CHAPTER 2

Petrography and geochemistry of selected lignite beds in the Gibbons Creek Mine (Manning Formation, Jackson Group, Paleocene) of east-central Texas

By Peter D. Warwick, Sharon S. Crowley, Leslie F. Ruppert, and James Pontolillo

U.S. Geological Survey, MS 956, Reston, VA 22092

ABSTRACT

This study examined the petrographic and geochemical characteristics of two lignite beds (3500 and 4500 beds, Manning Formation, Jackson Group, Eocene) that are mined at the Gibbons Creek mine in east-central Texas. The purpose of this study was to identify the relations among sample ash yield, coal petrography, and trace-element concentrations in lignite and adjoining rock layers of the Gibbons Creek mine. Particular interest is given to the distribution of 12 environmentally sensitive trace elements (As, Be, Cd, Cr, Co, Hg, Mn, Ni, Pb, Sb, Se, and U) that have been identified as potentially hazardous air pollutants (HAPs) in the United States Clean Air Act Amendments of 1990.

Eleven lignite, floor, and rock parting samples were collected from incremental channel samples of the 3500 and 4500 beds that were exposed in a highwall of pit A3 at the Gibbons Creek mine. Standard proximate and ultimate analyses were performed on all lignite samples, and lignite and rock samples were analyzed for 60 major, minor and trace elements. Representative splits of all lignite samples were ground and cast into pellets, and polished for petrographic analyses in blue-light fluorescence and reflected white light to determine liptinite, inertinite, and huminite maceral percentages.

The following observations summarize our results and conclusions about the geochemistry, petrography, and sedimentology of the 3500 and 4500 beds of the Gibbons Creek lignite deposit: (1) Average dry (db) ash yield for the two beds is 29.1%, average total sulfur content is 2.5%, and average calorific value is 7800 Btu (18.15 MJ/kg). Ash yields are greatest in the lower bench (59.33% db) of the 3500 bed and in the upper bench of the 4500 bed (74.61% db). (2) For lignite samples (on a whole-coal basis), the distributions of two of the HAPs (Pb and Sb) are positively related to ash yield, probably indicating an inorganic affinity for these elements. By using cluster analysis we found that Be and Cd were poorly associated with ash yield, indicating a possible organic affinity, and that Ni, Se, Hg, U, and Pb cluster with most of the rare-earth elements. (3) The dominance of the crypto-eugelinite maceral subgroup over the crypto-humotelinite subgroup suggests that all Gibbons Creek lignites were subjected to peat-forming conditions (either biogenic or chemical) conducive to the degradation of wood cellular material into matrix gels, or that original plant material was not very woody and was prone to formation of matrix gels. The latter idea is supported by pollen studies of Gibbons Creek lignite beds; results indicate that the peat was derived in part from marsh plants low in wood tissue. (4) The occurrence of siliceous sponge spicules in the lower benches of the 3500 bed suggests the original peat in this part of the bed was deposited in fresh, standing water. (5) The petrographic data indicate that the upper sample interval of the 3500 bed is greater in inertinite (3%) than the other samples studied. This increase in inertinite may be associated with oxidation or drying of the peat that formed the upper 3500 lignite bench.

INTRODUCTION

The purpose of this study is to examine the sedimentologic, petrographic, and geochemical characteristics of two lignite beds (Eocene, Manning Formation, Jackson Group) that are mined at the Gibbons Creek mine of east-central Texas (figs. 1 and 2). The lignite is used to produce electrical power for eastern Texas at mine-mouth pulverized coal generating stations. This study seeks to identify relationships between sample ash yield, coal petrography, and trace-element concentrations in lignite and adjoining rock layers. Particular interest is given to the distribution of twelve of the environmentally sensitive trace elements (As, Be, Cd, Cr, Co, Hg, Mn, Ni, Pb, Sb, Se, and U) that have been identified as possible hazardous air pollutants (HAPs) in the United States Clean Air Act Amendments of 1990.

