Evaluation of WRF-Chem Model Performance in Wind Field Simulation in Dust Storm April 2022 in ‎Khuzestan Province

Introduction

The dust phenomenon is one of the natural phenomenon that is formed by both human and natural factors that ‎causes adverse environmental consequences every year in many arid regions worldwide, including Iran.‎

Predicting the emission and transmission of dust and aerosols can be useful to mitigate harmful effects. ‎However, despite numerous studies, predicting dust events and their transmission remains challenging. Wind ‎speed, vegetation, and soil structure are the most important factors in local and regional dust generation. The ‎WRF-Chem model is a popular numerical model used for simulating wind fields, dust, and air pollution, and is ‎of interest to researchers worldwide.‎

The Khuzestan province located in the southwest of Iran is affected by both cross-border and internal dust due ‎to its geographical location. Wind plays a crucial role in the emission and transmission of dust to this region. ‎Due to the increasing number of dust days and their intensity in the province, predicting and simulating wind ‎and dust fields is of utmost importance. Therefore, finding the optimal configuration for the wind field during ‎dust events and transferring it to Khuzestan is necessary. Given the significance of this issue, the aim of this ‎study was to evaluate the wind field simulated by the WRF-Chem model under both dust and non-dust ‎conditions, and to determine the optimal configuration for each scenario.‎

Data and Method

To achieve this, meteorological and environmental data from 2020 to 2022 were collected from the Iran ‎meteorology and environmental organization. During the period of April 7-25, 2022, dust was recorded on ‎some days in Khuzestan province, while on other days, no dust was observed. In the second step, the wind field ‎of 700 hPa level was analyzed using GFS data in April 2022. The WRF-Chem model was run from 7 to 25 ‎April 2022 with GFS data for four different model configurations with two boundary layer scheme YSU and ‎MYJ as well as Lin and WMS6 cloud microphysics scheme for two horizontal resolution of 27 and 9 km.‎

Conclusion

The results of the model were compared with the initial GFS data and the direction and wind speed of 10 ‎meters observed. Statistical indicators and Taylor charts were also utilized.‎

The results show that, the highest number of dust days in Khuzestan province occurred in Bostan and Abadan in ‎‎2022. In the three months of May, March and July the highest number of dust days in the province was ‎obtained.‎

The wind filed of 700 hPa simulated by all four WRF-Chem model configurations is similar to the initial GFS ‎data, which indicates the good model performance in simulation of wind field in the selected area, although ‎there are differences in some details and in smaller scales.‎

The maximum and mean of estimated wind speed by configuration with the YSU boundary layer scheme is ‎lower than MYJ and is closer to observational data.‎

During the analysis of the two bias error indices (MB) and normalized mean growth error (NMGE), the lowest ‎values were observed at the Abadan, Mahshahr, and Omidieh stations, indicating excellent performance of the ‎WRF-Chem model in these areas. However, weaker results were obtained at the Bostan and Ahvaz stations.‎

Taylor diagram shows the good model performance in estimation of 10-meters wind in Abadan, Mahshahr, ‎Bostan, Ahvaz and Omidieh stations. The P1M6D1 and P1M2D1 have better results than P2M2D1 and ‎P2M6D1. In this way, the Taylor diagram shows the impact of the YSU boundary layer in estimating 10-‎meters wind better than MYJ.‎

Dust distribution obtained from WRF-Chem model and the dust mass observed in the image of the MODIS ‎sensor are in good harmony so that the dust formation center in eastern Syria and northwest of Iraq, eastern ‎Iraq and northern Saudi Arabia is well simulated by model. In the horizontal distribution of dust, the boundary ‎layer scheme has more effect than the microphysical scheme.‎