Regional estimates of acid mine drainage impact on ... - Springer Link

0 downloads 0 Views 981KB Size Report
streams in the Mid-Atlantic and Southeastern United States. The NSS ... of the streams in coal mining areas of the eastern United States (Kinney, 1964;.
REGIONAL STREAMS

ESTIMATES

OF ACID MINE DRAINAGE

IN THE MID-ATLANTIC

IMPACT ON

AND SOUTHEASTERN

UNITED

STATES

ALAN T. HERLIHY, PHILIP R. KAUFMANN, MARK E. MITCH Utah State University, c/o U.S. EPA Environmental Research Lab., 200 S W 35th St., Corvallis, OR 97333, U.S.A.

and DOUGLAS D. BROWN NSI Technology Services Corp., 200 S W 35th St., Corvallis, OR 97333, U.S.A.

(Received July 17, 1989; revised November 10, 1989) Abstract. The U.S. Environmental Protection Agency conducted the National Stream Survey (NSS) to provide unbiased estimates of the numbers and distribution of acidic and low acid neutralizing capacity streams in the Mid-Atlantic and Southeastern United States. The NSS employed a probability sample of 500 stream reaches to represent a target population of 64,300 stream reaches in the study area. All NSS samples were screened for acid mine drainage (AMD) influences, and population estimates of the regional extent of AMD impacts were made. Almost 10% of the stream reaches in the Northern Appalachians subregion were acidic during spring baseflow due to AMD. In the entire NSS, an estimated 4590 km (+_ 1670) of streams (2% of the total NSS length) were acidic due to AMD and another 5780 km (_+ 2090) of streams were strongly impacted, but not acidic. In subregions of the NSS with observed mine drainage effects, roughly the same number of streams were acidic during spring baseflow due to AMD as due to acidic deposition. The population estimates of mine drainage impact made in the NSS were similar to estimates made in previous surveys that attempted to census all of the streams in coal producing areas. These results demonstrate that a statistically based stream survey is a useful tool for evaluating regional water quality. 1. Introduction Acid m i n e d r a i n a g e ( A M D ) is a m a j o r water p o l l u t i o n c o n c e r n in m a n y m i n i n g regions o f the U n i t e d States (Barnes a n d R o m b e r g e r , 1968; K i m et al., 1982; Powell, 1988). W h e n m i n e spoils c o n t a i n i n g sulfide are exposed to air a n d water, the sulfide minerals are oxidized b y a series of microbial a n d chemical processes according to the following overall r e a c t i o n ( S t u m m a n d M o r g a n , 1970; D u g a n , 1985); 4 FeS2 + 15 02 + 14 H 2 0 --~ 4 Fe(OH)3 + 16 H ÷ + 8

8042-.

The p r o d u c t s of this reaction are carried into surface waters where they degrade water quality via acidification, heavy metal p o l l u t i o n , a n d s e d i m e n t a t i o n . Acid mine drainage waters are characterized by high heavy metal a n d sulfate c o n c e n t r a t i o n s , high conductivity, a n d low p H (Mills, 1985). Regional e s t i m a t i o n of s t r e a m w a t e r quality is a difficult p r o b l e m because of the c o n t i n u o u s , b r a n c h i n g n a t u r e of stream networks which makes e n u m e r a t i o n difficult. Stream p o p u l a t i o n e s t i m a t i o n requires either a complete census of the stream length in the s t u d y region or p r o b a b i l i t y s a m p l i n g from a n explicitly defined target Water, Air, and Soil Pollution 50: 91-107, 1990. © 1990Kluwer Academic Publishers. Printed in the Netherlands.

92

ALAN T. H E R L I H Y ET AL,

population of stream segments that can aggregate to form the total network. Previous estimates of the regional extent of the AMD problem have been made by compiling the work of the numerous state and federal agencies that have examined many of the streams in coal mining areas of the eastern United States (Kinney, 1964; Appalachian Regional Commission, 1969). Within these surveys, however, the sampling intensity (both spatially and temporally), the stream size definitions, data quality, analytical procedures, and the criteria used to define mine drainage degradation are different, depending on which agency or group did the field work. Also, there does not appear to be a consistently defined formulation for how the estimates of degraded stream length in the individual river basins were compiled from the census data. Thus comparisons between surveys and even between regions in the same survey contain an unknown element of bias. In addition, these surveys were performed over 20 yr ago, in the 1960s. Nevertheless, it would be informative to compare the results of these older 'census' type surveys to a recent, probability sample survey, such as the National Stream Survey. The U.S. Environmental Protection Agency's National Stream Survey (NSS) employed a randomized systematic sampling design to make unbiased estimates of the regional extent, location, and chemical characteristics of acidic and low acid neutralizing capacity (ANC) streams in areas of the Mid-Atlantic and Southeastern United States sensitive to acidic deposition (Figure 1). Streams acidic due to both acidic deposition and AMD were observed in the NSS probability sample (Kaufmann et al., 1988). Although the NSS was designed primarily from an acidic deposition effects standpoint, the NSS statistical design does allow the estimation of both numbers and length of any explicitly defined subpopulation of streams in the study area. All NSS samples were screened to classify streams impacted by AMD, and population estimates of the status and extent of AMD pollution were made for the sampled areas of the eastern United States.

