研究报告
张波,宋国君,周芳.基于PM2.5监测点空间聚类的关中五市空气污染区域识别[J].环境科学学报,2021,41(3):797-805
基于PM2.5监测点空间聚类的关中五市空气污染区域识别
- Regional delimitation of PM2.5 pollution: A case study of five cities in Guanzhong Plain
- 基金项目:国家重点研发计划项目(No.2017YFC0212501)
- 周芳
- 2. 首都经济贸易大学城市经济与公共管理学院, 北京 100070;3. 城市群系统演化与可持续发展的决策模拟研究北京市重点实验室, 北京 100070
- 摘要:区域大气污染联防联控是空气质量管理的重要举措,准确识别空气污染区域对联防联控措施有重大意义.本研究采用陕西省关中五市(西安、咸阳、宝鸡、渭南、铜川)国控和省控全部90个监测点的小时级PM2.5浓度监测数据,运用邻接约束层次聚类方法对监测点进行空间聚类,并利用泰森多边形和曲线平滑等技术识别空气污染区域.结果表明:①关中五市空气污染存在跨行政区划的区域性特征,本研究识别出2个特征显著不同的空气污染区域(区域1和区域2);②区域2的PM2.5浓度在统计上显著高于区域1,且重度和严重污染天数也显著高于区域1;③空气污染区域与地形特征关系密切,区域1均为高海拔区县,而区域2均为低海拔区县.依据空气污染区域的不同特征,在区域污染程度存在显著差异时,应当采取不同等级的污染防控措施,以减少对关中五市43%的国土面积、23个区县、639万人及3355亿元国内生产总值的影响,使区域空气污染防控措施更加科学、合理与精准.同时,空气污染区域的划分对缺失数据和不同空气污染等级表现稳健.
- Abstract:Regional air pollution prevention and control is a very important policy for air quality management in China, and accurate identification of air pollution regions is of great significance to the policy. Based on the PM2.5 concentration data of 90 state controlled and province controlled monitoring sites in 5 cities in Guanzhong Plain, this study used a spatial cluster algorithm with connection constrain to group all the sites and used voronoi diagram to generate the PM2.5 pollution regions. The result shows:①There exists two pollution regions in the five cities with significant different characteristics of PM2.5. ②Region 2 was much more polluted than region 1 and heavy polluted and severe polluted occurred more frequently. ③PM2.5 pollution regions were closely related to terrain features, region 1 was much higher than region 2 in altitude. Response level should be different when there was significant difference of PM2.5 pollution in these regions. 43% of area, 23 counties, 6.39 million people and 335.5 billion GDP could be less impact if the air quality management were implemented regionally according to the regional characteristic. The solution presented is robust of missing data and different grade of pollution.