PREVIOUS INVESTIGATIONS

The general geology of the Manning Formation and the Gibbons Creek lignite deposit has been reviewed in several studies (Russell, 1955; Eicher, 1985; Swaldi and Cezeaux, 1986; Yancey and Yancey, 1988; Yancey and Davidoff, 1991, 1994; Yancey and others, 1993; and Davidoff and Yancey, 1993). The Manning Formation is approximately 150 m thick and consists of interbedded sandstone and mudstone and five (or more in some areas) lignite zones (fig. 2). These strata gently dip about 1^o to 2^o toward the southeast and are offset locally by minor folding and faults with as much as 45 m of displacement. The lignite zone names have been designated in ascending order the 2000, 2500, 3500, 4500, and 5500 zones (Swaldi and Cezeaux, 1986). These zones contain as many as 12 minable lignite beds (generally <1 m thick) (Eicher, 1985; Swaldi and Cezeaux, 1986). The rocks closely associated with the lignite beds are dominated by siltstone and claystone that are usually burrowed and rooted. Yancey and Yancey (1988), Yancey and Davidoff (1991, 1994), and Davidoff and Yancey (1993), have described the lithostratigraphy and sequence stratigraphy of the Manning Formation based on subsurface and outcrop data. From outcrops of the Manning at the Lake Somerville spillway (fig. 1), Yancey (this volume) has recognized several fluctuations in depositional environments. The depositional environments identified range from shallow marine to fresh-water mires. In addition, Ruppert and others (1994) and Yancey (this volume) have described several tonstein beds that occur in the lignites of the 3500 zone and at the exposures at the Lake Somerville Spillway outcrop (fig. 1).

Three reports (Tewalt, 1986; Oman and Meissner, 1987; and Mukhopadhyay, 1989) have reviewed chemical and physical data from Manning Formation lignite samples. Tewalt (1986) described the characteristics of 126 samples from the Jackson Group trend in east Texas. Oman and Meissner (1987) reported average lignite chemical data from the Gulf Coast area that included Jackson Group samples from Texas, and Mukhopadhyay (1989) presented data from 20 samples collected mainly from drill core from an area about 100 km south of the Gibbons Creek mine. These data indicate that the mean dry (dry basis - db) ash yield for Jackson Group lignites ranges from 10 to 41%, with an overall mean of 36%. Mean concentrations for equilibrium moisture are 37% (Tewalt, 1986), and for total sulfur content are 2.2% (db) (Tewalt, 1986). Organic sulfur is the dominant sulfur type. Dry Btu values for the Jackson Group samples average 8,047 (18.7 MJ/kg) (Tewalt, 1986). Tewalt (1986) and Mukhopadhyay (1989) report the rank of Jackson Group deposits as lignite.

Mukhopadhyay (1989) first examined the petrographic characteristics of channel and lithotype samples collected from two locations at the Gibbons Creek mine and from each of the three lignite beds exposed at the Somerville Spillway (fig. 1). Unfortunately, geodetic locations and lignite bed names are not included in his report, so as a result comparisons to the results of this study are somewhat speculative.

Gennett and others (1986) studied several Manning Formation lignite samples collected near the Gibbons Creek mine and found an abundance of Momipites and Caprifoliipities-Salixipollenites. They suggested these pollen types were consistent with the deltaic depositional model proposed for the Manning Formation by Fisher and others (1970). Mukhopadhyay (1989) reported that samples obtained from the upper part of one of the lignite beds sampled at the Gibbons Creek mine contained abundant pollen grains of Nyssa. The lower part of the bed was rich in Engelhardtia, Calamuspollenites, Liliacidites, and Arecipites. Mukhopadhyay suggested that this transition represents a marsh depositional setting for the lower part of the bed and a more developed swamp/mire for the upper part of the bed. Using both palynological and petrographic data, Mukhopadhyay (1989) concluded that the Gibbons Creek and Lake Somerville lignite deposits formed in 1) marshes associated with lower delta plain settings (marsh environments, rich in densinite [or degraded plant material] and Engelhardtia), and 2) in the transition zone between lower and upper delta plain environments (swamp/mire environments rich in ulminite/textinite [or preserved woody material] and Nyssa). These interpretations were supported by the regional deltaic depositional models proposed for the Jackson Group of east Texas by Fisher and others (1970) and Kaiser (1982). N.O. Frederiksen (U.S.G.S., written commun., 1993) also found that the floor of the 3500 bed contained fern spores and some bisaccate pollen. The occurrence of the spore Azolla, a water fern, led Frederiksen to suggest that the floor rock was deposited in standing, shallow water.

