First estimation and evaluation of hourly biomass burning emissions in north American high latitudes

Smoke from wildfires across North American high latitudes can travel long distances, degrading regional air quality. Hourly fire emissions are a crucial input of air quality models. However, they are unavailable for fires at high latitudes. The Advanced Baseline Imager (ABI) onboard NOAA's Geostationary Operational Environmental Satellites (GOES)-R Series satellites detects fires across North America every 10 min at a nominal resolution of 2 km, offering a good opportunity to estimate hourly fire emissions. At high latitudes, the polar-orbiting Visible Infrared Imaging Radiometer Suite (VIIRS) sensor provides 375 m fire observations up to six times a day for the same locations. In this study, we estimated hourly fire emissions at high latitudes for the first time by blending the high-temporal resolution ABI fire radiative power (FRP) and fine-spatial resolution VIIRS FRP. First, we corrected the parallax issue in ABI active fire data by considering land surface elevation. Next, FRPs from ABI and VIIRS fire pixels were separately aggregated into 3 km grids, and ABI FRP was calibrated against VIIRS FRP to correct for underestimation at large view angles. Then, the calibrated ABI FRP and VIIRS FRP were fused to reconstruct FRP diurnal cycles with the help of FRP diurnal climatologies. Finally, hourly emissions for eleven species were estimated using the reconstructed FRP diurnal cycles for two years (2021 and 2022). The results suggest that fires emitted a total of ∼0.54 Tg and 0.4 Tg PM_2.5 (particulate matter with diameter < 2.5 μm) in 2021 and 2022, respectively, with over 86 % of emissions released from summertime boreal forest fires. Moreover, hourly emissions revealed a general diurnal pattern: emissions were limited in the early morning, peaked around 2:00 PM - 3:00 PM local time, and decreased in the evening. Notably, the diurnal pattern of emissions varied by region and season in timing of peak emission and emission magnitude. Furthermore, we evaluated hourly emissions estimates using carbon monoxide (CO) and nitrogen dioxide (NO₂) observations from the TROPOspheric Monitoring Instrument (TROPOMI) over 29 fresh smoke plumes. Evaluation results show that the hourly ABI-VIIRS CO estimates were significantly correlated with the TROPOMI-based CO estimates (R² = 0.94, P < 0.001) and both were generally comparable, with a difference of 17.5 %. The proposed algorithm in this study has been integrated into the near real-time hourly Regional ABI and VIIRS fire Emissions (RAVE) product and is expected to improve air quality forecasting.