Edges as hotspots and drivers of forest cover change in a tropical landscape

Forest cover changes are influenced by socioeconomic and biophysical dynamics. However, the relative importance of both social and biophysical factors to forest loss and gain remains in debate. Previous research has focused mainly in identifying determinants of deforestation, while less is known regarding forest regrowth, especially in tropical forests. Here we investigated the spatial pattern and determinants of forest cover changes within a landscape which is representative of the biophysical and socioeconomic conditions found in the Atlantic Forest, and also presents an exceptional conservation value within this global biodiversity hotspot. Events of deforestation and regrowth were identified by comparing land cover maps from aerial images taken in 1975 and 2010. We controlled the spatial autocorrelation among events and used a machine learning algorithm and a selection model approach to build linear models explaining forest loss and gain. Main determinants of both forest loss and gain were biophysical factors, although some of they may also reflect socioeconomic underlying processes. Distance from Strictly Protected Areas (SPAs) was the single socioeconomic determinant important to forest change, in loss events. Both deforestation and forest regrowth were strongly associated with forest edges. Deforestation occurred more intensively on concave or flat areas, low solar radiation conditions, near forests, and far from SPAs. Diversely, forest regrowth occurred mostly in high slopes, low elevation areas, high solar radiation and near to forest edges. Our findings reinforce the important role of topography and protected area on forest change, and highlight edge zones are “hotspots” of forest cover dynamics. .