegheny, Conemaugh, and Monongahela Groups, which are generally high in sulfur and ash content. The high-sulfur content of the coal and associated rocks in the northern coalfield makes these coals prone to acid mine drainage generation during mining. The dividing line between coalfields is the hinge line (Figure 1). The high-sulfur geology in the northern coalfield is also separated by the amount of carbonate or calcareous material in the rocks. The eastern part of the northern coalfield is characterized by low amounts of calcareous material or limestone in the strata, while the western part may have limestone or other alkalineproducing rocks associated with coal seams. Smith et al. (1976) separated these unique geologic settings into “surface mining provinces” (SMP on Figure 1 and Table 3). SMP 1 occurs in southern WV and is comprised of low sulfur and low carbonate rocks. SMP 2 occurs in the eastern part of the northern coalfield and contains coal beds and rocks with high sulfur and low carbonate content. SMP 3 is found in the western part of the northern coalfield with rocks containing high sulfur and high carbonate content. So the SMP concept provides a general prediction of the acid mine drainage potential of rocks disturbed in each of the provinces. SMP 2 is of special interest because it has the highest potential for acid mine drainage, but it is also the most geologically variable. Hence, 13 sites were in SMP 1, 25 in SMP 2, and 18 were in SMP 3. Coal seams extended from the Glenalum Tunnel coal seam found in the Pottsville Group (Kanawha Formation) upward to the Waynesburg coal seam at the top of the Monongahela Group (Table 4). Overburden Volume and Mass Overburden volume estimates from three methods were converted to mass and compared to actual truck weights on five differently sized and shaped sites. The PADEP method was consistently closer to the actual overburden mass than the other two methods, giving an average % error of only –4% (Table 5). The Ziem method generally overestimated overburden mass. The average % error for the Ziem method was 118%, with a maximum of 320%. The S&S was better than the Ziem method, with an average error of –3%, but consistently underestimated overburden masses and required considerably more time and effort than did the other two methods. The differences among the three methods were due, in part, to fundamental differences in how volumes were calculated. In the Ziem method, a single, average cross-sectional area perpendicular to the highwall was calculated from measurements and had to represent all topographic changes across the entire pit length. So changes in pit width, slopes, and highwall curvature all contributed error to the estimate. The BN (7-ha contour mine), DC (37-ha area mine), and HG (83-ha, long contour mine) sites had large over estimation errors with the Ziem method. The S&S method was an improvement in that some, but not all of this variability was accounted for by dividing the site into relatively homogenous, similarly-shaped sections. The sheer magnitude of measurements and the time consuming nature of this method made this a cumbersome technique, even though it was thought that this method would provide a more accurate method for estimating volume. The PADEP method was superior because a single measured or interpolated area parallel to the earth’s surface was multiplied by depth. Because the depth of any layer is significantly less than the total pit length, the average area only has to represent the actual area over a relatively short distance. The use of digital topographic maps and Geographic Information System based programming has the potential to refine the interpolated measurements and thus improve volume estimates. In spite of large differences in overburden mass estimates, this had a surprisingly small effect on NP and MPA estimates among the methods. By using the NP, MPA, and NP/MPA values of PADEP as a reference, only 6 of 30 estimates by the Ziem and S&S methods were off by more than 10%, but two were very far off (Table 5). Water Quality Prediction Eleven of the 56 sites gave net acid water (Table 6). Of the eleven sites that gave net acid water, eight were from Upper Freeport surface mines in Preston County (SMP 2), and these Upper Freeport sites gave the highest acid concentrations in water at our sites (Table 6). The