Forest fire emission estimates over South Asia using Suomi-NPP VIIRS-based thermal anomalies and emission inventory.

Emission estimates of carbon-containing greenhouse gases (CO, CH) and aerosols (PM) were made from forest fire across South Asia using Visible Infrared Imaging Radiometer Suite (VIIRS) based thermal anomalies and fire products. VIIRS 375 m I-band active fire product was selectively retrieved for the years 2012-2021 over forest cover across South Asia. Annual incidence of fire events across South Asia was 0.17 (±0.05) million (M) with robust spatio-temporal variation. Fire occurrences were mainly concentrated over the forest across Hindu Kush Himalayan region (HKH; 56%), Deccan Plateau (DP) and Central Highlands (CH; 34%). Monthly mean fire incidences emphasize February to May as a typical forest fire season, accounting 90% of annual fire counts. The highest fire pixel density (>1.5 km yr) was noted over the tropical dry/moist deciduous and tropical semi-evergreen forests. Strong diurnal nature of fire radiative power (FRP) was evident with >85% of FRP linked to daytime retrieval. VIIRS based Fire Emission Inventory (VFEI, Version 0) was followed to constitute regional emissions of PM and green house gases from forest fire. Forest fire accounted a yearly emission of 91.58 (±14.76) and 0.25 (±0.04) Tg yr CO and CH respectively, with 25.14 (±3.94) Tg of cumulative carbon release per year, i.e., roughly 1.3% of global fire-related carbon emission. Fire associated PM emission rate was 0.60 (±0.10) Tg yr, 95% of which emitted during peak fire season as was the case for carbon-containing gases. Forest fire across HKH (75%) and DP + CH (20%) predominately contribute to the regional carbon emission, while also accounting 68% (HKH) and 27% (DP + CH) of fire associated PM emission budget. With >70% of forest fires within South Asia being typically anthropogenic, forest fire appears to be a major sector of greenhouse gas and aerosols emissions, and necessitate planning and strict legalities to reduce emission load.