North African mineral dust sources: new insights from a combined analysis based on 3D dust aerosols distributions, surface winds and ancillary soil parameters

Abstract. Mineral dust aerosol is a key player in the climate system. Determining dust sources and the spatio-temporal variability of dust emission fluxes is essential to estimating the impact of dust on the atmospheric radiation budget, cloud and precipitation formation processes, the bio-productivity and ultimately the carbon cycle. Although much effort has been put into determining dust sources from satellite observations, geo-locating active dust sources is still challenging and uncertainties in space and time are evident. One major source of uncertainty is the lack of clear differentiation between near source dust aerosol and transported dust aerosol. In order to reduce this uncertainty, we use 3 dimensional information on the distribution of dust aerosol suspended in the atmosphere calculated from spectral measurements obtained by the Infrared Atmospheric Sounding Interferometer (IASI) by using the Mineral Aerosols Profiling from Infrared Radiance (MAPIR) algorithm. In addition to standard dust products from satellite observations, which provide 2 dimensional information on the horizontal distribution of dust, MAPIR allows for retrieving additional information on the vertical distribution of dust plumes. This ultimately enables us to separate between near source and transported dust plumes. Combined with information on near-surface wind speed and surface properties, low-altitude dust plumes can be assigned to dust emission events and low-altitude transport regimes can be excluded. Consequently, this technique will reduce the uncertainty in automatically geo-locating active dust sources. The findings of our study illustrate the spatio-temporal distribution of North African dust sources based on 9 years of data, allowing to observe a full seasonal cycle of dust emissions, differentiating morning and afternoon/evening emissions and providing a first glance at long-term changes. In addition, we compare the results of this new method to the results from Schepanski et al. (2012) which manually identify dust sources from Spinning Enhanced Visible and InfraRed Imager Red Green Blue (SEVIRI RGB) images. The comparison illustrates that each method has its strengths and weaknesses that must be taken into account when using the results. This study is of particular importance for understanding future environmental changes due to a changing climate.