Manganese removal from discharge waters is a serious problem in many states where Appalachian Basin coal is mined. We analyzed benthic macroinvertebrates and bacteria upstream and downstream from mining operations at the Queens Fork surface mining operations in Wayne County, western West Virginia, in order to assess the ecological effect of elevated instream Mn levels. Manganese data were also retrieved from the U.S. Geological Survey QWDATA database to compare mine wastewater Mn concentrations with those in other Wayne County surface and ground waters. The benthic macroinvertebrates included sensitive, facultative, and tolerant species. A small shift from more sensitive types upstream to more tolerant ones downstream is occurring throughout the ongoing benthic study. No statistically significant differences in abundance and number of taxa existed between the upstream and downstream localities. Rocks upstream and downstream from mining are coated black from Mn oxide precipitates. Glass microscope slides were placed in various sites to mimic rock surfaces and allow bacteria to colonize them for study. The major microbial taxon precipitating Mn upstream and downstream from mining was Leptothrix discophra, a species that is typical of natural settings having elevated Mn levels. Other unidentified Mn-fixing species dominated within the wastewater treatment ponds. Dissolved Mn concentrations in wells are higher than surface waters suggesting that many streams in Wayne County are fed by ground waters carrying elevated concentrations of Mn.
Manganese is one of the most difficult elements to remove from surface waters (10, 22, 3 1). Although dissolved Mn is not known to be toxic, and even blocks the toxic effect of H' (7), it has undesirable effects on water use. These include staining laundry and ceramic fixtures such as toilets where concentrations are greater than 0.05 mg/L (14). Federal regulations therefore control discharge limits. In drinking water sources, the secondary maximum contaminant level (SMCL) for Mn must not exceed 0. 05 mg/L (28). Acid-generating mines may discharge a maximum allowable 4 mg/L as long as the average of daily values for 30 consecutive discharge days does not exceed 2 mg/L (26). Discharges are also regulated by individual states. West Virginia's instream water quality limit is 1.0 mg/L (32).
Many studies have been conducted on Mn removal in the past, including those designed to evaluate chemical dynamics, experiment with packed columns of limestone, and evaluate passive treatment systems (2, 5, 8). Microbial remediation efforts include designed wetlands, microbial bioreactors, and pellets of mixed microbial cultures (1, 6, 29, 30).
In West Virginia, anomalous Mn occurs in a variety of geological settings. Southeastern West Virginia hosts strategic deposits of Mn in Lower Paleozoic rocks (11) (Fig. 1). Stream concentrate samples in eastern and northern parts of the state pinpoint local Mn concentrations (9). Mining operations in many Pennsylvanian age coals in the Appalachian Basin discharge Mn, along with a large variety of other major and trace elements (13, 15, 21).
Manganese removal is also a problem during coal mining in Wayne County, which is in western West Virginia. Pen Coal Corp. mines the 5-Block coal (Lower Kittanning-equivalent) along Kiah Creek (Fig. 1). Mn is concentrated in the sandstone overlying the coal in some areas of the operation (Fig. 2), similar to a situation that has been analyzed in Pennsylvania (20). Mine discharge at several localities is acidic, in the pH 2-3 range. Discharge is treated to raise alkalinity and precipitate Mn (Fig. 3).
Two studies were undertaken to learn about natural instrea m processes in the vicinity of coal mining along Kiah Creek. A long term benthic macroinvertebrate study was begun in 1995 to assess the ecological nature of the effect of elevated Mn levels upstream and downstream of mining operations. A 6-week-long microbial study was undertaken to learn the nature of the indigenous and adapted bacteria upstream and downstream of mining operations.
Study site: The study site lies near Kiahsville in Wayne County, WV (Fig. 1). The Queens Fork surface mining activities discharge into Kiah Creek. Kiah Creek is a tributary of the East Fork of Twelvepole Creek which flows into East Lynn Lake, and then into the Ohio River; East Lynn Lake is regulated by the Army Corps of Engineers. Discharge moves through a series of treatment ponds (Fig. 3). At the Pond 2 locality, mine discharge is directly treated at the inflow with CaO and NaOH and then enters Pond 2. Flocculates settle behind a barrier and clear water is discharged to Pond 2A. Additional NaOH is added at this point to elevate the pH to 11 to precipitate the remaining Mn to meet effluent limits. The outflow of Pond 2A discharges across large rocks and then into Kiah Creek. At the Pond 3 locality, mine discharge is treated at the inflow first with CaO and then with CaO and NaOH. The flocculates in Ponds 2A and 3A are periodically pumped up to drying cells and then permanently disposed.
