2014—2016年海口市空气质量概况及预报效果检验

Verification of air quality forecasted by CUACE model in Haikou City during 2014 to 2016

本文主要基于CUACE模式在海口市的预报产品,结合2014年3月—2017年2月海口市AQI、PM_(2.5)、PM_(10)和O_3的实况资料进行预报效果检验.结果表明,①近3年海口市空气质量等级主要以优和良为主,但仍有少部分天数以PM_(10)、PM_(2.5)和O_3为首要污染物,分别占所有首要污染物天数的27.6%、29.5%和42.9%,其中O_3上升幅度较快.②CUACE模式能较好的模拟出AQI和3类污染物浓度的变化特征,其中PM_(2.5)的预报值与实测值最为接近,而PM_(10)和O_3普遍偏低.③日平均浓度的预报效果检验表明,PM_(2.5)的标准误差(RMSE)最小,AQI和PM_(10)次之,O_3最大.3个时次预报平均偏差(MB)和归一化偏差(MNB)均为负值,表明CUACE模式预报的污染要素浓度均偏低于实测值.④海口市空气质量为优等级时,TS评分最高;无首要污染物时,首要污染物预报的TS评分最高,但首要污染物为PM_(2.5)、PM_(10)或O_3时,TS评分均偏低.

Based on the forecast product of CUACE model output,this study verified the forecast efficiency of AQI,PM2.5,PM10 and O3 in Haikou City by the observed data from March 2014 to February 2017.In observations,air quality in Haikou City was mainly kept at excellent and good in these three years and the primary pollutants were PM2.5,PM10 and O3,which accounted for 27.6%,29.5%and 42.9%,respectively.Basic features of AQI,PM2.5,PM10 and O3 simulated by CUACE model were close to observations in Haikou City.The model reasonably simulates the concentration of PM2.5 but relatively underestimate the concentration of PM10 and O3,which can also be reflected by the Mean Bias(MB)and Mean Normalized Bias(MNB)values shown by CUACE forecasted at the three times.In addition,the Root Mean Square Error(RMSE)is the smallest for the daily averaged concentration of PM2.5,following AQI and PM10,and is largest for O3.As a further step,the model performs the best when the air quality is excellent or there is no primary pollutant in Haikou City,as measured by the TS score.The TS score is relatively low when the primary pollutants were PM2.5,PM10 or O3.