Numerical air quality forecasting over eastern China: An operational application of WRF-Chem. (March 2017)

• Record Type: Journal Article
• Authors: Zhou, Guangqiang; Xu, Jianming; Xie, Ying; Chang, Luyu; Gao, Wei; Gu, Yixuan; Zhou, Ji
• Is Part Of: Atmospheric environment. Volume 153(2017)
• Page Start: 94
• Page End: 108
• Publication Date: 2017-03
• Subjects: Air quality modeling -- Air quality forecasting -- WRF-Chem -- PM2.5 -- Ozone; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2017.01.020
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: The Regional Atmospheric Environmental Modeling System for eastern China (RAEMS) is an operational numerical system to forecast near surface atmospheric pollutants such as PM2.5 and O3 over the eastern China region. This system was based on the fully online coupled weather research and forecasting/chemistry (WRF-Chem) model. Anthropogenic emissions were based on the multi-resolution emission inventory for China (MEIC), and biogenic emissions were online calculated using model of emissions of gases and aerosols from nature (MEGAN2). Authorized by the China Meteorological Administration (CMA), this system started to provide operational forecast in 2013. With a large domain covering eastern China, the system produces daily 72-hr forecast. In this work, a comprehensive evaluation was carried out against measurements for two full years (2014–2015). Evaluation results show that the RAEMS is skillful in forecasting temporal variation and spatial distribution of major air pollutants over the eastern China region. The performance is consistent in different forecast length of 24 h, 48 h, and 72 h. About half of cities have correlation coefficients greater than 0.6 for PM2.5 and 0.7 for daily maximum 8-h averaged (DM8H) ozone. The forecasted PM2.5 is generally in good agreement with observed concentrations, with most cities having normalized mean biases (NMB) within ±25%. Forecasted ozone diurnal variation is very similar to that of observed, and makes small peak time error for DM8H ozone. It also shows good capability in capturing ozone pollution as indicated by high critical success indexes (CSI). The modeling system also exhibits acceptable performance for PM10, NO2, SO2, and CO. Meanwhile, degraded performance for PM2.5 is found under heavy polluted conditions, and there is a general over estimation in ozone concentrations. Highlights: An operational application of WRF-Chem model was involved. Detailed evaluation was carried out of two years' operational run. Good performance was exhibited and bias sources were discussed.