Modelagem de distribuição potencial de espécies endémicas no Centro de Endemismo do Rovuma

Abstract

In the Rovuma Centre of Endemism (CER), changes in natural habitats caused by human activity, such as agriculture, mining and extraction of timber/forest resources, are driving deforestation and habitat fragmentation. These are considered the main threats to endemic species, such as Blepharis dunensis, Brachystegia oblonga and Scorodophloeus torrei. According to the IUCN Red List, these three species are in the threat categories: Critically Endangered -CR (Brachystegia oblonga) and Endangered-EN (Scorodophloeus torrei and Blepharis dunensis) and were included in these categories due to their limited distribution and small populations. This study aimed to: Assess the influence of socio-environmental conditions on the potential distribution of endemic species in the CER and identify areas of high environmental suitability for the species, in order to assist conservation actions. The MaxEnt species distribution modeling algorithm was used to relate the occurrence of species to socio-environmental variables that reflect different conditions of temperature, precipitation, and seasonality. The modeling results made it possible to identify environmental variables with a strong influence on the distribution of species, such as: maximum temperature of the hottest month (bio_05), soil density, soil texture, and soil pH. The resulting maps predict areas of high environmental suitability in the coastal strip of the CER, specifically in the districts of Moma, Pebane, Angoche, Mogincual, and Mossuril. New potential areas for the distribution of these species were identified, such as the Ilha de Moçambique and Mogovolas districts. The increase in the suitable area does not necessarily mean a benefit for the species, since anthropogenic changes in land cover may prevent the arrival and establishment of populations in the new suitable areas. The predictive models of the modeling were evaluated by the AUC index (area under the curve) and by the influence of the variables on the distribution of the species, these models presented a satisfactory performance of the AUC indexes that vary from 0.857 to 0.957 respectively.(TRADUÇAO NOSSA)