摘要
利用兰州市2007-2009年供暖期PM2.5实测数据,统计分析观测期间PM2.5的超标率、超标倍数、浓度水平以及典型天气(浮尘、沙尘暴和灰霾)下PM2.5浓度日变化特征,并结合同期气象数据,运用多元逐步回归分析的方法建立兰州市供暖期PM2.5浓度预报模型,最后对模型进行验证。结果表明:兰州市2007-2009年供暖期PM2.5污染十分严重,2007-2009年PM2.5的超标率分别为59.2%、67.9%和68.8%,最大超标倍数分别达到2.88、3.17和3.60;2007-2009年供暖全期PM2.5浓度日变化呈"双峰双谷"型,浮尘、沙尘暴和霾天气下PM2.5浓度均远高于平均值,霾天气下PM2.5日均浓度值最高;兰州市供暖期PM2.5浓度预报模型的预报值与实测值的变化趋势基本一致,但对高浓度PM2.5模拟结果较差,PM2.5预报浓度准确率和等级准确率分别为70.2%和81.8%。
The diurnal variation of PM2.5 and its levels during heating period from 2007 to 2009 in Lanzhou were analyzed by statistic analysis method, and PM2.5 concentration forecast model was established with the meteorological data of the corresponding period by using multiple stepwise regression analysis method. Results showed that PM2.5 pollution was extremely serious, with the exceeding standard rates of PM2.5 in 2007, 2008 and 2009 as 59.2%, 67.9% and 68.8% respectively, while the maximum exceeding limit multiples as 2.88, 3.17 and 3.60 respectively. The curve of diurnal variation of PM2.5 showed two peak values and two valley values. PM2.5 concentrations in floating dust, sand storm and ash haze weather were much higher than average value with PM2.5 daily average concentration in ash haze weather as the highest. PM2.5 concentration variation trend of forecasting value was basically the same as the measured value, but the simulation result of high concentration PM2.5 was poor. The forecast accuracy and level forecast accuracy of PM2.5 were 70.2% and 81.8% respectively.
出处
《环境科学与技术》
CAS
CSCD
北大核心
2014年第4期80-84,共5页
Environmental Science & Technology
基金
国家自然科学基金重点项目(91025015)
关键词
PM2
5
供暖期
浓度监测
逐步回归
预报模型
PM2.5
heating period
concentration monitoring
stepwise regression
forecast model
