Improving estimation of a record-breaking east Asian dust storm emission with lagged aerosol Ångström exponent observations

Abstract

A record-breaking east Asian dust storm over recent years occurred in March 2021. The Ångström exponent (AE), which measures the wavelength dependence of aerosol optical thickness (AOT), is significantly sensitive to large aerosols such as dust. Due to the lack of observations during dust storms and the accuracy of the satellite-retrieved AE depending on the instrument and retrieval algorithm, it is possible to estimate the dust storm emission using the time-lagged ground-based AE observations. In this study, the hourly AEs observed by the Aerosol Robotic Network (AERONET) are assimilated with the ensemble Kalman smoother (EnKS) and Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) to optimize simulated dust emissions from 14 to 23 March 2021. The results demonstrate that the additional inclusion of AE can optimize the size distribution of dust emissions and the associated total flux depending on the covariance between time-lagged AE observations and simulated dust emissions in each size bin. Compared to the experiment only assimilating AOT, validation by independent observations from the Skynet Observation NETwork (SONET) shows that assimilating additional AE information reduces the root mean square error (RMSE) of simulated AOT and AE by approximately 17 % and 61 %, respectively. The temporal variation in both simulated AOT and AE is improved through assimilating additional AE information. The assimilation of AOT and AE also makes the magnitude and variations in aerosol vertical extinctions more comparable to the independent Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) observations in both westward and eastward pathways of dust transport. The optimized dust emissions in the Gobi Desert during this period is estimated to be 52.63 Tg and reached a peak value of 3837 kt h−1 at 07:00 UTC on 14 March.