Estimation of biomass-burning emissions by fusing the fire radiative power retrievals from polar-orbiting and geostationary satellites across the conterminous United States

Biomass burning is an important source of atmospheric greenhouse gases and aerosols, and its emissions can be estimated using Fire Radiative Power (FRP) retrievals from polar-orbiting and geostationary satellites. Accurate and timely estimation of biomass-burning emissions (BBE) requires high-spatiotemporal-resolution FRP that is characterized by accurate diurnal FRP cycle. This study is to estimate hourly reliable BBE in a 0.25 degrees x 0.3125 degrees grid across the conterminous United States (CONUS) to be used in chemical transport models for air quality forecast. To do this, this study for the first time fused FRP retrievals from the Geostationary Operational Environmental Satellite (GOES) with those from Moderate Resolution Imaging Spectroradiometer (MODIS) Collection 6 after GOES FRP was angularly adjusted and was further calibrated against MODIS FRP. The FRP data was obtained from Terra and Aqua MODIS 1 km active fire products with fire observations of four times a day and from 4 km GOES WF_ABBA (WildFire Automated Biomass Burning Algorithm) fire products for GOES-W (GOES-11 and 15) and GOES-E (GOES-13) with observations every 5-15 min across the CONUS from 2011 to 2015. The diurnal FRP cycles at an interval of 15 min for a grid were reconstructed using the ecosystem-specific diurnal FRP climatology and actually available MODIS-GOES fused FRP, which were applied to estimate hourly BBE across the CONUS. The results indicate that the reconstructed diurnal FRP cycle varied significantly in magnitude and shape among 45 CONUS ecosystems. The biomass burning released 717 Gg particulate matter smaller than 2.5 mu m in diameter (PM2.5) in the CONUS each year; however, it presented significant temporal (diurnal, seasonal, and interannual) and spatial variations. Finally, the BBE estimates were evaluated using available data sources and compared well (a difference of similar to 4%) with emissions derived from Landsat burned areas in the western CONUS and with hourly carbon monoxide emissions simulated using a biogeochemical model over the Rim Fire in California (difference < 1%). The BBE estimates showed similar seasonal variation to six available BBE inventories but with variable magnitude.