Drivers of fire foci in a protected area in the Brazilian Amazon: a geospatial approach

Abstract Forest fires in the Amazon, particularly in protected areas, pose significant challenges related to CO 2 emissions and deforestation. This study aimed to evaluate fire foci and their driving factors in the Jamanxim National Forest (FLONA Jamanxim), located in the southeastern Amazon, using geographic information systems (GIS) to inform the decision-making process for environmental management. The methodology involved two phases: (I) processing data on fire foci, land-use and land-cover changes (LULCCs), and climatic variables in the pivotal years 2004, 2007, 2020, and 2022; (II) data analysis using the Mann–Kendall test and principal component analysis (PCA) to understand the trends and relationships between fire and LULC and climate variables. A total of 5012 fire foci were recorded, with the highest numbers in 2004 and 2007 during neutral and warm phase of the El Niño–Southern Oscillation (ENSO), respectively. Throughout the study period, the months of July–September (dry season months) showed the greatest concentration of fires. LULCC analysis indicated pasture expansion towards FLONA’s center. Spatial fire foci patterns were further analyzed through hotspot mapping, revealing concentration zones near pasture expansion areas. The Mann–Kendall test revealed a significant trend in fire foci over pasture ( p = 0.04). PCA showed that fire foci were linked to maximum temperatures (TMMXs), wind speed (VS), and vapor pressure deficit (VPD) and inversely correlated with soil moisture and evapotranspiration. These findings, derived from geospatial and statistical analyses, highlight that human activities are at the origin of fire foci, underscoring the need for improved management and monitoring of FLONA Jamanxim.