Ruppert and others (1994), reported that a 2-cm-thick claystone layer (fig. 3) in the upper part of the 3500 bed at the Gibbons Creek mine is an altered volcanic ash parting (tonstein). This layer, as well as a layer at the base of the middle lignite bed exposed at the Lake Somerville Spillway, contains kaolinite and accessory quartz, euhedral to subhedral zircons, feldspars, and Ca-Al phosphates (crandallite?) which indicates a volcanic ash-fall origin. Yancey (this volume) also identified a reworked volcanic ash from the base of the 3500 lignite bed at the Gibbons Creek Mine. Yancey suggests that the 3500 - 4500 lignite interval might be correlative to the upper part of the Lake Somerville interval because of the abundance of volcanic ash at both locations. Ruppert and others (unpublished data), also noted the occurrence of siliceous sponge spicules in the volcanic parting identified in the upper part of the 3500 bed. A freshwater environment of deposition was suggested for the ash layer because of the presence of the spicules.

Jackson Group near-surface lignite resources for east Texas, including the region containing the Gibbons Creek deposits, have been estimated to be about 4.5 billion short tons (Kaiser and others, 1980). The Keystone Coal Industry Manual (Maclean Hunter Publishing Company, 1994) reports the total resources for the Gibbons Creek mine holdings to be about 130 million short tons.

ACKNOWLEDGEMENTS

The authors wish to thank Randy Harris (Texas Municipal Power Agency) and Jan Horbaczewski (Navasota Mining Company, Inc.) for access to the Gibbons Creek mine and for permission to collect samples.

METHODS

Lignite, rock parting, and floor samples for this study were collected from incremental channel samples of the 3500 and 4500 beds (fig. 3) exposed in a highwall at an active opencast pit (A3) in the Gibbons Creek lignite mine (fig. 1). Incremental bench samples were collected in order to determine the vertical variability of physical and chemical characteristics of the lignite bed. The megascopic character of the lignite beds was used to determine the sample-interval size (fig. 3; appendix I).

Standard proximate and ultimate analyses were performed on all lignite samples and the results are recorded in table 1 and figure 4. Lignite and rock samples were analyzed for 60 elements using an analytical suite consisting of inductively coupled plasma (ICP) - atomic emission spectroscopy, ICP - mass spectroscopy, atomic absorption spectroscopy, X-ray fluorescence, and molecular absorption spectroscopy (tables 1 and 2).

Representative splits of all lignite samples were ground and cast into pellets, and polished for petrographic analyses according to the procedures outlined by ASTM (1993). The pellets were etched by acidified potassium permanganate (Stach and others, 1982) which permitted detailed subdivision of the huminite maceral group. An initial petrographic examination of the pellets was made in blue-light fluorescence by counting 1000 point counts on duplicate pellets to determine maceral percentages within the liptinite group. Subsequently, 1000 point counts were made in reflected white light to determine inertinite and huminite maceral percentages. A total of 22 maceral types was counted and recorded as percentages on an ash-free basis (fig. 5, table 3). The low-rank coal petrography terms used in this paper follow the nomenclature described in ICCP (1963, 1971), Stach and others (1982), and Warwick and Stanton (1988).