Sites were chosen upstream (control) and downstream from these discharge points for benthic, macroinvertebrate and microbial analyses. The invertebrate studies were conducted on Kiah Creek just below its confluence with Witcher Fork (upstream control) and downstream from mining operations at Queens Fork (Fig. 1). Microbial studies were undertaken on samples collected from mine discharge, within treatment ponds, at pond discharges, downstream from ponds, as well as in Parker Branch upstream from mining operations (Fig. 1).
A sandstone sample was collected from the backfill of the Queens Fork surface mines for laboratory microbial experiments. The gray and brown sandstone overlies the 5-Block coal and contains thin lenses of coal and streaks of black/brown Mn oxide.
Benthic macroinvertebrates: Standard kick-net and Surber samplers (EPA's Rapid Bioassessment Protocol) were used in both upstream and downstream localities to collect benthic. macroinvertebrates (aquatic insects, molluscs, and worms). Insects were identified to lowest practical taxonomic level, and trophic relationships and pollution sensitivities were assessed using standard methods (17).
Rock chemistry: Chemical analyses of materials disturbed by mining was undertaken to locate the sources of easily soluble Mn so as to be able to avoid them or handle them specially during mining. Paste pH, net neutralization potential, total Mn, and easily leachable Mn were analyzed according to standard procedures (3). Details of treatment are discussed in Maggard (this volume).
Water chemistry: Water samples were collected from upstream and downstream localities during benthic macroinvertebrate studies and measured on site for water temperature, pH, conductivity, and dissolved oxygen (DO). Water samples were appropriately preserved and returned to the lab for further analysis. Water chemistry was analyzed using current EPA-approved analytical methodology.
Manganese data for surface and wells in Wayne County were retrieved from the USGS water-quality database (QWDATA). The data were analyzed with the STATIT statistical package.
Microbial analysis: Three types of samples were collected for analysis by light microscopy:
(1) Flocculates and precipitates collected in the field (May 7, 1996) were analyzed. Motile (moving) and non-motile bacteria, cyanobacteria, algae, fungi, protozoans, and other organic matter were noted.
(2) Glass microscope slide sets were left at various locations) upstream and downstream of mining operations to learn about the benthic microbial community (see 18 and 24 for techniques). Multiple slide sets were left for 6 weeks (May and June 1996). After retrieval, the slides were analyzed for colorless, iron-encrusted, and manganese-encrusted microorganisms that settled on the slides. The presence of oxidized Mn was tested with o-Tolidine (19).
(3) In the laboratory, a crushed sample of Mn oxide-coated sandstone was placed along with a sterile glass slide set in a loosely-covered beaker filled with sterile deionized water to allow indigenous bacteria from the sandstone to settle on the slides. The slides were retrieved after 3 months and the benthic microbial community was analyzed.
Microbial statistics. The number of holdfasts, which are attachment structures of the bacterium Leptothrix discophora, were counted in equal areas on replicate slides left upstream and downstream from mining.
Mineralogy: Two slides that were particularly heavily coated with Mn from the final discharge water were subjected to X-ray powder diffractometry (Debye-Scherrer camera). This technique uses a few micrograms of powdered sample.
SEM: Selected slides were analyzed by scanning electron microscopy (SEM) using a JEOL JSM-840 scanning microscope equipped with a Princeton Gamma Tech X-ray energy dispersive analyzer (EDAX). (The use of brand name does not imply endorsement by the USGS.)
Studies conducted upstream and downstream from mining operations found similarities and dissimilarities in all parameters.
Benthic macroinvertebrates. Benthic macroinvertebrate communities were analyzed according to their abundance, taxa. diversity, and tolerance to pollution (Table 1). The presence of pollution sensitive individuals is important for the purposes of understanding the potential role of Mn in the environment. By combining the three upstream sampling dates and comparing them to the three combined downstream dates, the two sites appear relatively similar. When upstream dates are combined, a total of 1, 148 aquatic insects (29 taxa) were collected. Of this, 16.9 % of the abundance were pollution sensitive (11 taxa); 14.4 % were pollution facultative (9 taxa); and 68.8 % were pollution tolerant (9 taxa). When downstream dates are combined, a total of 1,478 aquatic insects (26 taxa) were collected. Of this 13.9 % were pollution sensitive (10 taxa); 7.6 % were pollution facultative (9 taxa); and 78.4 % were pollution tolerant (7 taxa).
Rock chemisry. The sources of easily soluble Mn were located (Fig. 2). Coarse-grained sandstone in proximity to coal and deficient in neutralization potential contained the most soluble Mn. A strong correlation between potential acidity, which is calculated from total sulfur content, and easily soluble Mn was noted (see 16).
