Figure 8.28 Paleogeographic reconstruction of the Pittsburgh sandstone and Redstone limestone time. The river in the southwestern corner of Pennsylvania is the sandstone channel shown in Figure 8.27. The Redstone limestone was deposited in freshwater lakes (see Figure 8.29). The shale overburden above the Pittsburgh coal shown in Figure 8.26 would have been deposited on "mudflats." From Donaldson and Shumaker (1979).

Blue Lick Coal to Sewickley Coal Overburden - The Pittsburgh coal often has a rider coal positioned approximately 5 to 50 ft (1.5 to 16 m) above it. This rider coal is usually thin and not economically significant. However, in Somerset County the coal at this stratigraphic horizon, named the Blue Lick, attains mineable thickness. The Blue Lick coal is usually overlain by calcareous units and produces alkaline drainage. Blue Lick coal overburden NP values are shown in OB-C, Figure 8.26. OB-C shows only a small amount of strata with sulfur greater than 0.5 percent. There is a zone of calcareous rock, including limestone, above the Blue Lick rider coal.

Figure 8.29 Isopach map of Redstone limestone in southern Pennsylvania and northern West Virginia. From Linger (1979).

The Redstone/Sewickley interval is shown in Figure 8.32. This zone contains strata with sulfur greater than 0.5 percent, however, abundant limestone and other calcareous rocks are present. Although limestone occurs above the Sewickley in OB-3, this log is not typical of overburden further west in Pennsylvania. To the west this interval is typically occupied by a very thick limestone, known as the Benwood. The Benwood is the thickest limestone in the Monongahela Group and can be 60 or 70 ft (18 to 21 m) thick. Figure 8.33 is a facies map showing the Benwood interval at 55 feet (16.8 m) above the Sewickly coal.

Pittsburgh Formation Limestones - Figure 8.33 illustrates the lateral persistence that many Monongahela lacustrine limestones exhibit. The lakes in which these laterally extensive limestones were deposited had to have been very large. For example the Benwood limestone above the Sewickley coal covers more than 4,000 square miles (10,000 km²) (Berryhill, et al.,), and Eggleston (1993) notes that the Redstone limestone extends from Somerset County, PA to as far south as Cabell County, WV, and as far west as Morgan County, OH. Figure 8.32 illustrates the NP for the Redstone and Fishpot limestones, and an unnamed limestone above the Sewickley coal. The highest NP for the Fishpot in OB-3 is 735 ppt.

Eggleston (1993) concludes that the Redstone limestone was deposited in a "very shallow lake that was subject to periodic subaerial exposure during drier periods." As evidence of drying she refers to desiccation breccia, root traces, and lack of original bedding. Evidence of shallow depositional conditions include rounded intraclasts, broken and nested shells, and bioturbated limestone. Berryhill et al. (1971) arrived at similar conclusions, using similar evidence, concerning the depositional environment of other limestones in the Monongahela and Dunkard Groups. Berryhill et al. suggest that the limestones were deposited in shallow lakes, where water depth "probably never exceeded a few feet" and that "(n)either fossil evidence nor physical properties of the rocks indicates any influence of marine conditions" (Berryhill et al., 1971, p. 34). In addition to the limestones, the sandstones and shales of the Monongahela Group are also often calcareous (Figure 8.2). The environmental interpretations of Eggleston (1993) are consistent with the paleoclimatic interpretations of Cecil et al. (1985) and Donaldson et al. (1985) for this stratigraphic interval.

Berryhill et al. (1971, p. 16) found that of 50 limestone samples from the Monongahela and Dunkard Groups, 34 were classified as limestone or magnesium limestone (< 10% dolomite, > 90% calcite), 13 were dolomitic limestone (10% to 50% dolomite, remainder being calcite) and three were calcitic dolomite (50% to 90% dolomite). The clay content is generally greater than 10%, thus the limestones are classified as marly. The ratio of calcium to magnesium varies both from bed to bed and also vertically and laterally within a single bed.

Figure 8.31 Logs and chemistry of Pittsburgh overburden at Site 19 in Salem Township, Westmoreland County. Relationship with paleogeography is shown on Figure 8.28.

The Dunkard is generally composed of fine-grained clastics which are frequently calcareous. Thick lacustrine limestones are especially prevalent in the Washington Formation. The only significant interval with sandstone is above the Waynesburg coal (Figure 8.2). This sandstone is often, but not always, calcareous. The Dunkard Group data represented in Figure 8.2 is all from Greene County. A comparison of this data with the study by Berryhill et al. (1971) for Washington County shows some differences in abundance of lithologies. For example Berryhill et al. show that 33% of the Group is limestone, whereas this study found only 6.2% limestone. This is probably largely attributable to a facies change between Washington and Greene Counties. Skema (personal communication, 1997) reports that the Upper Washington and Lower Washington Limestones, which are thick in Washington County are barely present in Greene County.

Figure 8.32 Overburden and chemistry of the Redstone and Sewickley interval at Site 18 in Elklick and Summit Townships, Somerset County. Note the presence of freshwater limestones and abundant calcareous shales.

Figure 8.34 Overburden and chemistry of the Waynesburg and Waynesburg ‘A’ coals at Site 20 in Greene Township, Greene County. Note the variability in NPl within the sandstone.

As mentioned, the only coal routinely mined in the Dunkard is the Waynesburg. For purposes of this study the Waynesburg Formation has been divided into upper and lower members. This is because in recent years mining has generally been restricted to only the Waynesburg, thus not disturbing strata above the Waynesburg ‘A’ coal. Also, the Waynesburg overburden is lithologically different. As shown in Figure 8.2, much of the Waynesburg overburden is sandstone, and much of this sandstone is calcareous. The Waynesburg ‘A’ and its overburden are notorious for producing poor quality water. Figure 8.34 shows the Waynesburg coal overburden. The shale above the coal can be high sulfur as illustrated in drill hole Fox #3. The three drill logs in Figure 8.34 illustrate the variability in overburden chemistry for this unit. Of the three holes, the sandstone is only calcareous in PS-3. The finer-grained calcareous zones, near the tops of drill holes Fox #3 and Fox #4, would not be encountered unless more than 50 or 60 ft (15 or 20 m) of cover is mined. Where the sandstone is calcareous, mines produce alkaline drainage, where it is not calcareous mining results in acid mine drainage.

The stratigraphy of the anthracite region of eastern Pennsylvania has not been studied as extensively as that of Pennsylvania’s bituminous coal region. Geologic and mining engineering work done in the anthracite region over the past 150 years, however, documents some significant stratigraphic differences between the anthracite and bituminous coal regions. The anthracite region is comprised of four coal fields as shown on Figure 8.4: The Northern Field, the Eastern Middle Field, the Western Middle Field, and the Southern Anthracite Field.

The four anthracite coal fields are located within the Valley and Ridge Physiographic Province and the orogenic activity in this province since the Pennsylvanian Period has resulted in: (a) the increase in rank of the coals due to metamorphism (as compared to time-equivalent coal beds in the Appalachian/Allegheny Plateau Province of the bituminous region), and (b) the preservation of the anthracite coal fields within synclinal basins which are essentially surrounded by sandstone/conglomerate ridges that are more resistant to erosion than the coal and associated finer-grained sedimentary rocks. A comprehensive description of the geologic history of the north-central Appalachians, is contained in Faill (1997a, 1997b, 1998a, 1998b). The most recent orogenic episode, the Alleghenian, commenced in the Early Permian (Faill, 1997b). Faill (1997a, p. 552) states that "(l)ate in the Allegheny orogeny, rock thrust northward over the Carboniferous rocks in the Anthracite region of northeastern Pennsylvania caused anthracitization of the underlying coals."