other three acid sites were a Waynesburg coal mine in Monongalia County (SMP 3), a Kittanning coal mine in Clay County (SMP 1), and a Pittsburgh coal mine in Upshur County (SMP 3). So 73% of the acid sites were from SMP 2, 18% were from SMP 3, and 9% were from SMP 1. Not all SMP 2 surface mines produced acid drainage. Four operations in Preston County (SMP 2) mined the Bakerstown coal, which all produced very alkaline post-mining water quality, and the other three sites in Preston County were alkaline-producing Kittanning and Pittsburgh coal surface mines. In comparing NNP with NP/MPA ratio for each site, there was no difference in the interpretation on its acidor alkaline-producing status. For example, all negative NNP values showed <1 NP/MPA ratio, and a range of 0.4 to 17.4 mt/1000 mt NNP occurs in the 1 to 2 NP/MPA ratio. Therefore, for convenience, the rest of the ABA site categories will use the NP/MPA ratio The prevailing thought is that sites with overburden NP/MPA ratios of <1 should produce acid mine drainage, while ratios of >2 should produce net alkaline drainage. Those between 1 and 2 could generate either acid, alkaline, or neutral drainage (Perry, 1998). Table 6 lists the sites according to NP/MPA ratio with lines separating NP/MPA ratio categories of <1, 1-2, and >2. In our study, eight sites had NP/MPA ratios of <1, and six of these sites produced net acid water (Table 6). The two sites that were not acid producers were from Fayette and Clay Counties (SMP 1) and both had very low total MPA and total NP, and had only slightly negative NNP. All ABA values for these two sites suggest that the overburden would not affect water quality significantly, and indeed the water quality is only slightly alkaline. In fact, the thirteen sites in SMP 1 (southern WV) gave post-mining water qualities of –15 to 228 mg/L CaCO3, but most were in the range of 5 to 60 mg/L CaCO3. Of the eight sites that gave NP/MPA ratios between 1 and 2, only one produced net acid water, and only slightly so. This Al site was a Middle Kittanning mine in Clay County (SMP 1). The other six sites with NP/MPA ratios between 1 and 2 were in northern and southern WV counties and included all SMPs. Forty of the 56 sites had NP/MPA ratios >2. Of these sites, four produced net acid water. The Cr, SH, D2, and HP sites had NP/MPA ratios of 2.2, 3.1, 3.6, and 8.9, respectively. It is hard to conceive that ratios of 3.1 and 3.6, and especially 8.9 could produce acid drainage. Three of the sites were from Upper Freeport Preston County mines (Cr, SH and D2, and all in SMP 2), while the other was a Redstone/Pittsburgh coal mine from Upshur County (HP, and borders SMP 2 and 3). The 1-ha Cr site was mined relatively slowly with the surface mine pits remaining open for long periods (personal communication from past inspectors and operators). Perry et al. (1997) noted that acid mine drainage was enhanced on potentially acid-producing sites when mining and reclamation was done slowly. Quickly reclaiming disturbed areas prone to acid mine drainage decreases the amount of time pyritic material is exposed to oxidation and weathering. The slow mining of the Cr site may have allowed excessive oxidation of pyritic materials, thereby creating a larger problem with acid drainage than if the site had been mined more quickly. The 2-ha SH surface mine removed the down dip coal outcrop of a 15-ha Upper Freeport underground mine. Therefore, the acid water draining the underground mine passes through the reclaimed backfill of the SH surface mine thereby generating acid seeps in the slopes at SH. Without this underground mine water influence, based on inspector opinions, this site would not be producing acid mine drainage. The 6-ha D2 site produced net acid drainage for the first seven years after mining and reclamation, but has been producing alkaline drainage for the past two years. It is evident that acid salts in the overburden were released quickly as water moved through the backfill. Over time, as the salts were leached and no more acid was generated, the acid drainage has been gradually overcome by the alkaline-producing potential of the backfill material. The 7-ha HP site is only slightly acidic with primarily manganese in the water. It is not adjacent to an underground mine, nor does it appear that coal refuse was placed on the site, nor was