Bar graphs (fig. 6) show the vertical change in the concentration of the HAPs trace-elements (whole-coal basis). To test for trace-element associations and affinities with the organic or inorganic fractions, correlation coefficients (r) for all constituents were calculated on a whole-coal basis to show the strength of the relationships. All correlation coefficients reported in this paper have a 5% probability error. Also, cluster analysis (Parks, 1966) using unweighted pair-group averages and squared Euclidean distances was used to determine similarities between variables that included 1) ash, elemental, and oxide concentrations, and 2) the petrographic and chemical characteristics of each sample (fig. 7). R-mode cluster analysis (clustering of variables having similar influences on the composition of the samples) was performed on the geochemical data, and Q-mode cluster analysis (clustering of variables having similar compositions) was performed on the petrographic data.

RESULTS

SEDIMENTOLOGICAL CHARACTERISTICS

Several lithologic types are found associated with the 3500 and 4500 beds. In close association with the lignite beds are mudstone and claystone layers that are burrowed, rooted, and contain varying amounts of organic material (fig. 3; appendix I). The interval between the 3500 and 4500 beds contains a 3.5 m thick cross-bedded sandstone unit that coarsens upward in grain size. This unit is underlain by a 1.5-m-thick intensely burrowed sandstone unit (appendix I). These combined sandstone units probably represent lower and upper shoreface deposits respectively and the over and underlying lignite beds and associated fine-grained clastic rocks represent supratidal mires that developed during times of relative shoreline stability. Similar depositional environments have been described for the Gibbons Creek interval by Yancey (this volume).

GEOCHEMICAL AND PHYSICAL CHARACTERISTICS

Proximate analyses of the 8 lignite samples collected from the Gibbons Creek mine indicate that the lower bench of the 3500 bed and the top bench of the 4500 bed contains greater than 50% ash (db; fig. 4). The other samples combined average dry ash yield is 24%; total sulfur content averages 2.8% (dominated by organic sulfur, 2.3%); and dry calorific values average 7800 Btu (18.15 MJ/kg) (table l; fig. 4). As-received moisture content averages 50% for the 6 samples with less than 50% ash content (db). Total sulfur content is greatest in samples 3500.3 (3.81% db) and 3500.4 (5.42%) of the lower bench of the 3500 bed.

Trace element data from the channel sample sites are listed in table 1 and the distribution of the 12 HAPs elements (on a whole coal basis) for both beds are plotted on figure 6. Visual analysis of these bar graphs shows that some of the HAPs elements (As, Hg, Ni, Se) are concentrated in the organic-rich (or lignite) samples. Some HAPs elements appear to be concentrated in the volcanic ash parting (3500.P1 - Cd, Sb) and the lower parting (3500.P2 - Cr, Pb, Sb) of the 3500 bed. The concentration of HAPs elements is generally greater in the 4500 bed than found in the 3500 bed. Of the 12 HAPs elements, only Pb (r = .68) and Sb (r = .88) had significant positive correlation with ash yield. The concentration of most of the major oxides correlated with lignite ash yield (SiO₂, r = .98; Al₂O₃, r = .90; MgO; Na₂O, r = .76; P₂O₅, r = .83; and K₂O, r = .71).

ORGANIC PETROLOGY

The 3500 and 4500 lignite beds generally exhibit a dull, somewhat massive texture near their bases. The texture grades upwards into attrital-rich lignite with increasing xylinitic bands (3-4 cm thick) towards the top of the individual benches. Locally compressed logs are exposed in cross section on the exposed face of the lignite benches (appendix I). The upper part of the lower lignite bench of the 3500 bed (sample interval 3500.3, fig. 3 and appendix I) appears to contain the greatest percentage of xylinitic banding when compared to the other benches. The 3500 and the 4500 beds are dominated by the huminite maceral group (92%) followed by the liptinite (7%), and inertinite (1%) maceral groups (table 3; fig. 5). The three dominant huminite maceral subgroups in decreasing order of abundance are: crypto-eugelinite (79%), crypto-humotelinite (11%), and crypto-detrogelinite (>1%). The liptinite maceral group is dominated by liptodetrinite (2%), followed by sporinite (1%) and bituminite (1%). The inertinite maceral group is primarily composed of inertodetrinite with trace amounts of macrinite, sclerotinite, and fusinite. Noticeable vertical trends in the petrographic data include increased crypto-eugelinite in samples from the bottom of the 3500 bed and in the 4500 bed, and an increased crypto-humotelinite and liptinite content in the upper part of the 3500 bed (fig. 5). The increased crypto-humotelinite concentrations correlate with the increase of xylinitic layers recorded in the megascopic descriptions of the lignite beds (appendix I).

