北京区域环境气象数值预报系统及PM_(2.5)预报检验

Beijing Regional Environmental Meteorology Prediction System and Its Performance Test of PM_(2.5) Concentration

基于北京地区快速更新循环同化预报系统(BJ-RUC)、WRF-Chem模式和优选的能见度参数化方案,建立了北京区域环境气象数值预报系统。对2014年全年PM_(2.5)浓度、能见度和APEC(Asia-Pacific Economic Cooperation)期间预报效果检验结果表明:该系统对京津冀及周边地区PM_(2.5)浓度的预报效果较好,大部分站点的相关系数在0.6以上,特别足北京的部分站点可达0.8以上,预报结果相比观测总体偏低,随着预报时效的延长,24 h之后预报效果略有下降。相比人工观测,能见度预报结果与自动观测能见度更加接近,对持续性低能见度过程预报与实况吻合较好,对于小时能见度低于10 km的分级检验显示,预报准确率从77%左右逐级下降,2 km以下在40%左右。2014年APEC期间,系统很好地预报出北京地区空气质量指数、PM_(2.5)浓度和能见度的时空演变特征,为APEC期间环境气象预报服务提供了有力的技术支撑。

Beijing Regional Environmental Meteorology Prediction System（BREMPS） is established by coupling BJ-RUC,WRF-Chem and preferred visibility parameterization scheme.The performance test with observations in 2014 and Asia-Pacific Economic Cooperation（APEC） period shows that BREMPS has a good forecasting ability for two important elements in air quality and haze forecasting in Beijing and surrounding area,PM（2.5） concentration and visibility.Correlation coefficients of PM（2.5） between forecasted and observed values reaches above 0.6 at most sites,and even reaches above 0.8 at some sites in Beijing.Forecasted values generally underestimate the PM（2.5） concentrations with a regional average normalized mean bias of-15%.The forecast performance shows slight decrease after 24 forecasted hours.Comparing with the regional average,the forecast performance is best in Beijing urban area and northern of Hebei.The forecasted PM（2.5） concentration agrees well with the observation in Beijing area.Correlation coefficients of PM（2.5）concentrations between forecasted and observed values in 48 forecasted hours are about 0.77 in urban area.The normalized mean bias is generally in a range of-26%.The correlation coefficient in rural area is higher than that in urban area.However,the mean bias is also higher in rural area.That is probably attributed to the inaccuracy of the emission information in these areas.The forecast performance is better in spring,autumn and winter,during which the correlation coefficient between forecasted and observed values mostly ranges from 0.7 to 0.9,and the normalized mean bias is within 20%.The forecasted visibility is closer to automatic measurements than artificial observations.Forecasted values are in good agreement with observations during sustained low visibility synoptic processes.For the hourly visibility lower than10 km,the accuracy of forecast is 77%,which decreases with the reduction of visibility and reaches 40%when the visibility is lower than 2 km.BREMPS shows good forecast performance during the APEC period,when the temporal evolutions of AQI and visibility,and spatial distribution of PM（2.5） concentrations are well forecasted,providing strong support for the environmental meteorology forecast service.