2. National Stream Survey Design The NSS is a randomized systematic sample of 500 stream reaches designed to estimate the characteristics of a target population of 64 300 stream reaches in acidsensitive subregions of the eastern United States (Figure 1). The sampling unit in the NSS is a stream reach, defined as a blue-line headwater segment or a segment between confluences on U.S. Geological Survey l:250000-scale topographic maps. Each sampled reach was assigned a weight (inversely proportional to its selection probability) equivalent to the number of reaches it represents in the map population (see Kaufmann et al. [1988] for a detailed description of the statistical design). During site selection, stream reaches with watershed areas greater than 155 km 2 and reaches within mapped urban areas were excluded from the NSS target population. Streamwater chemical data were collected in the spring, between March 15 and May 15 of 1986 (1985 in the Southern Blue Ridge Pilot Survey subregion). Water

ACIDMINEDRAINAGEIMPACTSIN THE EASTERNU,S.

93

SUBREGIONS OF THE NATIONAL STREAM SURVEY-PHASE I

Northern Appalachians (2Cn)

Valley and Ridge(2Bn) Poconos/Catskills ( I D)

Southern Blue Ridge (2As)

"

Mid-Atlantic Coastal Plain (3B)

Ozarks/Ouachitas ( 2 D ~

~

~edm!nt

Florida Southern Appalachians (2X) Fig. 1. Subregions of the National Stream Survey, subregion boundaries were drawn around areas with surface waters expected to have ANC's less than 400 ~,eq L -1 based on national alkalinity maps (Omernik and Powers, 1983). The Poconos/Catskills, Valley and Ridge, Northern Appalachians, and Mid-Atlantic Coastal Plain subregions are considered to be in the Mid-Atlantic portion of the NSS. The remaining subregions are considered to be in the Southeast portion of the NSS.

samples were taken f r o m both the upstream confluence (or headwater) and d o w n s t r e a m confluence (as shown on the 1:250000 scale maps) of each sample reach and are referred to as the ' u p s t r e a m reach end' and the ' d o w n s t r e a m reach end'. Samples were not collected during, or for 24 hr after, precipitation events in order to avoid possible storm influences. All samples were kept at 4 °C immediately after sampling and were stabilized (anion aliquots were filtered, base cation and metal aliquots were filtered and acidified with nitric acid) at a central processing l a b o r a t o r y within 36 hr of collection. Acid and base neutralizing capacities ( A N C and B N C ) were measured by G r a n analysis o f acid and base titration data; p H by potentiometric analysis o f a closed headspace sample; SO42- and Cl- by ion c h r o m a t o g r a p h y ; and Fe, Mn, and base cations by atomic absorption spec-

94

ALAN T. H E R L 1 H Y ET AL.

troscopy (Kaufmann et al., 1988). Inorganic monomeric AI was calculated as the difference between total monomeric and organic (nonexchangeable) monomeric A1 measured by pyrocatechol violet colorimetry and the use of a strong cation exchange column (Hillman et al., 1987). Standardized protocols specified in a quality assurance plan were followed during sample handling, chemical analysis, and data base manipulation (Cougan et al., 1988). In order to make regional population estimates, a single spring baseflow chemical 'index' value was calculated for each stream reach end by averaging the chemical data from multiple visits within the spring baseflow period (three observations in the Southern Blue Ridge Pilot subregion, two observations in the Mid-Atlantic subregions, and one observation in the remaining Southeast subregions).

3. Mine Drainage Screening and Classification Streams were classified as impacted by acid mine drainage based on field observations, mapped information, and chemical screening. Streams were first screened by examining chemical data to identify sites with chemical signatures indicative of mine drainage impact. The field and map data for these sites were then examined to confirm the presence of mining activity. Only sites with mining activity in their watersheds were considered to have mine drainage impacts. Mine drainage impacted stream reach ends were divided into three classes (Figure 2): - acidic due to AMD - nonacidic, strongly impacted by AMD - nonacidic, weakly impacted by AMD. All NSS stream reach ends that were acidic (index ANC _< 0 txeq L -1) were intensively screened for the presence of AMD. Acidic streams with index sulfate concentrations greater than 300 ~seq L -1 in the Mid-Atlantic and 200 ~eq L -1 in the Southeast were considered possible AMD impacted sites. These threshold values were chosen after examination of predicted subregional steady-state streamwater sulfate concentrations, assuming only atmospheric sources of sulfate. From these data, the best estimate of steady-state streamwater sulfate concentration in streams of the Mid-Atlantic subregions is 158 to 219 Iseq L -1 (Kaufmann et al., 1988). A value of 300 ~eq L -1 was chosen as a criteria to separate streams whose sulfate comes mainly from the atmosphere from streams whose sulfate source is primarily internal (from watershed sources). A similar line of reasoning applies to the selection of 200 p.eq L -~ as a criteria value in Southeast streams. Predicted streamwater sulfate concentrations in the Southeast (96 to 146 I~eq L -~ are lower than in the Mid-Atlantic, so a lower cutoff value was established (Kaufmann et al., 1988). Other researchers have used similar sulfate thresholds to screen for mine drainage effects. In a stream survey in Appalachia, Biesecker and George (1966) considered streams that had sulfate concentrations greater than 400 Iseq L -~ to be impacted by mine drainage. A bimodal distribution of streamwater sulfate was observed in the 47 streams of the nationwide U.S. Geological Survey hydrologic benchmark

ACIDMINEDRAINAGEIMPACTSIN THE EASTERNU.S.

