Satellite-Based Assessment of Fire Regimes Across Angola’s Protected Areas (2015–2025): Climatic Drivers, Vegetation Dynamics and Spatial Patterns

Protected Areas play a critical ecological, socio-economic and cultural role in Africa, particularly in Angola, where increasing anthropogenic pressures and recurrent wildfires threaten ecosystem integrity and biodiversity conservation. This study analysed the spatio-temporal dynamics of fire regimes across Angola’s Protected Areas (APAs) between 2015 and 2025 and evaluated the influence of climatic variables on fire occurrence. Spatial datasets from the World Database on Protected Areas were combined with MODIS burned area (MCD64A1) and active fire products (MCD14A1), TerraClimate precipitation and temperature datasets, MODIS vegetation indices (NDVI and EVI), and MODIS land cover products (MCD12Q1). Except for protected area boundaries, all datasets were retrieved and pre-processed within the Google Earth Engine environment, while data harmonisation and resampling to a 500 m spatial resolution were conducted in Google Colab. Annual time-series analyses and descriptive statistics were applied to characterise fire frequency, seasonality and vegetation selectivity across the 14 APAs, followed by Pearson correlation analyses to assess relationships between fire occurrence and environmental variables. Results revealed higher fire frequency and burned area in the south-eastern and south-central regions, particularly within Mavinga National Park, Luengue-Luiana National Park, Cameia National Park, and the Luando Reserve. These patterns were associated with high biomass availability and the predominance of herbaceous vegetation and wooded savannas. Conversely, lower fire incidence was observed in the arid south-western regions and the humid tropical forests of Maiombe. NDVI and EVI analyses indicated reductions in vegetation vigour during 2019, 2021 and 2022, followed by recovery between 2023 and 2025, suggesting notable post-disturbance ecosystem resilience. Strong positive correlations were identified between burned area and fire frequency (R² = 0.96), precipitation (R² = 0.82), and temperature (R² = 0.80). These findings demonstrate the effectiveness of MODIS-based fire monitoring for identifying critical fire-prone zones and supporting evidence-based conservation and fire management strategies within APAs.