###
DOI:
本文二维码信息
全国网格化多模式集成空气质量预报的初步建立
张天航, 迟茜元, 张碧辉, 张恒德, 江琪, 王继康, 饶晓琴, 谢超, 吕梦瑶, 安林昌, 南洋
(国家气象中心)
Development of gridding multi-model ensemble air quality forecast in China
Zhang Tianhang, CHI Xiyuan, ZHANG Bihui, ZHANG Hengde, JIANG Qi, WANG Jikang, RAO Xiaoqin, XIE Chao, LV Mengyao, AN Linchang, NAN Yang
(National Meteorogical Centre)
摘要
相似文献
本文已被:浏览 28次   下载 5
投稿时间:2019-04-29    修订日期:2020-02-11
中文摘要: 为降低单个模式预报的不确定性和提高多模式集成空气质量预报系统的精细化程度,笔者首先利用Cressman插值初步建立了我国0.25°×0.25°网格化污染物实况。再结合4套空气质量数值预报模式,通过均值集成、权重集成和多元线性回归集成分别逐格点建立了集成预报。最后在预报当天各单一模式和集成方法前50 d预报效果评估基础上,建立了最优集成预报。对2018年12月19~22日一次重污染过程中集成预报的PM2.5浓度评估结果显示:在污染较重时刻,最优集成预报与观测之间的NMB值在重污染地区保持在?20%~40%之内,对污染程度为良及以上区域的预报范围相较于单个模式更接近观测。整个过程中,最优集成在大部分污染区域与观测之间的NMB值为?20%~20%,RMSE值为35~75 μg m-3,R值大于0.4。相较于所有单一模式和其他集成方法,最优集成在全国最多的格点有着较高的总体评分。在污染最重区域的8个城市,最优集成预报的污染过程平均开始和结束时间分别比观测早1.8和6.9个小时。此外,笔者还提出未来需融合卫星反演和地表观测来提高网格化污染物实况的精细化程度,利用降尺度、主客观融合和滚动订正等方法进一步提高网格化多模式集成空气质量预报的准确率。
Abstract:To decrease the forecast uncertainties of single models and improve the refinement of multi-model ensemble air quality forecast system, the gridding observed pollution concentrations with resolution of 0.25°×0.25° was firstly established by Cressman interpolation method. Then, combined with four numerical air quality forecast models, mean, weighted and multiple linear regression were established in each grid, respectively. Finally, on the basis of evaluation results of single models and ensemble methods in previous 50 days, a best ensemble was established. Evaluation results of PM2.5 concentrations in a severer pollution process during December 19 to 22, 2018 showed that at time with heavy pollution, the NMB values between best ensemble forecast and observations could also be maintained between ?20% and 40%. And the forecast coverage area by best ensemble with good and above pollution was closer to observation than those of single models. During the whole process, the NMB, RMSE and R values between forecasted PM2.5 concentrations by best ensemble and observation were from ?20% to 20%, from 35 to 75 μg m-3 and higher than 0.4, respectively, in most polluted areas. Among all single models and ensemble methods, number of girds over China with high total scores was the largest in best ensemble. In eight cities located in the most polluted region, the average start and end time of the pollution process by best ensemble forecast was 1.8 and 6.9 hours earlier than observation, respectively. In addition, we proposed that pollutant concentrations retrieval by satellite and observed from ground should be fused to improve the refinement of gridding observed pollution concentrations. And the methods of scale reduction, subjective and objective fusion and rolling correction should be used to further improve the forecast accuracy of gridding multi-model ensemble air quality forecast.
文章编号:201904290186     中图分类号:    文献标志码:
基金项目:国家重点研发计划项目(2016YFC0203301)、中国气象局预报关键项目〔YBGJXM(2018)06-2〕、中国气象局气象预报业务关键技术发展专项(YBGJXM2018-7A)、国家气象中心青年基金(Q201808)
引用文本:
张天航,迟茜元,张碧辉,张恒德,江琪,王继康,饶晓琴,谢超,吕梦瑶,安林昌,南洋,0.[en_title][J].Meteor Mon,():-.
Zhang Tianhang,CHI Xiyuan,ZHANG Bihui,ZHANG Hengde,JIANG Qi,WANG Jikang,RAO Xiaoqin,XIE Chao,LV Mengyao,AN Linchang,NAN Yang,0.Development of gridding multi-model ensemble air quality forecast in China[J].Meteor Mon,():-.