A finite-volume numerical model for the simulation of dust transport in the atmosphere

Since dust phenomenon has become one of the major problems in many parts of the world, a large number of models have been developed in order to predict the concentration of dust particles in the atmosphere. The current study is devoted to present a model to simulate transport and deposition of dust particles using several schemes of the finite volume method. The outputs of each of these schemes are compared with each other, both quantitatively and qualitatively in two case studies. The data needed to run the model, including the wind are data derived from the Weather Research and Forecasting (WRF) model output and the GOCART emission scheme is used to calculate the vertical dust flux from surface. Comparing the model outputs with the satellite images available and the HYSPLIT model output show that both the areas covered with dust and the emission of dust are simulated correctly. The dust patterns obtained using the various schemes examined are reasonably similar to each other, considering the fact that the wind field and the emission scheme are similar.
Two cases of the dust storm that affected the extensive regions of Iraq and Iran are studied and simulated in this paper. In the first case, the dust emission starts from the eastern part of Syria on the 18th of June 2012. According to the satellite images, the dust that transported southeastward enters Iraq and subsequently affects the west and southwest regions of Iran, including parts of the Persian Gulf. The second case is related to a strong dust storm which sets up in Iraq, due to a synoptic system active on the 31st of August 2015. Dust emission is seen in the satellite images over Syria–Iraq border on the 31st of August 2015. The dust intensity increased during the next 24 hours, and was observed like a cyclone in the eastern and central parts of Iraq. Afterwards, the dust entered Iran’s borders and was extended towards the Persian Gulf, and finally covered all parts of the Persian Gulf.
The performance of the flux limiter, the second- and the third-order UNO, the second- and fourth-order Bott finite volume schemes have been examined in terms of numerical accuracy and computational cost. The numerical accuracy has been determined by comparing the dust concentrations obtained by the model with the corresponding results of the WRF-Chem. The dust concentration patterns obtained by all of the schemes are in overall agreement with each other even after 72 hours of integration, the differences being mainly in the damping caused by the schemes and their computational costs. Whereas the highest damping is observed for the upstream scheme, the fourth-order Bott exhibits the least damping followed by the third-order UNO. The latter two schemes are, however, involved high volumes of computation and may not be cost effective. Considering both numerical accuracy in terms of damping and the computational cost, the second-order UNO scheme offers promising results. With the quantitative comparison carried out, it can be concluded that the second-order UNO scheme shows the most correlation coefficient with the WRF-Chem model output and is the most appropriate scheme, among the schemes examined, for the dust operating model.