利用多角度偏振数据反演华北地区灰霾气溶胶

摘要

For haze aerosol, the retrieval from multi-angle polarized data was studied. Based on ground-based measurements, we simulated multi-angle polarized satellite signal of the haze, and it was shown that by near infrared data, the concentration of heavy haze was recognized well. Then, aimed to multi-angle polarized signal of POLDER3, in consideration of aerosol components of the haze, the haze was retrieved by best fitting method of polarized reflectance. The experiment in Feb. 23, 2010 showed that, by our algorithm, the aerosol optical depth (AOD) of haze can be retrieved well. The AOD of our algorithm from 2009 to 2010 was close to that of ground-based measurements (correlation coefficient was about 0.9), and the retrieval error was about 10%. But AOD of our algorithm was lower than that of ground-based measurements, and the reasons may be the underestimation of coarse particle for multi-angle polarized signal, the dated data of land classification, the assumption that haze signal is linear added by that of aerorol componets, and so on.%针对灰霾气溶胶，利用多角度偏振数据进行了反演研究.在地面观测数据的基础上，模拟了卫星观测的灰霾多角度偏振信号，发现利用近红外波段偏振信息能更好地区分灰霾的轻重程度.在此基础上，针对POLDER3的多角度偏振数据，考虑灰霾气溶胶的组分构成，采用偏振反射率最优拟合的方法，建立了华北地区灰霾气溶胶的反演算法.2010年2月23日的算法实验表明，本算法在重霾发生时能较好地区分灰霾的轻重程度.而2009~2010年的反演结果与地面观测结果有较好的相关性(相关系数在0.9左右)，灰霾时AOD的反演误差大约在10%左右.但本算法反演结果要整体低于地面观测结果，原因可能是偏振信号对粗粒子不敏感以及地表类型数据更新较慢、灰霾各气溶胶组分的线性组合假定.