The lignite sample collected from above the tonstein parting (described by Ruppert and others, 1994) in the 3500 bed, has the greatest inertinite content. Although the average concentration of inertinite for all samples is relatively small (1%), the inertinite concentration of sample 3500-1 is greater than 3%.

Abundant siliceous sponge spicules were noted in the samples from the lower part of the 3500 bed (intervals 3000.4 and 3500.5; figs. 3, 8). These silicious grains are elongate and have pitted surfaces. Although Ruppert and others (unpublished data) observed similar spicules in the volcanic ash layer near the top of the 3500 bed, no spicules were observed in the lignite samples from above and below the ash layer. Yancey (this volume) has also described sponge spicules from the clastic rocks associated with the lignites exposed at the Lake Somerville spillway section.

CLUSTER ANALYSIS

Data (on a whole-coal-basis) that included elemental concentrations, major oxide percentages, and ash yield percentages were grouped by cluster analysis (fig. 7A). From the elemental, major oxide, and ash data set, three major clusters were formed by the analysis. In Group I, HAPs elements Be and Cd cluster together. In Group II, Ni, Se, Hg, U, and Pb cluster with most of the rare-earth elements. In Group III, most of the oxides and As, Mn, Cr, and Sb cluster with ash yield. Elements having a possible volcanic origin (Ce, Nb, Ta, Th, Zr; Crowley and others, 1989; Zielinski, 1985) are divided between Groups II and III.

Samples were clusterd into three clusters identified by petrographic data (fig. 7B). In Group C, the samples from the 4500 bed cluster together. Group B is comprised of samples from the upper and lower parts of the 3500 bed. The third cluster (Group A) represents samples from the middle part of the 3500 bed.

DISCUSSION

The 3500 and the 4500 lignite beds contain a large amount of ash (mean ash yield is 29%, db) that tends to be concentrated in the lower part of the 3500 bed and in the 4500 bed. Visual inspection of the HAPs trace-element distributions (whole-coal basis; fig. 6) indicates that there may be some correlation between ash yield and trace element concentration, but correlation coefficients indicate that a significant relationship exists only between ash and Pb and Sb.

The cluster analysis of elemental data (fig. 7A) show HAPs elements Be and Cd are least associated with ash yield. Finkelman (1994) suggests that Be has an organic affinity and that Cd is most commonly associated with sphalerite. It is not clear why Cd concentrations in the Gibbons Creek lignite samples do not cluster with ash yield as would be expected if Cd concentrations were related to the occurrence of sphalerite. The correlation between concentrations of Cd and Zn are not significant. Cr clusters with ash yield and suggests an inorganic affinity for this element, an association also reported by Finkelman (1994). The clustering of the other HAPs elements is inconclusive. Correlation coefficients indicate that Pb has a significant association (r = .6) with ash indicating an inorganic affinity. Finkelman (1994) suggested Pb is usually associated with the occurrence of galena in coal.

Some inferences can be drawn from the petrographic data. The dominance of the crypto-eugelinite maceral subgroup over the crypto-huminite subgroup indicates that all Gibbons Creek lignites were subjected to peat-forming conditions (either biogenic or chemical) causing degradation of wood cellular material into matrix gels. An alternative possibility is that the plants that formed these lignite benches were less woody and more prone to formation of matrix gels. The latter possibility is supported by the work of Mukhopadhyay (1989) who suggested that the pollen in the Gibbons Creek mires was derived in part from marsh plants that are generally not rich in woody material.

