Abstract This study investigates the spatio-temporal variability of wildfires in Mozambique from 2001 to 2019, focusing on the relationships between climate dynamics, land cover, and ocean-atmosphere interactions (ENSO). Using remote sensing data from the MODIS sensor, we analyzed fire occurrences, burned areas, vegetation indices (NDVI and EVI), and land use/land cover patterns. These were integrated with ERA5-Land climate reanalysis data, including precipitation and air temperature. Results reveal strong seasonality in wildfire activity, predominantly from June to September (JJAS), with higher intensity in the Central and Northern regions. While both regions had the highest number of fire events, the Northern region experienced the largest burned areas, mainly in savanna ecosystems. The South exhibited isolated but severe fire events, particularly in natural grasslands. Trend analyses showed a stable fire regime in the Central and Northern zones, whereas the South displayed a statistically significant upward trend, especially in open shrubland. A decoupling between fire ignition and spread was also observed: the highest number of ignitions occurred during ENSO-neutral years, while the largest burned areas were associated with El Niño and La Niña events. This suggests that fire spread is more influenced by dry biomass availability than ignition frequency. Fire dynamics are modulated by landscape fragmentation, vegetation types, and land use practices. Tools such as the Fire Potential Index (PFIv2) and vegetation indices demonstrated strong potential for identifying fire-prone areas, reinforcing their relevance for fire risk monitoring and mitigation in tropical environments.
Sadique et al. (Tue,) studied this question.