Premining water quality has been used as a mine drainage prediction tool in Pennsylvania since at least the mid-1980s (Brady and Hornberger 1990). Brady et al. (1996) and Hawkins et al. (1996) have provided an in-depth examination of the physical and chemical characteristics of shallow groundwater in unmined regions of the northern Appalachian Plateau.

The bituminous coal field of Pennsylvania lies within the Appalachian Plateau physiographic province. The unglaciated portions of the Appalachian Plateau vary from moderate to high relief (300 > 1000 ft; 90 > 300 m) in the east and northeast, to low to moderate relief (100 to 600 ft; 30 to 180 m) in the western portion of the state. The higher relief areas are marked by deep, V-shaped valleys, and the lower relief areas are marked by a smooth, undulating surface with numerous narrow, relatively shallow valleys (Berg et al., 1989). The rock strata consist of sandstone, shale, siltstone, claystone, limestone and coal, which is horizontal to slightly dipping. This chapter will address only the unglaciated portions of the plateau.

Natural water quality in shallow groundwater flow systems of the Appalachian Plateau results from the influences of three factors: the chemistry/mineralogy of the rock the water contacts, flow path and water contact time.

Of the many solutes in groundwater, we will concentrate in this chapter on bicarbonate alkalinity because it can be compared to overburden neutralization potential (an estimate of carbonate content), and because it can be used as a mine drainage quality predictor. Other water quality parameters will be addressed where appropriate. The water quality can be evaluated from springs and wells.

The recharge area for most springs is the weathered/leached regolith or highly fractured and weathered bedrock (Figure 10.1). The depth of weathering is enhanced by horizontal bedding-plane separations, caused by unloading, and vertical fractures, caused by stress-relief (Hawkins, et al. 1996). The weathered zone is typically 20 to 40 ft (6 to 12 m) thick, but can extend deeper along fractures and is characterized by a red, yellow, or brown color resulting from iron oxidation. The weathered regolith and bedrock has been depleted of the most readily leachable (e.g., carbonate) and oxidizable (e.g., pyrite) minerals. Water flowing through the weathered zone has a short residence time because it is near the surface, has limited thickness, and has higher permeability which is induced by physical (fracturing) and chemical (leaching and oxidation) weathering. Hawkins et al. (1996) have estimated the residence time as days to weeks. Both the scarcity of readily weatherable minerals and short residence time result in spring water quality that is typically dilute.

Figure 10.1 Schematic cross section showing conceptual shallow groundwater flow model, which includes the near-surface weathered-rock zone and the deeper unweathered-rock zone (modified from Hawkins, et al., 1996).

Below the weathered regolith is a zone of largely unweathered bedrock (Figure 10.1). Weathering is restricted to some fractures and bedding-plane separations. Unweathered rock can contain readily soluble minerals, in particular carbonates, if they are part of the rock composition. The water flowing through the rock has a longer residence time because of lower permeabilities, and the slower flow rate allows longer contact with soluble minerals. The permeability is orders of magnitude less than that for the weathered zone and the residence time is in the order of years (Hawkins et al., 1996). Groundwater in passing through unweathered rock will typically have higher dissolved solids than water emanating from the weathered-rock/regolith zone.

To illustrate the relationship between rock chemistry and water chemistry, several specific sites will be discussed. The locations of these sites are shown on Figure 10.2.

Figure 10.2 Map of southwestern Pennsylvania showing location of sites discussed in text.

The quality of groundwater from both weathered rock/regolith systems and unweathered rock systems can help to identify the presence or absence of carbonates within the area to be mined. Water chemistry can also shed light on the groundwater flow system (see Chapter 2).

The three study areas discussed in this chapter were selected because they had:

Additionally, the sites were selected because they are isolated hill tops where the only recharge is from precipitation.

The acid-base accounting variables "neutralization potential" (NP) and "maximum potential acidity" (MPA) and their derivative "net neutralization potential" (NNP) are discussed in chapters 6 and 11. Units are traditionally reported as tons CaCO₃ /1000 tons of material (e.g., equivalent to parts per thousand, ppt, CaCO₃). Average NPs and MPAs were determined for each overburden drill-hole by using thickness weighting and assuming each hole represented a column of constant diameter, following the methods described in Smith and Brady (1990).

All water quality sample locations were represented by multiple samples. Medians were determined for pH. Other water quality parameters discussed in this paper, unless specified otherwise, are mean concentrations.

Mine A is the Kauffman mine where various research studies have been and are being conducted (e.g., Abate, 1993; Evans, 1994; Rose et al., 1995; Hawkins et al., 1996). Mine site topography, locations of drill holes and water-sample points for this mine are shown on Figure 10.3. This map also shows alkalinity and specific conductivity of wells and springs associated with the lower Kittanning (LK) coal seam, plus a few springs flowing from the Clarion #2 cropline. All the wells are completed as 10-cm-diameter piezometers that are open for 1.5 m at the interval of the lower Kittanning (LK) coal. Figure 10.3 also shows summaries of acid-base accounting data for overburden drill holes completed to the lower Kittanning coal.

Figure 10.3 clearly indicates that spring water quality is significantly lower in alkalinity and specific conductance than well water. Alkalinity of springs is typically less than 5 mg/L and conductivity less than 50 m S/cm. In contrast, wells in the western portion of the hill typically have much higher alkalinity (> 70 mg/L) and specific conductivity (~ 200 m S/cm). It is also apparent that there is a regional distribution of well water alkalinity. Wells on the eastern portion of the area have lower alkalinity than those for the western portion. This pattern is consistent with the higher NPs reflected in the overburden data on the western portion of the site (Figure 10.3). Piezometers completed within the hill of the Kauffman site indicate a steep downward flow gradient at the horizon of the LK coal. Between the lower permeability of the unweathered rock zone and the downward component of much of the flow, little water from the unweathered rock zone is contributed to the springs. The groundwater quality is related to overburden NP along the flow path.

The geologic controls on distribution of calcareous units are complex on this site. Observations of highwalls show that the calcareous material occurs as calcite and minor siderite as cement between sandstone grains in trough bottoms of some channel sandstones (V. Skema, 1995, personal communication). These calcareous trough deposits are laterally discontinuous and occur only in the basal portion (bottom 3 to 5 m) of a thick (18 to 24 m) channel-sandstone unit. Observations of drill cores also suggest that some of the calcareous cement in sandstone is associated with vertical and bedding plane fractures. However, most fractures are not associated with calcareous minerals.

Overburden holes and wells 470 ft (140 m) or less apart were paired for comparisons of alkalinity of groundwater with overburden NP. These two parameters show a strong positive relationship (Table 10.1; Figure 10.4), with alkalinity (mg/l) being four times the NP value (tons/1000 tons). A similar relationship is seen between specific conductance and NP (Figure 10.5). As would be expected, specific conductance and alkalinity show a strong positive relationship (r = 0.94). These relationships imply that dissolution of calcareous minerals, where present, has a significant impact on groundwater chemistry.

Figure 10.4 Alkalinity of groundwater associated with the Lower Kittanning coal seam as a function of neutralization potential at mine site a.

Table 10.1 Groundwater chemistry and acid-base accounting rock chemistry data for 11 pairs of water sample points and overburden sample points.

Sulfate, a conservative aqueous ion, is a weathering product of pyrite. Therefore, a comparison was made between MPA and sulfate (Figure 10.6). MPA was also compared with specific conductance (Figure 10.7). MPA shows no relation to sulfate in nearby groundwater and has, if anything, an inverse relation to specific conductance. With one exception, sulfate was generally less than 35 mg/L. The exception is the result of anthropogenic influences. The inverse relationship of sulfate with specific conductance is an artifact of the site data. Specific conductance is strongly related to alkalinity, and alkalinity to NP. And at this site the high NP overburden holes tend to have low MPA,

Figure 10.5 Specific conductance of groundwater associated with the Lower Kittanning coal seam as a function of neutralization potential at mine site A.

Figure 10.6 Sulfate of groundwater associated with the lower Kittanning coal seam as a function of maximum potential acidity at mine site A.

whereas high MPA holes tend to have low NP (Figure 10.8).

Brady et al. (1996) provide a detailed study of one of the springs, GR-413 (Figure 10.3). A hydrograph of this spring (Figure 10.9) illustrates the relationship between rainfall and spring flow. Almost immediately after rainfall there is a sharp peak in flow resulting from surface runoff. This is followed a few days later by a broader peak that is interpreted to result from flow through the shallow weathered zone. The hydrograph illustrates the rapid flow-through time for groundwater in the shallow weathered-rock zone. The broad peak in flow occurs 3 to 4 days after rainfall, and continues to tail off for a week or so. The spring flows year-round, thus storage must exist in this shallow zone to account for the year-round flow. The spring water is similar in chemistry to the other coal-cropline springs.

Figure 10.7 Specific conductance of groundwater associated with the lower Kittanning coal seam as a function of maximum potential acidity at mine site A.

Figure 10.8 Comparisons between neutralization potential, alkalinity, and maximum potential acidity at mine site A.

Figure 10.9 Spring GR-413 response to individual storm events. See Figure 10.3 for location of spring. The initial spikes are from surface runoff (from Abate, 1993).

The principal coal seam that was mined is the upper Kittanning (UK). Field pH and specific conductance were measured prior to mining in several uncased drill holes which were completed down to the UK seam. Additionally numerous UK cropline springs were sampled during collection of background data. Figure 10.10 is a map showing the Upper Kittanning coal cropline and water sample points.

Figure 10.10 Location of coal cropline, drill holes and water sample points, and water quality data at mine site B.

The specific conductivity of the crop springs and shallow (near crop) wells have lower values (38 to 62 m S/cm) than the drill holes located toward the middle of the hill where depth to coal is greatest (158 to 221 m S/cm). The pH of the springs is also lower than the pH of the wells. NP is negligible in areas with less than 9 to 12 m of cover, evidently because of weathering (Figure 10.11). The higher pH and conductance increase are both coincident with an increase in NP (Figures 10.11 and 10.12). The water in deeper drill holes exhibits higher dissolved solids (reflected as specific conductance) than in cropline springs and holes with shallow cover, due to the increasing abundance of calcareous minerals with increasing overburden. The high NP strata are freshwater limestones and calcareous shales. The geology of this area is addressed in Brady et al. (1988), where the mine site is referred to as area "C."

Figure 10.11 NP and pH as a function of depth from the surface to the bottom of the upper Kittanning coal seam for overburden holes (NP), and drill holes and springs (pH). Site is located in Fayette County. The scale of the x-axis is in meters.

Crop springs have alkalinity between 1 and 9 mg/L. Unfortunately alkalinity was not sampled in the drill holes, but it is a near certainty that the alkalinities were higher, judging from elevated pH and conductance. This hill top has been mined and the mine is producing drainage with an average alkalinity of 380 mg/L.

Mine Site C (Figure 10.13) is an example of a site where there is no significant calcareous overburden rock within the proposed mine area. This deficiency is shown by the low NPs of the two overburden holes (OB-A and OB-C), and by the low alkalinity of crop-

Figure 10.12 NP and specific conductance (m S/cm) as a function of depth from the surface to the bottom of the upper Kittanning coal seam for overburden holes (NP), and drill holes and springs (specific conductance). Site is located in Fayette County. The scale of the x-axis is in meters.

line springs, small country bank mines, and water collected from the two overburden holes which were uncased down to the coal. Near the actual area mined, the highest alkalinity is 11 mg/L from cropline spring SP-14. Two of the springs below the coal cropline have alkalinity as high as 31 mg/L. Apparently calcareous strata exist below the coal.

The two overburden holes encountered predominately sandstone, with coal at 54 ft (16 m) and 77 ft (23 m) for OB-A and OB-C respectively. The highest neutralization potential in OB-A is 19 ppt CaCO₃, the highest in OB-C is 15 ppt CaCO₃. The highest percent sulfur (excluding the coal) in OB-A was a 1 ft (0.3 m) binder within the coal bed that had 2.0 percent sulfur (% S). A 3 ft-thick (one m-thick) shale overlying the coal contained 0.5 % S. The highest percent sulfur in OB-C was the shale below the coal, which had 1.7% S. The next highest sulfur in the overburden is only 0.2%. The overburden shows the presence of acid-producing strata, but lacks alkalinity-producing strata.

Table 10.2 compares rock chemistry and water chemistry in the two overburden drill holes. The low alkalinity of the water agrees with the low NP of the overburden data. Because of the lack of naturally occurring calcareous rocks, an average of 45 T CaCO₃/ac (1 x 10⁵ kg/ha) was brought to the site. Additionally, the material with percent sulfur greater than

0.5 % was selectively placed in "pods" that were located such that they would be above the postmining water table. Twenty tons/acre of the alkaline material was placed on the pit floor, with most of the remainder mixed into the high-sulfur spoil pods. indications are that the amount of alkaline material added was inadequate and the site is producing acidic drainage. Table 10.3 shows water quality from one of the postmining discharges from this mine.

Table 10.2 Overburden and water quality comparisons for drill holes OB A and OB B at mine site C.

Table 10.3 Postmining water quality at CD4A.

