The purpose of this project is to maintain the performance of forecasting systems and to facilitate daily forecast of air quality over the Taiwan region. The main work included the maintainance and operation of the mesoscale numerical forecasting model, the China dust dynamic numerical model, the local dust dynamic model, and the Community Multi-scale Air Quality Model. Both the characteristics of atmospheric variables from the dynamic numerical model’s output, and the concentrations of ozone, PM10 and PM2.5 computed from the air quality model have been analyzed. The result from the previous models would help improving the performance of the dynamic numerical model in the future. Currently, the sensitive factors analysis of is performed by comparing the concentrations of O3, PM10 and PM2.5 with that from the 73 EPA monitoring stations. For general pollution analysis, the modeled concentrations and diurnal variations of the pollutants are compared with the observation of EPA monitoring stations and seven air quality districts, respectively. A severesmog event originated from the mainland China, and a serious air pollution event due to poor local dispersion, are being analyzed. To help the daily forecasting in EPA, special-duty manpower is assigned to work in EPA for organizing and integrating all of the information to offer the referral for the pre-warning decision, and supporting the organization learning. The expert opinion from project director is provided to EPA during the Mainland China severe dust events forecasting meeting. In this year, the performance of the operation models is improved. Especially in the ozone, PM10 and PM2.5 forecasting capabilities. The preliminary analysis for misalignment of model performance on air quality simulation has provided guidance for adjustment in the dust model. The modified TWKOSA version has been submitted since July 2017.

air-quality forecast, dynamic numerical model, Mainland China and local riverside mineral dust