Development of the Global Fire Emissions Database (GFED): Toward reconciliation of top-down and bottom-up constraints on fire contributions to variability and trends in carbonaceous aerosol

Over the past decade, multiple lines of evidence from satellite and surface observations indicate that fire inventories often underestimate emissions of organic carbon aerosols. The cause of this bias has not been conclusively identified, in part, because uncertainties remain considerable with respect to regional and global estimates of burned area, fuel consumption, and biome-specific emission factors. Here we describe variations in organic carbon aerosol emissions among different fire inventories, including the Global Fire Emissions Database. Toward the development of the 5^th version of GFED (GFED5), we provide a preliminary assessment of how new observational constraints on burned area and emission factors are helping to resolve this important discrepancy. For burned area, we describe how new observations from Landsat8 and Sentinel 2 have improved our ability to measure difficult-to-detect fires. We also describe new observations from boreal forest and tropical peatland ecosystems that suggest levels of smoldering combustion (and organic aerosol production) may have been previously underestimated. Individual fire size and spread rate estimates from the Global Fire Atlas v1, in addition, provide a new means to extrapolate variations in combustion completeness and emission factors across different landscapes and through time. Using GFED in combination with a global atmospheric model (GEOS-Chem), we then assess fire contributions to variability and trends in aerosol optical depth and surface PM2.5 concentrations.