95

no ~ NON MINE DRAINAG~

;IDIFIED BY AMD

no ] >

yes\~

yes..[

no I NON MINE DRAINAGE

n o ~

NONACIDIC, WEAKLY AMD IMPACTED

° The SO42-cutoff was 300 ,ueq/L in the Mid-Atlantic and 200 ,aeq/L in the Southeastsubregions. Fig. 2. Flowchartsummarizing the NSS acid mine drainage classification scheme.

network. The group of streams with SO4 z- concentrations less than 200 ~seq L -I presumably derived their sulfate from atmospheric deposition, whereas the group of streams with sulfate above 280 txeq L -I had watershed sources of sulfate (Campbell and Turk, 1988). Kleinmann et al. (1988), using the data of Dyer (1982a, b, c), calculated that the average sulfate concentrations of 92 first order perennial streams

96

ALAN T. HERLIHY ET AL.

with no surface mines in coal mining counties of Pennsylvania, West Virginia, and Tennessee were 485, 427, and 192 ~eq L -1, respectively. In streams with surface mines present in their watersheds, average sulfate concentrations were an order of magnitude higher. Very few (9%) acidic NSS stream samples had sulfate concentrations between 250 and 500 geq L -1. Thus the selection of an arbitrary cutoff value in this range has only a small impact on the AMD population estimates. In most acidic streams, the dominant source of sulfate was clearly either atmospheric (stream SO42- less than 250 ~eq L -1) or from watershed sources (stream S Q 2greater than 500 txeq L-l). In the NSS data, 20 stream reach ends were found to have acidic index chemistry (ANC < 0 iseq L -1) and SO42- concentrations above the threshold value. In 17 of these sites, there was evidence of mining impact, such as the presence of surface mines, noted on aerial photographs, topographic maps, or in site comments made during sampling visits. In 10 of these 17 sites, the chemical evidence was overwhelmingly in support of acid mine drainage classification; water quality met or exceeded all five of the common indicators of AMD ( [ 8 0 4 2 - ] > 1560 ~teq L -1, pH < 6, [Fe] > 9.0 ~M, [Mn] > 9.1 ~M, and BNC > ANC; Herb et al., 1981). Classification of stream sites with weaker AMD chemical signatures was confirmed with field observations of mining activity after revisiting the sites in the spring of 1987. Thus, these 17 sites were classified into the 'acidic due to AMD' group. The remaining three, acidic high-sulfate reach ends all had high dissolved organic carbon (DOC) concentrations (17 to 32 mg L -1) and high color values (110 to 163 Platinum Cobalt Units), and were located in the Mid-Atlantic Coastal Plain in New Jersey. There was no evidence of mining activity and these streams were probably acidic due to organic acids. Thus these three sites were not classified as 'acidic due to AMD'. Also added to the 'acidic due to AMD' group were 7 stream reach ends for which water samples were not collected for chemical analyses. In the NSS, water samples for complete chemical analysis were not collected from 7 stream reach ends that had an in situ conductivity greater than 500 gS cm -1 because they were considered to be too perturbed to be of use in an analysis of acidic deposition effects. Due to their high conductivity, all of these perturbed reach ends were screened for the possibility of AMD. All seven sites had low in situ pH (3.5 to 5.2), and examination of aerial photographs and topographic maps indicated the presence of surface mining activity in their watersheds. Because these sites were not sampled for complete water chemistry, checking for other acid mine drainage chemical signatures was not possible. The available evidence does indicate that these seven reach ends were acidic due to acid mine drainage and they were classified as such. All stream reach ends with nonacidic (ANC > 0) index chemistry and sulfate concentrations greater than 300 ixeq L -~ in the Mid-Atlantic and 200 Ixeq L 1 in the Southeast were screened for possible inclusion into the nonacidic, mine drainageimpacted groups. Those high-sulfate streams that had surface mines or mining symbols in their watersheds on 1:24 000-scale U.S. Geological Survey maps a n d / o r

ACID MINE D R A I N A G E IMPACTS IN T H E EASTERN U.S,

97

those in which the field sampling crew had noted the presence of mining activity were classified as being nonacidic, mine drainage impacted (i.e. a stream affected by A M D , but the stream itself is not acidic). These streams were then divided into two groups: those with sulface concentrations less than 1000 ~eq L -1 (considered weakly impacted by mine drainage) and those with sulfate concentrations greater than 1000 p.eq L -1 (considered strongly impacted by mine drainage). The rationale for this classification is that the streams in both groups are known to have a watershed source of sulfate, and the presence of mines in the watershed makes mining activity the likely sulfate source. The choice of 1000 p.eq L 1 as a criterion for separating weak from strong mine drainage effects is an arbitrary division. Dissolved Fe concentration was not used as a screening factor in the A M D classification because some of the less acidic A M D streams had very low Fe concentrations. Fe concentrations ranged from 0.1 to 379 ~M in the streams acidic due to A M D . In the whole NSS (excluding the A M D streams) over 20% of the upstream reach ends had Fe concentrations greater than 3 p.M (Sale et al., 1988). Thus a Fe cutoff value would not have worked as an A M D screen in the NSS. Sulfate is a better indicator of A M D than Fe because sulfate is a much more conservative ion. Bencala et al. (1987) found that sulfate was an excellent conservative tracer of A M D in a stream system in Colorado. Very few processes act to remove sulfate from solution in stream water. On the other hand, Fe rapidly precipitates out of solution as iron hydroxides ('yellow boy') as streamwater p H increases downstream from the A M D source. It was impractical in a large synoptic survey like the NSS to do an in-depth field reconnaissance of each study watershed. Although it is possible that mining activity might not show up on the maps or aerial photographs used to evaluate the NSS sites (due to small underground mines or reclamation), we feel that this will have only a minor impact on the A M D population estimates. All of the acidic NSS stream sites were examined in detail and those with sulfate concentrations below the A M D threshold criteria showed no evidence of mining activity in the watershed above the sample point. Sulfate concentrations in these streams were very close to the expected sulfate concentration assuming evaporative enrichment of sulfate in deposition, indicating little or no watershed sources of sulfate (Kaufmann et al., 1988). Given that sulfate is a good conservative tracer of A M D (Bencala et al., 1987), the lack of watershed sulfate sources indicate that if hidden mines were present they were having little or no impact on the observed stream chemistry. Assessing the source of sulfate in streams with sulfate concentrations above the sulfate threshold is more problematic. There are other sources of streamwater sulfate besides mining activity (e.g. agriculture, natural sulfide weathering, road cuts through pyritic bedrock). The acidic, high sulfate streams were examined in great detail and it was clear that except for the three sites in the Coastal Plain in New Jersey, all of the acidic, high sulfate sites were impacted by mining activity. The nonacidic, high sulfate systems were about evenly divided between those with and without mining evidence. Those sites without mining evidence typically were associated with

