The Gigantic African Dust Intrusion to the Caribbean Basin and southern U.S. in June 2020: An analysis of MODIS and CALIOP remote sensing observations and GEOS model simulations

Abstract:

In late June of 2020, gigantic African dust plumes intruded into Caribbean Sea and affected a large swath of regions in the Americas, including the southern and northern Caribbean Basin, the Gulf of Mexico, the Central America, the tropical eastern Pacific Ocean, and the southeastern U.S. This paper characterizes this dust event using the MODIS and CALIOP remote sensing observations and the GEOS model simulations complemented by meteorological fields from the reanalysis and surface PM2.5 concentrations from the EPA air quality network. Our analysis of the full record of MODIS aerosol measurements reveals that this “Godzilla” dust event is historic in the recent two decades, with the peak aerosol optical depth up to 1.6 in the Greater Caribbean Basin and 0.8 in the Gulf of Mexico. The episode elevated the surface PM2.5 concentrations to a level exceeding the EPA air quality standard for up to three days in a number of sampling sites across the southeastern U.S. Our analysis of MERRA2 meteorological fields suggests that the intense amount of dust observed in the Greater Caribbean Basin and beyond was a result of dust accumulation spanning over several days in the eastern Atlantic Ocean and subsequent rapid westward transport controlled by anomalous drifting of the Bermuda-Azores High. We also characterize the evolution of three-dimensional structures of the dust plumes with the improved MODIS retrievals of non-spherical dust optical depth and fine-mode fraction, CALIOP observations of vertical profiles, and GEOS replay simulations of aerosol. Although GEOS model well captures the observed trans-Atlantic movement of the gigantic dust plumes, the simulated dust optical depth is a factor of 2-5 lower than the MODIS retrievals in the Greater Caribbean Basin and the Gulf of Mexico. Contrary to what is revealed by MODIS observations, this dust event is not registered as a historic event in the context of GEOS multi-decadal simulations from 2000 onward. The GEOS simulated dust layer altitude is also lower than the CALIOP observations. Possible factors contributing to the large observation-model differences will be discussed.