Treatment of Acid Mine Drainage by Four
Vertical Flow Wetlands in Pennsylvania

Acid mine drainage (AMD) is a serious problem in many watersheds where coal is mined. Passive treatments, such as wetlands and anoxic limestone drains (ALDs), have been developed, but these technologies show varying treatment efficiencies. A new passive treatment technique is a vertical flow wetland or successive alkalinity producing system (SAPS). Four SAPS in Pennsylvania were studied to determine changes in water chemistry from inflow to outflow. The Howe Bridge SAPS removed about 130 mg/L (40%) of the inflow acidity concentration and about 100 mg/L (60%) Fe. The Filson 1 SAPS removed 68 mg/L (26%) acidity, 20 mg/L (83%) Fe, and 6 mg/L (35%) Al. The Sommerville SAPS removed 112 mg/L (31%) acidity, exported Fe, and removed 13 mg/L (30%) Al. The McKinley SAPS removed 54 mg/L (91%) acidity, and 5 mg/L (90%) Fe. Acid removal rates at our four sites were 17 (HB), 52 (Filson1), 18 (Sommerville) and 11 (McKinley) g of acid per m² of surface wetland area per day (g/m²/day). Calcium (Ca) concentrations in the SAPS effluents were increased between 8 and 57 mg/L at these sites. Equilibrators, which were inserted into compost layers to evaluate redox conditions at our sites, showed that reducing conditions were generally found at 60-cm compost depths and oxidized conditions were found at 30-cm compost depths. Deeply oxidized zones substantiated observations that channel flow was occurring through some parts of the compost. The Howe Bridge site has not declined in treatment efficiency over a six-year treatment life. The SAPS construction costs were equal to about seven years of NaOH chemical treatment costs and 30 years of lime treatment costs. So, if the SAPS treatment longevity is seven years or greater and comparable effluent water quality was achieved, the SAPS construction was cost effective compared to NaOH chemical treatment. Construction recommendations for SAPS include a minimum of 50 cm of compost thickness, periodic replacement or addition of fresh compost material, and increasing the number of drainage pipes underlying the limestone.

Introduction

Acid mine drainage (AMD) forms when sulfide minerals such as pyrite (FeS₂) are exposed to oxygen and water during mining and other large scale land disturbances. Acid mine drainage is characterized by high sulfate concentrations, high levels of dissolved metals, and pH generally <4.5. In 1995, the Environmental Protection Agency (US EPA 1995) estimated that about 5,000 km of streams were impacted by AMD in the northern Appalachian area of the United States (Pennsylvania, Maryland, Ohio, and West Virginia). Historically, AMD treatment is accomplished by adding a strong base to neutralize acidity, raise pH, and precipitate metals. Lime (in various hydrated forms), sodium hydroxide, sodium carbonate, and anhydrous ammonia are the four primary chemicals used for AMD treatment. Although effective, chemical treatment is expensive when the cost of equipment, chemicals, and manpower is considered (Phipps et al. 1991), and responsibility for treatment may be a long term liability. In 1990, the United States coal industry spent over $1 million per day on active treatment of AMD (Kleinmann 1990).

Recently, a variety of passive treatment systems have been developed that do not require continuous chemical inputs and that take advantage of naturally occurring chemical and biological processes to cleanse contaminated mine waters (Hedin et al. 1994). The primary passive technologies include constructed wetlands, anoxic limestone drains (ALD), and vertical flow systems such as successive alkalinity producing systems (SAPS).

Aerobic wetlands are generally used to collect water and provide residence time and aeration so metals in the water can precipitate. The water in this case usually has net alkalinity. Wetland plants encourage more uniform flow and help to introduce fresh organic matter to the wetland annually. For net alkaline waters, aerobic wetlands can be sized to remove iron at the rate of 10 to 20 g/m²/day (Hedin et al. 1994). Cost to build these wetlands is usually estimated at $10 per m² without plants and $20 per m² with plants. (The cost figures for active and passive treatment systems used in this paper are those used by the West Virginia Division of Environmental Protection to estimate construction costs).

