摘要:
PM2.5是影响开封地区空气质量的首要污染物,利用卫星遥感手段可以快速获得PM2.5浓度的空间分布.通过采用过境开封市的GF-1卫星数据,获取气溶胶光学厚度,结合地面PM2.5监测数据与边界层高度、相对湿度和气温等辅助数据,采用多元线性回归,建立了基于GF-1的PM2.5遥感反演模型.研究表明,2015年6—9月GF-1数据反演得到的PM2.5浓度与地面监测结果较为接近,且有较高的相关性;加入地理加权回归能明显提高模型精度,较好地反映PM2.5的空间分布;但在PM2.5浓度较高时,该模型会出现低估现象.
Abstract:
PM2. 5 is the key air pollution for air quality of Kaifeng City. With remote sensing technology, the distribution of PM2. 5 concentration could be determined quickly. In this paper, the authors collected the aerosol optical depth (AOD) of GF -1, height of planetary boundary layer (HPBL), relative humidity (RH) and air temperature ( AT) over Kaifeng City and then, with multiple regression analysis, revised the coefficients of all variables. After that, the authors built the PM2. 5 retrieving model from GF-1 in Kaifeng City. The validation from June to September in 2015 showed that the PM2. 5 concentration from remote sensing was similar to that from four ground-level monitoring sites, and the correlation coefficient was higher than 0. 8. The result of geographically weighted regression ( GWR) was obviously better than that of no GWR. Nevertheless, when PM2. 5 concentration was high, the model would underestimate PM2. 5 concentration.
下载: