Roles of benthic algae in the structure, function, and assessment of stream ecosystems affected by acid mine drainage

Coal has played a critical role in the advancement of society and technology for centuries, especially since the industrial revolution. As new technologies emerged, coal became widely used in the steel making process as coke for blast furnaces and to generate electricity for homes and factories by the end of the 1800s (United States Department of Energy (US DOE) 2013). These developments, along with shifts in national policy, increased the demand and subsequent production of coal in the USA from the late 1800s through the twentieth century (Fig. 1). Despite this rampant growth, the first federal regulation of coal mining for environmental protection in the USA was not passed until the Surface Mining Control and Restoration Act of 1977, after >38.7 billion metric tons of coal had already been mined. Sparse regulation prior to 1977 has led to a legacy of environmental impacts, acid mine drainage (AMD), and degraded stream ecosystems due to abandoned underground and surface mines, piles of discarded coal wastes, and tailings. AMD results from the oxidation of sulfide minerals, primarily FeS₂, when coal wastes are exposed to air, water, and bacterial processes (for extended details of AMD formation processes, see United States Environmental Protection Agency (US EPA) 1994). This process generates sulfuric acid leading to streams with pH as low as 2 and high concentrations of soluble metals leached from soils, rocks, and coal refuse—mostly large amounts of Al, Fe, and Mn, but also Zn, Cu, As, and other toxic metals (US EPA 1994). These chemical stressors negatively affect stream organisms, which are also impacted by iron and aluminum hydroxide precipitation that smothers benthic habitats as pH increases above ~3.5 and 5.0 respectively (Younger et al. 2002, Harding and Boothroyd 2004).

Abandoned mine drainage impairs >12,000 km of streams (14% of all stream km) in the Appalachian region of the eastern USA (US EPA 1997), a distance equivalent to driving a lap around the USA from Boston to Atlanta, San Diego, Seattle, and back to Boston, or from Lisbon, Portugal to Beijing, China (estimated from Bing Maps; www.bing.com/maps). This legacy of mine drainage is a concern for many countries that have historically mined coal, such as the United Kingdom (Zheng et al. 2007), South Africa (Bell et al. 2001), Australia (Wright and Burgin 2009), New Zealand (Hogsden and Harding 2012), Korea (Lee et al. 2013), China (Wu et al. 2009), and Brazil (Freitas et al. 2011). Studies of metal mines around the world find similar stream impacts due to associated sulfide minerals, making them informative to studies in coal regions (Niyogi et al. 1999, Luís et al. 2009). Some currently permissible coal mining practices in the USA, such as mountaintop removal, continue to negatively impact stream ecosystems and human health (Palmer et al. 2010, Lindberg et al. 2011). Legacy effects of coal mining and AMD also have economic and social justice ramifications manifested as lower housing and land values, personal and community distress, lost tourism, polluted and reduced availability of water, and loss of ecosystem services, such as recreational opportunities, food, and biodiversity (Blakeney and Marshall 2009, Li et al. 2011, Mishra et al. 2012, Morrice and Colagiuri 2013).

To meet the goals of improving and sustaining aquatic resources that will benefit communities and downstream ecosystems, management strategies need to focus on understanding and restoring the fundamental ecological structure and function of AMD-affected streams. Ecological assessments are critical to informing management efforts and for measuring the effectiveness of restoration (Yoder and Rankin 1998, Karr and Yoder 2004). As the trophic base of primary production in streams, algal-dominated periphyton communities are particularly effective indicators of anthropogenic stressors. Measures of community structure and functions, which are vital to ecosystem processes like nutrient cycling and energy flow, are needed for assessing AMD impacts on streams and quantifying progress toward remediation. In addition, algae provide valuable ecological information that could be missed by other organism groups when assessing the full AMD impact gradient (Freund and Petty 2007, Smucker and Vis 2009). Here, we synthesize the state of knowledge regarding how AMD-associated stressors affect (i) algal communities and their use as ecological indicators, (ii) their functional roles in stream ecosystems, and ultimately, (iii) how these findings inform management decisions and evaluation of restoration effectiveness. This review is intended to inform ongoing basic and applied research and to highlight the value of using benthic algae of streams in a biomonitoring framework.

Patterns of Algal Communities and Their Use as Indicators

Algal community structure is the end result of each species' evolutionary history and traits (e.g., physiology, growth form, and life history) that enable it to disperse, colonize, reproduce, and persist under certain abiotic and biotic conditions (e.g., niche opportunities, competition, disease, disturbance, grazing, water quality, etc.). Communicating ecological results from these complex, species-rich algal communities in relevant terms is important for informing managers, agencies, and decision makers who need to act on well-supported scientific data (Stevenson 2014). Species diversity and biological metrics (e.g., percent tolerant, sensitive, or acidophilic taxa, certain taxonomic groups, functional guilds) are widely used tools in stream assessments and management decisions (Karr 1991, Reavie et al. 2010, Potapova and Carlisle 2011). Metrics can be used separately or combined into an index that summarizes stressor effects on components of community structure and valued ecological attributes (Hill et al. 2000a, Wang et al. 2005, Zalack et al. 2010, Schowe and Harding in press). Biomonitoring is crucial to assessment programs because it gives quantifiable ecological outcomes that provide context for determining and managing land use effects and water quality impacts to ecosystems.

Measures of diversity are widely used in biomonitoring programs, and although the relationship between algal diversity and human stressors can be murky at times (Stevenson and Pan 1999), a clear and consistent pattern has emerged with regards to AMD (Hogsden and Harding 2012). As the severity of AMD increases, total algal species richness and diversity decline, as do diversity of diatom, soft-bodied, and macroalgae groups individually (Verb and Vis 2000, 2001, 2005, Niyogi et al. 2002, Griffith et al. 2005, Bray et al. 2008, Kim et al. 2008, Smucker and Vis 2009). Communities are particularly species depauperate in streams with pH <3.5–4.0 (DeNicola 2000, Kim et al. 2008, Schowe et al. 2013, Smucker and Vis 2013). However, further research is needed to describe the shape of this decline in diversity and community turnover, which could be used to inform water quality criteria and management goals. Bray et al. (2008) found a linear loss of total algal taxa with greater AMD stress. Schowe et al. (2013) identified a threshold at pH ~3.5, below which diatom diversity remained extremely low. Although Kim et al. (2008) showed a linear fit, their scatter plot may indicate a threshold around pH 6.5–7.0, below which fewer than 15 diatom taxa per site were observed. Variation in patterns of diversity loss could be due to stressors other than pH at intermediate to low AMD severity or complex watershed settings with natural geologic variability or agricultural and urban stressors, as has been reported for macroinvertebrates (Merovich and Petty 2010, Merriam et al. 2011). As acidity and metals decrease, conductivity along with nutrients and alkalinity from other land uses and calcareous geology (if present) can become prominent factors affecting algal diversity.

While biodiversity is a common focus of assessment programs, the identity, abundance, and traits of algal taxa provide important context for evaluating the impacts of stressors on stream ecosystems and for developing assessment tools, such as metrics and indices, to be used in resource management. An emerging multivariate approach that measures the average similarity (e.g., Bray-Curtis coefficient) of a site's community composition (species and abundance) to algal communities at regional reference sites can effectively quantify stressor severity (Wang et al. 2005, Stevenson et al. 2008). This approach is related to the conceptual model of the biological condition gradient (Davies and Jackson 2006), in which community attributes and functions of high quality streams are lost as anthropogenic stressors increase (Fig. 2). Regional reference sites represent the expected community in least disturbed and best attainable conditions (Stoddard et al. 2006, Herlihy et al. 2008). Measuring the losses of reference site community structure and reference taxa have been particularly effective in assessments of AMD severity (Smucker and Vis 2009, Zalack et al. 2010, Schowe and Harding in press). Plotting various metrics against percent similarity to reference sites helps identify potential reasons for changes in community structure. These types of analyses can inform development of management plans and targets for restoration. Using other multivariate (e.g., Lavoie et al. 2006) and predictive model (e.g., Gevrey et al. 2004, Carlisle et al. 2008) approaches also may be worth exploring in AMD-affected streams.

Algal metrics and indices of biological condition identify how components of the community change and they can be used to summarize and communicate ecological effects to various stakeholders, such as watershed groups, regulatory agencies, and policymakers (Barbour et al. 1999, Herlihy et al. 2008). The relative abundances of acidophilic diatoms, which increase with decreasing pH, has been one of the most widely used and highly effective metrics for assessing AMD impact severity, largely due to its direct relationship to the predominant stressor of acidity (Hill et al. 2000a, 2003, Hamsher et al. 2004, de la Peña and Barreiro 2009, Zalack et al. 2010, Schowe and Harding in press). As AMD severity increases, acidophilic diatoms begin to flourish and acid sensitive taxa become overwhelmed by the elevated proton influx (H⁺ ions) and/or a taxon's inability to manage or reduce metal uptake across cellular membranes (Gross 2000, Novis and Harding 2007). Smucker and Vis (2010) identified a potential threshold near 75 mg CaCO₃ · L⁻¹ alkalinity, below which the relative abundance of acidophilic diatoms increased, but further research is needed to substantiate this observation. Knowledge of naturally acidic streams may also be important to consider if they occur in close proximity to AMD; otherwise, funds and efforts might be misdirected toward trying to improve a system that is naturally acidic (e.g., as a result of bogs or geology). Although this is uncommon in the USA, it can be an issue in other countries (Schowe et al. 2013). Naturally acidic streams are often dominated by taxa typical of low pH, but diatoms tend to be buffered against toxic effects of metals because of high concentrations of humic substances, and higher diversity can be sustained when compared to anthropogenically acidified streams (e.g., acid precipitation in Passy 2006).

Despite the effectiveness of the widely used acidophilic metric, additional metrics are useful in moderately AMD-affected streams, which can be difficult to classify (Verb and Vis 2000, Hamsher et al. 2004, Bray et al. 2008). Additional diatom metrics can be informative when other factors beyond pH become important, such as conductivity, nutrients, and metals (Hill et al. 2000a, 2003, Zalack et al. 2010). Metrics such as percent Cymbella (Wang et al. 2005), species richness or diversity, indicator species, and increased similarity to regional reference site communities can provide finer resolution of AMD impacts and overall stream conditions, because they respond to stressors other than just acidity (Smucker and Vis 2009, Zalack et al. 2010). In addition, while indices using genus level taxonomy can be effective, species level taxonomy can be more sensitive to moderate AMD impacts (Smucker and Vis 2009), as some genera have more species with a variety of ecological niches, such as Navicula and Nitzschia (Manoylov 2014). Indices developed specifically for AMD impacts can provide finer resolution of stressor severity than indices developed for other purposes or “general impairment” (Zalack et al. 2010, Smucker and Vis 2013). Scores from non-AMD specific indices could be affected by including metrics that respond to impairment in the direction opposite of that expected. As an example, relative abundances of cyanobacteria and eutraphentic taxa are expected to increase with greater impairment, but they decrease as AMD severity increases, indicating higher quality waters (Hamsher et al. 2004). These “misbehaving” metrics would likely be useful in complex settings (e.g., multiple drivers of stressors, such as agriculture or urbanization, or geology) and highlight the need for careful consideration of indices, metric scores, and other ecological information when making subsequent decisions.

The presence and percent of deformed diatom valves, which are strongly associated with metals, but can also be related to light (Schmid 1979), acidity, and silica availability (McFarland et al. 1997), might have potential for use as an ecological indicator (Yang and Duthie 1993, Sgro et al. 2007, Smith and Manoylov 2007, Kim et al. 2008, Morin et al. 2008a,b, Falasco et al. 2009), but this is often reported merely as an observation rather than as established empirical relationships with concentrations and types of metals. The extent to which deformities affect ecosystem processes or populations and communities is also unknown. Practical use of this metric in a biomonitoring framework could be confounded (i) by the amount of extracellular polymeric substances, which reduce the bioavailability and toxicity of metals (Gold et al. 2003), (ii) high phosphorus availability that allow algae to form polyphosphate bodies that reduce toxic effects by binding metals (Jensen et al. 1982, Twiss and Nalewajko 1992), and (iii) when communities are dominated by taxa less susceptible to deformities because of physiological mechanisms or their ability to produce phytochelatins (Falasco et al. 2009).

A growing body of research continues to find ecological and ecosystem simplification as a result of increasing anthropogenic stressors, such as those associated with AMD. Increased severity of AMD can reduce spatial heterogeneity of diatom community structure from the stream reach (Hollingsworth and Vis 2010) to the watershed scale (Smucker and Vis 2011a). Chemical stressors associated with AMD may override benefits of physical habitat heterogeneity for diatom diversity (Smucker and Vis 2011b). Increased AMD severity may also simplify the successional process of periphyton, which could affect the resilience and recovery of communities to pulsed disturbances, such as high flow events (Smucker and Vis 2013). As mentioned previously, community structure (richness and diversity) is greatly simplified as AMD severity increases (Hill et al. 2000b, Bray et al. 2008). While diatom communities in moderately AMD-impaired and restored streams can be temporally variable, temporal and seasonal turnover of diatom species and community structure tends to become less dynamic as AMD severity increases (Verb and Vis 2000, Bray et al. 2008, Smucker and Vis 2011c). Temporal variation in metric and index scores is reduced in AMD-affected streams compared to reference streams, another expression of simplified communities (Smucker and Vis 2011c, 2013). The increasing simplification of diatom community dynamics as AMD severity increases likely results from stressors being so severe that they reduce niche opportunities and override the effects of other abiotic and biotic factors important to structuring diatom communities at moderate to non-AMD affected sites, such as light, nutrients, physical habitat, and disturbance regimes (Fig. 2). These losses in algal community structure and dynamics are parts of an overall reduction in ecosystem complexity and function that has been observed in stream food webs and biogeochemistry (Bott et al. 2012, Hogsden and Harding 2012).

Effects of AMD on Functional Roles of Periphyton

The ecosystem functions and services of algal communities are end products of the physiologies and life strategies of taxa present in ecological communities. Developing, testing, and implementing functional indicators in assessments and management strategies is receiving increased interest, but still remains in its infancy and underutilized relative to the widespread use of metrics and indices based on community structure (Brooks et al. 2002, Nelson et al. 2013). A notable exception is phosphatase activity as a metric in a periphyton index of biotic integrity (Hill et al. 2000a). Yet, characterizing how stressors affect functional indicators will likely provide insights into the extent of ecosystem alteration caused by AMD (Niyogi et al. 2002) and identify impairment in ways that might not be evident from structural assessments alone (Crossey and La Point 1988, Brooks et al. 2002). Incorporating functional measures will result in a fuller understanding of the impacts of AMD on the ecosystem scale and can help guide management decisions (Young et al. 2008).

Biomass and rates of primary production strongly respond to AMD, which has direct and indirect effects on periphyton function (Niyogi et al. 1999). In studies focused solely on pH, benthic algal biomass can increase (Muller 1980, Mulholland et al. 1986) or decrease (Kinross et al. 1993) with increasing acidity. However, metals add complexity to the biomass response to AMD (Verb and Vis 2000), which can create high biomass at severe AMD sites, a decrease in biomass as metals precipitate at intermediate pH, and then an increase as AMD conditions continue to improve (Fig. 2; Smucker and Vis 2011d). Iron and aluminum precipitation at ~pH 3.5 and 5, respectively, substantially reduce algal biomass (Niyogi et al. 1999, 2002, Hill et al. 2000b, Smucker and Vis 2011d, Bott et al. 2012). Periphyton communities in streams experiencing metal hydroxide precipitation are directly shaded by metal flocculation, prevented from adhering to the substratum, and face perpetual metal precipitation (Niyogi et al. 2002), which reduces biomass and primary productivity (Hill et al. 1997, Bott et al. 2012, DeNicola et al. 2012).

Acidified streams without heavy metal precipitation are often dominated by a few tolerant green algae, especially at high light availability, and diatom taxa capable of generating substantial biomass (Mulholland et al. 1986, Niyogi et al. 2002, Bott et al. 2012). The increase in primary producer biomass and benthic production are likely further enhanced indirectly by reduced grazing pressure rather than a stimulation of primary production by acidic conditions (Mulholland et al. 1986, Niyogi et al. 2002, DeNicola et al. 2012). Furthermore, stream acidification alters primary consumer assemblages, increases their reliance on allochthonous food resources, and reduces food web complexity (Layer et al. 2013, Hogsden and Harding 2014). Reduced palatability of algal taxa, such as Mougeotia, in acidified streams can shift the diets of macroinvertebrate consumers toward coarse detritus (Junger and Planas 1993). Chironomidae and Leuctridae ¹³C and ¹⁵N signatures have shown either a shift in diet from periphyton to allochthonous food or within-family species replacements by taxa having heavier detritus diets (Traister et al. 2013).

Beyond the physical stress of precipitates, AMD also alters the chemical environment that affects periphyton function, and despite greater biomass at low pH than at intermediate pH, benthic biomass typically does not reach levels observed in non-AMD impacted streams (Verb and Vis 2000, Sabater et al. 2003, Simmons et al. 2005, Smucker and Vis 2011d, DeNicola et al. 2012). A decrease in pH can inhibit photosynthesis by reducing the pool of total inorganic carbon (Mulholland et al. 1986, Turner et al. 1994, Planas 1996), and could be one possible explanation for reductions in productivity and biomass in AMD impaired streams. Long filaments of some green algae, such as Klebsormidium and Mougeotia, can be abundant in AMD streams (Stevens et al. 2001, Sabater et al. 2003), potentially due in part to being strong competitors for low DIC in acidic environments (Vinebrooke 1996). In addition to reductions in the carbon pool, elevated concentrations of metals regularly found in AMD streams can negatively affect primary production by forcing algae to divert resources toward reducing their toxic effects (Genter 1996).

AMD can greatly alter nutrient dynamics in streams. Extracellular enzyme activities have been successfully used as indicators of nutrient limitation in aquatic environments (Smucker and Vis 2011d, Bott et al. 2012). Rates of phosphorus turnover through extracellular enzyme activity are elevated in streams experiencing metal hydroxide precipitation (Mulholland et al. 1986, Smucker and Vis 2011d, Bott et al. 2012) as adsorption of phosphorus to metal precipitates reduces the available phosphorus pool, potentially limiting algal productivity and increasing the need to access organically-bound P (Moore and Miller 1994). Acidification also inhibits nitrogen uptake rates and nitrification in AMD-affected streams (Niyogi et al. 2003, Bott et al. 2012).

The functional roles of periphytic heterotrophs are affected by AMD stress, which can lead to altered metabolic rates. Microbial respiration is notably slowed with increasing metal deposition (Hill et al. 1997, Niyogi et al. 2001, Young et al. 2008). Algae are important sources of carbon for heterotrophic taxa in periphyton communities, and when algal biomass is reduced, the demand for allochthonous carbon increases in AMD impacted streams, as evident in carbon acquiring enzymes (Smucker and Vis 2011d). Heterotrophy tends to increase as AMD severity increases, as indicated by the ratio of gross primary production to community respiration (Crossey and La Point 1988, DeNicola et al. 2012). Niyogi et al. (2013) found that absolute microbial respiration and leaf breakdown rates were suppressed in streams with metal precipitates, but then increased to rates similar to reference streams when pH was <3. Hogsden and Harding (2013) found that leaf litter breakdown and microbial respiration were suppressed in AMD-impacted streams when compared to circumneutral streams. No significant differences existed between AMD-impacted and naturally acidic streams, and reductions in leaf litter breakdown rates were driven primarily by the extirpation of shredder invertebrates (Hogsden and Harding 2013). Increased heterotrophy also could be affected by reduced photosynthetic efficiency due to metal stress (Tlili et al. 2011, Luís et al. 2013), greater senescence of algal biomass as AMD severity increases, and less severe effects of AMD on community respiration processes (DeNicola et al. 2012). In addition, experimental microcosms suggest that losses in species richness can decrease algal biomass and rates of oxygen production in aquatic ecosystems, which also affects the amount of carbon sequestered and stored as biomass (Power and Cardinale 2009).

Assessment of AMD Restoration Using Periphyton

Increased awareness of AMD's negative effects and consequences for society contributed to changes in policy and governance in the 1970s. Research by Verb and Vis (2000) showed that new regulations led to noticeable improvements in water quality and diatom condition following new laws requiring mining companies to regrade mine spoil, replace topsoil, and revegetate affected areas (Ohio Revised Code 1513, 1972). Diatom communities were even more similar to reference site communities following the additional requirements of proper disposal of mine tailings and sludge and the treatment of mine water discharge (Surface Mining Control and Reclamation Act, 1977). While these new regulations were critical for reducing impacts to streams going forward, substantial efforts were, and continue to be, needed to restore stream ecosystems left impacted by preregulation mining.

Pollution by AMD is a continuous and chronic problem that can take centuries to improve without human intervention (Banwart and Malmstrom 2001, Raymond and Oh 2009). Therefore, technologies have been developed to ameliorate AMD effects on downstream water chemistry, physical environment and subsequently the aquatic biota. These remediation strategies are placed into two broad categories, passive and active treatment. Passive treatments are employed when the amount of acidity to be neutralized is relatively low and are typically one pond/wetland or a series of ponds, such as successive alkaline-producing systems, limestone or steel slag leach beds. These systems add alkalinity and retain metal precipitates, resulting in downstream water with less acidity and sediments (Skousen et al. 2000). Active treatment systems are utilized when the amount of acidity would quickly overwhelm a passive treatment system either producing too much sediment and/or acidity. Dosers, which consist of a silo of alkaline material, typically calcium oxide, and a mechanism to add a prescribed amount of this material into a treatment channel or directly into the stream, are often used to treat AMD. Both active and passive treatment systems have a considerable cost either for construction and/or maintenance (Skousen et al. 2000). Therefore, agencies and watershed groups typically monitor these systems to determine their efficacy in producing the desired effect on stream water quality and biotic recovery.

Remediation of AMD using both active and passive systems is relatively recent. These systems are normally monitored with periodic water chemistry testing and indices measuring assemblage diversity for fish and macroinvertebrates (NPS 2011, Kruse et al. 2012). Periphyton has been less often employed, but the few studies available suggest periphyton structure may provide valuable insights into stream biotic diversity and recovery potential. Study of periphytic diatom structure has shown a relationship of the AMD-diatom index of biotic integrity (DIBI) scores with efficacy of remediation efforts (Pool et al. 2013). In this study, the researchers found an increase in scores indicating higher quality communities downstream of a doser in an AMD-contaminated watershed with no other substantial downstream AMD inputs (Fig. 3A). However, in a watershed with a doser and passive treatment sites, yet additional downstream AMD sources, the index indicated continued impairment in areas of the stream (Fig. 3B). Near the doser where precipitates are likely a major stressor, index scores indicated poor condition; further downstream scores improved to fair conditions, then decreased to poor after considerable acid input (Pool et al. 2013). These data emphasized that the diatom index data very effectively tracked AMD inputs and indicated a need for further remediation efforts.

A study of diatom diversity downstream of reclamation efforts focused on the ability of stream reaches to support a reference diatom community through a tile transplant study (Gray and Vis 2013). This study concluded that recovery was site specific with the sites farthest upstream and downstream of the doser showing signs of improvement, while the sites in between were most impacted and failed to reestablish reference diatom communities (Gray and Vis 2013). In Smucker and Vis (2009), NaOH treatment of AMD from an underground mine pool greatly reduced acidity, but diatom metrics still indicated impaired conditions downstream likely due to extremely high conductivities that supported communities more typical of brackish conditions than freshwater (Smucker et al. 2008). Lastly, recovery from the physical effects of AMD sedimentation was mimicked and studied by placing iron oxide armored rocks from an AMD stream into an unimpacted stream (DeNicola and Stapleton 2002). Results showed no difference in diatom diversity or density between the rocks previously coated with AMD precipitates and controls suggesting if remediation efforts could improve the water quality and remove precipitates, the hard surfaces would be suitable for recolonization (DeNicola and Stapleton 2002).

Beyond the documentation of periphyton community diversity, recent studies are emphasizing the measurement of functional metrics to gage remediation and provide more insights into stream function and alteration to ecosystem services. DeNicola et al. (2012), studying headwater streams in Pennsylvania, and Simmons et al. (2005), studying streams in West Virginia that were remediated with various types of passive treatment systems, showed a significant increase in algal biomass as measured by chl a downstream of treatment sites as compared to AMD-impacted locations. In Pennsylvania, the increase in chl a was still less than the reference sites, indicating incomplete recovery, but Simmons et al. (2005) reported only slightly lower chl a in treated streams that was not statistically different from reference sites. DeNicola et al. (2012) showed that net primary productivity was also negatively affected by AMD and improved with downstream reclamation.

Bott et al. (2012) measured numerous metrics associated with periphyton function including biomass, gross primary productivity (GPP), extracellular enzyme activity and ecosystem respiration over a 24-hour period (ER₂₄) in AMD impacted, remediated and reference streams in both bituminous and anthracite coal regions of Pennsylvania. Remediation seemed to improve algal growth and ecosystem respiration to levels comparable with reference sites. However, GPP, ER₂₄, and measures of nitrogen cycling rates were still lower than those in reference streams, indicating that restoration was incomplete (Bott et al. 2012). In addition, these researchers found that elevated enzyme activities suggested that the stream communities in the reclaimed anthracite area were still more stressed than those in the bituminous region.

Algae also may provide services that enhance restoration. As reviewed by Das et al. (2009), algae can be directly and indirectly involved in alkalinity generation through nitrate assimilation and production of extracellular polysaccharides that serve as a carbon source for sulfate reducing bacteria if treatment systems include anoxic zones. Various algae can absorb or adsorb metals such as Fe and Mn (Das et al. 2009, references therein). For example, green algae-cyanobacterial mats can enhance Mn precipitation in oxidation ponds (Phillips et al. 1995). More research examining naturally occurring algal communities in passive treatment systems is needed to fully understand their role and potential to enhance remediation.

The Future of AMD and Algal Assessments

Financial and technical limitations are the two biggest barriers to ongoing restoration. Currently, the majority of funding for restoration projects comes from the Abandoned Mine Land (AML) Reclamation program, which is funded by fees placed on coal production (currently $0.286 per metric ton of surface-mined coal and $0.125 per metric ton of underground-mined coal), and to a lesser degree Clean Water Act, EPA Superfund, and state funds. While the AML provides a substantial source of funding, the Office of Surface Mining reported that “of the $8.2 billion of high priority coal related problems in the AML inventory, $6.6 billion, 80%, have yet to be reclaimed” (Costello 2003).

Viewing AMD wastes as a valued resource could help spur further restoration efforts and technological advances. Among many possible uses, metals in particular have the potential to be used commercially as pigments in paints and glazes, as corrosion inhibitors, and as raw materials put back into the manufacturing process (Hedin 2003, PA DEP 2013). In addition, harvested iron rich, dried AMD sludge, which is difficult and costly to dispose, shows promise to one day be a low cost adsorbent that efficiently removes phosphorus from agricultural and municipal wastewater (Sibrell and Tucker 2012). This technology would improve streams elsewhere and create economic motivation for restoring AMD-affected streams. Further research and development is needed (Costello 2003, PA DEP 2013), but technological advances, innovation, and making the harvesting and uses of AMD wastes economically viable could create incentives for private investors, inject more money into restoration, and create jobs. Partnerships among diverse stakeholders, such as private land owners, local businesses, watershed and nonprofit groups, industry, academia, and government agencies will be critical to leveraging multiple resources and interests that work toward the common goal of improving the social, economic, and environmental wellbeing of regions affected by the legacy of unregulated coal mining.

Algal assemblages are not only highly effective indicators that can benefit monitoring programs assessing AMD severity and the effectiveness of management practices, but they are also critical to ecosystem processes. Algae connect physical and chemical habitat conditions to primary production in streams. How communities are shaped by species' traits, niche opportunities, and their response to stressors have direct ecosystem consequences for biogeochemical processes and the uptake and retention of nutrients and carbon (Cardinale et al. 2011, Hooper et al. 2012). Further development and testing of functional metrics will likely complement and improve stream assessments based solely on traditional measures of community structure, but better understanding of the links between community structure, diversity, and outcomes of ecosystem processes is imperative. While many restoration actions are effective, complicating factors can arise, and adaptive management will be needed to create and sustain the desired states of ecosystems. Ongoing monitoring, analysis, and re-evaluation of potential strategies will aid this process. Examining a suite of metrics and community responses should help inform management efforts that seek to restore valued ecological attributes, aquatic life uses, and ecosystem services that depend on re-establishing the structure and function of ecological communities in ways that resemble those in best attainable regional reference streams (Fig. 2).

Comments on an earlier draft from Jan Stevenson and Kalina Manoylov are greatly appreciated. We also thank Greg Pond, Anne Kuhn, Bryan Milstead, Peg Pelletier, Glen Thursby, Dean DeNicola, and an anonymous reviewer for thoughtful comments on this manuscript. This submission is ORD Tracking Number ORD-004074. This manuscript has been reviewed by the Atlantic Ecology Division and approved for publication. Approval does not signify that contents necessarily reflect the views and policies of the Agency.