Much of the Southern and Western Middle Fields has been geologically mapped by Wood and associates (e.g., the Minersville Quadrangle, Wood, et al., 1968, and related cross sections). The maps depict the synclinoria and other complex geologic structures. The geologic structure and stratigraphy of the Southern Anthracite Field are described in Wood et al. (1969) and the depositional and structural history of the entire Anthracite Region are presented in Wood et al. (1986). The complexity of the geologic structure, resulting in nearly vertical beds of coal and other rocks in some areas of the anthracite fields, has impeded the acquisition of stratigraphic data from routine exploration drilling. Detailed mine maps of the abandoned underground mines and cross-sections through vertical shafts and nearly horizontal tunnels have added to the understanding of the structure and stratigraphy of the anthracite coal fields, however most stratigraphic efforts have been directed toward coal seam delineation.

The Pennsylvanian age rocks of the anthracite region of Pennsylvania have been divided into two major formations, the Pottsville and the Llewellyn. Generalized columnar sections of the Pottsville and Llewellyn Formations are shown on Figure 8.35.

The Pottsville Formation ranges in thickness from a maximum of approximately 1600 ft (490 m) in the Southern Field to less than 100 ft (30 m) in the Northern Field. The Tumbling Run and Schuylkill Members of the Formation are not present in the Northern Anthracite Field (Wood et al., 1969, 1986; Meckel, 1967, 1970; and Edmunds et al. 1979, 1998).

The Pottsville Formation contains up to 14 coal beds in some areas, but most are relatively discontinuous and only a few persist outside of the Southern Field (Edmunds et al. 1998). Figure 8.35 shows the mineable coals of the Pottsville Formation. The Lykens Valley Coal Numbers 4 through 7 are within the Tumbling Run Member; the Lykens Valley Coal Numbers 1 through 3 are within the Schuylkill Member; and the Scotty Steel and Little Buck Mountain Coals are within the Sharp Mountain Member of the Pottsville Formation (Figure 8.35). The base of the Buck Mountain Coal is considered the top of the Pottsville Formation in eastern Pennsylvania; however, the Buck Mountain coal is generally correlated with the lower Kittanning Coal within the lower Allegheny Group in western Pennsylvania (see Edmunds et al., 1998). The type section of the Pottsville Formation (located near Pottsville) is described by C.D. White (1900) and more recently by Wood et al. (1956) and Levine and Slingerland (1987).

The Pottsville Formation in eastern Pennsylvania is entirely of a nonmarine depositional environment (Edmunds et al., 1998). As in western Pennsylvania, the dominant lithology of the Pottsville Group is sandstone and conglomerate; but the Pottsville Formation of the Anthracite region contains significant pebble conglomerates derived from an orogenic source area to the southeast (Meckel, 1967, 1970; Edmunds et al. 1998; and Faill, 1997b). The Tumbling Run Member is composed of approximately 55% conglomerate and conglomeratic sandstone, about 30% fine- to coarse-grained sandstone, and about 15% shale and siltstone. Conglomerate and conglomeratic sandstone comprise about 50% of the Schuylkill Member, and the sandstone in the member ranges from very fine to very coarse, constituting approximately 30% of the member. The Sharp Mountain Member in most of the Southern Anthracite Field is composed of about 45% conglomerate, 25% conglomeratic sandstone, 15% sandstone, 5% siltstone, 9.5% shale, and 0.5% anthracite (Wood et al. 1969, 1986). The carbonate content of the rocks has not been determined.

The Llewellyn Formation is as much as 3500 feet (1050 m) thick. The maximum known thickness of the

Pennsylvanian in Pennsylvania is approximately 4400 ft (1340 m) near the town of Llewellyn in Schuylkill County (Edmunds et al., 1998). The Llewellyn Formation contains up to 40 mineable coals (Edmunds et al., 1998), most of which are shown on Figure 8.35. The thickest and most persistent coals occur in the lower part of the Llewellyn Formation, particularly the Mammoth coal zone. The Mammoth coal zone typically contains 20 ft (6 m) of coal and thicknesses of 40 ft to 60 ft (12 to 18 m) are not unusual. A local thickness of greater than 125 ft (38 m) has been reported in the Western Middle Field. This was attributed to structural thickening in the trough of the syncline. The nomenclature and stratigraphy of the coal bearing rocks of the Llewellyn Formation in the Northern Anthracite Field are different than in the Southern and Middle Fields (Figure 8.35).

The dominant lithology of the Llewellyn Formation is sandstone, including conglomerate units, as in the Pottsville Formation. According to Edmunds et al. (1998, p. 159): "Lithologically, the Llewellyn is a complex, heterogeneous sequence of subgraywacke clastics, ranging from conglomerate to clay shale and containing numerous coal beds. Conglomerates and sandstones dominate". The Llewellyn Formation in the Southern and Middle Fields is believed to be entirely continental in depositional environment (i.e., lacking any marine beds). The Llewellyn Formation in the Northern Field, however, contains one known marine bed, the Mill Creek Limestone (Figure 8.35). I.C. White (1903) suggested that the Mill Creek was correlative with the Ames limestone of western Pennsylvania. This belief is generally held to the present. The

Mill Creek Limestone is a one- to three-ft (0.3 m to 1 m), richly fossiliferous marine limestone (Chow, 1951). The Llewellyn Formation contains several other nonmarine limestones in the Northern Field including the Cannal and Hillman limestones (Chow, 1951, and Edmunds et al., 1998). Additionally, Inners and Fabiny (1997) have identified calcareous paleosols ("calcrete") in the uppermost Llewellyn Formation in the Northern Field. They have tentatively correlated this portion of the stratigraphy with the Conemaugh of western Pennsylvania. The calcretes are indicative of "seasonally semi-arid conditions" (Inners and Fabiny, 1997, p. 85). The calcretes are often associated with siderite. Kochanov (1997) has found calcareous sandstones in the lower part of the Llewellyn in the Northern Field. Siderite nodules are also common in the lower Llewellyn (Kochanov, 1997, personal communication).

The identification and mapping of limestone and other calcareous rocks in the Southern and Middle Fields have not been reported in the literature; however, some large mine pool discharges such as the Wadesville Colliery (see Table 8.14), have alkalinity of several hundred milligrams per liter, which must be attributed to some carbonate minerals in the overburden. Discharges in the Eastern Middle Field have little if any alkalinity (see Table 8.14). This strongly suggests a lack of calcareous rock in this coal field. Study of carbonate minerals and identification of calcareous lithologic units in the Southern and Middle Fields is needed.

Based on permits issued by the Department of Environmental Protection in glaciated regions in northwestern Pennsylvania, glacial sediments are most likely to be encountered where mining of the Brookville/Clarion, lower Kittanning and middle Kittanning coals takes place. These sediments affect mine water quality because they contain calcareous clasts. Lower Freeport permits have also been issued in glaciated regions. For Clarion/Brookville coal mines, it can be difficult to assess the impact of glacial sediments on postmining water quality because the Vanport limestone is often part of the overburden. Figure 8.36 shows the area of overlap of glacial sediments, Vanport limestone, and coal areas. The Vanport limestone becomes thinner and more discontinuous near its northern limit.

Glacial overburden in surface coal mines can exert a different influence on mine drainage quality than does bedrock. The texture, composition (mineralogy), and structure of glacial overburden differs from that of bedrock. Much information on the glacial geology and sediment composition exists for northwestern Pennsylvania.

Figure 8.36 Map showing the locations of glacial borders, Vanport limestone, and Pennsylvanian Period rocks and location of sites with glacial overburden that are discussed in the text.

Because of a lack of carbonate source rock, glacial sediments in northeastern Pennsylvania are typically low in carbonate. The only exception is in an area near the Delaware River, east of the anthracite fields (William D. Sevon, personal communication, 1994; Epstein, 1969). Several overburden holes have encountered till in the semi-anthracite Bernice Basin, Sullivan County. The till ranges in thickness from 0 ft (m) to 32 ft (9.8 m). The highest NP and sulfur (out of 32 samples) was 3 ppt CaCO₃ and 0.04 percent respectively, i.e., negligible. There are no overburden analyses of glacial sediments in the anthracite coal fields. Duane Braun (Bloomsburg University, personal communication, November 18, 1997) has done field work in the northern Field and has not observed any till reacting visibly to HCl. Available data suggests that glacial sediments are generally not a source of NP in northeastern Pennsylvania. Therefore, glacial sediments in northeastern Pennsylvania will not be discussed further.

Northwestern Pennsylvania has been glaciated a number of times during the Pleistocene. At least 8 tills have been identified on the Appalachian Plateau of northwestern Pennsylvania (White and others, 1969) and northeastern Ohio (White, 1982) (Table 8.3). The glaciations range in age from pre-Illinoian (> 500,000 ya) to late Wisconsinan (about 15,000 ya). Generally, each advance was less extensive than the previous. The last glacier to advance into Pennsylvania, the Ashtabula advance, stopped at the northwestern edge of the Appalachian Plateau, just south of Lake Erie. The Titusville glaciation deposited the bulk of the glacial sediment in northwestern Pennsylvania. Stacking of multiple Titusville sheets makes up the bulk of the main end moraine, the Kent Moraine (White and others, 1969; White, 1982).

Table 8.3 Pleistocene tills, from youngest to oldest and carbonate content in the clay fraction of those tills in northwestern Pennsylvania. Data from White and others (1969) and White (1982).

Individual tills in northwestern Pennsylvania are generally only a few feet (~1 m) to a few tens of feet (10s of m) thick, but can be over 100 feet (30 m) thick (White, 1982). Multiple tills of different composition may be encountered in the overburden during mining. White et al. (1969) show this for the Ambrosia mine near Grove City, PA (Figure 8.37). The Mapledale, Titusville, Keefer and Kent tills are present. The upper portion of the Mapledale and Titusville tills are leached of carbonates, where the unweathered portions contain calcareous minerals. Figure 8.38 shows the variable carbonate contents of tills in mines near the Ambrosia site.

Figure 8.38 Distribution of neutralization potential for glacial overburden from mine sites southeast of Grove City. Sites are located at "A" on Figure 8.36.

The compositions of sediments from different glaciations are determined by the source materials eroded by the glacier and incorporated into the sediments. Glaciers advancing into northwestern Pennsylvania advanced from the Lake Erie basin. The bedrock in the basin, which was the source for the glacial sediments, consists of a large amount of limestone and dolomite. Therefore, the sediment contained in the glaciers was high in carbonate minerals (calcite and dolomite) when it entered Pennsylvania. Early glaciers advanced far to the south of the Lake Erie basin over a surface of weathered siliceous bedrock and soil, resulting in dilution of the carbonate rich sediment by carbonate poor sediment. The resulting glacial deposits are relatively low in carbonate. Later glaciers advanced less far out of the Lake Erie basin over unweathered bedrock (the weathered horizons having been removed by previous glaciations), and over previously deposited glacial sediments. These glacial sediments were diluted less by the siliceous bedrock south of the Erie basin and are relatively high in carbonate, more characteristic of the source limestone and dolomite rocks in the Lake Erie basin (White, 1982).

Till in the coal fields, near the glacial margins, is probably lower in carbonate content than the average for that till. Because of the difference in source materials for succeeding glaciations, carbonate content generally increases from older tills to younger tills (White et al., 1969) (Table 8.3).

The Mapledale Till is the only till that has such a low carbonate content that unweathered till will not react visibly with dilute hydrochloric acid. It is completely leached of carbonates in the clay fraction in its outcrop area (Gross, 1967). Unfortunately, from a surface mining standpoint, only the older glaciations, mainly the Mapledale, Titusville, and Kent, advanced far enough to reach the coal fields of Pennsylvania. The higher carbonate tills (Ashtabula, Hiram) occur mainly in Crawford and Erie Counties (Figure 8.36). Although the Kent and Titusville Tills, in addition to the Mapledale Till, occur beyond the Kent Moraine, little neutralization benefit will be realized from glacial sediments beyond the Kent Moraine because they are all thin, discontinuous, and weathered. Fortunately, overburden data suggests that the amount of iron sulfide in glacial sediments is negligible (Figure 8.39), thus enabling the till to contribute a generally net positive influence on mine water quality. The highest sulfur in any of the tills is 0.38% in OB 3-Spag. All other till samples had less than 0.25% sulfur.

By comparison, in the Illinois Basin, most of the coal field is within the glaciated region and the bulk of the tills have carbonate contents greater than 10% in the clay fraction, and up to 64% in the coarse sand fraction (Fleeger, 1980), similar to the Pennsylvania tills near Lake Erie. The presence of glacial deposits is much more significant to the coal mining industry for the prevention of AMD in Illinois than in Pennsylvania. Although coal and associated strata in the Illinois Basin are high sulfur (Maksimovic and Mowrey, 1993), acid streams (pH < 6) were only identified south of the glacial border (Hoffman and Wetzel, 1993, 1995).

Tills have characteristic grain size distributions (Shepps and others, 1959). All have a significant amount of fine-grained material, that is, sand size and smaller. Groundwater movement through glacial sediments differs from that through consolidated bedrock in the Appalachian Plateau. The majority of groundwater movement through bedrock is along fractures and bedding planes. In glacial sediments, much more of the groundwater movement is intergranular, although some movement does occur through fractures in dense tills. The small grain size, poor sorting, and lack of fractures and bedding planes in tills results in low permeability. Water moves through tills very slowly, increasing the amount of time that the groundwater remains in contact with the mineral grains. Like the increased surface area, the increased residence time

allows for greater reactivity in tills than in shallow, fractured bedrock.

The neutralizing capability of a till varies considerably and is reported in the literature in various ways. Published carbonate values for tills of northwestern Pennsylvania are for the clay fraction only, and not representative of the till as a whole. In addition, the carbonate content of any size fraction varies areally due to dilution by local bedrock, and vertically due to dilution and weathering (White, 1969). The published mean carbonate contents for the clay fraction of tills likely to be encountered in mines in northwestern Pennsylvania are based on limited numbers of sample analyses (3 Kent, 9 Titusville, and 9 Mapledale) (Gross, 1967). For these reasons, converting published carbonate values for tills in northwestern Pennsylvania into NP values is not valid, and site-specific testing and analyses are required.

Four mines near Grove City in Mercer County (Location A on Figure 8.36), including the Ambrosia mine, demonstrate the variability that till can have. All are within the Kent Moraine. All mined the Brookville/Clarion coals. About 50% of the glacial overburden in the Ambrosia Mine (Figure 8.37) is calcareous. Two nearby sites, Oddfellow Mine (Figure 8.36 and Figure 8.38) and Brothers 3 Mine (Figure 8.37), have very little NP > 30, even though the depth of the glacial overburden is up to 60 feet (18 m). The fourth site, McCoy Mine (Figure 8.38), did have significant NP in the glacial overburden.

Carbonate minerals in tills with NP less than 30 ppt CaCO3 can probably dissolve to produce alkalinity. Calcareous minerals in till probably react more completely than those in bedrock because of the greater surface area of particles in till. Because of the source area (Erie Basin) the NP of glacial sediments is less likely to be complicated by siderite, and reflects actual neutralizing potential.

