A grid independence study to select computational parameters in dust storm prediction models: A sensitive analysis

Abstract

Grid options are computational parameters that lead to proper numerical accuracy in air quality simulation and their selection is challenging. Therefore, the present study intends to provide a framework for selecting grid options. In this paper, the impact of five computational parameters such as time step, modeling duration, computational domain, and vertical and horizontal resolution on predicting a severe dust storm was investigated by WRF-Chem. A two-step algorithm based on the grid independence framework was proposed to select the appropriate grid options. In total, >22 different configurations were considered to investigate the effect of these options. Also, grid options were prioritized by wins in the league (WIL) theory to evaluate factors affecting numerical results. The results showed that the choice of computational parameters has a significant effect on prediction. Inappropriate computational domain and modeling duration cannot accurately predict the intensity of the dust storm and these two parameters have the highest effect on the numerical modeling. The results of WIL approach showed the proposed algorithm was able to appropriately estimate the dust storm. The proposed framework can be used not only in predicting dust storms but also in predicting weather conditions and other aspects of air quality.