Work

Reconstructing savannah wildfire events using a deep learning framework

Year: 2025

Type: article

Abstract: This research presents a method to produce fire reconstructions with a high spatial and temporal resolution by downscaling geostationary satellite observations using a deep learning segmentation framework. This was achieved by training a U-Net on low earth orbit active fire detections paired with synchronous observations from the geostationary satellite Himawari-9/AHI resampled to the resolution of the 500m 0.64 µm RED channel. Geostationary satellites provide a means for continuous monitoring of fire behaviour but are constrained by the low spatial resolution of the infrared channels used to detect active fires. Localising fire activity using information from the higher spatial resolution solar reflective channels of geostationary satellites enables detailed fire progression mapping comparable to the spatial resolution of low earth orbit systems. Six case study fires in the northern Australian savannahs are reconstructed with their lifecycles compared to the burn scar mapped by the Northern Australia and Rangelands Fire Information (NAFI), with F1-scores ranging from 0.82 to 0.95. Model predictions synchronous to VIIRS active fire detections during the selected case study fires were used to test performance during case study events. The results indicate a high positive detection rate (75%) for detections with a fire radiative power (FRP) above 15.0 MW during day-time and 4.7 MW at night-time which degraded with decreasing FRP as the fire signal becomes increasingly difficult to detect within the 2km instantaneous field of view (IFOV) of the AHI infrared channels against the high day-time background temperatures within the northern Australian savannahs.

Source: Remote Sensing Applications Society and Environment

DOI: 10.1016/j.rsase.2025.101837

Authors: Simon Ramsey, Karin Reinke, Andrew Edwards, Simon Jones

Institutions: RMIT University, Charles Darwin University

Language: English

Topic: Fire Detection and Safety Systems

Subfield: Safety, Risk, Reliability and Quality

Field: Engineering

Domain: Physical Sciences

SDG: Climate action

Open Access status: hybrid

Estimated APC paid: $2,640