UC Davis

Dust aerosols have several impacts on the Earth system, including their influence on weather and climate through their interactions with clouds and radiation. Yet, dust modeling has long been a challenge in the research and operational communities. In particular, many weather and climate models underestimate dust long-range transport and dust-radiative effect. In this study, we aim to improve modeled dust representation over the Sahara Desert and the eastern Atlantic Ocean by integrating non-spherical (tri-axial ellipsoidal) dust particle shape and globally- and locally-representative observationally-constrained dust emitted size distributions into our dust model—the Weather Research and Forecasting model coupled with an online dust module (WRF-Dust). Based on these model modifications, we conducted two sets of numerical experiments over North Africa in 2016. Our numerical experiments on dust particle shape reveal that accounting for dust non-sphericity augments overall modeled aerosol optical depth (AOD) and promotes the westward extension of low Angstrom exponent (AE) values, thanks to the reduced dust sedimentation and enhanced dust optical properties under non-spherical conditions. As AE is inversely related to the average particle size, the non-spherical dust simulated results suggest that a greater fraction of large particles is transported from the Saharan source regions into the Atlantic Ocean, but the amount still falls short. Model evaluation against the AErosol RObotic NETwork (AERONET) observations shows that in the non-spherical shape simulation, there are 7.80% and 1.30% improvements in the root mean square error (RMSE) for AE and AOD, respectively, and 15.30% and 10.21% improvements in the mean bias for AE and AOD, respectively. By constraining dust emitted size distribution, we have further improved the magnitude and spatial distribution of AOD and AE, as indicated by the reduced error against a series of satellite and ground-based observations. Moreover, in a case study of Hurricane Earl (2010), by applying the best model configuration (i.e., the setting with the best overall model performance), we have slightly improved the storm intensity in minimum sea level pressure and maximum 10-m winds, although the improvements are not significant, requiring further adjustments in modeled dust. This study demonstrates the importance of emitted dust size distribution and particle shape in accurate dust simulation, which has far-reaching implications in numerous weather systems, such as Atlantic hurricanes, convective storms, African easterly waves (AEWs) and the African easterly jet (AEJ).