Improving GOES Advanced Baseline Imager (ABI) aerosol optical depth (AOD) retrievals using an empirical bias correction algorithm

Abstract

Abstract. The Advanced Baseline Imager (ABI) on board the Geostationary Operational Environmental Satellite-R (GOES-R) series enables retrieval of aerosol optical depth (AOD) from geostationary satellites using a multiband algorithm similar to those of polar-orbiting satellites' sensors, such as the Moderate Resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS). However, this work demonstrates that the current version of GOES-16 (GOES-East) ABI AOD has diurnally varying biases due to limitations in the land surface reflectance relationships between the 0.47 µm band and the 2.2 µm band and between the 0.64 µm band and 2.2 µm band used in the ABI AOD retrieval algorithm, which vary with the Sun–satellite geometry and NDVI (normalized difference vegetation index). To reduce these biases, an empirical bias correction algorithm has been developed based on the lowest observed ABI AOD of an adjacent 30 d period and the background AOD at each time step and at each pixel. The bias correction algorithm improves the performance of ABI AOD compared to AErosol RObotic NETwork (AERONET) AOD, especially for the high and medium (top 2) quality ABI AOD. AOD data for the period 6 August to 31 December 2018 are used to evaluate the bias correction algorithm. After bias correction, the correlation between the top 2 quality ABI AOD and AERONET AOD improves from 0.87 to 0.91, the mean bias improves from 0.04 to 0.00, and root-mean-square error (RMSE) improves from 0.09 to 0.05. These results for the bias-corrected top 2 qualities ABI AOD are comparable to those of the corrected high-quality ABI AOD. By using the top 2 qualities of ABI AOD in conjunction with the bias correction algorithm, the areal coverage of ABI AOD is increased by about 100 % without loss of data accuracy.

Daily and Hourly Surface PM2.5 Estimation From Satellite AOD

“…Diurnal variation of AOD bias were found due to the errors in surface reflectance relationships. Therefore, a bias correction algorithm was applied to remove the biases before using them to estimate PM2.5 (Zhang et al, 2020). In the bias correction algorithm, the AOD bias is obtained through analysis of time series of ZHANG AND KONDRAGUNTA past 30 days of AOD retrievals.…”

“…The current GOES Advanced Baseline Imager (ABI) contains similar spectral bands as those from MODIS/VIIRS and consequently is expected to retrieve AOD with similar accuracy as MODIS (Laszlo et al., 2008; Zhang et al., 2020). For example, bias corrected ABI AOD compares well against Aerosol Robotic Network (AERONET) AOD with zero mean bias and 0.05 Root mean square error (RMSE) over CONUS (Zhang et al, 2020). Using ABI AOD, PM2.5 can be estimated with higher temporal resolution to track regional/local transport of aerosols.…”

“…For example, Geostationary Operational Environmental Satellites (GOES) can provide full disk images every 15 or 10 min and can provide Continental United States (CONUS) images every 5 minutes (Schmit et al., 2005). The current GOES Advanced Baseline Imager (ABI) contains similar spectral bands as those from MODIS/VIIRS and consequently is expected to retrieve AOD with similar accuracy as MODIS (Laszlo et al., 2008; Zhang et al., 2020). For example, bias corrected ABI AOD compares well against Aerosol Robotic Network (AERONET) AOD with zero mean bias and 0.05 Root mean square error (RMSE) over CONUS (Zhang et al, 2020).…”

Quasi‐Global Maps of Daily Aerosol Optical Depth From a Ring of Five Geostationary Meteorological Satellites Using AERUS‐GEO

“…Furthermore, i-AERUS-GEO will exploit the full capacity of SEVIRI in terms of temporal resolution with the retrieval of AOD every 15 min. This work will join the efforts that are currently underway to retrieve the diurnal evolution of aerosol properties from geostationary missions such as MSG (Luffarelli & Govaerts, 2019), Himawari (Gupta et al, 2019;Lyapustin et al, 2019;Yoshida et al, 2018), GOES (Zhang et al, 2020), and GOCI (Lennartson et al, 2018). Information on subdaily AOD variations is key in many studies such as the calculation of aerosol radiative forcing (Xu et al, 2016).…”

Automated Low-Cost LED-Based Sun Photometer for City Scale Distributed Measurements

“…However, the spatial resolution for this type of satellite is poorer than that from lower orbit satellites. In addition, GOES-16 and Himawuari-8 have reported uncertainties for satellite products much higher than those from ground measurements [23][24][25][26][27]. In particular, for cities, the uncertainties of satellite products skyrocket due to the changes in the reflecto-properties of ground [28].…”