A spatio-temporal analysis of fire occurrence patterns in the Brazilian Amazon

Area of Science:

• Ecology
• Environmental Science
• Forestry

Background:

• Wildfires in the Amazon profoundly affect forest structure and the carbon cycle.
• Effective management strategies necessitate a thorough understanding of fire occurrence patterns.

Purpose of the Study:

• To analyze spatio-temporal changes in wildfire occurrence patterns in the Brazilian Amazon.
• To model the intensity function of fire occurrence using a dynamic Log Gaussian Cox process.

Main Methods:

• Utilized a novel spatio-temporal point process framework.
• Employed a dynamic Log Gaussian Cox process to model fire intensity.
• Decomposed the intensity function into trend, seasonality, cycles, covariates, and spatial effects.

Main Results:

• Observed a significant decrease in long-term fire occurrence trends until 2012, followed by an increase.
• Attributed these changes to governance measures and market mechanisms.
• Successfully captured spatial variability, linking higher fire rates to regions with more frequent dry seasons.

Conclusions:

• The study provides critical insights into the drivers of Amazonian wildfire patterns.
• Understanding these patterns is essential for developing targeted conservation and management interventions.
• The findings highlight the interplay between environmental factors and socio-economic influences on wildfire dynamics.