The overlap of the main bituminous coal field and glaciated region is mainly in Beaver, Lawrence, Butler, Mercer, and Venango Counties (Figure 8.36). The Slippery Rock, Mapledale, Titusville, and Kent Tills are present in these areas. Rare thin patches of Lavery Till may also be encountered in northwestern Lawrence County.

In glaciated areas of Pennsylvania, postmining water quality is frequently good, presumably due to the carbonate content of glacial sediments. Overburden for the Spagnolo mine is shown in Figure 8.39, and the location of this and the Owens mine are shown in Figure 8.36. Water quality at the Spagnolo Mine, at the time of this writing, has not changed significantly from the premining background quality Table 8.4. However, the Spagnolo Mine reclamation was completed about the time this was written. Thus, there has not been sufficient time to collect and evaluate the postmining water quality.

The Owens Mine, about 6,000 feet (1,830 m) from the Spagnolo Mine (Location B on Figure 8.36) was mined in the late 1960's and early 1970's, and left partially unreclaimed. No acid-base accounting data are available for this site. Its geology, however, is likely similar to the Spagnolo Mine, although, as has been previously shown at Grove City, there can be rapid lateral variability in the mineralogy of glacial sediment. Both are located within the Kent Moraine, have similar drift thickness (Schiner and Kimmel, 1976), and mine the same coal. Postmining water from the Owens mine in 1990 was higher in specific conductance, alkalinity, and sulfates than the pre- or during-mining water from the Spagnolo site Table 8.4. The Owens water is characteristic of neutralized mine drainage. Other parameters are largely the same as the Spagnolo site water quality. Not all mines in glacial overburden will have alkaline drainage. Older tills, especially in their outcrop area, are low in carbonate and may not exert much neutralization effect.

In summary, glacial overburden can be beneficial in preventing acid mine drainage if it is calcareous. Because of the small grain size, unlithified nature, and the source of carbonates in glacial sediments, the NP determinations of glacial overburden probably more accurately reflect the ability of the glacial sediments to prevent and neutralize acid mine drainage than is sometimes the case with bedrock overburden. Site specific data are required to determine the NP of glacial sediments because of their variability in the texture and composition due to dilution and weathering.

Several lithologic and mineralogic trends for the coal-bearing rocks of western Pennsylvania are evident from examination of Figure 8.2. The clay/shale content increases upward from a low of about 25% in the Pottsville Group to highs of 70 to 80% in the Dunkard Group. In general, there is a reverse relationship with sandstone, with sandstone increasing from ~20% or less in the higher Groups (with the exception of the lower Waynesburg Formation) to a high of 50% in the Pottsville. The percent calcareous rocks increases upward. The Pottsville Group has less than 1% calcareous rock. The lower Allegheny is less calcareous (6%) than the upper Allegheny Group (14%). Rocks of the Casselman Formation and higher are greater than 50% calcareous. The percentage of sandstone that is calcareous is generally similar to the percentage of overall rock that is calcareous. The uppermost Dunkard Group rocks contain the least amount of siderite, whereas over 20% of the rocks in the Allegheny Group and Glenshaw Formation are sideritic. Within the Allegheny and Pottsville Groups sideritic rocks are more abundant than calcareous rocks (Figure 8.2).

Some stratigraphic trends are also worth discussing. The Allegheny Group was divided into upper and lower Allegheny based on the presence of marine and brackish zones below the Johnstown limestone and the occurrence of freshwater limestones in the upper Allegheny, which are rare or absent in the lower Allegheny. The Conemaugh Group, like the Allegheny is divided such that the Glenshaw Formation contains most of the marine rocks, whereas the Casselman is primarily of freshwater origin. Mineable coals are generally restricted to the Allegheny and Monongahela Groups, although coals in the other groups are occasionally mineable.

The important geologic relationships, from a mine drainage standpoint, are the location of marine zones and distribution of calcareous rocks. The marine rocks are frequently associated with high sulfur strata, but can also have calcareous zones. Brackish rocks tend to have high sulfur and a lack of calcareous minerals. The conchostracan-bearing rocks frequently have less sulfur than their marine and brackish counterparts. Truly freshwater sediments tend to have calcareous minerals and limestone. These topics are discussed below in detail.

Little is known about the overburden mineralogy of the anthracite region. The Llewellyn Formation of the Northern Field is known to contain calcareous rocks. Some of the water quality reflects this source of alkalinity. Although stratigraphic studies have not yet identified calcareous rocks in the Southern and Western Middle Fields, highly alkaline mine waters imply that calcareous rocks exist. The lack of alkaline drainage in the Eastern Middle Field suggests a dearth of calcareous strata in this anthracite field.

Pleistocene glaciation in the bituminous coal field only came as far south as the northwestern corner of the field. Tills, where calcareous, can contribute significantly to the alkalinity of postmining water quality. Tills in the Northern Anthracite Field are not calcareous and would not contribute alkalinity to water.

The only places where the stratigraphy is important from a mining standpoint is where these strata will be disturbed by surface mining. Ultimately, postmining water quality is related to the mineralogy of these rocks.

Coal overburden is composed of many different minerals in varying abundance. As mentioned previously, mining accelerates the weathering of these minerals by exposing fresh rock surfaces, and from a mine drainage or soil reclamation standpoint, two groups of minerals are overwhelmingly important. These are the acid-forming iron sulfides and the acid- neutralizing carbonates. The weathering of these two groups of minerals dictate whether the mine spoil will produce alkaline or acidic water. These minerals need only be present in an abundance of a few percent or less to be significant. By comparison the other 95% or so of overburden minerals play only a minor role.

Iron sulfide (pyrite) oxidizes and dissolves to create acidic water. Calcareous minerals (Ca-rich carbonates, such as calcite) can dissolve to neutralize and inhibit acid production (Chapter 1). The inhibitory effect is probably due to neutral pH conditions reducing the catalytic effects of pyrite oxidizing bacteria, and the virtual elimination of pyrite oxidation by Fe⁺³ because of low solubility at near neutral pH. An example of the interaction between pyrite and a calcareous mineral (calcite) can be represented by:

FeS₂ + 2 CaCO₃ + 3.75 O₂ + 1.5 H₂O Þ

Fe (OH)₃ + 2 SO₄^2- + 2 Ca²⁺ + 2 CO_{2 (g)} (8.1)

In the above equation the acid generated by the pyrite is neutralized by the calcite. The ultimate products of this reaction are iron hydroxide, sulfate, calcium and carbon dioxide gas. Other equations can be written and other products produced, and these are discussed in Chapter 1. For the present chapter it is simply necessary to recognize that weathering of pyrite makes acid, and dissolution of calcareous minerals will neutralize acid and possibly inhibit pyrite oxidation.

In this section we will discuss the influence of forms of sulfur, pyrite morphology, and pyrite genesis as they relate to the production of acid-sulfate weathering products. Forms of sulfur that occur in coal overburden are sulfide, sulfate and organic. As discussed above, iron sulfide minerals (principally pyrite) are the culprit in the formation of acid mine drainage. Pyrite occurs in several crystal morphologies, ranging from micron-size framboids to millimeter (or larger) size euhedral crystals and coatings. Pyrite genesis has been suggested as a factor influencing pyrite reactivity, for example sedimentary pyrite is more reactive than hydrothermal pyrite (Hammack et al., 1988). Pyrite associated with Pennsylvania’s bituminous coal seams and overburden is sedimentary in origin.

Chemical Forms of Sulfur (Sulfur Mineralogy) in Overburden Rock - When overburden is analyzed, weight percent total sulfur is generally determined as a means of estimating pyritic sulfur and thus the acid-producing potential of the rock. Forms of sulfur can be determined as described in Noll et al. (1988), however because of difficulties with analytical methods, added cost of analysis, and the fact that most sulfur in overburden rock is pyritic, typically only total sulfur is determined.

