Estimates of Surface PM2.5 using Merged GOES-16/17 ABI Aerosol Optical Depths for 2020 California Fires
Abstract
NOAA has been generating near real time surface PM2.5 using aerosol optical depths (AODs) from Moderate Resolution Imaging Spectroradiometer (MODIS) since 2008. The surface PM2.5 algorithm relied on 4-km grid specific climatological regression model parameters, which were found to be inadequate when day to day variations differed from climatology. NOAA recently modified its algorithm to dynamically vary the regression model parameters by using the Geographically Weighted Regression (GWR) technique. NOAA is also moving towards generating hourly near real time surface PM2.5 product using GOES-16 and GOES-17 Advanced Baseline Imager (ABI) AODs for ingestion into Environmental Protection Agency (EPA) AirNow system that is accessed daily by more than 150 million users. The new surface PM2.5 algorithm merges GOES-16 and GOES-17 ABI AOD data spatially into 2-km grids covering the Continental United States (CONUS) and temporally average the 5-minute observations into hourly composite. Prior to using the ABI AOD data, they are corrected to remove biases associated with viewing and solar geometry. We tested the algorithm for the 2020 extreme fires in the western US and found that the algorithm captures the spatial temporal variability. Between August and September 2020 when fires raged across the western US, estimated and observed hourly PM2.5 concentrations reached more than 400 µg/m3; diurnally, hourly PM2.5 concentrations peaked around 18 UTC. Compared to ground observations of surface PM2.5, the estimated values are found to have a good correlation (r2 = 0.61) with a slope of 0.83, bias of -0.02 and root mean square error of 43 µg/m3.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFM.A22B..05K