Figure 10.13 includes water quality for some springs, up slope and to the north of the mine site, that have higher alkalinity (35 to 52 mg/L) than cropline springs. The recharge of these springs is stratigraphically and topographically higher than the upper Freeport coal. The Glenshaw Formation, which is the unit from which these springs arise, is known to contain several calcareous marine zones (Flint, 1973). These higher alkalinities may indicate that there is some calcareous material preserved within the shallow flow system or the springs may be receiving some water form a deeper source. Although no deeper well data exists for the Glenshaw Formation in the immediate area of the mine site, a 130-ft (40-m) deep well 2 miles (3 km) north-northeast of the site, which penetrates the lower portion of the Glenshaw Formation, has alkalinity of 106 mg/L, specific conductance of 260 m S/cm, and sulfate of 9 mg/L (Tom McElroy, 1996 personal communication). these data are consistent with the findings of McElroy (in preparation), that wells within the Glenshaw Formation have higher alkalinity than springs associated with the same strata.

All sulfate concentrations shown on Figure 10.13, whether from a well or a spring, and regardless of stratigraphic or topographic position, are less than 40 mg/L.

Climate in the Appalachian Basin from Pennsylvania to Alabama is fairly similar. Average annual temperature in the north is about 50° F (~10° C) and above 60° F (>15° C) in the south. Precipitation also increases from north to south, with rainfall averaging 35 inches (90 cm) in northern Pennsylvania to more than 55 inches (140 cm) in Alabama (Weeks et al., 1968). Annually the infiltration rate of water is sufficient that groundwater is continually flushed through the rock strata from recharge to discharge points. This flow tends to leach out the soluble products of the more weatherable minerals.

Geology throughout the Appalachian Plateau is also similar in many respects. The rocks are predominately flat-lying, consisting of sandstone, conglomerate, siltstone, shale, claystone, limestone and coal (Miller et al., 1968). Hills have been subjected to stress-relief and unloading forces which have intensified fracturing and bedding-plane separations, and subsequently, weathering of the rocks. This physical weathering, coupled with chemical weathering, has resulted in higher permeabilities within the weathered zone (Hawkins et al., 1996). These geologic and climatologic factors result in similar groundwater hydrologic characteristics throughout the plateau.

Numerous studies have investigated the groundwater hydrologic characteristics on the Appalachian Plateau, but very few studies have related the hydrology, groundwater chemistry and rock chemistry. Powell and Larson (1985) investigated water quality in an unmined watershed in a coal-producing region of the Appalachian Plateau of southwestern Virginia. The rock strata are in the Pennsylvanian Norton Formation. The most common carbonate present was siderite, following by calcite and dolomite. Minor amounts of pyrite were generally associated with coal and adjacent rocks. They observed that water from springs typically had lower concentrations of alkalinity and dissolved solids than water from dug and drilled wells Figure 10.14 is a plot of alkalinity as a function of bottom depth of sample point. The springs typically have lower alkalinity than dug wells, which have lower alkalinity than drilled wells. With one exception sulfate was less than 40 mg/L, for springs, dug wells and drilled wells. In fact only 6 of 32 sample points had sulfate above 20 mg/L. This is consistent with observations of sulfate for groundwater not impacted by mining in Pennsylvania.

Figure 10.14 Alkalinity as a function of depth of groundwater sample source in Buchanan Co., VA. Springs are shown at zero depth. Data from Powell and Larson (1985).

Recent studies in Somerset (McElroy, in preparation) and Indiana (Williams and McElroy, 1997) Counties, Pennsylvania show spring water to be routinely more dilute than well water. This is true throughout the Pennsylvanian Period strata in these counties.

Powell and Larson (1985) envision two groundwater flow systems, one of which "moves under and through the weathered rock or soil layer along the surface of the consolidated rock." The other system "flows through rock fractures and provides the main source of domestic supply" (most domestic supplies were drilled wells). Water in the shallow flow system flows relatively rapidly through rocks thoroughly leached of calcareous minerals. A few of the springs, and many of the deeper wells have high alkalinity (>100 mg/L) (Figure 10.14). The two springs with highest alkalinity occur in stream valleys, and they probably issue from deeper flow systems. The spring with the highest alkalinity occurs at the lowest elevation of any of the sample points. Springs with high alkalinity are the exception rather than the rule, indicating that most springs are from shallow groundwater sources.

Singh et al. (1982) in a study of the Mahoning Sandstone in Preston Co., WV, found a weathered zone of about 20 ft (6 m) deep which was low in sulfur and "exchangeable bases" (i.e., Ca and Mg). This indicates that both pyrite and calcareous minerals were removed by weathering within this zone. Brady et al. (1988), in a study that included the area of Mine C of this study also documented the loss of calcareous minerals within ~20 ft (6 m) of the surface. These findings on surface rock weathering are consistent with the current study.

The literature review presented by Hawkins et al. (1996) demonstrates the consistency of groundwater hydrologic properties throughout the Appalachian Plateau. Although less research has been done on rock and water chemistry, that which has been done shows results similar to those in PA. It appears that the principals observed in Pennsylvania are applicable to much of the Appalachian Plateau.

The combined observations of Powell and Larson (1985), Singh et al. (1982), Brady et al. (1988) and this study provide an explanation for the differences in water quality often observed between wells and springs from the same stratigraphic horizon. There are two separate shallow flow systems; a near-surface zone that is chemically and physically weathered, and a deeper unweathered-rock zone. Figure 10.1 is the conceptual model of these groundwater flow zones.

Surface mining in Pennsylvania generally occurs within one hundred feet (30 m) or less of the surface and most mines are in groundwater recharge areas (e.g., hilltops). A common misconception has been that water quality from cropline springs in unmined areas is typical of water associated with the coal seam. As illustrated at Mine Sites A and B, water associated with coal-cropline springs can be much more dilute than water from the same coal seam under deeper overburden cover. The cropline springs and shallow wells represent water flowing through the near-surface weathered-rock zone. This weathered-rock zone is quite permeable due to chemical and physical weathering. Typically calcareous rocks are absent or negligible within this weathered zone. The rapid flow-through time for the water along with the leached nature of the weathered rock results in water that lacks alkalinity and contains low dissolved solids.

Wells penetrating deeper overburden are completed in unweathered or less weathered rock with lower permeability. Groundwater flow is primarily along fractures and bedding planes. The combination of calcareous minerals, and longer residence time for the groundwater, results in significant dissolution of calcareous minerals forming bicarbonate alkalinity. Downward flow and substantially lower hydraulic conductivities probably result in little of this water reaching the cropline springs.

Mine Site C is an example of a site that lacks significant calcareous rocks. This situation persists at shallow cover and at deep cover (maximum overburden that would be mined). The lack of calcareous rocks is confirmed by the chemistry of the two overburden test holes and water quality associated with springs, country bank mines and the overburden drill holes. Wells penetrating strata that lack calcareous minerals will exhibit low alkalinity.

From Mine Sites A and B there appears to be a direct relationship between the amount of calcareous material preserved in the overburden and the alkalinity, conductivity, and pH of the groundwater. Whether the relationships observed in Figures 10.4, 10.5, 10.11, and 10.12 are site specific or more universal is not known. More sites were NP and alkalinity comparisons can be made must be investigated. The general rule of thumb, however, that alkalinity is higher in areas with calcareous rocks compared to areas without calcareous rocks is certainly true.

No relationship seems to exist between MPA (i.e., percent sulfur) and sulfate. This is probably because of the very limited oxidation of pyrite under saturated conditions. Calcareous minerals are rather soluble in groundwater, whereas pyrite is not. The acid in acid mine drainage is not produced by a simple dissolution process, but by an oxidation process. Pyrite in unmined areas remains largely in an unoxidized state. Premining alkalinity in deeper drill holes provides a second confirmatory tool, along with acid-base accounting NP, to determine the relative presence or absence of calcareous rock and its distribution within the proposed mine area.

Work in other parts of the Appalachian Plateau suggests that the observations regarding groundwater chemistry and hydrology in Pennsylvania are probably applicable to other portions of the Appalachian Plateau.

All of the Pennsylvania sites studied were isolated hill tops within groundwater recharge areas. In all cases the coal outcropped on the sides of the hill. The dominant flow systems for all mines are relatively shallow, with the deepest monitoring wells (completed to the coal seam) on the order of 120 ft (35 m) deep. Where conditions are similar to those given in the examples above, groundwater alkalinity reflects the presence or absence, and relative abundance of calcareous rock.

Water quality is directly related to the flow path, the dissolution of minerals contacted by the groundwater, and the contact time of the water with the rock. Cropline springs and shallow wells (6 to 9 m deep) that have low or no alkalinity are indicative of shallow leached/weathered overburden. No significant calcareous strata (measured as NP) are likely to occur within this zone. Where calcareous rocks are present, such as some deeper cover situations, the calcareous minerals will dissolve in the water and can be measured as alkalinity. Low alkalinity in well or spring water indicates the absence of calcareous strata within the groundwater flow path for that well or spring. It might be expected that sulfate would reflect the amount of pyrite that is present, but there is no relationship between the amount of pyrite (in terms of MPA) and sulfate concentrations, thus indicating that pyrite oxidation prior to mining is negligible.

These findings have several important implications. These are:

Much of the content of this chapter appeared in a paper by Brady et al. (1996), although an early draft of this chapter predated the paper. This chapter benefited greatly from the insights and advise of the paper’s co-authors Arthur W. Rose and Jay Hawkins, and Michael DiMatteo’s help with the maps. The chapter has further benefited from reviews by A.W. Rose and Thomas McElroy.