Acidic atmospheric deposition and soil acidification have been shown to affect health, growth, and regeneration of important forest tree species in the northeastern United States (e.g. sugar maple, red spruce), but their effects on forest understory plant community composition and diversity are not well understood. Biodiversity is positively associated with ecosystem stability and most of the plant biodiversity of temperate forest ecosystems is concentrated in the understory. We studied how forest understory vegetation composition and richness varied across a well-defined gradient in acidic atmospheric deposition (S, N) and soil acidification in the Adirondack Mountains of the northeastern United States. We surveyed overstory and understory species composition, plant community richness, soils, and light in northern hardwood forests on 50 plots (20 × 50 m) in 20 small watersheds. Differences in understory community composition across watersheds were analyzed using Non-metric Multidimensional Scaling. Ordinations were compared to environmental drivers including soil nutrients and pH, historical N and S deposition, topography, soil moisture, and canopy cover. We used a systematic process to build OLS regression models predicting species richness with up to two of these drivers. We performed diagnostics to evaluate adherence to model assumptions and ranked resultant models by AICc and Adjusted-R2.
Our NMS ordination was significant (p = 0.002; stress = 11.23) with axis 1 explaining 63% of variation in community composition. Species richness was positively correlated to axis 1 (Pearson r = 0.60) as were calciphilic species (sugar maple, r = 0.72). Soil variables associated with these species and low soil acidity were also positively correlated with axis 1 (% base saturation, r = 0.87, exchangeable Ca, r = 0.94, pH, r = 0.90). Conversely, 3 year averages of atmospheric deposition of S and N preceding soil sampling were negatively correlated with axis 1 (r < -0.52) and directly so with % base saturation (Spearman r < -0.73, p < 0.001). In the best regression model, soil pH and C:N positively predicted species richness (model p < 0.001, Adjusted-R2 = 0.60). A single variable of exchangeable Ca + Mg positively predicted richness in a follow-up model (p < 0.001, Adjusted-R2 = 0.52). We conclude that soil nutrients and pH in our study area influence understory community composition and richness and are negatively associated with acid deposition. As deposition has decreased and soils may be recovering, understory community composition may begin to change and richness to increase in Adirondack hardwood forests.