Although pyrite may comprise only a few percent, or even a fraction of a percent, of the overburden rock, its importance to postmining water quality far outweighs its seemingly minor presence. An overburden that averages just a fraction of a percent sulfur, in the absence of neutralizing rocks, can create significant postmining water quality or revegetation problems if not dealt with properly.

Sulfide Sulfur - Two iron-sulfide minerals occur in bituminous coal and overburden. They are pyrite and marcasite. Both have the chemical formula FeS₂ and are 53.4 percent sulfur with the remainder being iron, but the two minerals have different crystallinity. For simplicity we will refer to iron sulfide minerals as pyrite.

The oxidation of pyrite results in the production of sulfuric acid and iron. This acid can dissolve other minerals and thus release undesirable ions such as aluminum and manganese. The oxidation and acid-weathering processes are discussed in Chapter 1. A summary equation for the process can be described as follows:

FeS₂ + 3.75 O₂ + 3.5 H₂O Þ

Fe (OH)₃ + 2 SO₄^-2 + 4 H⁺ (8.2)

As seen above, one of the products of pyrite oxidation is aqueous sulfate. Under evaporative conditions sulfate minerals can form.

Sulfate Sulfur - Sulfate minerals are generally secondary weathering products of pyrite oxidation. Nordstrom (1982) shows the sequence by which these minerals can form from pyrite (Figure 8.40). Many sulfate minerals have been identified in Pennsylvania overburden (Table 8.5). They are divided into acid-producing and non-acid-producing. These minerals (with the exception of barite) are typically very soluble and transient in the humid east. They form during dry periods and then are flushed into the groundwater system during precipitation events. The phases that contain aluminum or iron are essentially stored acidity and will produce acid when dissolved in water. Gypsum, which is not acid forming, is relatively uncommon in Pennsylvania, whereas other sulfate minerals such as pickeringite and halotrichite occur more commonly.

The dissolution of coquimbite will be given as an example of how dissolution of sulfate minerals can create acid:

Fe₂³⁺(SO₄)_3· 9H₂O Þ

2Fe(OH)₃ + 3 SO₄^2- + 3 H₂O + 6 H⁺ (8.3)

Furthermore, Cravotta (1991, 1994) showed how the dissolution of roemerite could oxidize pyrite in the absence of oxygen, and thus create acid.

FeS₂ + 7 Fe²⁺Fe₂³⁺(SO₄)_4· 14H₂O Þ

22 Fe²⁺ + 30 SO₄^2- + 16 H⁺ + 9 H₂O (8.4)

Figure 8.40 The overall sequence of mineral reactions for pyrite oxidation showing the relationships between oxidizing agents, catalysts and mineral products. From Nordstrom (1982).

The bottom line is that in Pennsylvania sulfate sulfur should be assumed to be acid producing unless the mineral can be verified as a non-acid producing one.

No systematic work has been done in Pennsylvania on pyrite weathering products, so the list in Table 8.5 is by no means exhaustive. Sulfate minerals are common to coal mine environments and have been reported from (among other places) Texas (Dixon et al., 1982) and Indiana (Bayless and Olyphant, 1982). Dixon et al. (1982) provide a list of 26 sulfate minerals "related or potentially related to lignite mine spoil..." A more complete list for Pennsylvania coal overburden would likely include most of the same minerals.

Organic Sulfur - Organic sulfur is sulfur that is tied up in organic molecules. This sulfur can originate by two processes. It can be associated with the original plant material, and it can be complexed with organic molecules during diagenesis. Casagrande et al. (1989) concluded that organic sulfur is not acid forming. Harvey and Dollhopf (1986) concluded that some forms of organic sulfur are acid producing, although the amount of acid produced in two leaching tests showed that organic sulfur contributed only 6 and 19 percent of total acidity. They showed, however, that theoretically some forms of organic sulfur should be capable of creating more acid than pyrite, whereas other forms are comparatively inert. Few, if any, other studies have been conducted on acid potential of organic sulfur.

Table 8.5 Secondary sulfate minerals identified in western Pennsylvania mine spoil and overburden. (Minerals from L. Chubb and R. Smith (PA Geologic Survey, personal communications, various years), Cravotta (1991, 1994), and observations by the authors. Mineral chemistries are from Roberts et al. (1990).

One factor complicating organic sulfur estimates is that organic sulfur is usually determined by "difference" (Noll et al., 1988). That is, total weight percent sulfur minus pyritic sulfur and sulfate sulfur. If pyritic sulfur or sulfate sulfur are incorrectly determined during analysis, organic sulfur will also be incorrect. In any event, acid production from organic sulfur is probably much smaller than from sulfide or sulfate sulfur.

It might be expected that rocks with high organic content would have greater amounts of organic sulfur compared with rocks with lesser amounts of organic matter. This appears to be true for coals, but not so for other rock types. A study by Brady and Smith (1990) examined the percentage of pyritic sulfur in relation to total sulfur for various rocks. Carbonalith (i.e., an organic-rich rock) did not have significantly different pyritic sulfur in relation to total sulfur than other rock types such as shale, sandstone and clay. The lack of higher amounts of organic sulfur in organic-rich rocks may be allied to the relationship between organic matter and pyrite formation, which is discussed below.

Pyrite Morphology - A considerable effort has been expended over the years looking at pyrite morphology and attempting to relate this to acid mine drainage generation. Some of the earliest work is by Caruccio (e.g. 1970), however numerous other individuals have also examined this issue. This seems like a logical thing to look at because pyrite is what causes acid mine drainage.

Morrison (1988) defined nine "classes" of pyrite morphology, end members being framboids and euhedral crystal structures. Framboids tend to be small (< 1 micron size), and euhedral are generally larger (tens to thousands of microns). For any given percentage of sulfur, framboids would have a proportionally larger surface area than euhedral crystals. Other classification systems have also been discussed (e.g., Arora et al., 1978; Hawkins, 1984). Caruccio (1970) and Morrison (1988) found a relationship between relative surface area and acid production, with the small particles more reactive than large particles. Hornberger and associates (1981,1985) found a statistically significant difference in the abundance of framboidal pyrite (i.e. framboids and framboidal clusters <10 microns) in lower Kittanning coal samples from marine and freshwater paleoenvironments. There were almost no discrete framboids in point counts of the freshwater samples, but the marine samples had many. However, some of the freshwater samples produced as much acidity and sulfate in leaching tests as the marine samples. One LK sample from a brackish paleoenvironment produced by far the highest acidity and sulfate of all of the LK coal samples tested, despite having a dearth of framboids (and total and pyritic sulfur contents similar to the marine samples). It seems reasonable that surface area would have an effect on reactivity, but it is not by any means the only factor controlling reactivity.

Discussion on Sulfur Minerals and their Relation to Acidic Water - Typically in Pennsylvania determination of total sulfur will adequately serve as a proxy for acid potential. This is because it includes the sulfur from acid-generating sulfide and sulfate minerals and typically the amount of organic sulfur in overburden rock is insignificant. In locations where gypsum, and other sulfur-bearing non-acid-forming materials are abundant, accurate determination of sulfide sulfur should provide a better prediction of acid potential.

This section will discuss the factors involved in the formation of pyrite. These factors are responsible for the amount of pyrite present in a stratigraphic interval, and thus that rock’s acid-producing potential. Pyrite consists of iron and sulfur, so clearly these ingredients are required. Other factors that influence the creation of pyrite are the presence of organic matter decomposable by sulfate-reducing bacteria (essential for creating reducing conditions), sedimentation rate and bioturbation.

Sulfur and Iron - Sulfate, the typical aqueous form of sulfur, is the source of sulfur in most sediments. The sulfur in sulfate is oxidized and must be reduced to sulfide to become part of pyrite. Ocean water contains about 2700 mg/L SO₄. Freshwater, on the other hand, is typically low in sulfate. An average river transports about 8 mg/L sulfate (Hem, 1985), although rivers draining arid and semi-arid regions can have much higher sulfate (hundreds of mg/L). Another source of sulfate that has been suggested as a source of sulfur for pyrite is sulfate from eroded evaporites. Erosion of upstream evaporites containing gypsum has been espoused by Querol et al. (1991) as the source of sulfur for some high sulfur coals in Spain and by Gibling et al. (1989) for high sulfur coals in the Maritime basin of Canada. Most studies of sulfur in rocks have dealt with marine sediments, where the water column provides a nearly unlimited supply of sulfate sulfur. Few studies are available for freshwater rocks and the formation of pyrite. The few that do exist seem to indicate that sulfur is low in freshwater sediments.

The concentration of iron in ocean water is very low (<.01 mg/L), whereas iron associated with fresh water is comparatively high (~0.7 mg/L average) (Hem, 1985). Iron can also be transported as detridal iron-bearing minerals (e.g., iron oxides) and bound to clay minerals. Thus iron can be transported into a marine or marginal marine environment by iron-bearing water and sediments.

It has been suggested that the most optimum environment for pyrite formation would be a brackish environment, where freshwater high in iron mixes with marine water high in sulfate (Guber, 1972). The subject of paleoenvironmental controls on pyrite formation are discussed later in this chapter.

Organic Carbon and its Relationship to Sulfur - Sulfate can only be reduced to sulfide in sediments by bacterial activity in the absence of oxygen. The bacteria oxidize organic carbon to provide energy for their metabolic processes and sulfate is the terminal electron accepting reactant. The sulfate is reduced to sulfide. Suitable reducing environments are typically found below the sediment-water interface. Although pyrite is the end product, intermediate iron sulfide minerals are created first (e.g., mackinawite (FeS); greigite (Fe₃S₄)). A schematic of this process is shown in Figure 8.41 (Goldhaber and Kaplan, 1982). Pons et al. (1982) have described this process with the following overall equation (CH₂O represents organic matter):

Fe₂O₃ + 4 SO₄^2- + 8 CH₂O + ½ O₂ Þ

2 FeS₂ + 8 HCO₃^- + 4 H₂O (8.5)

Figure 8.41 Pathway of sedimentary pyrite formation. From Goldhaber and Kaplan (1982).

A positive linear relationship has been shown between percent organic carbon and percent sulfur for Recent and Pleistocene marine sediments (Goldhaber and Kaplan, 1982; Raiswell and Berner, 1986). Fairly strong positive linear relationships exist for marine sediments back in time to at least the Jurassic (Holocene and Pleistocene, correlation coefficient (r) = 0.95; Upper Cretaceous, r = 0.73; Jurassic, r = 0.91), but more random scatter seems to be the case during the lower Carboniferous of Europe (Mississippian Period in the US) with r = 0.39 (Raiswell and Berner, 1986). The reason for the poor correlation between carbon and sulfur during the Carboniferous is not discussed. But the correlation coefficient of 0.39 indicates that carbon abundance explains only 15% of the variation in sulfur, while carbon explains 90% of the variation of sulfur in the Holocene and Pleistocene samples.

Despite the importance of pyrite and the abundance of organic matter in coal-bearing rocks of Pennsylvanian age, studies of the relationship between organic carbon and sulfur in these rocks are rare. No studies are known from Pennsylvania. Gilb (1987) looked at the relationship between organic carbon and sulfur in the Pennsylvanian Breathitt Formation of eastern Kentucky. Gilb separated the lithologies into carbonaceous shale (organic carbon between 5 and 30%), mudrock with calcareous concretions, and mudrock without calcareous concretions. Chesnut (1981) concluded that mudrocks with calcareous concretions were of marine origin. Gilb proceeded on this assumption, plus the assumption that the mudrocks lacking concretions are freshwater. The carbonaceous shale shows a positive relationship (slope = 0.13) between percent sulfur and percent organic carbon (% Org. C), with a correlation coefficient (r) of 0.60. The mudrocks with calcareous concretions also have a positive slope (0.6) and r = 0.72. The relationship between sulfur and organic carbon in mudrocks not having calcareous concretions is essentially random. Of 38 samples, all but four are less than 0.4% sulfur and most are less than 0.1% sulfur. It must also be remembered that these rocks were defined as "freshwater" based on negative evidence (lack of calcareous concretions), so the few high sulfur rocks (and perhaps some of the low sulfur rocks) may have a marine origin. Some of the samples Gilb included in his study had carbon less than 1%. Berner (1984) cautions that the use of sulfur to carbon ratios as a paleoenvironmental indicator should only be used for sediments with organic carbon greater than 1%.

Studies of pyrite formation in freshwater sediments are few in number. Most suggest that sulfur values will be low because sulfate availability is limited to only a small percentage of that present in marine waters (e.g., Berner and Raiswell, 1984; Davison, 1988). Berner and Raiswell (1984) have suggested that the carbon to sulfur ratio (C/S) can be used as a means of distinguishing freshwater from marine sediments. They found that marine rocks are characterized by C/S values of 0.5 to 5 and that freshwater rocks had C/S values >10 (Figure 8.42). This difference is attributed to the low sulfate levels of freshwater being limiting on pyrite production. They cautioned that this method should (1) only be applied to rocks with greater than 1% organic carbon, (2) not be applied to nearly pure limestones because low amounts of iron probably limited pyrite formation and (3) not be applied to coal because organic sulfur could be mistaken for pyritic sulfur in total sulfur analyses and the "superabundance of organic matter" would create a situation where pyrite formation becomes limited by iron availability. Also "...at such high organic-carbon concentrations, high C/S ratios can result even at high salinities." Gilb (1987) adds a further caution regarding the use of C/S ratios for coal. He points out that many swamp environments have low pH. This is not conducive to sulfate-reducing bacteria, which prefer neutral water conditions. "Therefore peat deposited in swamps having low pH may form low sulfur coal regardless of sulfate and iron availability" (p. 13). It would be interesting to see the results of a study of C/S for Pennsylvania’s Pennsylvanian rocks and compare these results with the other studies.

Figure 8.42 Frequency distribution of C/S weight ratios (plotted on a logarithmic X-axis) for British Carboniferous siltstones, mudstones and shales that were determined to be marine or non-marine by independent means. From Berner and Raiswell (1984).

The Berner and Raiswell (1984) study included modern freshwater lake sediments and British Carboniferous freshwater sediments. The freshwater sediments (modern and ancient) were, with few exceptions, low in sulfur (<0.5%). This fact differs from what we see in Pennsylvania, where presumed "freshwater rocks," especially organic-rich rocks, frequently exceed 0.5 % sulfur (for example see Figures 8.15, 8.17, 8.18, 8.19, 8.20, 8.21, 8.26, and 8.31). As discussed earlier, the upper Allegheny "freshwater rocks" are believed to have been deposited in a slightly-saline, marginally-brackish environment. For a comparison with marine and brackish rocks from the Allegheny and Conemaugh Groups, see Figures 8.7, 8.12, 8.13, 8.14, 8.23, and 8.24. It does appear, however, within the Allegheny Group and Conemaugh Group that high-sulfur, marginally-brackish strata overlying coals are typically not as thick as the high-sulfur strata associated with brackish and marine environments. For example the high sulfur (>0.5% S) associated with the UK, LF and UF in Figures 8.15, 8.17, 8.18, 8.19, 8.20, 8.21, and 8.26 is one to 5 ft (0.30 to 1.5 m) thick, whereas the thickness of the high-sulfur strata in marine and brackish environments (Figures 8.7, 8.12, 8.13, 8.14, 8.23, and 8.24) can be up to 30 ft (10 m). Freshwater rocks associated with coals of the Monongahela and Dunkard Groups (Figures 8.31, 8.32, and 8.34) seem to have thicker sequences of high-sulfur strata than the Allegheny overburden. Since there seems to be no evidence of marine influence within this section, this may be due to more arid conditions resulting in higher sulfate levels in surface water and groundwater. A modern relationship between aridity and high dissolved solids, including sulfate, has been noted in the arid and semi-arid western United States. Hem (1985) notes that in semi-arid regions, the amount of sulfate is large in proportion to the water volume in which it is carried away, thus elevated sulfate.

