A near real-time web-system for predicting fire spread across the Cerrado biome

Area of Science:

• Environmental Science
• Geospatial Analysis
• Climate Change Impact

Background:

• Wildfires pose a significant threat, exacerbated by climate change, necessitating advanced tools for effective management.
• Vast regions like Brazil require robust territorial intelligence for wildfire prevention and rapid response.

Purpose of the Study:

• To develop and implement a fire-spread prediction system tailored for the Brazilian Cerrado, a biome highly susceptible to wildfires.
• To provide actionable data for public policies focused on wildfire prevention and combat.

Main Methods:

• The system integrates satellite data and hot pixel information to generate daily maps of fuel loads, vegetation moisture, and burning probability.
• Fire spread simulations are conducted at two resolutions: 25 ha for the entire Cerrado and 0.04 ha for nine specific conservation units.
• Model performance is evaluated based on spatial match, achieving 65-89% accuracy.

Main Results:

• The developed system provides thrice-daily fire spread predictions for the Brazilian Cerrado.
• High spatial accuracy (65-89%) was achieved in fire spread simulations across different resolutions.
• An interactive web platform disseminates model results and ancillary data, including historical fire data and CO2 emissions.

Conclusions:

• The fire-spread prediction system offers crucial territorial intelligence for enhancing wildfire prevention and response strategies in the Cerrado.
• The interactive platform is actively used for daily operations in selected conservation units, demonstrating practical application.
• This technological advancement is vital for mitigating the impact of climate change-driven wildfires in South America.