统计分析成都市空气质量指数(AQI)日数据与6项指标小时浓度数据,结合大气能见度与相对湿度日数据,采用非线性拟合等方法对成都市AQI与雾霾天气的关系进行研究,并分别分析能见度与相对湿度、AQI之间的线性和非线性关系.结果表明,成都市现阶段AQI和颗粒物污染密切相关,2013年和2014年AQI变化趋势大致相同,4-9月份较低,以良和轻度污染为主,秋冬季大气污染是一年中的高峰期;2014年的逐月AQI普遍较2013年低,一定程度上反映了空气质量的改善.2013年和2014年平均大气能见度变化趋势也大致相同,均表现为春夏较好,而秋冬较差,秋冬季是雾霾多发的季节;随着相对湿度的逐渐增加,大气能见度与AQI的非线性相关系数逐渐升高,具有较好的相关性;在不同的空气质量等级下,相对湿度与能见度的非线性相关系数RN普遍高于线性相关系数RL.本研究所得出的AQI与雾霾的非线性模型在判别霾日程度上准确度一般,而在判别雾日方面准确度极高,全年平均准确率达到98％以上,尤其在春季表现得更为明显,拟合值和观测结果吻合度达到100％.同时,该模型在判别霾日或非霾日方面较准确,尤其在雾霾多发的秋冬季节,非霾日的准确度最高分别可达89.44％和92.78％,霾日的准确度最高分别可达88.89％和85％.在季节判断上,分季度模型比全年模型更加准确.%The daily data of Chengdu AQI and the hourly concentration data of six indexes combined with the data of the air visibility and the relative humidity are collected and analyzed in this paper.Besides,this paper adopts the non-linear fitting methods to study the relationship between the Chengdu AQI and haze,and respectively discusses the relationship between the visibility and relative humidity,and the relationship between the visibility and AQI.The results show that the present AQI index conditions of Chengdu are closely related with the particle pollution,and that the trends of AQI indexes in 2013 and 2014 are basically the same,for between April and September in these two years the pollution is mostly mild and the indexes are both rather low;however the indexes reach its peak in the autumn and winter.Moreover,the monthly AQI in 2014 is generally lower than that in 2013,thus reflecting the improvement of air quality in a way.Besides,the trends of the average air visibility of these two years are also basically the same with better air quality in spring and summer while with poor air conditions in autumn and winter.Moreover,with the gradual increase of relative humidity,the nonlinear interaction coefficient between the air visibility and AQI is gradually increasing and is provided with relatively preferable relativity.Under different air quality degrees,the nonlinear interaction coefficient between the relative humidity and the visibility is generally higher than the linearly dependent coefficient between them.Besides,the obtained nonlinear model of AQI and haze has common accuracy in judging haze days but has high degree of accuracy in judging fog days with the annual average accuracy rate reaching more than ninety-eight percent.This high accuracy can be embodied apparently in autumn and winter when the fitting value one hundred percent coincides with the observed results.Simultaneously,this model is relatively accurate in judging haze days and non-haze days.In autumn and winter especially when haze is prevalent,the accuracy reaches eighty-nine point four four percent and ninety-two point seven eight percent.And the accuracy in judging haze days can even reach as high as eighty-eight point eight nine percent and eighty-five percent.Moreover,the quarterly model is more accurate than the yearly model in the seasonal judgment.
展开▼
