NORTH
BRANCH OF THE POTOMAC RIVER: RESULTS OF TWO YEARS OF LIME DOSINGJoseph
E. MillsMaryland Department of the Environment
Bureau of Mines
160 S. Water Street
Frostburg, Maryland 21532INTRODUCTION
The North Branch of the Potomac
River (North Branch), from its headwaters to Jennings Randolph Lake, has been
adversely affected by acid mine drainage (AMD) from abandoned coal mines since
the last decade of the 19th century. This problem is most serious in the upper
reaches, where drainage from abandoned coal mines represent a high percentage
of the mainstem flow. The North Branch watershed consists of seventy
subwatersheds, twenty of which produce AMD. Fifty-two sites produce AMD, with
almost ninety percent of the acid loading being produced by fifteen sites from
four subwatersheds and two direct discharges (see Attachment 1). The major acid
producing sites in Maryland are underground mines in Laurel Run (Kempton Mine)
and Three Forks Run (Vindex Mine), and direct discharges to the North Branch at
Shalimar and Kitzmiller. Sites located on Abram Creek and Stony River are the
major acid producers in West Virginia.
Many streams in the North Branch
watershed have historically been completely devoid of aqua life due to high
concentrations of acid, dissolved metals and sediments from AMD. The
destruction of aquatic habitat and the reduction of aesthetic values has
greatly restricted recreational use of the North Branch.
The Bloomington Dam was completed
in 198 1, and formed the Jennings Randolph Lake. Bloomington was constructed
with the ability to mix different qualities of water from within the lake to
achieve a consistently good quality effluent discharging to the mainstem of the
North Branch. This "high" quality discharge from the dam allowed for
the establishment of a sport fishery below the dam, but due to the input of
highly acidic waters into the dam, the dam effluent was often of marginal
quality. The establishment of this downstream fishery renewed interest in the
problems that still existed in the upper reaches of the North Branch. In 1992,
the Maryland Bureau of Mines (BOM) began looking for ways to eliminate the
impacts of the AMD on the North Branch above the dam which would also improve
the quality of the water in the lake and allow for more consistent, higher
quality discharges from the dam. By October of 1993, four lime dosers were
installed in the North Branch watershed with the purpose of treating the acid
impacts of unabated abandoned surface and deep mine discharges and restoring
the river to a recreational fishery.
In 1989, each stream in the North
Branch watershed above Jennings Randolph Lake was intensively evaluated for
water quality by "The North Branch Potomac River Abandoned Mine Drainage
Study" (North Branch Study). This study was performed under a joint
contract funded by the Federal Office of Surface Mining, the Maryland
Department of Natural Resources, and the West Virginia Division of
Environmental Protection. The main objective of this study was to locate and
evaluate the major sources of AMD in the North Branch watershed. During this
study, water quality samples were collected at the nodes of each stream during
high, low and normal flow periods.
Additional pre-monitoring data was
collected prior to installation of the lime dosers. Sixteen sampling stations
were selected for collection of chemical background data (May 1991 through
October 1992). Biological background data were collected at fifteen of the
sixteen stations (May 1991 and April 1992) (see Attachment 1). These data were
used to evaluate the effectiveness of AMD treatment after abatement
technologies were implemented in the watershed (see Attachment 2).
In 1991, fishery management
objectives were defined for the North Branch, in consultation with both
Maryland and West Virginia fisheries management personnel. These objectives
were determined based upon biota management potential of each segment of the
North Branch, and upon biological tolerance limits for each species of interest
as well as human population densities and accessibility to the river.
Available AMD technologies were
evaluated to determine their ability to remediate the major AMD sources in
Maryland, factoring in funding limitations, and the physical and chemical
extremes at each source. Lime dosing was determined to be the best, most cost
acceptable alternative for the restoration of a fisheries in the North Branch
(International Science and Technology, Inc., 1990).
In 1992, optimum deployment of lime
dosers to achieve water quality targets in the North Branch was determined with
a model using data from the North Branch Study and the fishery management
objectives and running selected treatment scenarios (see Attachment 1). The
doser location scenario which met the most water quality targets in Maryland
tributaries and the North Branch mainstem, located lime dosers on Laurel Run
(Kempton Mine), at Gorman, MD on the North Branch mainstem, on Lostland Run and
on Three Forks Run (the Three Forks Run doser was relocated to Kitzmiller, MD -
the proximity of one site to the other had little effect on the model).
