Evaluation of numerical method combinations for transport-dominated problems in the chemistry transport model (CHIMERE): Ozone prediction and Computing optimization study

Abstract Ozone concentration generally affects a health system, including human health, and an ecosystem, including plant properties and soil processes. Based on high-performance computing, this study used the transport chemistry model CHIMERE to model ozone pollution in the city of Agadir. The aim is to investigate the integrating horizontal and vertical transport effect on ozone concentration and computation time. The findings are validated experimentally by measurements and also compared with other studies. The results highlighted four main points: First, the CHIMERE ozone modeling remains acceptable with a correlation coefficient of 70%. Second, the combination Adv00 based on the Upwind method gives good results in terms of ozone concentration with a significant implementation time, equivalent to 80 minutes for 360 hours of prediction, which means an optimization in computational performance. Third, it is required to optimize all possible parts of the modeling process to reduce costs and time. Fourth, detailed local emission information is recommended to get a clearer picture of the correction of the CHIMERE output. The present work can serve as guidance for the CHIMERE settings in the case of limited computational infrastructure for long-term studies and studies using a high spatial resolution to predict polluted air.