Horizontal flow, anaerobic wetlands encourage interaction of water with organic-rich substrates. The wetland substrate usually contains a layer of limestone in the bottom of the wetland or the limestone may be mixed with the organic matter. Wetland plants are transplanted into the organic substrate. These systems are used when the water has net acidity, so alkalinity must be generated in the wetland and introduced to the net acid water in order to accomplish significant precipitation of dissolved metals.

Several treatment mechanisms are enhanced in anaerobic wetlands compared to aerobic wetlands, including formation and precipitation of metal sulfides, metal exchange and complexation reactions, microbially-generated alkalinity due to reduction reactions, and continuous formation of carbonate alkalinity due to limestone dissolution under anoxic conditions. Microbial mechanisms of alkalinity production are likely to be of critical importance to long term AMD treatment. However, Wieder (1992) documents that the mechanism and efficiency of AMD treatment varies seasonally and with wetland age, presumably because of microbial effects.

Hedin et al. (1994) suggest that anaerobic wetlands treating net acid waters be sized using a factor of 3.5 g of acidity removal per m² per day (g/m²/day) or 5 g of iron per m² per day. Cost for building anaerobic wetlands usually range between $40 to $50 per m² without plants and $50 to $75 per m² with plants.

Anoxic limestone drains (ALDs) are buried cells or trenches of limestone into which anoxic water is introduced. The limestone dissolves in the mine water and adds alkalinity (Watzlaf and Hedin 1993). Under anoxic conditions, the limestone does not coat or armor with Fe hydroxides because Fe²⁺ does not precipitate as Fe(OH)₂ at pH <8.0. The effluent pH of ALDs is typically between 6 and 7. The sole function of an ALD is to convert net acidic mine water to net alkaline water by adding bicarbonate alkalinity. The removal of metals within an ALD is not intended and has the potential to significantly reduce the permeabilty of the drain resulting in premature failure.

Longevity of treatment is a concern for ALDs, especially in terms of water flow through the limestone. If appreciable dissolved Fe³⁺ and Al³⁺ are present, clogging of limestone pores with Al and Fe hydroxides has been observed (Faulkner and Skousen 1994, Watzlaf et al. 1994).

Sizing criteria for ALDs are generally based on the acid load per year and then multiplying this acid amount by an expected number of years of treatment (usually 20 years). This number is the amount of limestone needed to neutralize the acid load. A cell or trench capable of holding this amount of limestone is then dug in the ground. Others have simply suggested that 8 Mg of limestone are needed to treat a liter per minute of AMD (30 tons per gpm of acid water). Cost for building ALDs can be estimated by using $45 per Mg of limestone.

Any strategy or system which could inexpensively raise pH while removing Al, and either precipitate Fe³⁺ or reduce Fe³⁺ to Fe²⁺ prior to entering an ALD may broaden treatment application. Ferrous iron does not armor limestone at the pH values attained in ALDs (Turner and McCoy 1990). Such a system was developed by Kepler and McCleary (1994) and called a successive alkalinity producing system (SAPS) or vertical flow wetland. Vertical flow wetlands incorporate the use of organic matter similar to anaerobic wetlands but also include a limestone bed similar to an ALD. In vertical flow, anaerobic wetlands, water flows downward, from a pond established on top of the system, through organic matter and through limestone before flowing out of the system through drain pipes. The principle treatment mechanisms are: 1) the oxidized water becomes more reduced as it flows through the organic matter, and Fe and Al may be removed from the water through exchange and filtering with organic matter; and 2) the water, under more reducing conditions and containing less metals, flows downward and contacts limestone, which then can add bicarbonate alkalinity.

Compared with horizontal-flow, anaerobic wetlands, vertical flow systems greatly increase the interaction of water with organic matter and limestone. In a typical SAPS, acid water is ponded at depths of 50 to 150 cm over 15 to 50 cm of an organic compost, which is underlain by 50 to 200 cm of limestone. Below or within the limestone is a series of drainage pipes that convey the water into a pond where metals are precipitated. The hydraulic head drives ponded water through the anaerobic organic compost, where oxygen is consumed and ferric iron is reduced to ferrous iron. Sulfate reduction and Fe sulfide precipitation can also occur in the compost. After aeration and metal precipitation in a pond or wetland, water retaining net acidity can be passed through additional SAPS.

