The Atlantic Forest biome is annually exposed to forest fires that damage thousands of hectares of forest, promote the formation of forest fragments, the destruction of biodiversity, soil compaction, runoff and silting of water bodies. Thus, the prediction and suppression of fire outbreaks are important to minimize the damage caused by fire. In this sense, this study aimed to analyze and model the risk of forest fires occurring in the Rio Preto National Forest and its buffer zone using the Fuzzy artificial intelligence technique. To do so, the land use and occupation, proximity to roads, slope and relief orientation variables were used to compose the model. Thus, the influence of each variable on burning episodes was determined with the aid of geographic information systems (GIS), as well as the spatial distribution of each of the risk classes (very low, low, moderate and high). a historical series of fires between the years 2010 and 2020 was used to perform a comparative analysis of the model. The results showed that the study area does not present worrisome risks regarding the occurrence of fires, since it is mostly covered by very low and low risk classes. Thus, it can be concluded that the application of Fuzzy modeling enables evaluating the spatial distribution of fire risk classes for the protected area areas, for which the proposed comparative analysis indicated the model’s effectiveness.
Comparative analysis of adaptive and neutral markers of Drosophila mediopunctata
populations dispersed among forest fragments
