Combinational Optimization of Physical Parameterization Schemes to Improve Air Quality Prediction Using Intelligent Optimization System

  The environmental problems related to air pollution have been increasing, especially in East Asia, due to human activities, including high energy consumption and rapid economic growth. In order to address the air pollution problems, it is essential not only to improve the national and local air pollution control measures but also to enhance the air quality forecast skill through a numerical prediction system. The performance of numerical air quality prediction is significantly dependent on the land surface and the PBL parameterization schemes in a coupled atmosphere-chemistry prediction system, such as the Weather Research and Forecasting (WRF) model coupled with Chemistry (WRF-Chem).

  In this study, to improve the air quality prediction performance in East Asia, we built an intelligent optimization system by coupling the micro-genetic algorithm (μGA) and the WRF-Chem model --- the WRF-Chem-μGA system. This system can find an optimal set of physical parameterization schemes in WRF-Chem to improve the air quality forecasting.

 Before optimization, we selected several cases by considering the synoptic weather patterns according to the sources and the transport routes of the sand dust storms that affected Korea. As a preliminary study, we aim to obtain the optimal set of the land surface and PBL schemes via the intelligent optimization system for each case, which is the most suitable for predicting some Asian sand dust storm (SDS) events over Korea. Overall, our preliminary results show that the WRF-Chem with the optimized set of parameterization schemes produces better results than that with non-optimized scheme sets in forecasting the selected SDS events in East Asia.