Sizing of SAPS is based on acid removal similar to anaerobic wetlands. For acid water, 30 g of acid can be removed for every m² of surface area per day (30 g/m²/day) or about a ten-fold increase in acid removal vs the removal for anaerobic wetlands (Ben Faulkner, personal communication). Costs for building SAPS range from $60 to $75 per m².

Kepler and McCleary (1994) reported data for three SAPS in Pennsylvania. The Howe Bridge SAPS (1500 m²) reduced acidity of a 140-L/min flow from 321 mg/L to 93 mg/L as CaCO₃ (71% decrease or 30.6 g/m²/day). This is five to ten times the acid removal rate of horizontal flow wetlands. The Schnepp Road SAPS (1200 m²) decreased acidity from 84 to 5 mg/L as CaCO₃ (94% decrease and 10.8 g/m²/day acid removal rate) and removed all 19 mg/L of ferric iron.

In a West Virginia SAPS, Kepler and McCleary (1997) found that Al in the AMD precipitated as aluminum hydroxide in the limestone. Their drainage design incorporated a flushing system, which allowed for precipitated Al to be flushed from the pipes thereby maintaining hydraulic conductivity through the limestone and pipes. The Greendale SAPS (572 m²) received 25 L/min of acid mine drainage, and increased the pH from 2.8 to 6.5, changed the water from a net acid water (925 mg/L as CaCO₃) to a net alkaline water (150 mg/L as CaCO₃ and 58 g/m²/day acid removal), and reduced Fe from 40 to 35 mg/L and Al from 140 to <1 mg/L.

The SAPS are a relatively new passive treatment technology and widely varying results have been reported. Many questions remain as to the proper sizing criteria, depth of water, thickness and life of the compost, short circuiting and channeling of water, piping system design, and plugging with hydroxides. Therefore, the objectives of this study were to determine the influent and effluent water chemistry of four SAPS, to evaluate treatment efficiency and longevity, and to provide construction recommendations for future SAPS.

Materials and Methods

Site Descriptions

Four SAPS, located in western Pennsylvania, were chosen for study. The Howe Bridge, Filson 1, and McKinley SAPS are located in the Mill Creek Watershed in Jefferson County, and are part of an organized effort to remediate AMD impacts to Mill Creek, a tributary of the Clarion River. The Sommerville SAPS is located in Clearfield County. Water from this SAPS flows into Moose Creek. These four SAPS were chosen because they were constructed at different times and had varying designs (Table 1). Sizing of each system was based on water chemistry and flow. Residence time for water in each system was based on the surface area of the system, standing water depth, and a 40% porosity within the organic matter and limestone.

The Howe Bridge system was constructed in November 1991. Two groundwater seeps are treated by two ALDs. The water leaving the ALDs flows into two sedimentation ponds where large amounts of Fe floc accumulate. The effluent from the second pond discharges into a constructed wetland. The first half of the wetland has a shallow water layer over mushroom compost. The second half of the wetland is a SAPS. The SAPS has 190 cm of standing water on top of 45 cm of mushroom compost, which overlies 45 cm of limestone. The total surface area of the SAPS is 1500 m². The underlying piping system consists of two, 10-cm flexible perforated pipes beneath the limestone. The water then discharges through a wooded area before flowing into Mill Creek.

The Filson 1 SAPS was constructed in the summer of 1994. The AMD flows out of coal mine spoils into the first SAPS (Filson 2). The outflow of this system flows into a settling pond, which then discharges into a second SAPS (Filson 1). The second SAPS is the system where water measurements were taken. Filson 1 has a surface area of 420 m². The SAPS has 125 cm standing water over 40 cm of spent mushroom compost, which overlies 65 cm of limestone. The underlying piping system consists of one, 15-cm flexible perforated pipe near the downstream side of the SAPS. The effluent water discharges into a wooded area where it meanders through a natural wetland before entering Little Mill Creek.

The Sommerville SAPS was constructed in the summer of 1995 on a reclaimed surface mine site and intercepts AMD that flows from diffuse sources in coal mine spoils. The seeps are collected in a ditch and the water flows into a holding pond. The pond discharges into a SAPS containing 100 cm of standing water over 35 cm of horse manure compost, which overlies 40 cm of limestone. An additional 288 tons of limestone dust were incorporated into the compost to increase its alkalinity. The SAPS has a surface area of 1350 m². The underlying piping system consists of one, 15-cm perforated pipe extending the length of the system with four lateral pipes coming off of the main trunk. The treated water then flows into Moose Creek.

The McKinley SAPS was constructed in the fall of 1996 and has a surface area of 600 m². The SAPS has standing water of 40 cm over 15 cm of mushroom compost, which overlies 40 cm of limestone. The underlying piping consists of one, 30-m long, 10-cm diameter perforated pipe under the limestone. The system discharges into a settling pond and the effluent goes directly into an unnamed tributary of Mill Creek.

Sampling Techniques

Influent and effluent water samples were collected monthly from December 1996 through November 1997. The influent samples were taken at stations directly before the AMD entered the surface of the SAPS. The effluent samples were taken from pipes that were attached to the underdrain system of each SAPS. Influent and effluent flows were measured with V-notch weirs or by the use of a bucket and stopwatch. Data are reported at both high flow and low flow conditions. The high flow was defined as the greatest inflow to the system during the 12-month sampling period, while the low flow was the smallest inflow to the system during the 12-month sampling period. Field pH and temperature were measured at the time of collection with an Orion portable pH tester and a standard thermometer. Two unfiltered water samples were taken at each sampling point: a 500-mL sample was taken for general water chemistry (pH, acidity, alkalinity, and sulfate), and a 150-mL sample was collected, acidified to pH <2 with 5 ml 10% HNO₃, and used for elemental analyses.

Water analysis was performed at the Pennsylvania Department of Environmental Protection water test lab in Harrisburg using standard EPA techniques (Clesceri et al. 1998). Water pH, alkalinity and acidity were determined by a Metrohm pH Stat Titrino Titration System (Brinkman Instruments, Westbury, NY). Acidity and alkalinity values were converted to CaCO₃ equivalents. Sulfate was determined with an Alpkem Flow System (Pulse Instrumentation, Saskatchewan, Canada). Al, Fe, and Ca were determined by atomic absorption spectrophotometry (AAS).

The oxidation status and pH of the compost layer were determined using an apparatus called an equilibrator. The equilibrators were constructed of 340-cm-long, 2.5-cm by 10-cm (1x4 inch) wooden stakes. One side of the stake had a 110-cm length of 1.8 cm (½-inch) PVC pipe attached to the stake with screws and covered with silicone caulking. The PVC pipe was cut in half to create a concave surface, which held an agar-Fe sulfide suspension. A 2% agar solution was prepared in the field using distilled water. Agar is a gelatinous material derived from seaweed and is used as a gelling agent. It dissolves above 50^oC and begins to harden below 35^oC. Iron monosulfide was made by adding 1 ml of 2.4 M Na₂S to 250 ml soluton of 28 mM FeSO₄. The Fe monosulfide was mixed with the 2% agar until a homogenous suspension was obtained (Edenborn et al. 1993) and it was then poured into the concave PVC pipe and allowed to solidify. As stated by Edenborn et al. (1993), the oxidation zone was determined by color changes in the suspension along the length of the equilibrator. If the suspension color remained black, the redox conditions at that depth were assumed to be reduced. If the suspension was red, orange, yellow or a clear color, the redox conditons at that depth were assumed to be oxidized.

Equilibrators were equally-spaced at intervals of 2.5 to 3.0 m in three SAPS. Because of different surface areas, each SAPS had a different number of equilibrators: Howe Bridge had 126, Filson 1 had 56, and Sommerville had 85. McKinley was not suitable for equilibrator testing because the shallow depth of the compost layer would not support the equilibrators. Equilibrators were placed into the organic layer for 48 hours and then removed. Immediately after removal, a determination was made at four depths along the equilibrator (60, 50, 40, and 30 cm) based on the color of the agar-sulfide gel whether the water at that depth was oxidized or reduced.

