Comparison of emission models used for large scale simulation of the mineral dust cycle

Three different emission models of mineral dust of increasing complexity have been compared to satellite observations of mineral dust over the Western Sahara, in order to evaluate their capability to simulate the frequency of the dust events. The first model is a linear function of the wind velocity. It provides a very poor agreement with the observations (8%). The second one is a function of the wind velocity raised to the third power with a unique threshold for the dust emission, weighted by a soil texture parameter. With regards to the limited number of input parameters required by such a model, the agreement with the observations is quite good (61%) but the location and intensity of the dust sources are not accurately simulated. The third model is based on a physical description of the main processes of dust emission. It also utilizes mapped inputs of parameters characterizing the surface features in the tested area. The agreement with the observations is very good (79%) and the model provides a fine description of the dust sources location and intensity. From comparison with observations, it is stated that a simple wind power model is not appropriate to simulate the mineral dust emissions. A single threshold source model provides simulation of the present dust emissions with a reasonable confidence level. Moreover, the input parameters required by such a model are available to perform global simulations of the mineral dust cycle, even if its application may be limited to climatological studies for the present time. The best agreement with the satellite observations is obtained by using a physically based dust emission model. However the application of such models on large scale is limited by the availability of the input data characterizing the surface features of the desert areas.