Spatial and temporal analysis of wild peat fire on island peatlands using remote sensing data

Abstract

Indonesia has the largest areas of tropical peatland, which is becoming vulnerable to fires. Smoke haze crises due to forest fires that occur annually have plagued many parts of Southeast Asia between 2013-2015 with devastating effects on human health and economy. Fires spread fast on island peatlands in eastern Sumatra because their small catchments drain quickly into the sea. This research dived into the spatial and temporal patterns of peat wildfires in the Peat Hydrological Unit of Bengkalis Island. Drawing on remote sensing sources such as MODIS-C6 Terra/Aqua thermal hotspots from 2001 to 2020 and Landsat images, the findings revealed two clear yearly fire peaks in Bengkalis: Period I (January-March) and Period II (June-August). Hotspot density was concentrated in deep peat areas 57.3% of detections are from the 500-700 cm thickness class, and 93.1% from peat >300 cm. During 20 years, 343 fire-days burned a total of 56,202 ha; the largest single occurrence involved a burnt area of 19,999 ha in 2014. Monthly rainfall is the dominant climatic control, exhibiting a clear inverse relationship with hotspot frequency; rainfall below 100 mm/month defines elevated fire-risk conditions. These findings establish that MODIS hotspot monitoring combined with Landsat burned area mapping provides a robust long-term fire assessment framework, and that rainfall thresholds offer a practical operational basis for fire early warning systems on island peatlands.

Keywords

• hotspot
• island peatland
• Landsat
• peat fire
• remote sensing