Sedimentation Rate and Bioturbation (Open vs. Closed System) - Sedimentation rate and bioturbation can influence whether the sediment just below the sediment/water interface is an open or a closed system with respect to sulfate in the overlying water column. Bloch and Krouse (1992), studying rocks of the Cretaceous period in western Canada, found the highest sulfur concentrations (>2.75%) associated with bioturbated marine mudstones and siltstones, when compared to mudrocks that were not bioturbated (0.1 to ~2%). Bioturbation irrigated the upper portions of the sediment with sulfate from the overlying water column, thus creating an open-system source of sulfur. They also suggested that some of the low sulfur sediments were due to fast sedimentation rate, resulting in a more closed system where sulfate would be limited to that entrapped within the sediment pore-water. Bloch and Krouse calculated, assuming a porosity of 75% for the original sediment and 100% retention of sulfur (from sulfate) within the sediment, that a closed marine sediment would be limited to 0.1% sulfur. Berner (1970, p. 2) gives a higher figure, saying that "simple burial of sea water and reduction of all included sulfate can provide only about 0.3 percent pyrite sulfur..." For this to be true, the porosity of the sediment would be around 90%. Another important factor is that pyrite sulfur probably only represents a small fraction of the total hydrogen sulfide, H₂S, produced by sulfate reduction (Dean and Arthur, 1989). Jorgensen (1977) in studying a marine sediment in the coastal waters of Denmark found that only 10% of H₂S was fixed as iron sulfide, the remainder was lost by diffusion and reoxidation. The point Berner (1970) and Bloch and Krouse (1992) make is that for sulfur to be higher than a few tenths of a percent, diffusion and/or irrigation by burrowing organisms must occur to permit sufficient sulfate into the sediment to achieve high sulfur values. Movement of H₂S away from sites with high organic matter to other sites with available iron could result in pyrite formation in the absence of organic matter. This may explain the high sulfur in some sandstones that "cut" down to the coal and overlay the coal. Typically the high sulfur zones occur near the base of the sandstone. Drill Hole A-5 and OB-4 in Figure 8.13 are examples of such a sandstone. The bottom 5 ft (1.5 m) of the sandstone in A-5 has greater than 0.5% S, with the interval from 3 to 5 ft (1 m to 1.5 m) above the coal having greater than 8% S. The base of the sandstone in OB-4 has 6.15% S.

Goldhaber and Kaplan (1982) also recognize the potential importance of bioturbation to increase the amount of sulfate available for reduction. They, however, present an extended argument to show that sulfate reduction rates are greater at higher sedimentation rates. Higher sedimentation rates tend to help preserve organic carbon, a component necessary for bacterial conversion of sulfate to sulfide. Their comparisons of sedimentation rate showing higher sulfide with higher sedimentation rates, are made with H₂S and "acid volatile" sulfides, not pyrite. Goldhaber and Kaplan do not discuss how these rates correspond to final pyrite concentrations. Berner (1984) concluded that "rapid burial enables relatively reactive (organic) compounds and more organic matter in general to become available for bacterial sulfate reduction at depth...These considerations help to explain why there is a crude correlation between sedimentation rate on the one hand, and organic matter and pyrite contents on the other" (p. 607).

Curtis and Spears (1968) looked at iron minerals, principally pyrite and siderite, in sediments of the Carboniferous Period coal measures and of the Jurassic Period in England. Pyrite and siderite both require reducing conditions. In the Carboniferous and Jurassic sediments they observed that pyrite and siderite were not randomly distributed in the sediment, but that siderite-bearing mudstones occurred above shales and mudstones containing pyrite. They concluded that slow sedimentation rates in marine sediments favored pyrite formation (more availability of sulfur) and fast rates favored siderite formation.

It is obvious from the conflicting conclusions that more work on the subject of pyrite formation as a function of sedimentation rates is warranted. High sedimentation rates would tend to create a "closed" system which, in the absence of bioturbation, would result in little input of sulfate from the overlying water column, and limited sulfate available for additional sulfate reduction. High sedimentation rates, however, preserve organic matter that is necessary for bacterial reduction of sulfate. In any event, it appears that bioturbation is important in irrigating the sediment such that additional sulfate is available for reduction.

Discussion on Formation of Pyrite - Although there is a clear positive relationship between percent organic carbon and percent sulfur in Recent and Pleistocene marine sediments, the relationships between these parameters (plus percent iron) in older marine sediments is complex. Raiswell and Berner (1986) examined British Carboniferous marine sediments and found that there was considerable scatter and a low correlation coefficient for C/S ratio. They however suggest that the overall C/S ratio remained relatively constant from the Devonian to Tertiary (~2). It is somewhat higher in more recent sediments (~2.8), and lower (<1.0) in older Paleozoic sediments.

Gild's (1987) findings on the relationships between organic carbon and sulfur are similar to those by Berner and Raiswell (1984). Even though no studies relating sulfur to organic carbon have been performed on Pennsylvanian rocks in Pennsylvania, the general findings of Gilb (1987) and Berner and Raiswell (1984) are probably applicable. Typically there will be a relationship between percent organic carbon and percent sulfur for marine rocks; higher sulfur values will be found in marine mudstones than in freshwater mudstones; and carbonaceous rocks will typically contain more pyrite than noncarbonaceous rocks for any given paleoenvironment. More discussion on paleoenvironmental and rock-type controls on sulfur will be presented later.

The higher the content of organic matter the darker the rock tends to be. If a mudstone is known to be of marine or brackish origin and it is dark in color, there is a good chance that it is also high in sulfur. Carbonaceous rocks (> 5% organic carbon) may be high in sulfur, at least relative to other rocks, regardless of paleoenvironment. This can be useful in helping to identify potentially high sulfur rocks in the field or in drill cuttings/cores.

Carbonate minerals play an extremely important role in determining postmining water chemistry. They not only neutralize acidic water created by pyrite oxidation, but as discussed above, there is evidence that they also inhibit pyrite oxidation (Hornberger et al., 1981; Williams et al., 1982; Perry and Brady, 1995). These processes are discussed more fully in Chapters 1 and 11. Brady, et al. (1994) determined that the presence of as little as 1% to 3% carbonate (on a mass weighted basis) on a mine site can determine whether that mine produces alkaline or acid water. Although pyrite is clearly necessary to form acid mine drainage, the relationship between the amount of pyrite present and water quality parameters (e.g., acidity) is only evident where carbonates were absent. Neutralization potential, a measure primarily of the carbonate content of the overburden, relates positively to the alkalinity of postmining water.

A knowledge of the distribution, amount, and type of carbonates present on a mine site is extremely important in predicting the potential for postmining problems and in designing prevention plans.

Carbonate Mineralogy - The most common carbonate minerals found in coal mine overburden are listed in Table 8.6. Carbonate minerals are often not "pure" end members, but form solid solution series with cation substitution.

