A. Change in sapling height across species and treatments (two-way ANCOVA). The species × treatment interaction was not significant. B. Watering treatment significantly affected growth (F _2,162 = 48.46, p < 0.001); saplings in the AMD water grew less than those in the partially treated water or the control (Tukey-adjusted p < 0.001 indicated by *); growth in the partially treated water did not differ from the control (Tukey-adjusted p = 0.070). Bars in A and B are raw means (±1 SE). C. ANCOVA-adjusted mean growth (± 1 SE) by species. Regardless of watering treatment, growth differed among species (F _4,162 = 5.25, p < 0.001); this was driven primarily by the slower growth of tulip poplar (Tukey-adjusted comparisons). Bars sharing a letter do not differ (Tukey-adjusted p > 0.05).

Acid mine drainage (AMD), generated primarily by pyrite oxidation, produces acidic waters enriched in dissolved metals (Rose and Cravotta 1998). In streams affected by mine drainage, low pH, metal toxicity, and iron and aluminum hydroxide precipitates degrade aquatic habitat and reduce the abundance and diversity of macroinvertebrates and fish, often eliminating sensitive taxa (Cherry et al. 2001; Hogsden and Harding 2012). These impacts can extend into adjacent riparian habitats where reduced emergence of aquatic insects limits food resources for terrestrial species, such as spiders (Paetzold et al. 2011; Schmidt et al. 2013). More broadly, coal surface mining can alter terrestrial ecosystems by limiting forest establishment and recovery (Zipper et al. 2011).

Passive treatment systems are constructed features, such as aerobic wetlands, limestone drains, and vertical flow wetlands, that improve water quality using gravity flow and natural chemical and biological processes rather than continuous chemical additions (Skousen et al. 2017). The passive system examined here consists of an aerobic wetland followed by three vertical flow wetlands (McGuire et al. 2025). In the aerobic wetland, iron oxidizes and settles out as iron oxyhydroxides. In vertical flow wetlands, water passes through an organic layer (often spent mushroom compost) that consumes oxygen and supports sulfate-reducing bacteria, and then through underlying limestone, which adds alkalinity, raises pH, and causes aluminum to precipitate (Skousen et al. 2017). Treatment success is typically judged by changes in pH, alkalinity, and dissolved metals (McGuire et al. 2025). Studies that use biological indicators of water quality often focus on benthic macroinvertebrates, measuring changes in taxon richness, abundance, or the return of metal-sensitive taxa (e.g. Kotalik et al. 2023; Dean et al. 2025), while plant responses are less frequently measured (but see Wu et al. 2022). Because reforestation is critical for long-term watershed recovery, tree growth offers a direct test of whether treated water can support vegetation.

Our study site is located in the Shamokin Creek watershed, which lies within the Western Middle Anthracite coalfield of Pennsylvania, where numerous abandoned mine discharges affect water quality (Cravotta and Kirby 2003). The region of the watershed where this passive system is located is primarily forested, with little agricultural land (Cravotta and Kirby 2003). At the study site, a portion of the AMD discharge is diverted through the treatment system before entering the creek, allowing us to collect both untreated influent and partially treated water exiting the system.

We compared growth of five native Pennsylvania tree species irrigated with untreated AMD, partially treated water from the system, or tap water under greenhouse conditions. Species were selected to represent common native hardwoods with varying tolerance to AMD-associated conditions (see Ashby and Kolar 1998; Groninger et al. 2006; Evans et al. 2013; Nkongolo et al. 2017; Zhang et al. 2022), with black locust commonly used in reclamation efforts (Ashby et al. 1985). We hypothesized that untreated AMD water would reduce sapling growth relative to partially treated and control water.

Sapling growth responded strongly to watering treatment, with no evidence of a species × treatment interaction ( Fig. 1A ; F _8,162 = 1.02, p = 0.422), indicating a consistent response across species. Growth differed significantly among watering treatments ( Fig. 1B ; F _2,162 = 48.46, p < 0.001), with saplings irrigated with untreated AMD (pre-passive system) exhibiting reduced growth relative to those receiving partially treated water (post-passive system) or tap water (Tukey-adjusted p < 0.001). Growth in the post-passive treatment did not differ from the control (Tukey-adjusted p = 0.070), indicating that partial treatment largely removed the growth reduction caused by untreated AMD. Across species, growth also varied ( Fig. 1C ; F _4,162 = 5.25, p < 0.001), driven mainly by reduced growth in tulip poplar relative to black locust (Tukey-adjusted p = 0.017) and pin oak (Tukey-adjusted p < 0.001). Initial height was a significant covariate (F _1,162 = 9.61, p = 0.002): taller saplings grew slightly less, and this relationship did not differ among species (species × initial height: F _4,156 = 0.75, p = 0.561).

Water chemistry data collected by McGuire et al. (2025) at the same site in November 2023 compare influent (pre-passive treatment) to water exiting the final north standpipe (post-passive treatment), providing context for growth responses in our study. Water collected from the final standpipe had a higher pH (4.1 to 5.8) and substantially lower dissolved aluminum (Al; 2.54 to 0.11 mg/L), a metal whose solubility and phytotoxicity are enhanced under acidic conditions (Ur Rahman et al. 2024). Aluminum is known to inhibit root elongation, shoot growth, and nutrient uptake (Kochian 1995; Ur Rahman et al. 2024); its reduction in post-passive treatment water may have alleviated a key limitation on plant growth. Dissolved iron (Fe) concentrations were also reduced following treatment (14.72 to 3.20 mg/L), which may have further contributed to improved growth by decreasing metal stress and limiting iron precipitation on root surfaces. Concentrations of nickel (Ni) and cobalt (Co) were only partially reduced (83.0 to 46.9 µg/L and 55.1 to 33.1 µg/L, respectively), zinc (Zn) declined only modestly (106 to 90 µg/L), and manganese (Mn) increased slightly (2.53 to 2.84 mg/L). The improvement in growth following passive treatment was consistent with the reduced acidity and dissolved aluminum concentrations reported at this site, although water chemistry was not measured during the greenhouse experiment.

Together, these results indicate that partial treatment primarily improves plant performance by mitigating some of the chemical stressors associated with AMD, particularly low pH and elevated aluminum. Remaining metals did not detectably limit plant growth during the timeframe of this study. This finding suggests that passive treatment can reduce the adverse effects of AMD on the growth of native tree saplings, although responses may vary with water chemistry, plant species, and exposure duration.