Sensoriamento remoto aplicado ao estudo dos incêndios florestais no Brasil: uma revisão da literatura

Wildfires in Brazil represent a socioecological crisis primarily driven by deforestation. Scientific evidence demonstrates that the majority of these events are anthropogenic in origin, directly correlated with forest clearing for the expansion of the agricultural frontier. Remote sensing has emerged as a fundamental methodological tool for monitoring and analyzing wildfires. However, the rapid growth in scientific output has generated a fragmented body of literature, hindering effective synthesis. This bibliometric and systematic review analyzes the scientific literature on the topic to map its structure, evolution, and knowledge gaps, aiming to guide future research and policy. A total of 240 articles from Web of Science and Scopus (1997–2024) were examined using a hybrid approach, combining quantitative analysis via R’s Bibliometrix package with qualitative synthesis. Results reveal exponential growth in scientific production since 2018, with Brazil as the leading country and central node of international collaboration, and the National Institute for Space Research (INPE) as the most productive institution. Term co-occurrence analysis reveals a clear conceptual dichotomy in the field, organized into three thematic clusters. The first cluster centers on the Cerrado biome, examining fire through the lens of ecology and Integrated Fire Management (IFM). In contrast, the other two clusters focus on the Amazon, where fire is framed as a tool for deforestation and forest degradation, linked to selective logging and agricultural expansion. Methodologically, the field has transitioned from reactive mapping of burned areas toward predictive and risk modeling, increasingly employing machine learning and deep learning algorithms. We conclude that, despite the field’s technological maturity, critical knowledge gaps persist. Geographically, key biomes such as the Pantanal, Caatinga, and Pampa remain understudied. Thematically, there is insufficient integration of socioeconomic drivers underlying fire incidence. Future research must therefore adopt a more integrative, interdisciplinary, and predictive approach. Fusing multisensor remote sensing data with socioeconomic variables is essential to develop effective fire management policies tailored to Brazil’s diverse and complex ecological contexts.