During 1992, the BOM actively
pursued a donation from The Conservation Fund, Inc. of Arlington, VA, of four
lime dosers (two of the dosers - both water powered, dry powder, CaC03 dosers
-were used in the AMD abatement project, while the remaining two - both small
automated, wet slurry dosers - were put into storage for possible use at a
later date). In December, 1992, one of the donated dosers was installed on
Lostland Run, while the second doser was installed at Kitzmiller, MD to treat
AMD that discharges directly into the North Branch. Several months after
operation, studies conducted by the U.S. Bureau of Mines showed that the doser
located at Kitzmiller did not treat the AMD as well as expected due to AMD
chemistry and to the relative flatness of the receiving stream and the lack of
agitation or mixing action. In August 1994, at the suggestion of the U.S.
Bureau of Mines, the doser located at Kitzmiller was modified to use Pebble
Quick Lime, CaO, instead of limestone. This modification was deemed successful,
since the pH below the doser, in the channel where the AMD enters the North
Branch, has remained near 7.0.
In 1993, the BOM developed
specifications for two large automated, dry powder, CaCO3 dosers and
solicited bids for these machines. In October, 1993, these large capacity
machines were installed on Laurel Run (Kempton Mine) and on the mainstem of the
North Branch at Gorman, MD.
One year after installation of the
four doser system, post-doser monitoring of water quality was initiated.
Post-doser monthly water quality samples were collected during 1995 at the 16
pre-doser monitoring stations and post-doser biological samples were collected
at the 15 pre-doser biological monitoring stations in April 1995. These data
were compared to the pre-doser data collected in 1991 and 1992. These results
are the theme of this paper.
The following is a summary by
station of the results obtained during this study. With a few exceptions, that
will be noted, only the chemical results that are statistically significant
will be discussed. Relative doser locations and brief descriptions of their
functions are included in this section.
Sixteen stations were monitored for
water quality during the pre-doser and post-doser phases of the project (see
Attachment 1). Stream water samples were collected in syringes for closed pH
and in cubitainers for analysis of acid neutralizing capacity (ANC), iron,
manganese, aluminum, calcium, magnesium, sulfate, sodium, and conductivity.
The pre-doser and post-doser data
were subjected to univariate statistics (ANOVA analysis). Multivariate
statistics (preliminary factor analysis) were also analyzed but will not be
discussed in this paper (Morgan et al., 1996, draft report). The ANOVA analyses
indicate significant changes in many water quality parameters over the course
of the study at all stations sampled.
Pond Station
This control station is located on
the North Branch above the town of Henry, WV, above any influence of the lime
dosers. The pH (6.76 to 6.48 ) of this station stayed about the same, and ANC
(314 to 239 m eq/1) at this station also showed no significant
change. A significant decrease in aluminum and a significant increase in
magnesium did occur at this station. Mean aluminum was 0.29 mg/l before and 0.
15 mg/l after, and mean magnesium was 4.2 mg/I before and 18.1 mg/l after
dosing. Comparison of other analytes showed no significant difference.
This station is downstream of the
Henry Pond station at the town of Henry WV, upstream of the dosers and also
served as a control. The pH of this station remained unchanged, 6.75. Significant
increases occurred in ANC (263 to 433 m eq/l), iron (0. 19 to 0.50 mg/1), and
manganese 0.47 to 0.89 mg/1), while sulfates (843 to 441 mg/1) decreased
significantly. These changes may be attributable to upstream AMD treatment at
active mining sites. Other analytes showed no significant change.
The first doser in the system is
located on Laurel Run. Laurel Run, which is highly impacted by the AMD from the
Kempton Mine discharge, enters the North Branch below the Henry Railroad station
and above the next mainstream sampling station at Wilson, MD. This doser was
sited here to eliminate the impacts that Laurel Run exerts on the North Branch.
The doser is located on Laurel Run, approximately 2.5 stream miles below the
AMD discharge at the Kempton Borehole and Airshaft, and approximately 2.5
stream miles above the confluence of Laurel Run with the North Branch. This
doser delivers the largest quantity of neutralizing agent of the four dosers to
the North Branch system.
This sampling station is located on
Laurel Run about I mile downstream of the doser. Comparing pre-doser and
post-doser data, dramatic changes in water quality are evident at this site
with significant increases in pH (3.24 to 4.2 1), ANC (-871 to 62 m
eq/l), calcium (28 to 87 mg/1) and magnesium (16 to 30 mg/l)and significant
decreases in iron (7.8 to 3.1 mg/1) and aluminum (8.3 to 2.8 mg/1).
The next downstream, mainstem
station is located approximately 1000 feet below the confluence of Sand Spring
Run with the North Branch. Dramatic changes in pH (4.46 to 6.87) and ANC (-28
to 375 m eq/l) were noted as was a significant reduction in aluminum
concentrations (1.7 to 0.05 mg/1). These data show the river changing from
biologically intolerable conditions to conditions that are more favorable to
aquatic life.
The next downstream sampling
station is located at Bayard, WV. At this station dramatic changes in pH were
achieved (5.00 to 6.94) and ANC (23 to 392 m eq/1). Aluminum and
sulfates also decreased significantly (1.7 to 0.42 mg/l and 692 and 361 mg/l
respectively). Iron increased significantly in concentration after dosing was
initiated, with a change from an average pre-doser concentration of 0. 17 mg/I to
an average concentration after doser operation of 0.47 mg/l.
This station is located about 3
miles downstream of the Bayard station and about 6 miles below the confluence
of the North Branch with Laurel Run. All concentrations of analytes were
similar to the concentrations measured at the Bayard Station. Significant
changes were measured in pH (5.16 to 7.03), ANC (33.2 to 375.3 m
eq/l) and aluminum (1.8 to 0.37 mg/1). Before and after iron concentrations at
this station (0. 17 to 0.42 mg/1) were about the same as they were at the
Bayard station.
The Gorman doser is the next doser
downstream in the doser system. It is located near the town of Gorman, MD, and
it discharges directly into the North Branch. Its main purpose is to
"level out" pH fluctuations in the river and to add additional ANC to
the North Branch.
The Steyer sampling station is
located approximately 2 miles downstream of the Gorman doser. This station is
about 300 feet below the confluence of the North Branch with Steyer Run, a very
high quality stream. Statistically, only ANC and iron had any significant
changes during the study period, with ANC increasing from 265 to 467 m
eq/1 and iron increasing from 0.099 to 0.30 mg/l. The relatively high quality
of the water at this station before initiation of the doser project may be due,
in part, to a sampling error. The close proximity of the sample station to the
confluence of the streams may not allow for adequate mixing of the waters prior
to sampling.
The next sampling station is
located approximately 3 miles downstream of the Gorman doser at Bradshaw, WV,
just upstream of the confluence with Snowy River, At this station the pH
increased from 5.76 to 7.21 during the study. ANC (92 to 444 m
eq/l) and iron (0. 12 to 0.27 mg/1) both increased significantly.
The next sampling station is
located at Schell, WV, above the confluence of Trout-Laurel Run with the North
Branch. This station exhibited significant increases in pH (6.62 to 7.26), ANC
(115 to 377 m eq/1) and iron (0.01 to 0.20 mg/1) and a
significant decrease in sulfate (470 to 215 mg/1).
The third doser in the system is
located on the South Prong of Lostland Run, a slightly AMD impacted tributary
to the North Branch. This doser is situated about 2.5 miles upstream of the
confluence. Its main purpose is to add buffering ability to Lostland Run and to
carry excess ANC into the North Branch.
This sampling station is located on
Lostland Run approximately I mile downstream of the doser, below the confluence
of the North and South Prongs of Lostland. This station exhibited a significant
increase in pH (7.02 to 7.21) and significant decreases in calcium (70 to 36
mg/1), sulfate (226 to 56 mg/1) and conductivity (380 to 154 m
S/cm).
The next mainstem station is
located below the confluence of Lostland Run with the North Branch. The water
quality measurements were the about the same when compared to water quality
data collected at the previous upstream mainstem station at Schell. The cause
of the water quality improvements at this station are not easy to separate
between the doser on Lostland Run and the dosers upstream on the North Branch.
This station exhibited significant increases in pH (6.76 to 7.3 1), ANC (160 to
356 m eq/]), iron (0.081 to 0. 18 mg/1) and aluminum (0. 10 to 0.24
mg/l) after operation of the lime dosers.
This station is located just above
the confluence of the North Branch with Abram Creek. This station exhibited
significant increases in pH (6.78 to 7.08), ANC (120 to 350 m
eq/1), iron (0.087 to 0.22 mg/1), and aluminum (0.098 to 0.27 mg/1), and a
significant decrease in sulfate (423 to 194 mg/1).
The Kitzmiller station is located
along the mainstem of the North Branch, approximately 2 miles below the
Shallmar station and the confluence of Abram Creek with the North Branch
mainstem. Significant increases in pH (6.67 to 7.02), ANC (78 to 283 m
eq/l), and aluminum ( 0. 13 to 0.33 mg/1) and a significant decrease in sulfate
(405 to 186 mg/1) were measured after operation of the dosers.
The Kitzmiller doser is located on
an AMD drainage that discharges directly into the North Branch. There are no
sample stations below this doser on the North Branch mainstem until Luke, a
station 8 miles below the Jennings Randolph Dam.
The last mainstem. station in this
study is located near the town of Luke, MD, approximately 8 miles below
Jennings Randolph Lake, and above the confluence of the Savage River. Three
analytes showed significant increases at this station, ANC (79 to 131 m
eq/l), iron (0.093 to 0.28 mg/1) and magnesium (11.5 to 24.8 mg/1). This
station is located below the dam which forms Jennings Randolph Lake and the
flow is manually controlled. These data may reflect the ability of the dam
operator to manipulate the quality of the effluent.
This sample station serves as an
AMD control. It is located on Three Forks Run, a severely degraded, AMD
impacted stream, with a pH of 3.13. Three Forks Run is located at the
headwaters of Jennings Randolph Lake. This station receives no treatment from
the lime dosing system and, as expected, showed no significant changes in water
quality during this study.
The Savage River Station was set up
as a chemical and biological control station. This station is located upstream
from the North Branch on Savage River, a high quality, recreationally important
stream that enters the North Branch near Luke, MD. This station receives no
treatment from the lime dosers. This station showed a significant increase in
magnesium (4.0 to 19.5 mg/1) and significant decreases in pH (6.69 to 6.27) and
sulfate (19 to 13 mg/1) during the study.
Biological sampling was conducted
at the same stations as water sample were collected with the exception of the
Luke station. Baseline samples (pre-doser) were collected in May 1991 and April
1992. The doser system was essentially completed and operational by late 1993.
By early summer of 1994, the lime doser system was operating smoothly with
minimal down time. In April 1995, allowing I year for colonization, the
post-doser biological sampling was conducted.
The initial results for macroinvertebrate
response to improved water quality indicate that recolonization is slow,
although definitely occurring at some stations, but no consistent pattern of
improvement was observed for the mainstem stations from pre-doser sampling to
the post-doser sampling
The results of the fish collections
indicated that recolonization of fish population is also slow. The "Index
of Biotic Integrity", IBI, consisting of ten separate metrics, was used to
compare changes in fish communities (Jacobson et al. 1992). The Savage River
control station did not exceed an IBI of 50, due primarily to the lack of
darter species. The uppermost North Branch station showed a decent fish
assemblage, with an 1131 of about 45. The remainder of the North Branch mainstem
stations showed poor fish communities, with no stations exceeding an IBI of 30.
The mainstem stations at Wilson, Bayard, and Steyer exhibited slight decreases
in the IBI during the study period. The station on Lostland Run and the
mainstem stations below Lostland Run exhibited the same IBI for all three years
of sampling, in spite of the of dosing activities. The stations at Bradshaw and
Schell both showed slight improvement in the IBI.
Habitat is a major factor for
aquatic community potential, with both quality and quantity of habitat
influencing the structure and function of biological communities. Habitat
quality was assessed using metrics developed by Plafkin et al. (1989) and
modified by Kazyak and Jacobson (1994). Overall habitat scores were very poor
at the North Branch mainstream stations, both before and after operation of the
dosers. This was due in large part to embeddedness caused by past AMD deposits
creating poor substrate habitat. There were no major changes in habitat structure
at any station during the study period
Biotic restoration of a disturbed
watershed is usually based on the remediation of water quality (Yount and Niemi
1990), and is impossible without adequate water quality. Following water
quality restoration, improvements in the biotic community are usually observed,
especially if a number of refugia are present within the watershed. Biotic
recovery in a lotic system is a function of the availability and accessibility
of unaffected upstream and downstream refugia. However, disturbances of a
physical nature (such as embeddedness and sediment deposition) are usually more
difficult to remediate (Niemi et at, 1990).
There has been an improvement in
water quality within the North Branch watershed since the installation and
operation of the time dosers. All mainstem stations located below the dosers
showed improvement in mean pH values. The monthly sampling showed that the
target pH values set forth during the planning stages of the project were met
and exceeded at every station on the mainstem. Prior to the doser project, all
mainstem stations below the confluence with Laurel Run, except the Steyer Run
station (265 m eq/l) had ANC levels less than 200
"q/1 (from 28 to 160 m eq/l), an ANC level indicative of poor
buffering capacity in the system and considered by many to be the lower extreme
of good quality habitat. After the doser project, all mainstem stations except
the Kitzmiller station (283 m eq/l), had ANC levels in excess of 300 m
eq/l (350 to 467 m eq/1). All mainstem stations below the
dosers also showed increased magnesium and decreased sulfate and conductivity.
The increases in pH and ANC at the
mainstem stations are not directly related to increases in calcium in the
system, but appear to be driven by reductions in sulfate and increases in
magnesium. Preliminary work on the sediments associated with the dosers
indicates that the primary mineral formed is ettringite, calcium aluminum
sulfate hydroxide [Ca6A12(SO4)3(OH)12
25H2O]. The information of this mineral may explain why there is no
pattern of calcium increases in the North Branch system, since it is being tied
up with sulfate and aluminum, and precipitated out of the water column.
Aluminum concentrations dropped at
all mainstem stations except, the Lostland Run/North Branch, Shallmar, and
Kitzmiller stations. The decreases in aluminum at most mainstem stations is due
to the formation of ettringite and other insoluble aluminum compounds. The
increases in concentrations of aluminum at Kitzmiller, Shalimar and below
Lostland Run although significant, are very low.
Iron concentrations increased at
aft mainstem stations below Laurel Run, with many increases being statistically
significant. All mainstem, post doser, mean iron concentrations were 0.50 mg/I
or less except for the station at Wilson which was 0.66 mg/l.
Many years of AMD deposits have
created an extremely embedded substrate; this situation, will correct itself
over time if water quality remains good. The impacts of increased sediment
deposition on biotic recovery are currently being studied. The elimination of
the sediment deposits from the system, under current conditions, is monetarily
impossible. These two characteristics, embeddedness and sediment deposits, may
be the two major factors retarding the biotic recovery of the North Branch.
Some additional concerns, center on potential temperature and/or gradient
limitations that may also inhibit the biotic recovery of the system.
An additional factor to consider in
the slow biological recovery of the North Branch is that upstream movement of
fishes is blocked and upstream movement of benthic invertebrates is slowed by
Bloomington Dam. All colonization of fishes would have to come from present
refugia above the darn and most invertebrate colonization will also come from
above the dam. An assessment of the refugia resources along the North Branch is
necessary to determine the adequacy of the refugia in the watershed.
Britt, D. L. 1990. Preliminary
design for an acid mine drainage mitigation demonstration program.
International Science and Technology, Inc., Sterling, Va. Prepared for: State
of Maryland, Department of Natural Resources, Water Resources Administration,
Bureau of Mines, Frostburg, MD.
Jacobson, P., P. Kazyak, A.
Janicki, D. Wade, H. Wilson, and R. P. Morgan 11. 1992 Feasibility of using on
index of biotic integrity (1131) approach for synthesizing data from a Maryland
biological stream survey. Prepared for: State of Maryland, Department of
Natural Resources, Tidewater Administration, Chesapeake Bay Research and
Monitoring Division, Annapolis, MD.
Morgan, R. P. II, C. K. Murray, K.
M. Meager, D. M. Gates. 1995. Biological and chemical evaluation of the North
Branch of the Potomac River restoration. Prepared for: State of Maryland,
Department of the Environment, Water Management Administration, Mining Program,
Bureau of Mines, Frostburg, MD.
Niemi, G. J., P. DeVore, N.
Detenbeck, D. Taylor, A. Lima, J. Pastor, J. D. Yount, and R. J. Naiman. 1990.
Overview of case studies on recovery of aquatic systems from disturbances.
Environmental Management 14:571-587.
Plafkin, J. L., M. T. Barbour, K.
D. Porter, S. K. Gross, R. M. Hughes. 1989. Rapid bioassessment protocols for
use in streams and rivers. Benthic Macroinvertebrates and fish. EPA/444/4-8900
1. Office on Water, U. S. Environmental Protection Agency, Washington, D.C.
Yount, J. D. and G. J. Niemi. 1990.
Recovery of lotic communities and ecosystems from disturbance - a narrative
review of case studies. Environmental Management 14:547-569.
