Drivers of tree demographic processes in forest fragments of the Brazilian Atlantic forest

In heterogeneous landscapes, such as the Brazilian Atlantic Forest, forest fragments are likely to undergo highly uncertain successional pathways because anthropogenic and environmental factors strongly affect demographic processes such as recruitment, growth, and mortality. We used data from 134 permanent plots (0.01-0.10 ha) across seven forest fragments in Minas Gerais, Brazil, to assess how anthropogenic variables and environmental factors drive demographic processes in Atlantic forest fragments. For each plot, we quantified anthropogenic variables related to land-use history (LUH) and estimated levels of fragmentation based on forest cover, forest fragment size, and distance to the forest edge. We also measured environmental factors related to topography (i. e., elevation and slope), soil chemical characteristics (i.e., soil pH in H2O, exchangeable cations, P content and soil organic matter), and climate (i.e., average annual precipitation and climatic water deficit [CWD]). We used plot basal area (BA) as a proxy for forest age. We quantified the effect of anthropogenic and environmental factors on relative recruitment, growth, and mortality rates using linear mixed-effects models (LMM). We found that anthropogenic variables (LUH and BA) were stronger predictors of relative recruitment (p-value < 0.001) than environmental variables. For relative growth, soil, elevation, and BA (p-value < 0.001) have the strongest effect. CWD (p-value < 0.001) was the most important predictor for mortality, suggesting that dry periods in-crease mortality rates in the studied fragments. Our findings suggest that the land-use history of individual fragments, combined with a higher frequency of drought events leading to increased mortality and turnover of individuals, will increase differences in demographic processes and successional pathways among fragments, thereby increasing the heterogeneity of these fragmented landscapes.