Connectivity value of Atlantic forest fragments: pathways towards enhancing biodiversity conservation

An accurate combination of remote sensing-based techniques with predictive modeling is critical to evaluate the effects of fragmentation on forest fragments’ connectivity. Our study aimed to classify the landscape and evaluate its ecology in a highly fragmented region within the Atlantic Forest. We combined satellite image-based data, and spatial data of a region in Southern Brazil with Random Forest (RF) supervised algorithms. Using this combination, we defined the land use and native forest classes and estimated landscape metrics. Overall, native forest fragments varied widely, especially regarding their areas (ha), shape, and the number of core areas. Among all fragments, the one in the Rio Doce State Forest Park (PERD) holds the largest area, most regular shape, and the highest number of core habitats in the study region. However, most (64.4