Inverse Modeling to Improve Emission Inventory for PM10 Forecasting in East Asia Region Focusing on Korea.

Abstract:

The aerosol transports from China and Mongolia along the Northwestern wind have large influence on the air quality in Korea and the assessment of the emission in the East Asia region is an important factor in air quality forecasting in Korea.

In order to obtain working PM10 emission inventory for the PM10 forecast modeling over East Asia, the Bayesian approach with CAMx (Comprehensive Air-quality Model with extension) forward model was applied. The surface observations of PM10 from EANET (Acid Deposition Monitoring Network in East Asia), API (Air Pollution Index) sites over China and AAQMS (Ambient Air Quality Monitoring Stations) in Korea were used for the inverse modelling. The predicted PM10 concentrations with a priori emission were compared with observations at monitoring sites in China and Korea. The comparison showed that PM10 concentrations with a priori emissions were generally under-predicted. The result also indicated that anthropogenic PM10 emissions in the industrialized and urbanized areas in China were under-estimated in particular.

Optimized a posteriori PM10 emissions over East Asia from inverse modelling analysis ware proposed. A posteriori PM10 emissions were much lower than the a priori emission where the soil dust emissions were prevailing. This implied that the dust emission module still had large uncertainty and it was necessary to further research on the improvement of in-line emission modelling for the soil dust. In contrast, a posteriori anthropogenic emissions from industrialized areas such as Beijing and Shenyang sites were slightly higher than a priori emission at regions. Especially, a posteriori PM10 emissions increased in Korea and in Northeast region of China.

The predictions of PM10 with proposed a posteriori emission showed better agreement with the observations, implying that the inverse modelling minimized the discrepancies in the model estimation by improving PM10 emissions in East Asia. Further details of inverse modeling results and future research will be discussed in the presentation.