基于K-means聚类分析法的肇庆市干季PM2.5污染天气分型研究

通过分析肇庆市2013—2018年国控大气环境监测站的PM_(2.5)连续监测数据,发现肇庆市区PM_(2.5)浓度在干季(10月—次年4月)明显高于其余月份,轻度以上污染基本发生在干季,且PM_(2.5)浓度对年总浓度贡献达70.8%.基于Era-interim再分析资料采用K-means聚类分析法对2013—2018年干季逐日的海平面气压和10 m水平风进行分型,揭示了肇庆市易出现PM_(2.5)污染的6种大气环流形势,包括冷锋前部(CF)、变性高压脊(THR)、脊后槽前型(BRFT)、高压底后部(HSW)、弱冷高压脊(HR)和台风外围型(TP).2013—2016年易污染天气型影响天数呈明显减少趋势,2017—2018年呈增加趋势.不同天气型PM_(2.5)浓度与局地气象要素相关性不一致,其中CF、HR、HSW、TP天气型与湿度相关性最好,THR与风速、BRFT与气压相关性最好.PM_(2.5)污染除BRFT天气型主要以本地排放累积影响为主,其余易污染天气型存在不同尺度的外来输送影响,HSW、HR主要来自广州、清远、韶关, CF主要来自佛山、中山,THR来自广州、清远、佛山.同一污染天气型在不同月份的污染影响差异较大,其中HSW、THR污染型主要影响1月和10月,CF为1月和12月,HR为2月和12月,TP为10月,BRFT为1月和10—11月.不同年份的同一月份造成不同程度的PM_(2.5)污染除了排放影响,还与天气环流类型和同一天气型下的局地气象要素密切相关.

Objective synoptic classification on PM2.5 pollution during dry season based on K-means in Zhaoqing

By analyzing the hourly average PM_2.5 concentration data from the state-controlled atmospheric environmental monitoring stations from year 2013 to 2018 in Zhaoqing, it was found that PM_2.5 concentration in dry season (from October to next April) was significantly higher than other months, and the exceeding standard days almost occurred in above months. PM_2.5 mass concentration in these seven months contributed 70.8% to the total PM_2.5 mass concentration. Based on the sea level pressure and 10 m wind of Era-interim reanalysis dataset, K-means cluster analysis method was applied for synoptic weather classification for the dry seasons of year 2013 to 2018. Six weather types, including front of cold front(CF), transformed cold high ridge(THR), frontal low trough and behind ridge(BRFT), south-west to high(HSW), weak cold high ridge(HR) and peripheral subsidence of typhoon (TP), favor the accumulation of air pollutants, resulting in high PM_2.5 pollution. The number of days controlled by the above six circulation patterns showed a significant decrease before 2016, but an increase during 2017 and 2018. The correlation between PM_2.5 concentration and local meteorological parameters was inconsistent in different synoptic situations. Under the control of BFRT, meteorological condition was mainly conductive for accumulation of local air pollutants. However, under the control of other five weather patterns, regional transport can be an important source of PM_2.5 pollutants in Zhaoqing. Under HSW and HR, particulate pollutants mainly came from Guangzhou, Shaoguan and Qingyuan, while from Foshan and Zhongshan under CF, and from Guangzhou, Qingyuan and Foshan under THR. The same weather type could have different effects on air pollution in different months. PM_2.5 pollution was more severe significantly in January and October under HSW and THR, and January and December under CF, February and December under HR, October under TP, January and October to November under BRFT. Different degrees of PM_2.5 pollution in the same month of different years resulted from not only emissions, but also different weather types and local meteorological factors.