Results and Discussion

The Howe Bridge SAPS had a high influent pH due to the water being pre-treated by ALDs, but the water still contained an average acid concentration of 330 mg/L as CaCO₃ (Table 2), predominately in the ferrous iron form. After passage through the SAPS, acidity decreased by 130 mg/L or 40% (Table 3), while alkalinity rose about 15 mg/L as CaCO₃ (Tables 2 and 3). Iron concentrations decreased by 50 to 69%, and Ca levels increased by 30 mg/L (Table 3). The Howe Bridge SAPS had a low flow per area (.09 L/min/m²) and a noticeably longer residence time (27 days) than the other SAPS (2 to 6 days). The standing water depth of 190 cm at this site was much more than the shallower water depths at the other sites (0.40- to 125-cm water depths).

The acid removal rate at this site was 17.4 g/m²/day (Table 4), which is much higher than anaerobic, horizontal flow wetlands (3 to 5 g/m²/day), but lower than the acid removal rate early in the life of this system (30 g/m²/day, as reported by Kepler and McCleary, 1994). The cost for acid removal by this SAPS over the seven-year treatment period is $940 per Mg of acid . If the same amount of acid is removed during the next three years (treatment for ten years), the cost for acid removal would be $458 per Mg. The SAPS construction costs at $62,500 (Table 1) were equal to about seven years of NaOH chemical treatment costs ($8900/yr, Table 4) and about 30 years of lime chemical treatment costs ($2090/yr).

The water at Filson 1 increased in pH from 3.9 to 4.9 at high flow and decreased in acidity by 46 to 90 mg/L (19 to 35%). About 15 mg/L alkalinity as CaCO₃ was produced. About 90% of the Fe and 20 to 50% of the Al were removed. This site had the highest flow per area of wetland (.52 L/min/m²), the lowest ratio of limestone to acid amounts per year, and the shortest residence time of any of our sites, all of which should decrease the treatment effectiveness and longevity of this system. Also, the Filson 1 SAPS had some initial problems after construction, so some modifications were made. In this process, the compost layer was compacted and subsequently hindered downward water flow. The compaction was obvious because we had difficulty inserting the equilibrators into the compost throughout the wetland, and we also found that 50% of the flow exited the SAPS over the spillway rather than through the underlying pipes. Even with only 50% of the flow moving through the organic matter and limestone, removal of acidity, Fe and Al was high for the total pollution load in the water. Acid removal rates at Filson 1 were the highest of our sites at 51.5 g/m²/day, which is similar to the 58 g/m²/day rate at the Greendale SAPS (Kepler and McCleary 1997) and the 47 g/m²/day at the Douglas reclamation site (Skousen et al. 1999).

The SAPS construction cost at Filson 1 was $23,000. Acid removal cost was $798 per Mg over four years, but if the acid removal rate continued for ten years, the cost of treatment would be $291 per Mg of acid. Chemical treatment costs were calculated to be about $7400/yr for NaOH and $1740 for lime. So about three years of treatment by the SAPS would be required to break even with the NaOH treatment and about 13 years of treatment to recover lime treatment costs.

The SAPS at Sommerville caused an increase in pH, decreased acidity by 70 to 150 mg/L (about 30%), and produced 15 mg/L of alkalinity as CaCO₃. Iron and sulfate concentrations increased from influent to effluent at this site. The high Al concentration of this water was reduced by about 30% and Ca levels increased by 40 to 55 mg/L.

The acid removal rate at Sommerville was 18 g/m²/day, which is similar to other SAPS reported in Kepler and McCleary (1994). Since this site was only three years old, the cost of acid treatment was $1,921 per Mg. If the treatment continued for ten years, the cost would be $576 per Mg. Annual chemical costs at this site were estimated to be $8424 for NaOH and $1980 for lime, which means that about six years of NaOH treatment equates to the construction cost of the SAPS and about 26 years for lime.

Sommerville also had water flow problems, which may have decreased its treatment efficiency. The Al concentration of the influent exceeded 40 mg/L and a thick amorphous precipitate developed at the compost surface. The compost in the system was also not evenly spread. In the middle of the system, only about 10 cm of compost were present, compared to a 60-cm-thick layer around the outsides of the system. A tracer study was conducted at this site, using Rhodomine dye, in an attempt to determine if preferential flow was occurring within the system. The test was conducted twice. First, the dye was placed at the inflow and observed. The second test was conducted by first freezing the dye in a plastic bag and placing it in the area of the system where preferential flow was suspected. Both dye tests were observed for 4 hours and both showed preferential flow through the middle of the system where it contained only 10 cm of compost. The preferential flow within the system decreased treatment because the majority of water flowed through a small portion in the middle and the entire system was not being fully utilized for treatment.

The SAPS at McKinley showed the most dramatic treatment results. The water pH increased from 4 to greater than 6, 40 to 70 mg/L acidity was almost completely removed, and the water became net alkaline. This site had a similar .09 L/min/m² flow per wetland area as Howe Bridge, but it had a much lower acid load per year and a much higher ratio of limestone amount to acid load than any of our sites. The acid removal rate was 11.2 g/m²/day, which was the lowest of our four sites, but the acid load to be removed was much lower at McKinley than the other sites. The cost of acid removal was very high ($3,437 per Mg) over two years of treatment, but would be $687 per Mg over ten years of treatment. The construction cost of the system would be equaled after seven years of NaOH treatment and 31 years of lime treatment.

McKinley had good treatment success due to the low flow and low metal concentrations of the influent water. It was also the youngest system. The Howe Bridge system was also more efficient in acid removal rate immediately after construction based on values reported by Kepler and McCleary (1994). Most of the treatment at McKinley appears to have come from limestone dissolution (Ca increased by 57 mg/L) since only about 15 cm of compost was present. This system exhibited the degree of success in terms of pH increase that can occur with SAPS when AMD contains low acidity and low metals in the influent water. Since this site is the youngest of the systems we studied, the good treatment success may be short-lived due to the precipitation of Fe and Al in the limestone.

One interesting result from the water analysis at all sites was the increase in sulfate concentrations from influent to effluent (Table 2). The SAPS should remove sulfate through bacterial sulfate reduction in the organic layer by forming hydrogen sulfides and metal monosulfides. Since residence time of these SAPS was between 2.4 and 27 days, we assumed that some sulfide production would occur and that effluent sulfate concentrations would be lower than influent concentrations. Moreover, Fe and Al hydroxide flocs produced during AMD treatment (in either passive or chemical treatment) almost always contain some sulfate in the mix, forming various types of aluminum and iron hydroxy sulfates (like jarosite, HFe₃(SO₄)₂(OH)₆, and jurbanite, AlOHSO₄). So the increase in sulfate in effluent water was unexpected. One explanation could be that colorless sulfur bacteria found in most AMD wetlands accumulate sulfide granules when reducing conditions occur, but the bacteria release these granules when oxidizing conditions prevail, which then produce sulfate (Eleanor Robbins, personal communication).

Equilibrators were placed in the compost layer at three sites to estimate redox conditions within the compost layer. The McKinley SAPS only had 15 cm of organic material and equilibrators could not be placed in this shallow layer. The equilibrator data showed that the SAPS organic layers were reduced at 60 cm in >80% of the equilibrators (Figure 1). In fact, Summerville, the newest site, had reducing conditions in all equilibrators at the 60-cm depth. At shallower depths in the substrate, all sites showed an increase in oxidizing conditions nearer the surface (the suspension changed color in the equilibrator). Filson 1 and Sommerville substrates were almost completely oxidized at the 30-cm depth.

Figure 2 shows that at Howe Bridge much of the flow through the compost and limestone occurred above the pipe near the bottom of the figure, and this was also indicated by observations of greater flow coming from that pipe’s outlet. The Filson 1 SAPS (Figure 3) appeared to have more uniform flow through the system since there were no areas of distinct flow through the system. Again, the compost in this system was more compact and 50% of the water flowed out of the wetland over the spillway. The Sommerville SAPS was shown by dye tests to have preferential flow in the center of the wetland through a thinner compost layer. This finding was substantiated by the equilibrators (Figure 4) showing that the majority of flow was moving downward in the middle of the wetland.