there any other unusual reclamation practice that may have caused an otherwise alkaline-producing site to generate acid mine drainage. It is possible that non-alkaline-producing carbonates were counted as alkaline-producing carbonates in this overburden, which gave erroneously high NP numbers. Figures 2 through 6 graphically show some of the ABA overburden parameters versus net alkalinity of post-mining water quality. Figure 2 showed no relationship (R = .02) between MPA in mt/1000 mt and net alkalinity of post-mining water. Acid mine drainage occurred on sites with very low MPA (2.7 mt/1000 mt) to high MPA (20.1 mt/1000 mt). And alkaline drainage occurred on sites with 20 to 45 mt/1000 mt MPA. So the critical parameter for post-mining water quality prediction was not total sulfur or potential acidity calculated from total sulfur measurements. Brady et al. (1994) similarly found no relationship between MPA and post-mining water quality. diPretoro and Rauch (1988) suggested that sites with NP <20 would produce acid drainage, while Brady et al. (1994) indicated this value was <10. Total NP correlated with net alkalinity in post-mining water on our 56 sites (Figure 3, R = .12). Nineteen sites had total NP <20 mt/1000 mt with only six having acid water (32%). Twelve sites had NP values <10 mt/1000 mt and only four (33%) produced acid drainage (Table 7). Five sites out of 37 with NP >21 mt/1000 mt (14%) gave acid drainage, and two of these acid sites had total NP >40 mt/1000 mt. Therefore, 86% of sites with NP >21 mt/1000 mt gave alkaline drainage. From our data, NP was not a clear indicator for acid post-mining water quality. Total NNP combines MPA and NP into one variable, which was plotted against net alkalinity (R = .15, Figure 4). Past predictions have used >12 NNP (Brady et al., 1994; Perry and Brady, 1995), >15 NNP (Skousen et al., 1987), and >30 NNP (diPretoro and Rauch, 1988) as values that should produce net alkaline water. Values of <0 NNP (Brady et al., 1994; Perry and Brady, 1995; Skousen et al., 1987) and <10 NNP (diPretoro and Rauch, 1988; Erickson and Hedin, 1988) have been used as predictors of acid drainage. In our data, six sites with NNP values of –12.7 to –2.3 mt/1000 mt gave acid drainage, two sites with NNPs of –1.9 and –0.9 gave neutral drainage, and three sites with NNPs between 0.4 and 1.6 gave a mixture of drainage. Four sites (the same problem sites as noted above), all with >18 NNP, produced slightly acid drainage (59 to 16 mg/L net acidity). Again, if these four sites are excluded, all of the acid mine drainage sites had NNP <0.4 mt/1000 mt. Therefore, six of eight sites (75%) with <0 NNP produced acid drainage, while 32 of 36 sites (89%) with >12 NNP produced alkaline drainage (Table 7). Twelve sites between 0 and 12 NNP all produced alkaline drainage except one. The NP/MPA ratio is also used to predict acid drainage. Figure 5, plotting NP/MPA ratio against net alkalinity, shows the same trend as NNP versus net alkalinity (R = .12). The same four sites with acid water and high NP/MPA ratios (2.2 to 8.9) are apparent on this graph. Six acid drainage sites fit the general prediction pattern of NP/MPA ratios of <1 producing acid drainage, and one site with a 1.2 NP/MPA ratio also produced slightly acid drainage. So, in general, an NP/MPA ratio of <1 will produce mostly acid drainage sites, between 1-2 will produce mostly alkaline drainage sites, while NP/MPA ratios of >2 will produce alkaline drainage with a few exceptions. We wondered if any relationship existed between the size of the mine (equating to total amounts of overburden moved) and net alkalinity of post-mining water (Figure 6). All 19 sites >15 ha gave alkaline water, while eleven of the 37 sites 15 ha or less in size (30%) gave acid water. Small mines move less overburden and therefore have less chance of intercepting calcareous strata. In comparing our data to that of Brady et al. (1994), we found very similar results (Table 7). We did not find a good relationship with our post-mining water quality data on acid sites and NP, but better relationships existed between NNP and NP/MPA ratios. Summary and Conclusions Acid-base accounting values have been adopted to help in the prediction of post-mining water quality on surface mines. Calculation of ABA parameters are based on volume adjustments for overburden, and