Lógica Fuzzy Aplicada Para Análise de Riscos de Incêndios Florestais Para o Bioma Amazônia, Brasil

• Ano de defesa: 2024
• Autor(a) principal: Carvalho, Rita de Cássia Freire
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Doutorado em Ciências Florestais; Centro de Ciências Agrárias e Engenharias; UFES; Programa de Pós-Graduação em Ciências Florestais
• Palavras-chave em Português: Modelagem de risco; bioma Amazônia; geotecnologias; lógica Fuzzy; subject.br-rjbn; Área(s) do conhecimento do documento (Tabela CNPq)
• Link de acesso: http://repositorio.ufes.br/handle/10/17381

Forest fires have emerged as an environmental issue of relevance in recent years, demanding the attention of researchers. Forest ecosystems, such as the Amazon biome, are increasingly threatened by forest fires. In this context, the main objective of this study was to develop a model capable of predicting the location and quantification of forest fires in the Amazon biome, Brazil, using Fuzzy logic. For the development of the fuzzy logic-based fire risk model, the following variables were used: temperature, water deficiency, precipitation, altitude, slope, aspect, population density, highways, and land use/cover. Three forest fire risk modeling approaches were conducted to test different membership functions for the same variable, referred to in this study as Forest Fire Risk 1 (FFR 1), employing the linear increasing, linear decreasing, generalized bell-shaped, and Gaussian fuzzy membership functions; in Forest Fire Risk 2 (FFR 2) modeling, large and small fuzzy membership functions were used; and in Forest Fire Risk 3 (FFR 3) modeling, the large fuzzy membership function was applied to all variables. Subsequently, a fire density map was generated using fire scar data from the MapBiomas database, followed by a Pearson correlation analysis to check the consistency of the maps and identify the most effective model. Among the three modeling approaches, FFR 1 showed the highest Pearson correlation value; thus, it was determined to be the best model capable of predicting the location of forest fires. The Amazon biome is classified as having a medium risk of fire occurrence, which corresponds to 30.05% of the entire study region, an area of approximately 1,263,561.00 km². Therefore, it is understood that the forest fire risk modeling was effective in predicting and quantifying the risk of forest fires for the Brazilian Amazon biome.