98

ALAN T. H E R L I H Y ET AL.

agriculture, livestock, orchards, logging, highway maintenance or proximity to the coast. For the sake of a standardized, reproducible classification scheme we decided that field, map or aerial photograph evidence of mining activity was necessary before attributing the watershed source of sulfate to mining activity. We think it is unlikely that many (if any) of these nonacidic sites were impacted by mining activity in cases where no evidence of mining was detected. 4. Population Estimation Procedure The statistical sampling unit in the NSS is the stream reach. All NSS population estimates (number and length) are based on weighted extrapolation from characteristics measured on the individual sample reaches. For purposes of this study, a reach was considered 'acidic due to A M D ' if either or both of the reach ends were classified as 'acidic due to AMD'. Similarly, nonacidic stream reaches were considered strongly impacted by mine drainage if either reach end was classified as 'nonacidic, strongly impacted by mine drainage'. Each reach sampled in the NSS has an associated sample weight equivalent to the number of reaches it represents in the map population (Kaufmann et al., 1988). The number of reaches in the NSS target population impacted by AMD was estimated by summing all of the sample weights of the impacted stream reaches in the geographic area of interest. The length of stream reaches impacted by AMD was estimated by summing the product of each sample reach's weight and reach length. The standard error associated with each estimate was determined using the Horvitz-Thompson variance estimator as applied to the NSS design (Overton, 1986). Full details of the NSS population estimation, and variance calculation procedures are given in Kaufmann et al. (1988). Population chemical characteristics were calculated as weighted means and standard deviations using the observed chemical index values and sample weights. 5. Results and Discussion The NSS was designed to survey streams large enough to be perennial but small enough to be sensitive to acidic deposition in regions with expected low ANC surface waters. Thus the NSS estimates of AMD impacts pertain only to those streams that are in the NSS target population within the areas of the eastern United States actually surveyed by the NSS (Figure 1). The NSS only included streams that were large enough to appear on l:250000-scale United States Geological Survey maps but that had watersheds areas less than 155 km 2 (60 mi 2) and were not located inside mapped urban areas. Thus the numbers and length of small intermittent streams, large streams and rivers are not included in these NSS population estimates. The estimated median drainage area of streams in the NSS target population was 1.5 km 2 at the upstream reach end and 8.1 km 2 at the downstream reach end. The median reach length was 2.8 km and the majority of stream widths were between

ACID MINEDRAINAGEIMPACTSIN THE EASTERNU.S,

99

NATIONAL STREAM SURVEY SAMPLE REACHES IMPACTED BY ACID MINE DRAINAGE Acidic due to AMD

Nonacidic, strongly AMD impacted

Nonacidic, weakly AMD impacted

~Coal

Regions

Fig. 3. Location of NSS sample sites impacted by acid mine drainage; solid lines indicate the NSS subregion boundaries. The hatched area indicates areas of potential coal mining activity (U.S. Geological Survey, 1967). 1 and 7 m (Kaufmann e t al., 1988). The NSS did not sample all areas in the eastern United States where A M D impacts are known to occur. Large sections of Ohio, Kentucky and West Virginia containing major coal deposits were not sampled during the NSS (Figure 3). The NSS estimates apply only to the subregions of the eastern United States actually sampled (Figure 1).

5.1. LOCATION The locations of observed stream reaches impacted by mine drainage coincided

100

ALANT. HERLIHYET AL. TABLE I

Population estimates of the number (N) and length (L) of stream reaches impacted by acid mine drainage in areas of the Eastern United States sampled by the NSS. Numbers in parentheses are the standard errors of the population estimates

State

AMD

(% State in NSS)

N

Pennsylvania (75%) Maryland (95%) West Virginia (51%) Tennessee/Kentucky (20% / 1%) Total NSS

AMD SMD WMD

= =

SMD L(km)

N

Total State NSS Resource

WMD L(km)

660 (386)

1980 (1030) -

19 (18) -

66 (64) -

596 (284) 121 (121) 1380 (489)

1490 (719) 1120 (1120) 4590 (1670)

1250 (479) 364 (206) 1630 (521)

3280 (1380) 2430 (1570) 5780 (2090)

N 293 (203) 145 (145) 592 (289) 121 (121) 1150 (390)

L(km)

N

L(km)

968 (682) 182 (182) 1200 (603) 613 (611) 2960 (1090)

9 910 (1 360) 5 280 (1 110) 9 390 (1 190) 2 140 (435) 64 300 (2065)

33 400 (5 180) 12 600 (2920) 21 600 (3 140) 11 200 (2590) 224000 (9973)

no NSS samples observed in this group. acidic due to acid mine drainage. nonacidic, strongly mine drainage impacted. nonacidic, weakly mine drainage impacted.

with areas of coal deposits within the NSS study area (Figure 3). Streams impacted by AMD were found in four of the nine NSS subregions: the Poconos/Catskills, Valley and Ridge, Northern Appalachians, and Southern Appalachians (Figures 1 and 3). The isolated AMD impacted sample reach at the northern border of the Valley and Ridge subregion is located in the anthracite coal belt of eastern Pennsylvania. Nonacidic, mine drainage impacted streams were located mostly in the southern half of the NSS (southern West Virginia, Tennessee, and Kentucky). Most of the mine drainage impacted streams in the northern half of the NSS were acidic. The map in Figure 3 is a good indicator of where mine drainage impacted streams may be found; however, it should not be used to infer spatial density because the NSS sample weights have not been taken into account. 5.2. POPULATION ESTIMATES

There was a total of 224000 km of target stream length in the NSS area. In 1986, an estimated 4590 km (2.0%) of this stream length were acidic during spring baseflow due to AMD and an additional 5780 km (2.6%) of streams were nonacidic but strongly impacted by mine drainage (Table I). Another 2960 km (1.3%) of NSS target stream length were nonacidic and weakly impacted by mine drainage. In the area surveyed by the NSS, Pennsylvania had the highest number and length of streams acidic due to AMD. Out of the estimated 9910 stream reaches in the NSS area in Pennsylvania, 6.7% (660) were acidic due to AMD. In West Virginia and Tennessee, there was a large population of nonacidic, mine drainage

101

A C I D MINE D R A I N A G E IMPACTS IN T H E E A S T E R N U.S.

i m p a c t e d systems. T h e s t a n d a r d errors o f the i n d i v i d u a l state estimates are often quite large ( s o m e t i m e s a l m o s t 100% o f the estimate) due to the small s a m p l e size within each class in each state (Table I). These s t a n d a r d errors are a s s o c i a t e d o n l y with the e r r o r in e s t i m a t i n g the NSS target p o p u l a t i o n f r o m the o b s e r v e d characteristics o f the s a m p l e streams a n d are b a s e d on the s a m p l e weights a n d the NSS statistical design. T h e r e p o r t e d s t a n d a r d errors do n o t a c c o u n t for a n a l y t i c a l o r t e m p o r a l v a r i a b i l i t y . Previous w o r k with the NSS d a t a has shown, however, t h a t the m a i n source o f v a r i a b i l i t y in NSS p o p u l a t i o n estimates is a s s o c i a t e d with the s p a t i a l e x t r a p o l a t i o n ( K a u f m a n n et aI., 1988). NSS p o p u l a t i o n estimates are less precise w h e n the n u m b e r o f i m p a c t e d s a m p l e sites is very small. S t a n d a r d errors for s t r e a m s i m p a c t e d b y A M D in the t o t a l NSS were higher in m a g n i t u d e b u t p r o p o r t i o n a t e l y lower ( a b o u t 35% o f the estimate) t h a n the s t a n d a r d errors in the i n d i v i d u a l state values, due to larger s a m p l e sizes. T h u s the r e g i o n a l (entire NSS) estimates are m o r e reliable t h a n the i n d i v i d u a l state estimates. 5.3. CHEMICA~ CHARACTERISTICS A c i d i c , A M D - i m p a c t e d s t r e a m s h a d an average sulfate c o n c e n t r a t i o n o f 3270 txeq L -1 a n d p H o f 4.3 (Table II). Sulfate was the d o m i n a n t a n i o n in all m i n i n g i m p a c t e d streams. In a n o t h e r survey, Biesecker a n d G e o r g e (1966) s a m p l e d 318 sites in A p p a l a c h i a a n d f o u n d an average sulfate c o n c e n t r a t i o n o f 3330 ~eq L -1 in m i n e d r a i n a g e i m p a c t e d s t r e a m s in P e n n s y l v a n i a a n d Ohio. A v e r a g e s t r e a m w a t e r sulfate c o n c e n t r a t i o n s in o t h e r states were s o m e w h a t lower (West Virginia=2700 ~eq L -l, Tennessee=937 ~eq L-I). I n the NSS, c o n c e n t r a t i o n s of F e a n d M n were m u c h higher in the A M D i m p a c t e d s t r e a m s (102 ~ M a n d 44 ~tM) t h a n in the n o n a c i d i c , mine d r a i n a g e i m p a c t e d streams (all < 7 g M ) , b e c a u s e o f the i n c r e a s e d s o l u b i l i t y o f these metals at lower p H (Table II). F o r c o m p a r i s o n , K l e i n m a n n et al. (1988),

TABLE II Chemical characteristics (population mean + standard deviation) of streams acidic due to acid mine drainage (AMD), and nonacidic streams impacted by strong (SMD) and weak mine drainage (WMD) Chemical species

AMD a

SMD

WMD

Sulfate (geq L -t) pH Dissolved Fe (IsM) Dissolved Mn (IxM) Inorganic Monomeric A1 (~M) Conductivity(isS/cm) BNC(geq L -1) ANC (~eq L -1) Ca2++ Mg2+(geq L- 0 Cl-(~eqL 1)

3270 + 1840 4.3 + 0.4 120 + 133 43.7 ± 32.4 72.4 ± 59.2 377 +_ 199 629 _+ 449 -306 _+ 341 2480 + 1400 67.7 + 66.6

2270 ± 1110 7.1 ± 0.7 1.59 ± 3.06 6.97 ± 11.7 1.2 + 2.4 311 ± 125 66.0 +_ 40.9 623 ± 659 2450 _+ 956 49.9 ± 29.8

480 + 158 7.0 ± 0.4 0.67 +_ 1.26 2.12 ± 3.11 0.2 ± 0.2 92.1 +_ 29.3 47.4 + 11.7 269 + 194 719 + 291 57.2 + 39.0

a Chemical analysis were not performed on 7 of the 24 AMD classified sample reach ends (see section 3). These 7 reaches had the highest conductivity in the NSS and presumably the most severe AMD impact.

102

ALAN T. H E R L I H Y ET AL.

using the data of Dyer (1982a, b, c), found that in 181 perennial first order streams in Pennsylvania, West Virginia, and Tennessee that had surface mining in their watersheds, the average sulfate, Fe and Mn concentrations were 6860 ~eq L -1, 221 I~M and 220 IxM, respectively. Inorganic monomeric A1 (Alim) is the form of A1 most toxic to fish (Driscoll et al., 1980). The Alim concentration in the AMD streams was very high (72 ~M), and along with the low observed p H in these streams, would be toxic to most fish. In an experiment by Baker and Schofield (1982), no brook trout (Salvelinus fontinalis) postlarvae survived after two weeks exposure to water with an Alim concentration of 11 ~M and p H of 5.1. All of the NSS stream reach ends classified as being acidic due to AMD had pH < 5 and Alim ~ 12 ~M. Thus, it is unlikely that many trout (and probably any other fish) survive in these streams. It should be noted that while toxic conditions exist in the 'acidic due to A M D ' streams, conditions in the nonacidic classes of mine drainage impacted streams do not appear to be toxic (Table II). These streams were by definition nonacidic (all had pH > 6) and thus had low metal concentrations. The observed pH and heavy metal concentrations in these streams are well below the required monthly average effluent limits (pH=6 to 9, Fe < 54 p.M, Mn < 33 ~xM) for operating mines (U.S. Code, 1985). Although these nonacidic streams are probably not biologically toxic, it does not mean that they are unimpacted by mine drainage. These streams likely suffer from physical habitat and substrate alteration, sedimentation and altered chemical compositon. 5.4. COMPARISONOF ACID MINE DRAINAGEAND ACIDIC DEPOSITION Using NSS data, it is possible to compare the regional extent of acidic deposition and AMD effects on an explicitly defined stream population. Although the acidity from both AMD and acidic deposition is derived mainly from HzSO4, they are very different problems in terms of sources (point versus nonpoint), concentrations, and geographic location. AMD is a concentrated point source of acidity only found in mining regions of the country whereas acidic deposition is a more diffuse, widespread regional phenomenon. While these two types of acidic pollution may not be directly comparable, the population estimates presented in this section can be thought of as a snapshot in time indicating the extent of acidic streams attributable to either acidic deposition or AMD during spring baseflow of 1986. In the NSS, about the same number of streams were estimated to be acidic due to AMD as to acidic deposition in subregions where mine drainage effects were observed (Table III). In the Northern Appalachians subregion, almost 10% of the stream reaches were estimated to be acidic due to AMD, and another 5% were acidic due to atmospheric deposition. In the Poconos/Catskills and Valley and Ridge subregions, more streams were acidic due to acidic deposition than to AMD. In terms of length, an estimated 4460 km of streams in the entire NSS were acidic due to acidic deposition (Kaufmann et al., 1988), as opposed to 4590 km of streams acidic due to AMD (Table I). Thus, AMD appears to be as serious a problem

103

ACID MINE DRAINAGE IMPACTS IN THE EASTERN U.S.

TABLE III Estimated number and percentage of upstream (U) and downstream (D) reach ends in subregions of the NSS acidic due to acid mine drainage (AMD) and acidic deposition. See Figure 1 for NSS subregion locations

NSS subregion Poconos/Catskills Valley & Ridge Northern Appalachians Southern Appalachians Other NSS Subregions Total NSS

Reach end U D U D U D U D U D U D

AMD Number

9 318 754 928 121 1190 938

%

Acidic dep. a Number %

Total number

0.3 2.0 8.0 9.9 2.0 1.9 1.5

209 636 499 326 121 724 407 2190 733

3 510 3510 15 900 15 900 9 420 9 420 6130 6 130 29 300 29 300 64 300 64 300

6.0 4.0 5.3 3.5 2.0 2.5 1.4 3.4 1.1

- - No samples observed in this group. a = Streams classified as most likely to have been acidic during spring basefiow due to acidic deposition (Kaufmann et aL, 1988).

as acidic deposition to stream resources in mining regions of the eastern United States. It may even be more serious, considering that the water chemistry is more toxic in AMD impacted streams than in streams impacted by acidic deposition due to heavy metal pollution and sedimentation problems. For example, the mean pH, Fe, and inorganic monomeric A1 concentration in the streams acidic due to acidic deposition was 4.8, 1.1 IsM, and 6.7 ~tM (Kaufmann et al., 1988) as compared to 4.3, 102 txM, and 721aM in the streams acidic due to AMD (Table II). It should be emphasized, however, that streams with low ANC that were not acidic during spring baseflow many become acidic during storm events (Sharpe et al., 1984). An additional 15 500 km of streams in the NSS with ANC between 0 and 50 ~eq L -1 had atmospheric deposition as their major source of acid anions, and were likely to be extremely sensitive to acidic deposition (Kaufmann et al., 1988; Schindler, 1988). 5.5.

COMPARISON TO PREVIOUS ACID MINE DRAINAGE SURVEYS

A number of other studies have been conducted on the regional effects of mine drainage in the eastern United States (Appalachian Regional Commission, 1960; Kinney, 1964). To compare results of these surveys to NSS results, the NSS data were grouped into subsets according to river basin location (Table IV). Table IV was not constructed to show changes in AMD impacts in time over the last 20 yr. It was compiled to show the patterns in the extent of AMD impacts in three surveys of completely different design (probability sample versus census). Conclusions should not be drawn regarding changes in the status and extent of the AMD

104

ALAN T. H E R L I H Y ET AL.

TABLE IV Regional estimates of the length (km) of streams acidic due to acid mine drainage (AMD) and all mine drainage (MD) in surveys conducted by the NSS, Appalachian Regional Commission (ARC), and the United States Fish and Wildlife Service (USFWS) Stream length (km) NSS

ARC

River basin

AMD

MDa

Anthraciteb Tioga W, Branch Susquehanna Juniata Allegheny Monongahela N. Branch Potomac Kanawha Cumberland/Tennessee Entirely outside NSS basins Survey totals

849 78 745 90 1160 178 369 1120

156 1587 178 3 227 3 553

4590

10 370

849 78 745

USFWS

AMD

MD

AMD

912 89 1670 132 1700 2680 277 238 240 1290 9230

982 89 1 830 132 1 720 2 690 277 2 260 1070 5 850 16 900

1020 53 781 130 1510 1400 175 "~09 201c 4000 9480

No NSS samples with mine drainage observed in this basin. In the NSS, mine drainage impacts (MD) were estimated by summing the length of streams acidic due to AMD and nonacidic, strongly impacted by mine drainage (Table I). b The Anthracite coal belt in Eastern Pennsylvania is usually treated as a distinct entity. It is composed of flows from parts of the North Branch of the Susquehanna and Delaware River basins. C This USFWS estimate is only for the portion of the basin within the state of Tennessee, data for Kentucky was not subset by river basin. ARC = Estimates were made for all of Appalachia in the late 1960s by consulting with state and local agencies, reviewing reports, and field reconnaissance (Appalachian Regional Commission, 1969; Federal Water Pollution Control Admin., 1969). USFWS = Estimates were made in 1963 by summerizing reports from each State's Fish and Game Department (Kinney, 1964). a

p r o b l e m with time because the three surveys e m p l o y e d different sample designs, target p o p u l a t i o n s , sample area b o u n d a r i e s , a n d definitions of what is significant mine drainage. Also, m a i n s t e m rivers a n d u r b a n streams were excluded from the NSS target p o p u l a t i o n a n d are n o t included in a n y of the NSS estimates. G i v e n that the precision in the NSS estimate of the total A M D i m p a c t e d stream length was 36% (4590 k m +_ 1670), a n d that the precision of the estimates made in the other surveys is u n k n o w n , m e a n i n g f u l time t r e n d c o m p a r i s o n s from the data in Table IV can n o t be made. W i t h a few exceptions, NSS length estimates of streams acidic due to A M D agree quite well with estimates made by the U n i t e d States Fish a n d Wildlife Service ( U S F W S ) in 1963. Basin by basin, length estimates made by the A p p a l a c h i a n Regional C o m m i s s i o n ( A R C ) in the late 1960's are usually higher t h a n NSS or U S F W S estimates. Both the U S F W S a n d A R C estimates include all river b~isins affected by m i n e drainage in Appalachia. Large portions of some of the river basins in

ACID MINE D R A I N A G E IMPACTS IN T H E EASTERN U.S.

105

Table IV (Allegheny, Monongahela, Kanawha, and Cumberland/Tennessee), however, are not entirely within the NSS boundaries. This may be the reason why the estimate of AMD impacts in the Allegheny River basin is so much lower in the NSS than in the other surveys (Table IV). About half of the Allegheny River basin lies outside the NSS boundaries. The portion of the basin within the NSS does not appear representative of the basin as a whole with respect to AMD impacts probably due to heavy mining influences in the western portion of the basin outside the NSS boundaries. In both the NSS and ARC, almost all mine drainage impacted streams in river basins in the northeastern portion of Appalachia were acidic, whereas the majority of mine drainage impacted streams in the southern portion (Kanawha and Cumberland/Tennessee River basins) were not acidic (Table IV). A likely explanation for this pattern is the presence of calcareous rocks in the watersheds of the southern basins that would neutralize the AMD-acidity but would not change the sulfate concentration. High-alkalinity streams (from carbonate mineral weathering) have been observed in this part of Appalachia (Biesecker and George, 1966). Helsel (1983) found that aquifer rock type was a major factor influencing stream pH and alkalinity in mine drainage impacted streams in eastern Ohio. Alternatively, mine drainage production may be less intensive in the southern portion of the NSS due to less coal mining or lesser amounts of S-bearing material exposed per amount of coal mined (Biesecker and George, 1966).

6. Conclusions Acid mine drainage (AMD) is currently a serious water quality problem in streams in coal mining regions of the eastern United States. Almost 10% of the stream reaches in the National Stream Survey's (NSS) Northern Appalachians subregion were acidic during spring baseflow in 1986 due to AMD. In the entire NSS, 4590 (+ 1670) km of streams were acidic due to AMD and another 5780 (_+ 2090) km of streams were impacted by mine drainage (and had SO42- > 1000 ~xeqL -1) without being acidic. In subregions of the NSS with observed mine drainage effects, roughly the same number and length of streams were acidic due to AMD as to acidic deposition. The population estimates of mine drainage impact made using the NSS probability sampling design were similar to estimates made in previous census surveys. Overall, these results demonstrate that a statistically based stream survey employing relatively complete chemical characterization can be a useful tool for evaluating the regional impact of water pollution problems, quite apart from the original objective of that survey.

Acknowledgments The research described in this article has been funded by the U.S. Environmental Protection Agency. This document has been prepared at the EPA's Environmental

106

Research

ALAN T. H E R L I H Y ET AL.

Laboratory

in

Corvallis,

Oregon,

through

cooperative

agreement

(CR815168) with U t a h State University. It has been subjected to the Agency's peer a n d a d m i n i s t r a t i v e review a n d a p p r o v e d for p u b l i c a t i o n . M e n t i o n of trade names or commercial p r o d u c t s does n o t constitute e n d o r s e m e n t or r e c o m m e n d a t i o n for use. We t h a n k the n u m e r o u s people involved with the execution of the N a t i o n a l Stream Survey. We also t h a n k E McIntire, R. H o l d r e n , E Shaffer, D. Coffey, a n d S. Christie for helpful c o m m e n t s o n an earlier draft.