three methods were tested to determine which one should be used. The volume adjustment method developed by the PADEP was found to be the most accurate and reliable method to calculate overburden mass. Total MPA, NP, NNP, and NP/MPA ratios were determined for 56 surface coal mining sites and compared to post-mining water quality. The MPA showed a poor relationship, NP was intermediate, while NNP and NP/MPA ratio were equally good in predicting water quality from overburden. Eight of 56 sites had NP/MPA ratios of <1, and six of these eight sites (75%) produced acid drainage. The two remaining sites with NP/MPA ratios <1 were from southern WV and produced only slightly alkaline drainage. Eight sites had NP/MPA ratios of between 1 and 2, and only one of these eight sites (13%) produced acid drainage, and only slightly so. Thirty-six of 40 sites with NP/MPA ratios >2 produced acid drainage (90%), but all four had very high ratios and would not have been expected to produce acid drainage. From these data, all six sites in northern West Virginia (SMP 2 and 3) with NP/MPA ratios <1 produced acid drainage. For all sites with an NP/MPA ratio >1, 43 out of 48 (90%) produced alkaline drainage, but only two of the five acid sites with an NP/MPA ratio of >1 (Upper Freeport Preston County surface mines in SMP 2) gave post-mining acid water above 40 mg/L CaCO3. The other three acid drainage sites with NP/MPA ratios >1 produced only slightly acid water (-15 to -17 mg/L CaCO3 acidity). In general, an NP/MPA ratio of <1 will produce mostly acid drainage sites, between 1-2 will produce mostly alkaline drainage sites, while NP/MPA ratios of >2 will produce alkaline drainage with a few exceptions. Acknowledgments The authors express appreciation to Ron Hamric of Patriot Mining Company, and Steve Shaffer of Buffalo Coal Company for information on several sites. Literature Cited Barlow, J.A. 1974. Coal and coal mining in West Virginia. Coal Geology Bull. No. 2. West Virginia Geological and Economic Survey, Morgantown, WV. Brady, K.B.C., and C.A. Cravotta III. 1992. Acid base accounting: an improved method of interpreting overburden chemistry to predict quality of coal mine drainage. In: Thirteenth Annual West Virginia Surface Mine Drainage Task Force Symposium, April 8-9, 1992, Morgantown, WV. Brady, K.B.C., E.F. Perry, R.L. Beam, D.C. Bisko, M.D. Gardner, and J.M. Tarantino. 1994. Evaluation of acid base accounting to predict the quality of drainage at surface coal mines in Pennsylvania, U.S.A. p. 138-147. In: International Land Reclamation and Mine Drainage Conference and Third International Conference on the Abatement of Acidic Drainage, Vol. 1, U.S. Bureau of Mines IC 9183, Pittsburgh, PA. Caterpillar. 1991. Caterpillar performance handbook. 22 Ed. Peoria, IL. Cravotta, C.A. III, K.B.C. Brady, M.W. Smith, and R.L. Beam. 1990. Effectiveness of alkaline addition at surface mines in preventing or abating acid mine drainage: Part I. Geochemical considerations. p. 221-226. In: 1990 Mining and Reclamation Conference and Exhibition, April 23-26, 1990, Charleston, WV. diPretoro, R.S., and H.W. Rauch. 1988. Use of acid-base accounts in premining prediction of acid drainage potential: a new approach from Northern West Virginia. p. 1-10. In: Proceedings: Mine Drainage and Surface Mine Reclamation, Vol. 1, U.S. Bureau of Mines IC 9183, Pittsburgh, PA. Erickson, P.M., and R. Hedin. 1988. Evaluation of overburden analytical methods as a means to predict post-mining coal mine drainage quality. p. 11-19. In: Proceedings: Mine Drainage and Surface Mine Reclamation, Vol. 1, U.S. Bureau of Mines IC 9183, Pittsburgh, PA. Kania, T. 1998. Laboratory methods for acid-base accounting: An update. Chapter 6. In: Coal Mine Drainage Prediction and Pollution Prevention in Pennsylvania, PA Department of Environmental Protection. Harrisburg, PA. Perry, E. 1998. Interpretation of acid-base accounting. Chapter 11. In: Coal Mine Drainage Prediction and Pollution Prevention in Pennsylvania, PA Department of Environmental Protection. Harrisburg, PA. Perry, E.F., and K.B. Brady. 1995. Influence of neutralization potential on surface mine drainage quality in Pennsylvania. In: Sixteenth Annual Surface Mine Drainage Task Force Symposium, April 4-5, 1995, Morgantown, WV. Perry, E., M. Gardner, and R. Evans. 