Evaluation of the effect of soil moisture and wind speed on dust emission using aeronet, seviri, soil moisture and wind speed data

Dust emission has a large temporal and spatial variation making it extremely challenging to model. Combination of land surface model and remote sensing model are used for dust detection and monitoring in recent years. In this work, possibility of using ground measured wind speed (WS) data and satellite measured soil moisture (SM) data in AOT retrieval is investigated using artificial neural network (ANN) model. A combination of SEVIRI Brightness Temperature Differences/Brightness Temperature (BTD_3.9-10.8, BTD_8.7-10.8, BTD_10.8-12 and BT_3.9) is used as input and AERONET AOT (level 2) data at 0.5 μm as output for developing a base ANN model. Later, AMSR-E SM data and ground measured WS are employed as additional inputs to the base model to investigate their contribution on AOT retrieval. This improves the simulation accuracy of the ANN model in retrieving AOT. The R-square is increased from 0.70 to 0.76 while RMSE is reduced from 0.113 to 0.09.