References Appalachian Regional Commission: 1969, Acid Mine Drainage in Appalachia, Appalachian Regional Commission Report, Washington D.C. Baker, J. E and Schofield, C. L.: 1982, Water, Air, and Soil Pollut. 18, 289. Barnes, H. L. and Romberger, S. B.: 1968, Water Pollut. ControlFed. J. 40, 371. Bencala, K. E., McKnight, D. M., and Zellweger,G. W.: 1987, Water Resour. Res. 23, 827. Biesecker, J. E. and George, J. R.: 1966, Stream Quality in Appalachia as Related to Coal-mine Drainage, 1965, U.S. Geological Survey Circular 526, Washington D.C. Campbell, D. H. and Turk, J. T.: 1988, WaterResour. Res. 24, 871. Cougan, K. A., Sutton, D. W., Peck, D. V., Miller, V. J., and Pollard, J. E.: 1988, National Stream Survey - Phase I: Quality Assurance Report, EPA/600/4-88/018, U.S. Environmental Protection Agency, Las Vegas, Nevada. DriscoU, C. T., Baker, J. E, Bisogni, J. J., and Schofield, C. L.: 1980, Nature 284, 161. Dugan, E R.: 1985,Biochemical Ecology of Water Pollution, Plenum Press, New York. Dyer, K. L.: 1982a, Stream Water Quality in the Coal Region of Tennessee, U.S. Dept. Agric. Forest Service Tech. Rep. NE-77, Berea, KY. Dyer, K. L.: 1982b, Stream Water Quality in the Coal Region of Pennsylvania, U.S. Dept. Agric. Forest Service Tech. Rep. NE-76, Berea, KY. Dyer, K. L.: 1982c, Stream Water Quality in the Coal Region of West Virginia and Maryland, U.S. Dept. Agric. Forest Service Tech. Rep. NE-70, Berea, KY. Federal Water Pollution Control Administration: 1969, Stream Pollution by Coal Mine Drainage in Appalachia, U.S. Dept. Interior, Cincinnati, Ohio. Helsel, D. R.: 1983, Water Resource. Bull. 19, 881. Herb, W. J., Shaw, L. C., and Brown, D. E.: 1981, U.S. Geological Survey Water-Resources Investigations 81-538. Hillman, D. C., Pia, S. L., and Simon, S. J.: 1987, National Surface Water Survey: National Stream Survey Analytical Methods Manual, EPA/600/8-87/005, U.S. Environmental Protection Agency, Washington D.C. Kaufmann, E R., Herlihy, A. T., Elwood, J. W., Mitch, M. E., Overton, W. S., Sale, M. J., Messer, J. J., Cougan, K. A., Peck, D. V., Reckhow, K. H., Kinney, A. J., Christie, S. J., Brown, D. D., Hagley, C. A., and Jager, H. I.: 1988, Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United States. Volume I: Population Descriptions and Physico-chemical Relationships,

EPA/600/3-88/021a, U.S. Environmental Protection Agency, Washington, D.C. Kim, A. G., Heisey, B. S., Kleinmann, R. L. E, and Deul, M.: 1982, Acid Mine Drainage: Control and Abatement Research, U.S. Bureau of Mines Information Circular 8905, Pittsburgh, PA. Kinney, E. C.: 1964, Extent of Acid Mine Pollution in the United States Affecting Fish and Wildlife, U.S. Dept. Interior, Fish and Wildlife Circular 191, Washington, D.C. Kleinmann, R. L. E, Jones, J. R., and Erickson, P. M.: 1988, An Assessment of the Coal Mine Drainage Problem, Proceedings of the 10th annual Conf. of the Assoc. Abandoned Mine Land Programs, Wilkes-Barre, PA. Mills, A. L.: 1985, in D. Klein and R. L. Tate (eds.), Soil Reclamation Processes, Marcel Dekker, Inc., New York, pp. 35-81. Omernik, J. M. and Powers, C. F.: 1983,Annals Assoc. Amer. Geographers 73, 133.

ACIDMINEDRAINAGEIMPACTSIN THE EASTERNU.S.

107

Overton, W. S.: 1986, A Sampling Plan for Streams in the National Stream Survey, Technical Report 114, Department of Statistics Oregon State University, Corvallis, OR. Powell, J. D.: 1988, Environ. Geol. Water Sci. 11, 141. Sale, M. J., Kaufmann, E R., Jager, H. I., Coe, J. M., Cougan, K. A., Kinney, A. J., Mitch, M. E., and Overton, W. S.: 1988, Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United States. Volume II: Streams Sampled, Descriptive Statistics, and Compendium of Physical and ChemicalData, EPA/600/3-88/021b, U.S. Environmental Protection Agency, Washington D.C. Schindler, D. W.: 1988, Science 239, 149. Sharpe, W. E., DeWalle, D. R., Leibfried, R. T., Dincola, R. S., Kimmel, W. G., and Sherwin, L. S.: 1984, J. Environ. Qual. 13, 619. Stumm, W., and Morgan, J. J.: 1970, Aquatic Chemistry, Wiley-Interscience, New York. U.S. Code of Federal Regulations: 1985, Title 40 Protection of the Environment, Chapter 1 Part 434, Subpart C, Section 434.32, p. 245. U.S. Geological Survey: 1967, in The National Atlas of the United States of America, U.S. Geological Survey, Washington D.C., pp. 186-187.