It remains unclear how the multitude of species observed in forest communities coexistent over time. Species stability could be maintained by stochastic processes operating over macroecological scales, or through the partitioning of complex multidimensional niche-space. We explore the latter hypothesis by tracking seedling demography over four years in southwestern Pennsylvania. We ask 1) are seedling density, survival, and growth dependent on environmental and biotic processes, and are there differences among species in these relationships? 2) are there obvious interactive effects among these suites of variables that are influencing the regeneration layer in this forest? And 3) do seedling abundances reflect adult abundances, or are future compositional shifts like to occur? We established 208 1×10 m belt transects to sample dynamics of the understory seedling layer. From 2013 to 2016, we tagged approximately 15,000 individuals ≥7 cm tall and <1 cm DBH, and collected soil cores, hemispherical photographs, topographic data, and microhabitat information. We also erected exclusionary fences around plots to quantify the effects of white-tailed deer herbivory, and quantified neighborhood metrics to test conspecific vs. heterospecific density effects. We then used generalized linear models to test for the importance of these variables on key demographic attributes of the seedling layer.
Results/Conclusions
Microhabitat variables had the most important influence on seedling density and survival. Rock and fern cover in particular had a strong negative influence on understory seedling density (P < 0.001). Survival over four years was 6% greater in fenced versus unfenced plots, indicating an important effect of white-tailed deer herbivory. We also found conspecific seedling neighbor density to have a negative impact on survival (odds ratio = .97, P < 0.001), while heterospecific seedling density had a positive influence on survival (odds ratio = 1.1, P < 0.001), although both effects were quite small. In general, understory species were positively correlated with adult abundances, although these correlations were often weak (R2 < 10%), suggesting a likely future change in composition. Overall, our results point to multiple interacting influences on the demography of understory seedlings. Important and consistent differences observed across species indicates that temperate tree species may partition their environment in complex and surprising ways. Further research on the impacts of environmental variation on seedling distribution, growth, and survival will be necessary to fully assess how dynamics at early life stages influence future changes in composition and diversity.