The lower bench of the 3500 bed and all of the 4500 bed are high in ash and crypto-eugelinite indicating that these intervals were probably deposited under wetter conditions than those in which the middle intervals formed. This agrees with work by Frederiksen (written commun., 1993) that proposed that the floor of the 3500 bed was deposited in standing water. Cluster analysis of the petrographic data shows that the samples rich in crypto-eugelinite (Groups A and C, fig. 7B) cluster separately from those rich in crypto-humotelinite (Group B). Megascopic descriptions of the lignite benches (appendix I) indicated that the lower and upper benches of the 3500 bed and both benches of the 4500 bed contain less preserved xylinitic material than found in the middle benches of the 3500 bed, further indicating that a more robust vegetation comprised the paleomires of the upper benches of the 3500 bed. Perhaps the depositional setting for the upper part of the 3500 bed was more suitable for forest vegetation than the paleoenvironments for the lower benches of the 3500 bed and all of the 4500 bed.

The occurrence of sponge spicules (fig. 8) in the lower benches of the 3500 bed indicates the original peat was deposited in standing water. Although siliceous sponges are found in both marine and freshwater environments (Pennak, 1978), these spicules resemble those observed in freshwater peats from the Okefenokee (Andrejko and others, 1983) and Riau Province, Sumatra (Ruppert and others, 1993).

The occurrence of high inertinites adjacent to volcanic ash partings in coal has been observed in other coal beds (Crowley and others, 1989; 1994). Increases in inertinite content in the upper part of the 3500 bed (sample 3500.1) may have been associated with alteration of the peat by acids derived (mechanism described by Crowley and others, 1989; 1994) from the volcanic ash layer in the 3500 bed (3500.P1; Ruppert and others, 1994), but our sample interval included all of the upper part of the 3500 bed and was too large to study detailed changes in petrography next to the ash parting. Alternatively, the high inertinite content could have been caused by fire, oxidation and drying, or biologic alteration of the peat in the paleo-mire that formed the upper bench of the 3500 bed.

CONCLUSIONS

The following conclusions can be made about the geochemistry, petrography, and geology of the 3500 and 4500 beds of the Gibbons Creek lignite deposit:

1. Average dry ash yield for the two beds is 29.1%. Ash yields are greatest in the lower bench of the 3500 bed and in the upper bench of the 4500 bed. Total sulfur content is 2.5% (dominated by organic sulfur, 2.3%), and average calorific value is 7800 Btu (18.15 MJ/kg).
2. For lignite samples (on a whole-coal basis) the distribution of two of the HAPs trace elements (Pb and Sb) are positively related to ash yield and suggests an inorganic affinity for these elements in the lignite. Some HAPs elements (Cr, Cd, Pb, Se) appear to have elevated concentrations in the ash partings of the 3500 bed. The concentration of HAPs elements are generally greater in the 4500 bed than found in the 3500 bed. By using cluster analysis we found that Be and Cd were poorly associated with ash yield, indicating a possible organic affinity, and that Ni, Se, Hg, U, and Pb cluster with most of the rare-earth elements. Most of the major oxides (SiO₂, Al₂O₃, MgO, Na₂O, P₂O₅, and K₂0) have a significant correlation (r>.5) with ash yield.
3. The dominance of the eugelinite maceral subgroup over the huminite subgroup indicates that all Gibbons Creek lignites were subjected to peat-forming conditions (either biogenic or chemical) that enabled degradation of wood cellular material into matrix gels. An alternative possibility is that the plants that formed these lignite benches were less woody and more prone to formation of matrix gels. The latter hypothesis is supported by the work of Mukhopadhyay (1989) and N.O. Frederiksen (written commun., 1993) which suggests that the pollen in the lower part of Gibbons Creek lignite beds was derived in part from marsh plants that are less woody than plants in an upland forest environment. We found that the lower lignite bench of the 3500 bed contained the greatest amount of eugelinite.
4. The occurrence of siliceous sponge spicules in the lower benches of the 3500 bed indicates the original peat was deposited in fresh water. This is suggested because the spicules resemble those found in peats from the Okefenokee (Andrejko and others, 1983) and Riau Province, Sumatra (Ruppert and others, 1993).
5. The petrographic data indicates that the upper sample interval of the 3500 bed is greater in inertinite (3%) than the other samples. Increases in inertinite content in the upper part of the 3500 bed may have been associated with alteration of the peat by acids derived from the volcanic ash (Crowley and others, 1989; 1994). Alternatively, the high inertinite content could have been caused by fire, oxidation and drying, or biologic alteration of the peat in the paleo-mire that formed the upper bench of the 3500 bed.

