结合卫星遥感的PM2.5浓度估算与站点异常分析

针对大气污染难监测、异常数据难分析等问题，充分发挥气象卫星多时相观测优势，以N维代价函数算法为依托进行逐小时气溶胶光学厚度反演。进一步以随机森林回归算法为基础，结合气象参数、地面监测参数等辅助变量进行近地面PM2.5浓度估算，并据此开展时空分布与浓度异常分析。江苏省2021年1–6月PM2.5浓度遥感估算精度验证结果显示，其相关性精度达到94%，偏差为5.59μg/m3，证明以卫星遥感数据为基础进行PM2.5浓度估算具有高可靠性与可行性。通过其建立的全区域、多时相监测体系可有效进行PM2.5浓度的时空分布分析，明晰PM2.5分布状况，助力空气污染治理管控。通过卫星估算结果与地面站点监测结果的对比分析，准确识别大泉街道数据低报事件，验证人为干扰数据采集工作，实现星地数据双向监督。文章研究证明，卫星遥感技术可有效支撑近地面PM2.5浓度估算与数据异常分析，推动新时代生态文明建设。

PM2.5 Concentration Estimation and Anomaly Analysis of Station Combined with Satellite Remote Sensing

Aiming at the problem in air pollution monitoring and abnormal data analysis, with taking advantage of the multi-temporal observation advantages of meteorological satellites, the hourly aerosol optical depth inversion is carried out based on the N-dimensional cost function algorithm. Further, combined with auxiliary variables such as meteorological parameters and ground monitoring parameters, the concentration of PM2.5 near the ground is estimated based on the random forest regression algorithm. After that, the temporal & spatial distribution and concentration anomalies are analyzed accordingly. The verification results of remote sensing estimation accuracy of PM2.5 concentration in Jiangsu Province from January to June 2021 showed that the correlation accuracy reached 94%, with a deviation of 5.59μg/m³, it is proved that the estimation of PM2.5 concentration based on satellite remote sensing is highly reliable and feasible. Through its established regional and multi-temporal monitoring system, it can effectively analyze the temporal and spatial distribution of PM2.5 concentration, clarify the distribution of PM2.5, and help control air pollution. Through the comparison and analysis of satellite estimation and ground station monitoring results, accurately identify the under-reporting event of Daquan Street data, verify the human interference data collection work, and realize two-way monitoring of satellite-ground data. The research shows that satellite remote sensing technology can effectively support the near-ground PM2.5 concentration estimation and data anomaly analysis, and promote the construction of ecological civilization in the new era.