Table 8.6 Common Carbonate Minerals in Mine Overburden, listed in descending order of their capability to neutralize acid.

The above minerals, with the exception of siderite, will neutralize acid generated from pyrite oxidation. An example of this neutralization process is:

CaCO₃ + 2H⁺ Þ Ca²⁺ + CO₂ + H₂O (8.6)

Not all carbonates are created equal when it comes to neutralization of acid. Calcite is more soluble than dolomite although the overall dissolution is similar to that shown for calcite (Geidel, 1982). Both calcite and dolomite will neutralize acid, and potentially inhibit pyrite oxidation. Siderite is less soluble than calcite and dolomite, and its role as an acid-neutralizer is either limited or non-existent as discussed below.

Siderite is a very common mineral in coal measures (e.g., Texas - Senkayi et al. 1986; and Japan - Matsumoto, 1978), including those in Pennsylvania (Morrison et al, 1990a, 1990c; and Figure 8.2). Siderite will form preferentially to calcite where iron is abundant (Blatt et al., 1972). It forms under reducing conditions where sulfur availability (as sulfate) is limited (e.g., Curtis and Spears, 1968; Postma, 1982). Where reduced sulfur is abundant, pyrite will form in preference to siderite. Dissolution of siderite can create weakly acid conditions in a closed system (Barnes and Romberger, 1968; Cravotta et al., 1990):

FeCO₃ + 0.25 O₂ + 2.5 H₂O Þ

Fe(OH)₃ + H₂CO₃ (8.7)

Even in an open system, the weathering of siderite, because of the hydrolysis of iron, has no net neutralizing effect. As discussed in Chapter 6 (see also Skousen et al., 1997), tests for neutralizing potential can give misleading results if siderite is present. Siderite may cause the test to appear positive for neutralizers. Additionally, siderite can contribute iron to water, and is probably responsible for at least some, and possibly much of the Mn observed in some mine waters. Mn is a common substitute for iron in siderite (Morrison, et al., 1990c).

Curtis and Spears (1968) attributed whether pyrite or siderite formed to the relative anion concentrations. A factor that may contribute to the bicarbonate ion concentration is that bicarbonate is produced during sulfate reduction (equation 8.5). Presley and Kaplan, (1968) observed an increase in bicarbonate with a corresponding decrease in sulfate in some modern near-shore marine sediments. Under reducing conditions and a more or less closed system, sulfate reduction would occur first with iron fixed as pyrite; once the sulfate was consumed, and assuming iron is available, siderite could form. A variety of factors may result in no clear relationships between the relative amounts of pyrite and siderite. For example, pyrite forms under slightly acidic to alkaline conditions, thus bicarbonate may have been present in the pore waters even before sulfate reduction. Also, Presley and Kaplan (1968) found more biogenically produced bicarbonate than could be explained solely by sulfate reduction. They attributed this to other biologic processes such as fermentation. The potential loss of sulfide due to oxidation or diffusion and the presence of preexisting bicarbonate may also result in a lack of relationships. No work similar to that by Curtis and Spears (1968) has been done in the Pennsylvanian of Pennsylvania.

The presence of siderite does not by itself reflect a particular depositional environment. Siderite occurs in sediments of freshwater, brackish, and marine origin (Matsumoto, 1978; Mozley, 1989; Morrison, 1988). Siderite is normally an early diagenetic mineral, formed below the sediment-water interface. Siderite can also form in terrestrial soil forming processes and is associated with some underclays (Gardner et al., 1988).

Mozley (1989) looked at compositional variabilities of siderite from marine and freshwater sediments. He found that marine siderite is impure with no more than 90 mole percent (mol %) FeCO₃. It can have Mg substituting up to 41 mol %, and Ca up to 15 mol % in the siderite crystal lattice. Freshwater siderite is commonly very pure (> 90 mol % FeCO₃), and can have higher Mn concentrations compared to marine siderite. Unfortunately Mozeley did not discuss siderite from brackish environments, and did not study "(s)iderite from complex depositional sequences with intercalated marine and nonmarine strata (e.g., coal-measure siderite)...", because of the possible early mixing of pore waters from different paleoenvironments. No efforts similar to Mozley’s have been conducted in the Pennsylvanian of Pennsylvania. It would be interesting to see if siderite composition varied by paleoenvironment. A better understanding of siderite composition would be useful for a variety of purposes. First, impure siderite may offer some neutralizing capacity; second, it would be a step in determining to what extent siderite contributes to Mn in mine waters; and third it may offer insight into paleodepositional environments.

Distribution of Carbonates in the Pennsylvanian of Western Pennsylvania - Within the Allegheny Group in Pennsylvania, siderite is the most common carbonate mineral (Morrison, 1988; and Figure 8.2). It occurs throughout the Pennsylvanian of Pennsylvania, regardless of paleoenvironment (Figure 8.2). An important aspect of siderite, in terms of mine drainage, is that it is the only carbonate that normally occurs in brackish sediments (see Table 8.15 in Appendix). This, coupled with the high sulfur content of brackish rocks, often leads to poor quality water from mines developed in brackish sediments.

Calcite and dolomite occur in the freshwater limestones of the Allegheny, Conemaugh, Monongahela, and Dunkard Groups. Additionally, in the Monongahela and Dunkard Groups, calcite and dolomite are the most common carbonate minerals present in other rock types (Figure 8.2). Marine limestones and associated rocks (Allegheny and Conemaugh Groups) can have a high carbonate content. As shown previously the Vanport limestone can be over 90% CaCO₃.

Chemical composition data for the Vanport limestone and other Pennsylvanian limestones and dolomites are shown in Table 8.7. Data from a few Cambrian and Ordovician limestones and dolomites, including the Valentine limestone from Centre County are included for comparison, as these high-calcium limestones are well known for their purity and industrial significance (see chemical composition data and additional discussion in O’Neill, 1964, 1976; Rones, 1969; and Chapter 7 in this volume). Table 8.7 shows that some marine and freshwater limestones of the Pennsylvanian Period exceed 90% calcium carbonate, which is also evident from NPs exceeding 900 ppt on Figures 8.7, 8.18, 8.20, and 8.39. Chemical composition data for the freshwater Fishpot, Redstone and upper Freeport limestones are shown in Table 8.7. These freshwater limestone samples typically have lower calcium carbonate contents than the Vanport limestone samples included in Table 8.7. NPs in the range of 500 to 875 ppt CaCO₃ for some of these stratigraphic intervals are also shown in Figures 8.17, 8.18, 8.19, 8.20, 8.25, 8.26, and 8.32. The origins of limestones and other calcareous rocks will be discussed in more detail in a following section of this chapter.

Not much is known about limestones and other calcareous rocks of the Anthracite Region. As mentioned earlier and shown on Figure 8.35 one marine limestone and several freshwater limestones occur in the northern field. No analyses of these limestones are known. Siderite, at least in the form of nodules, is fairly common in the Anthracite Region. Another possible source of calcite is secondary mineralization along fractures.

The most effective minerals, in terms of acid neutralization, are the carbonates. Some other minerals, however, can also neutralize acid. For example silicate minerals can neutralize acid, but the reaction rates at near neutral pH tend to be slow. Quartz, the most common silicate, is for all practical purposes inert. Many of the other silicate minerals, however, can provide buffering at low pH’s (Stumm and Morgan, 1981; Lapakko, 1992). The following example shows how anorthite can neutralize acid.

CaAl₂Si₂O_{8 (s)} + 2 H⁺ + H₂O Þ

(anorthite)

(kaolinite) (8.8)

Another source of neutralization is cation exchange on clay particles. Clay particles tend to have negative charges at their surfaces. These attract positively charged ions. Thus, H⁺ can adhere to the clay surfaces and raise the pH of the solution.