Biomass burning emissions estimated with a global fire assimilation system based on observed fire radiative power

摘要

The Global Fire Assimilation System (GFASv1.0)calculates biomass burning emissions by assimilating FireRadiative Power (FRP) observations from the MODIS instrumentsonboard the Terra and Aqua satellites. It correctsfor gaps in the observations, which are mostly due to cloudcover, and filters spurious FRP observations of volcanoes,gas flares and other industrial activity. The combustion rateis subsequently calculated with land cover-specific conversionfactors. Emission factors for 40 gas-phase and aerosoltrace species have been compiled from a literature survey.The corresponding daily emissions have been calculated ona global 0.5°×0.5° grid from 2003 to the present. Generalconsistency with the Global Fire Emission Database version3.1 (GFED3.1) within its accuracy is achieved while maintainingthe advantages of an FRP-based approach: GFASv1.0makes use of the quantitative information on the combustionrate that is contained in the FRP observations, and itdetects fires in real time at high spatial and temporal resolution.GFASv1.0 indicates omission errors in GFED3.1 dueto undetected small fires. It also exhibits slightly longer fireseasons in South America and North Africa and a slightlyshorter fire season in Southeast Asia. GFASv1.0 has alreadybeen used for atmospheric reactive gas simulations in an independentstudy, which found good agreement with atmosphericobservations. We have performed simulations of theatmospheric aerosol distribution with and without the assimilationof MODIS aerosol optical depth (AOD). They indicatethat the emissions of particulate matter need to be boostedby a factor of 2–4 to reproduce the global distribution oforganic matter and black carbon. This discrepancy is alsoevident in the comparison of previously published top-downand bottom-up estimates. For the time being, a global enhancementof the particulate matter emissions by 3.4 is recommended.Validation with independent AOD and PM10 observationsrecorded during the Russian fires in summer 2010show that the global Monitoring Atmospheric Compositionand Change (MACC) aerosol model with GFASv1.0 aerosolemissions captures the smoke plume evolution well whenorganic matter and black carbon are enhanced by the recommendedfactor. In conjunction with the assimilation ofMODIS AOD, the use of GFASv1.0 with enhanced emissionfactors quantitatively improves the forecast of the aerosolload near the surface sufficiently to allow air quality warningswith a lead time of up to four days.