Prioritizing landscape connectivity of a tropical forest biodiversity hotspot in global change scenario

Abstract The probability that a propagule reaches, establishes and persists in a certain site is affected by the distance and quality of the environment. Fragmented landscapes promote the isolation of forests surrounded by a matrix that hinders or impedes the movement of species, affecting their distribution and threatening their conservation. Studies of landscape connectivity are essential to provide information for ecological conservation planning. Based on the classification of the landcover of the Rio Doce Basin - RDB and on the circuit theory, we used a habitat / non-habitat approach to assess the connectivity of the RDB to tree species. We built six resistance surface models based on habitat and non-habitat areas, using the GIS plug-in and Linkage Mapper to generate least cost paths maps. Three models explained the Jaccard similarity index matrix of 78 fragments used as a proxy of connectivity to test the models by GLMs, and one out the three was the simplest and most parsimonious. The map generated by the chosen model showed that the RDB is very fragmented but still has well-connected regions. The west to northwestern and southeastern portions of the RDB are well-connected and demand conservation of remaining fragments as well as the creation of reserves, while the center-north, east, and the far southwest of the basin are regions with greater resistance to connectivity as a result of anthropic pressures that reduced and fragmented the forests, requiring intervention through restoration projects to re-establish landscape connectivity.