Water chemistry at Howe Bridge from 1992 to 1993 (one year after construction) was obtained and compared to our data (six years after construction) to evaluate treatment efficiencies (Table 5). In the year following construction (1992-1993), acidity was reduced at an average of 45% compared to 39% in 1996-1997. The acid removal rate was 19.4 g/m²/day in 1992-1993 (this value conflicts with acid removal values in Kepler and McCleary 1994 and stated earlier) versus 17.4 g/m²/day in 1996-1997. Iron removal in 1992-1993 was 46% and the 1996-1997 iron removal rate was actually a little higher at 60%. Calcium increases were about the same during these two sampling periods. It appears that the Howe Bridge treatment efficiency has changed very little over the past six years.

The Howe Bridge SAPS has accumulated about 45 cm of Fe floc on the surface of the organic matter. The impact of this floc layer on treatment effectiveness is unclear. It may provide additional depth for enhanced reducing conditions at shallower depths, and it may cause additional surface area and exchange sites for dissolved metal attachment. But on the other hand, it may reduce or slow the downward movement of water through the substrate. The high percentage (55%) of equilibrators at Howe Bridge showing reducing conditions at 30- and 40-cm depths (Figure 1) promotes the idea that the floc layer promoted reducing conditions at shallow depths. The function and impact of accumulated floc in passive treatment systems should be studied further.

Conclusions and Recommendations

Acidity decreases between influent and effluent water for four SAPS varied between 54 and 129 mg/L (19 and 100%) in these four SAPS, Fe removal was between 0 and 129 mg/L (0 and 92%), and all systems increased in alkalinity and Ca concentrations. Acid removal rates were between 11 and 51 g/m²/day, which are all within the range of other SAPS reported in the literature. Treatment of water at the four sites varied because of flow rates through the system, influent water chemistry, water residence time, and system age. The McKinley SAPS showed AMD treatment to effluent limits because of low flow and low metal concentrations of influent water, a high ratio of limestone to acid per year, and a young age. The Howe Bridge SAPS demonstrated effective treatment for six years and no significant decline in acid and metal removal was seen since 1992. A thick compost layer enhanced reducing conditions, but compaction and uneven compost depths have caused short-circuiting and uneven water flow in some systems. SAPS are a relatively new passive treatment technique and continual monitoring of both influent and effluent water chemistry and metal removal mechanisms are needed.

From the data, a minimum of 50 to 60 cm of compost depth is needed to promote reducing conditions. The compost type is also important and should have the ability to decompose at a slow rate with the necessary carbon for the microbial community. A combination of mushroom compost and larger organic material, such as wood chips, may be an option. The larger material would decompose at a slower rate and provide organic material for a longer time period. The wood chips may also decrease compaction by encouraging particle separation and help to vary water flow paths.

Other ideas to increase the longevity of the compost may be to replenish the organic layer after two to three years. By adding even 5-10 cm of new material to the surface or by mixing this fresh material into the current layer, the organic layer may be reconditioned to perform in a way similar to when it was fresh. Laboratory studies have also shown that by decreasing flow, the organic layer may function after appearing nonfunctional (Stark et al. 1994). Flushing is another option, but the release of flushed materials from the system should be captured in settling ponds so the receiving stream is not inundated with metal-laden floc materials.

The piping system is another important component. By using only one or two pipes, preferential flow may occur. More pipes may cause water to move more uniformly downward through the compost and limestone, rather than moving toward one pipe within the substrate. Pipes at different levels in the limestone may also be helpful.

Water chemistry is important to treatment design. If Al is high, the system may be doomed to failure by premature clogging of void spaces in limestone and pipes. The flushing system as designed by Kepler and McCleary (1997) should be installed at all SAPS with high Al and used routinely to remove accumulated flocs. With high flows and high metals, more complicated designs may be needed that incorporate more treatment cells in a series with increased numbers of settling ponds. More complex systems are costly to build and a larger area is needed. So in these high acid and metal load situations, the cost effectiveness of passive systems may not be realized.

