Case Studies of Smoke Aerosol Optical Depth Retrievals from GOES-16 and GOES-17

The frequency and intensity of wildfires in North America in recent years has increased interest in the detection of smoke and the retrieval of aerosol properties from space-based platforms. This task is challenging because wildfire emissions vary with fuel and burn conditions, and smoke aerosols undergo changes in composition and optical properties during atmospheric transport and aging. To monitor these changes in smoke plumes over time, geostationary satellites provide by far the most suitable observations due to their excellent temporal resolution, making it important to evaluate the retrievals of smoke plumes and identify key elements in the retrieval process.

In this work, we study verified smoke plumes during summer 2021 using GOES-16 and GOES-17 operational aerosol optical depth (AOD) retrievals along with those from the Aerosol Robotic Network (AERONET). To test whether aging is discernable, we analyze three observational days at variable downwind distances in Montana, a case in Wisconsin of long-range-transport from Canada, and three cases in Colorado of aged smoke. These cases are over variable terrains, allowing us to probe the role of surface reflectances.

We found that the operational model retrieved AOD well, but frequently selected one of the non-smoke models. After testing combinations of phase functions and single-scattering albedo (SSA) in our radiative transfer model, our preliminary conclusion is that the assumed SSA currently built into the smoke models is too low for these cases, suggesting that aging had already occurred. None of the fires was closer than 100 km to the AERONET site, and the smoke was likely aged at least two hours. Field studies using aircraft, as well as theoretical models of smoke evolution, have shown that smoke has undergone significant changes after just a few hours in the atmosphere, in agreement with our findings. During these first studies, we confirmed that the surface reflectance plays a critical role in these over-land retrievals, despite seeking heavy smoke cases with total reflectances dominated by the aerosol signal. Accordingly, appropriate representations of the diurnally- and seasonally-varying surface reflectances are critical to accurate AOD retrievals.