REFERENCES CITED

Andrejko, M.J., Raymond, Robert, Jr., and Cohen, A.D., 1983, Biogenic silica in peats: Possible source for chertification in lignites, in Raymond, Robert, Jr., and Andrejko, M.J., eds., Mineral matter in peat: Its occurrence, form and distribution: Alamos, N.Mex., Los Alamos National Laboratory, p. 25-37.

ASTM [American Society for Testing and Materials], 1993, Annual Book of ASTM Standards, vol. 5.05 Gaseous Fuels, Coal and Coke., Philadelphia, Pa., ASTM, 526 pp.

Crowley, S.S., Stanton R.W., and Ryer, T.A., 1989, The effects of volcanic ash on the maceral and chemical composition of the C coal bed, Emery Coal Field, Utah: Organic Geochemistry, v. 14, p. 315-331.

Crowley, S.S., Dufek, D.A., Stanton, R.W., and Ryer, T.A., 1994, The effects of volcanic ash distributions on a peat-forming environment: environmental disruption and taphonomic consequences: Palaios, v. 9, p. 158-174.

Davidoff, A.J., and Yancey, T.E., 1993, Eustatic cyclicity in the Paleocene and Eocene: data from the Brazos River Valley, Texas: Tectonophysics, v. 222, p. 371-395.

Eicher, R.N., 1985, The stratigraphy of the Jackson Group, Grimes County, Texas: College Station, Texas A&M University, unpublished M.Sci. thesis, 181 p.

Finkelman, R.B., 1994, Modes of occurrence of potentially hazardous elements in coal: levels of confidence: Fuel Processing Technology, v. 39, p. 21-34.

Fisher, W.L., Proctor, C.V., Jr., Galloway, W.E., and Nagle, J.S., 1970, Depositional systems in the Jackson Group of Texas: Transactions of the Gulf Coast Association of Geological Societies, v. 20, p. 234-261.

Gennett, J.A., Raymond, Anne, and Parker, W.C., 1986, Changes in floral composition with depositional environment in Texas Jackson Group lignites: Gulf Coast Association of Geological Societies Transactions, v. 36, p. 449-456.

ICCP [International Committee for Coal Petrology], 1963, International handbook for coal petrology, 2nd edition: Paris, Centre National de la Recherche Scientifique, unpaginated.

ICCP [International Committee for Coal Petrology], 1971, International handbook for coal petrology--supplement to the 2nd edition: Paris, Centre National de la Recherche Scientifique, unpaginated.

Kaiser, W.R., 1982, Lignite depositional models, Texas Eocene: a regional approach to coal geology: in Shobert, H.H., ed., Proceedings of the basic coal science workshop: U.S. Department of Energy, Grand Forks Energy Technology Center, p. 10-67.

Kaiser, W.R., Ayers, W.B., Jr, and La Brie, L.W., 1980, Lignite resources in Texas: The University of Texas at Austin Bureau of Economic Geology Report of Investigations 104, 52 pp.

Maclean Hunter Publishing Company, 1994, Keystone Coal Industry Manual: Chicago, Illinois, Maclean Hunter Publishing Company, 527 p.

Mukhopadhyay, P.K., 1989, Organic petrology and organic geochemistry of Texas Tertiary coals in relation to depositional environment and hydrocarbon generation: The University of Texas at Austin, Bureau of Economic Geology, Report of Investigations 188, 118 pp.

Oman, C.L., and Meissner, C.R., Jr., 1987, Chemical analysis of Gulf Coast lignite samples with significant comparisons and interpretations of results, in R.B. Finkelman, D.J. Casagrande, and S.A. Benson, eds., Gulf Coast lignite geology, Reston, Va., Environmental and Coal Associates, p. 211-222.

