A comparison of dust forecast models with Earth observations from satellites and ground-based data during African dust events over the Greater Caribbean Basin

Abstract

The tropical Atlantic and the Caribbean are subject to annual African dust (AD) intrusions. The study used ground-based and satellite measurements from the summers of 2020-21 across eight locations in the Greater Caribbean Basin (GCB). This study examined dust surface presence and vertical distribution. In addition, the capacity of the dust forecast model in reproducing dust events and storms at the GCB. As a result, fourteen dust events were recorded during dust and non-dust days, differences in dust concentration and PM₁₀, as well as AOD datasets, were observed across the Caribbean islands, although they were not evident at distant sites. Correlations between dust concentration and PM₁₀, and/or AOD datasets were weak at three stations (Miami, Merida, and Camaguey), indicating a lower contribution of AD in these locations. Vertical profiles showed that transport mainly in the MBL was shorter and more localized, while in the SAL or in both layers, it was generally longer. Models used to overestimate the values compared with measurements by a factor of up to 10 on the surface and during storms. The models represented consistent differences in storms, locations, and variables with a specific pattern. The Godzilla event (June 2020) was the most intense and extensive. This study offers a new regional perspective, emphasizing the need for more integrated data to enhance our understanding of AD.