Evaluating Retrievals of Evolving Smoke Plumes from Geostationary Observations

Satellite remote sensing offers comprehensive observation of wildfires, with extensive coverage of emissions in space and time, including observing evolution of the properties of individual plumes. Using the ‘dual-view’ of two geostationary sensors provides additional constraints, with the potential to better observe detailed changes in smoke aerosol at high temporal resolution. Still, current satellite remote sensing retrieval algorithms require the use of pre-defined aerosol models and the accuracy of retrieved smoke properties strongly depends on the representativeness of these pre-defined models. In this work, we identified case studies to evaluate operational satellite retrievals of near-source smoke plumes. We show that, when validated against NASA’s Aerosol Robotic Network (AERONET), operational models perform well in characterizing aerosol optical depth, but they may miss the evolution of smoke absorption properties, affecting the use of such data in modeling smoke impacts on climate and Earth’s radiation budget. We further applied optical flow techniques to estimate the age of wildfire smoke along the plume, a factor that has significant influence on aerosol optical properties due to time-dependent processes like hygroscopic growth, dilution, and photochemistry. We implement a range of aerosol optical models in our assessment, helping us to determine the source of current operational model shortcomings and where improvements may be made.