Parks J.M., 1966, Cluster analysis applied to multivariate geological problems: Journal of Geology, v. 74(5), p. 703-715.

Pennak, R.W., 1978, Fresh-water invertebrates of the United States (2d ed.): New York, John Wiley and Sons, p. 80-97.

Ruppert, L.F., Warwick, P.D., Crowley, S.S., and Pontolillo, James, 1994, Tonsteins and clay-rich layers in coal-bearing intervals of the Eocene Manning Formation, east-central Texas. Gulf Coast Association of Geological Societies Transactions, v. 44, p. 649-656.

Ruppert, L.F., Neuzil, S.G., Cecil, C.B., and Kane, J.A., 1993, Inorganic constituents from samples of a domed and lacustrine peat, Sumatra, Indonesia, in Cobb, J.C., and Cecil, C.B., eds., Modern and ancient coal-forming environments: Geological Society of America Special Paper 286, p. 83-96.

Russell. W.L., 1955, Stratigraphy of the Wellborn and Manning formations in east-central Texas: Gulf Coast Association of Geological Societies Transactions, v. 5, p. 165-172.

Stach, E., Mackowsky, M.-Th., Teichmüller, M., Taylor, G.H., Chandra, D., and Teichmüller, R., 1982, Stach's textbook of coal petrology, third edition: Berlin, Gebrüder Borntraeger, 535 p.

Swaldi, Steve, and Cezeaux, Rod, 1986, Gibbons Creek lignite mine: Mining Magazine, v. 155, no. 6, p. 568-579.

Tewalt, S.J., 1986, Chemical characterization of Texas lignite: University of Texas at Austin, Bureau of Economic Geology Geological Circular 86-1, 54 p.

Warwick, P.D., and Stanton, R.W., 1988, Petrographic characteristics of the Wyodak-Anderson coal bed (Paleocene), Powder River Basin, Wyoming, U.S.A: Organic Geochemistry, v. 123, p. 389-399.

Yancey, T.E., and Davidoff, A.J., 1991, Paleogene sequence stratigraphy and lithostratigraphy in the Brazos River Valley, Texas: Gulf Coast Association of Geological Societies Field trip Guidebook, 41 Annual Meeting, 112 p.

Yancey, T.E., and Davidoff, A.J., 1994, Paleogene sequence stratigraphy of the Brazos River section, Texas: Gulf Coast Association of Geological Societies Field Guidebook, 104 p.

Yancey, T.E., and Yancey, E.S., 1988, The lower Tertiary of the Texas Gulf Coast, in Hayward, O.T., ed., South-Central Section of the Geological Society of America: Centennial Field Guide v. 4, p. 377-382.

Yancey, T.E., Davidoff, A.J., and Donaho, T.S., 1993, Depositional gradient analysis in transgressive systems tracts and highstand systems tracts, mid-late Eocene of the Brazos River Valley, Texas: Gulf Coast Association of Geological Societies Transactions, v. 43, 465-471.

Zielinski R. A., 1985, Element mobility during alteration of silicic ash to kaolinite--a study of tonstein: Sedimentology, v. 32, p. 567-579.

APPENDIX I. MEASURED SECTION AND SAMPLE INTERVALS OF PART OF THE MANNING FORMATION (JACKSON GROUP), EAST-CENTRAL TEXAS

The following section from A3 Pit of the 4500 and 3500 beds of the Gibbons Creek mine was measured by P.D. Warwick on April 20, 1993. Sample collection and description was by S.S. Crowley, L.F. Ruppert, and P.D. Warwick on the same day. Sample numbers are given in parentheses following the primary lithology. The approximate location of the section and samples is Latitude 30^o34.05N, and Longitude 96^o04.02W. The 4500 bed is somewhat weathered and within the surface weathering zone.

Chapter 1.

Chapter 3.

Created by the EERT WWW Staff.