Achknowledgment

References

Clesceri, L.S., A.E. Greenberg, and A.D. Eaton. 1998. Standard methods for the examination of water and wastewater (20th Ed). American Public Health Association. Washington, DC.

Edenborn, H.M., L.A. Brickett, D.H. Dvorak, and S.L. Edenborn. 1993. Monitoring iron and manganese diagenesis in constrcuted wetlands with continuous gradient gels. In: International Biohydrometallurgy Sumposium, Proceedings of the Minerals, Metals, and Materials Society, August 22-25, 1993, Jackson Hole, WY, 287-297.

Faulkner, B.B., and J.G. Skousen. 1994. Treatment of acid mine drainage by passive treatment systems. In: Proceedings, International Land Reclamation and Mine Drainage Conference, April 24-29, 1994, US Bureau of Mines SP 06A-94, Pittsburgh, PA, 250-257.

Hedin, R.S., R.W. Nairn, and R.L.P. Kleinmann. 1994. Passive treatment of coal mine drainage. US, Bureau of Mines Information Circular 9388, Pittsburgh, PA.

Kepler, D.A., and E.C. McCleary. 1994. Successive alkalinity-producing systems (SAPS) for the treatment of acidic mine drainage. In: Proceedings, International Land Reclamation and Mine Drainage Conference, April 24-29, 1994, USDI, Bureau of Mines SP 06A-94. Pittsburgh, PA, 195-204.

Kepler, D.A., and E.C. McCleary. 1997. Passive aluminum treatment successes. In: Proceedings, Eighteenth West Virginia Surface Mine Drainage Task Force Symposium, April 15-16, 1997, Morgantown, WV.

Kleinmann, R.L.P. 1990. Acid mine drainage in the United States. In: Proceedings, First Midwestern Region Reclamation Conference, Southern Illinois University, Carbondale, IL.

Phipps, T., J. Fletcher, W. Fiske, and J. Skousen. 1991. A methodology for evaluating the costs of alternative AMD treatment systems. In: Proceedings, Twelfth Annual West Virginia Surface Mine Drainage Task Force Symposium, April 3-4, 1991, Morgantown, WV.

Skousen, J., J. Gorman, and P. Ziemkiewicz. 1999. Long term effects of acid mine drainage remediation projects on stream quality. In: Proceedings, Twenthieth Annual West Virginia Surface Mine Drainage Task Force Symposium, April 13-14, 1999, Morgantown, WV.

Stark, L., W. Wenerick, F. Williams, S.E. Stevens, and P.J. Wuest. 1994. Restoring the capacity of spent mushroom compost to treat coal mine drainage by reducing the inflow rate: a microcosm experiment. Water, Air, and Soil Pollution 75, 405-420.

Turner, D., and D. McCoy. 1990. Anoxic alkaline drain treatment system, a low cost acid mine drainage treatment alternative. In: Proceedings, 1990 National Symposium on Mining, University of Kentucky, Lexington, KY.

U.S. Environmental Protection Agency. 1995. Streams and fisheries impacted by acid mine drainage in Maryland, Ohio, Pennsylvania, Virginia, and West Virginia. Map produced by EPA Region III, Wheeling, WV.

Watzlaf, G.R., and R.S. Hedin. 1993. A method for predicting the alkalinity generated by anoxic limestone drains. In: Proceedings, Fourteenth Annual West Virginia Surface Mine Drainage Task Force Symposium, April 27-28, 1993, Morgantown, WV.

Watzlaf, G.R., J.W. Kleinhenz, J.O. Odoski, and R.S. Hedin. 1994. The performance of the Jennings Environmental Center anoxic limestone drain. In: Proceedings, International Land Reclamation and Mine Drainage Conference, April 24-29, 1994, USDI, Bureau of Mines Special Publication SP 06A-94, Pittsburgh, PA, 427.

Wieder, R.K. 1992. The Kentucky wetlands project: A field study to evaluate man-made wetlands for acid coal mine drainage treatment. Final Report to the U.S. Office of Surface Mining, Villanova University, Villanova, PA.

Table 1.

Table 4:

Table 5: