[1] Mineral dust aerosols play an important role in the climate system. Coupled climateaerosol models are an important tool with which to quantify dust fluxes and the associated climate impact. Over the last decade or more, numerous models have been developed, both global and regional, but to date, there have been few attempts to compare the performance of these models. In this paper a comparison of five regional atmospheric models with dust modules is made, in terms of their simulation of meteorology, dust emission and transport. The intercomparison focuses on a 3-day dust event over the Bodélé depression in northern Chad, the world's single most important dust source.Simulations are compared to satellite data and in situ observations from the Bodélé Dust Experiment (BoDEx 2005). Overall, the models reproduce many of the key features of the meteorology and the large dust plumes that occur over the study domain. However, there is at least an order of magnitude range in model estimates of key quantities including dust concentration, dust burden, dust flux, and aerosol optical thickness. As such, there remains considerable uncertainty in model estimates of the dust cycle and its interaction with climate. This paper discusses the issues associated with partitioning various sources of model uncertainty.

[1] The Sahara desert is the world's primary source of mineral dust aerosols and is known to be an important but poorly understood component of the climate system. Climate models which incorporate dust modules have the potential to improve our understanding of the climate impacts of dust. In this study, the performance of the Regional Climate Model version 3 (RegCM3) with an active dust scheme is evaluated, using a major dust event of 6-11 March 2006 as a test case. To account for the distribution of preferential dust source regions, soil texture characteristics were modified in dust source regions identified from long-term SEVIRI satellite data. The dust event was associated with a pronounced cold outbreak of midlatitude air over the northern Sahara which produced anomalously strong northerly winds, which propagated from west to east over the Sahara during the study period. This resulted in dust mobilization from multiple dust sources across the domain. RegCM3 represents the space/time structure of near-surface meteorology well, although surface winds are underestimated in absolute terms. The experiment in which soils are modified provides a better representation of local dust sources and emission and resulting atmospheric optical thickness (AOT). In this experiment, model simulated dust flux exported from the Sahara to the Sahel and the tropical east Atlantic is estimated as 1.9 Tg d À1 . The dust event had a profound impact on the surface solar radiation budget of $À140 W m À2 per unit AOT (domain average). The shortwave radiative effect at the top of the atmosphere is $À10 W m À2 per unit AOT over the study domain. However, this is strongly dependent on surface albedo. The results also highlight how errors in model simulated circulation lead to errors in the position of the dust plume.