1997. Effects of acid material handling and disposal on coal mine drainage quality. p. 1007-1026. In: Fourth International Conference on Acid Rock Drainage, Vol. III, Vancouver, BC. Skousen, J.G. 2000. Static tests for predicting coal mine drainage. In: Acid Mine Drainage Prediction in the Eastern United States. Acid Drainage Technology Initiative (ADTI) Technical Report. National Mining Association and National Mine Land Reclamation Center, Morgantown, WV. Skousen, J., J. Renton, H. Brown, P.Evans, B. Leavitt, K. Brady, L. Cohen, and P. Ziemkiewicz. 1997. Neutralization potential of overburden samples containing siderite. J. Environ. Qual. 26:673-681. Skousen, J.G., J.C. Sencindiver, and R.M. Smith. 1987. A review of procedures for surface mining and reclamation in areas with acid-producing materials. National Research Center for Coal and Energy, National Mine Land Reclamation Center. Morgantown, WV. Skousen, J.G., R.M. Smith, and J.C. Sencindiver. 1990. Development of the acid-base account. Green Lands Vol. 20 (1): 32-37, West Virginia Mining and Reclamation Association, Charleston, WV. Smith, M.W., and K.B.C. Brady. 1990. Evaluation of acid-base accounting data using computer spreadsheets. p. 213-219. In: 1990 Mining and Reclamation Conference and Exhibition, April 2326, 1990, Charleston, WV. Smith, R.M., A.A. Sobek, T. Arkle, J.C. Sencindiver, and J.R. Freeman. 1976. Extensive overburden potentials for soil and water quality. EPA-600/2-76-184. U.S. Environmental Protection Agency, Cincinnati, OH. Sobek, A.A., J.G. Skousen, and S.E. Fisher, Jr. 2000. Chemical and physical properties of overburdens and minesoils. p. 77-104. In: Reclamation of Drastically Disturbed Lands, Agronomy No. 41, American Society of Agronomy, Madison, WI. Surface Mining Control and Reclamation Act. 1977. Public Law 95-87. U.S.C. 1201 et. seq. West Virginia Surface Mine Law. 1971. West Virginia Department of Mines and Minerals. Charleston, WV. Table 1: Summary of water quality prediction from studies in West Virginia and Pennsylvania (Perry, 1998). ABA Calculated Predicted Parameter Value Water Quality Source NP <20 net acid (mt/1000 mt) >40 net alkaline diPretoro and Rauch, 1988 <10 net acid >21 net alkaline Brady et al., 1994 NNP <10 net acid (mt/1000 mt) >30 net alkaline diPretoro and Rauch, 1988 <20 net acid >80 net alkaline Erickson and Hedin, 1988 <0 net acid >12 net alkaline Brady et al., 1994 <0 net acid >15 net alkaline Skousen et al., 1987 NP/MPA <1 net acid >2 net alkaline Perry, 1998 Table 2. Unit weight (mt/ha-m) of rocks used in this study to calculate volumes and mass of overburden materials (Caterpillar, 1991). Rock Type Unit Weight (mt/ha-m) Soil 18,174 Sandstone 28,122 Shale 32,324 SS/SH 30,228 Mudstone 32,627 Limestone 31,664 Coal 16,461 mt/Ha-m is a conversion from tons/Ac-ft (x 8.922) Table 3. Locations of our 56 study sites including the county, the surface mining province (SMP), and position in West Virginia. County Number of Sites SMP Northern or Southern WV Barbour 2 2 & 3 N Braxton 1 1 S Clay 2 1 S Fayette 2 1 S Grant 4 2 N Harrison 2 3 N Kanawha 1 1 S Mineral 1 2 N Monongalia 12 3 N Nicholas 3 1 S Preston 17 2 N Randolph 1 1 S Taylor 2 2 & 3 N Tucker 1 2 N Upshur 2 3 N Webster 3 1 S Table 4. Coal seams represented in the acid-base accounting study. Group Coal Seam Number of Sites Monongahela Waynesburg 8 Sewickley-Redstone 2 Redstone-Pittsburgh 6 Pittsburgh 3 Conemaugh Elklick 2 Harlem 2 Bakerstown 6 Allegheny Upper Freeport 13 Lower Freeport 1 Freeports + Kittannings 3 Kittannings (5 & 6 Block) 5 Kanawha Alma-Eagle 2 Gilbert-Eagle 1 Peerless 2 Glenalum Tunnel 1 __________________________________________________________________________________ Table 5. Comparison of overburden amounts (total metric tons of overburden, NP, MPA, and NP/MPA ratio) calculated by three different methods (Ziem, S&S, and PADEP) and the actual overburden amounts determined from haulage trucks. Site Method Total % Overburden Error NP MPA NP/MPA ___________________________________________________________________ (mt) ( mt/1000 mt ) CH1 Ziem 3,852,869 27 43.4 54.1 0.80 S&S 4,738,441 11 43.4 12.9 3.36 PADEP 4,708,563 12 41.9 13.6 3.08 Actual 5,325,274 BN Ziem 8,480,647 164 70.1 11.9 5.89 S&S 2,482,974 22 79.3 10.0 7.93 PADEP 3,453,915 8 70.8 12.0 5.92 Actual 3,205,621 KE Ziem 37,873,619 22 58.9 19.9 2.96 S&S 19,520,492 37 60.0 23.3 2.58 PADEP 30,190,263 3 62.8 19.8 3.16