设为首页 加入收藏 登录旧版
江苏省空气质量预报准确度及改进方法分析
Analysis on the Accuracy of Air Quality Forecast and Its Improvement Method in Jiangsu Province
投稿时间:2016-11-20  修订日期:2017-03-07
DOI:10.19316/j.issn.1002-6002.2017.06.03
中文关键词:  江苏省  空气质量预报  预报模型  结果检验  误差统计  预报准确度
英文关键词:Jiangsu province  air quality forecast  forecast model  performance evaluation  error statistics  prediction accuracy
基金项目:科技部国家环境保护公益性行业专项(20151015);国家科技支撑计划课题(2014BAC22B04);江苏省环境监测基金(1624)
作者单位
张璘 江苏省环境监测中心, 江苏 南京 210036 
张祥志 江苏省环境监测中心, 江苏 南京 210036 
茅晶晶* 江苏省环境监测中心, 江苏 南京 210036 
杨雪 江苏省苏协环境技术研究院, 江苏 南京 210036 
曹天慧 江苏省环境监测中心, 江苏 南京 210036;中科院大气物理研究所, 北京 100029 
杨金德 南京恩瑞特实业有限公司, 江苏 南京 211106 
徐秀莉 南京恩瑞特实业有限公司, 江苏 南京 211106 
通讯作者:茅晶晶*  江苏省环境监测中心, 江苏 南京 210036  
摘要点击次数: 110
全文下载次数: 1
中文摘要:
      介绍了江苏省重污染天气监测预报预警系统以及大气重污染预警会商流程,将2015年13个地级市的模式预报、人工预报结果分别与实际观测值进行比较。结果表明:人工预报更准确,PM2.5日均值、臭氧日最大8 h平均值、AQI 3个指标人工预报和实况的相关性分别比模式预报高出12.8%、0.3%、11.4%,平均标准误差(MNE)分别低20.7%、3.1%、23.1%。依据国家空气质量预报技术指南评分办法,对各市2015年全年空气质量级别为"良"时进行评分。通过开展07∶00预报更新,使2015年上半年空气质量预报级别得分平均提高了0.9分,全年级别得分平均提高了2.6分;通过改进模式预报参数,使PM2.5日均预报值、臭氧日最大8 h平均预报值、AQI预报值和实际观测值的相关性比上年同期分别提高26.0%、5.0%、33.9%,MNE分别降低3.6%、31.3%、7.6%。
英文摘要:
      This article introduced the general structure of the serious air pollution monitoring forecasting system and the conference discussion procedure of serious air pollution early warning in Jiangsu province. By comparing with the predicted data against observations,the results indicated that the predicted data by engineers were more accurate than air quality forecast model of all 13 prefecture-level cities in 2015. The correlations of daily average PM2.5, daily maximum 8-hour average ozone and AQI between predicted and observed data by engineers were 12.8%, 0.3% and 11.4% higher than model forecast, respectively. The mean normalized error (MNE) of predicted data against observations by engineers were 20.7%, 3.1% and 23.1% lower than model forecast, respectively. According to the national forecast technical guide scoring method, every cities' scores are calculated when AQI level is "good". Two measurements could improve prediction accuracy, one of them was to updating forecast data at 07 o'clock in the next day's morning, as a result, air quality forecast level increased 0.9 point in the first half year of 2015, and increased 2.6 points for the whole year. The second method was to optimizing forecast model's parameter. The correlations between observed data and predicted data of daily average PM2.5, daily maximum 8-hour average ozone and AQI increased 26.0%, 5.0% and 33.9% respectively, MNE of three indicators decreased 3.6%, 31.3% and 7.6% respectively.
查看全文  查看/发表评论  下载PDF阅读器