Water chemistry. Some generalizations can be drawn from the upstream and downstream water quality data for the April and October collection dates (Table 2). Seasonal variations were evident. When flow was low, there were significant increases in the concentrations of Ca, SO, Na, Fe, and Mg in the downstream direction. When flow was high, Na increased slightly; TDS, hardness, alkalinity, SO, Ca, and Mg increased by a factor of 2; and Mn increased by a factor of 10. Temperatures were lower downstream than upstream.
A statistical analysis of Mn data for streams and wells in Wayne Co. revealed some interesting trends (Table 3). Total Mn in streams ranged from < 0. 0 10 to 2 mg/L, and dissolved Mn ranged from < 0. 00 1 to 2 mg/L. Dissolved Mn in wells ranged from < 0. 00 1 to 1. 4 mg/L. A comparison of mean and median concentrations shows that total Mn is twice as high as dissolved Mn in streams of Wayne County. The average (mean) concentration of dissolved Mn in 86 wells in Wayne County is higher than that of streams.
Macroscopic observations of use to microbial analyses. Dependant on pH, the products of discharge, of chemical fixation by NaOH and CaO (Fig. 3), and of microbial precipitation are a colorful mixture of black Mn-rich, red Fe-rich, and white At-rich phases. Outflow from one Queens Fork mine is at a rock shelter that is coated red by thick iron-oxide precipitates. At the treatment pond discharges, rocks are coated black from Mn oxide. At the Parker Branch control site, rocks in the riffles are also coated black.
Microbial analysis. Numerous types of bacteria, along with cyanobacteria, algae, fungi, and protozoans settled on glass slides left in the upstream control site, at the mine outflow site, in the treatment ponds, downstream from the treatment ponds, and in the sandstone laboratory experiment (Table 4). Colorless, iron-coated, and Mn-coated cocci, rods (rod-shaped bacteria), and filaments were present on all slides (Figs. 4, 5 and 6). Diatoms were abundant on slides from each environment with the exception of the acid discharge from the mine and at the pipe that discharges NaOH.
Manganese oxide precipitates occurred as dark brown or black stains on the slides. The role of bacteria in Mn oxidation was evident by a wide variety of precipitates that were intimately associated with the bacteria. Brown films were laced with numerous colorless rods (Figs. 4,4, 4,8, and 5,17). Filamentous sheaths were brown (Figs. 4,7 and 5,20) or were becoming coated (Figs. 5,13 and 5,14). Individual rod-shaped bacteria were brown (Fig. 6,23). Specific bacteria having identifiable structures included clumps that resembled Siderocystis sp. (Fig. 4,5). The most distinct structures were the typical spherical or doughnut-shaped holdfasts of Leptothrix discophora that were collected upstream and downstream of mining operations, as well as at discharge locations (Figs. 4,9-12 and 5,18-19). These localities also had something quite unusual -- some of the holdfasts were intricately elaborate structures that radiated out from individual holdfasts (Fig. 4,6).
The Mn-precipitating microbial populations upstream predominantly formed holdfasts and brown-coated rods. Downstream, the population was much more varied and diatoms were extremely abundant. Downstream, Mn was also precipitated by coated filaments and cf. Siderocystis sp. Brown amorphous particles and brown films colonized by short rods were also present.
The milky brown flocculate in the treatment ponds, in contrast, was obviously a chemical precipitate of a paste-like substance (Fig. 4,2). The highly alkaline treatment ponds were not sterile environments, however. Instead, bacteria, cyanobacteria, algae, fungi, and protozoans lived in them (Table 5).
Microbial statistics. The average number of L. discophor a holdfasts on slides left upstream was 1, 130 holdfasts/cm. The average of downstream counts was 1, 100 holdfasts/cm2. If the colonization process were linear, then the average colonization rate upstream and downstream would be 28 holdfasts/cm2/day.
Mineralogy. Although the black and brown precipitates were granular and rarely birefringent using petrographic microscopy, the material was poorly crystallized and produced 6 diffuse lines by X-ray powder diffractometry. The mineral(s) that produced these lines could not be matched by any single phase listed in the Joint Powder Diffraction Data Commission.
SEM. Mn was concentrated in discrete areas on and around holdfasts (Fig. 6), but not on any analyzed filament. Mn, Ca, and S were the only elements present at the limit of analytical detection, along with Si from the diatoms.
Sandstone experiment . L eptothrix discophora holdfasts attached to glass slides in the laboratory sandstone experiment (Fig. 4,12). The average count on the slides was 40 holdfasts/cm2.
Although we have focused this paper on the chemical and microbial processes that remove Mn and the effects of elevated Mn on invertebrate communities, it is not possible to control chemical, physical, or biological variables in the field. Obviously, Mn is only one of a variety of potential controlling variables that govern the distribution of organisms in the streams.
The benthic macroinvertebrates included sensitive, facultative, and tolerant species. Among the sensitive insects, there were individuals that were found strictly upstream from mining operations and others that were found strictly downstream. Only a small shift from more sensitive types upstream to more tolerant ones downstream occurred throughout the ongoing benthic study. There were no statistically significant differences between the upstream and downstream sites for both abundance and number of taxa. For this reason, one can assume that negative impacts from mining have been kept to a minimum.
Analyses of water quality parameters showed that most increased in the downstream direction. Chemical analyses showed that downstream habitats are subjected to more SO,, Ca, Mg, and Mn than in upstream habitats. The concentrations vary with flow rates.
The statistical analysis of Mn data for streams and wells in Wayne County suggests the presence of dynamic trends in interactions among surface water, bed sediments, and ground water. Total Mn was twice as high as dissolved Mn in streams, which suggests a large component of Mn transported in streams is suspended rather than dissolved. Bed sediments, particularly rocks that are coated black, contain a large flux of Mn that could be transported during high streamflow events. Because the mean concentration of dissolved Mn in well water was higher than in streams, a relatively large proportion of Mn in Wayne County streams may be derived from Mn in ground waters that discharge to the streams. For both ground water and surface water in Wayne County, average concentrations of Mn exceed the U.S. EPA SMCL of 0. 05 mg/L. A total of 152 of 268 streams (5 6.7 %) analyzed for dissolved Mn had concentrations greater than or equal to the 0.05 mg/L SMCL for drinking water. Based upon the data available from the USGS water-quality database (QWDATA), only 6 of 268 sites (2.2 %) analyzed for dissolved Mn and only 4 of 187 sites (2.1 %) analyzed for total Mn had concentrations greater than the I mg/L West Virginia State standard for streams.
The microbial populations of Mn-rich flocculates are quite varied. The bacteria that fix or oxidize Mn onto solid glass surfaces, which are used to mimic rocks in riffles, form simple brown rods and simple Leptothr ix discophora holdfasts, as well as highly complex and elaborate holdfast structures. The simple holdfasts are structures that the rod-shaped bacteria use to attach to surfaces and are typical of those precipitating Mn in streams having elevated Mn concentrations elsewhere (4, 23, 24, 25). The elaborate holdfast structures are unusual and occur both upstream of mine operations and in mine treatment ponds. The highly efficient Metallogenium -type Mn-fixing bacteria isolated by Vail and others (1988) inorganic-matter-rich environments were not found. Lacking these, the thin coatings of Mn on rocks in swiftly moving water show that rocks serve, at the very least, as mildly efficient Mn-stripping structures. Rates of Mn removal by coatings on rocks need to be addressed in further studies.
Holdfasts of Leptothrix discophora also attached to glass slides in the 3-month-long laboratory sandstone experiment. These bacteria typically attach where anoxic, Mn-bearing ground water is present; they then proceed to oxidize the Mn on their holdfasts where water is oxygenated (23). Therefore, the sandstone Mn was probably in a reduced form and the deionized water of the experiment supplied oxygen. The results were not unexpected because this bacterium will colonize new surfaces as they become available near redox boundaries in beakers or test tubes filled with natural waters and placed in the laboratory (Robbins, pers. obs.). The fact that no organic matter was added to the sterile beaker water suggests either the bacterias' organic compound requirements are minimal or this strain may be able to oxidize Mn as an energy source, therefore being autotrophic in its nutrition.
This information about the different types of bacteria that precipitated oxidized Mn, and the different forms that the precipitates took, are very useful for explaining how Mn is bound in mine treatment situations. Evangelou and others (1992) recognized that Mn removal is catalyzed at mineral surfaces; the bacteria that participate in the catalysis in this West Virginia locality are dominated by L. discophora . Watzlaf (1988) stressed adsorption onto reactive surfaces as an important process in mine drainage treatment. In reality, these bacteria precipitate Mn on a wide variety of surfaces including non-reactive ones such as glass slides.
The current effluent limits for Mn in receiving streams that are not used for domestic purposes have been shown to be lower than can be supported by scientific research (7, 12, 27). Our analyses show that benthic macroinvertebrates do not appear to be affected in any significant manner by the levels of Mn in the control stream or in the mine discharge. Furthermore, it appears that the microbial population proliferates to take advantage of the Mn concentration in the water.
We would like to thank Mark Kazoo (USGS-WRY-WV) for retrieval and initial interpretation of Mn concentrations in Wayne County waters, and Bob Kleinmann for useful editorial comments.
