区域优化Peterson模型的气溶胶光学厚度估算分析

Estimating Aerosol Optical Depth Based on Regional Optimized Peterson Model

气溶胶光学厚度作为描述气溶胶光学特性的重要参数之一,被广泛应用于空气质量监测及辐射传输模型的大气订正等研究中。卫星遥感可快速反演获取大范围气溶胶信息,但其产品通常因云覆盖或暗目标算法等原因而存在空间覆盖率较低的问题,且产品时相受限于卫星过境时间。水平能见度作为描述气溶胶光学特性的另一重要参数,由分布广泛的气象台站一日8次固定时间多次发布。建立水平能见度与气溶胶光学厚度的转换关系,可实现对卫星反演气溶胶光学厚度的有益补充。本文利用2001-2009年的MODIS气溶胶光学厚度产品与中国华东地区71个气象台站的水平能见度数据,对描述两者转换关系的Peterson模型进行区域优化。采用分区域高斯曲线拟合的方法,对影响转换精度的主要参数气溶胶标高随时间变化规律开展研究和模拟。利用2010年数据对优化模型进行精度及区域适用性验证。结果表明,优化后模型的气溶胶光学厚度估算均方根误差为0.31,低于原模型误差;精度基本上与单站点优化模型一致,但在实用性方面优于单站点优化模型。

Aerosol optical depth （AOD） is one of the most important parameters in describing aerosol optical characteristics. It is widely used in researched on air quality monitoring and atmospheric correction based on radiative transfer model. A wide range of aerosol information can be fast obtained and retrieved using satellite remote sensing technology. While, satellite retrieved aerosol product has low spatial coverage because of the limitation of cloud coverage and dark target algorithm, and the time phase of the product is limited by satellite overpass time. Horizontal meteorological visibility （HMV） is another very important parameter in describing aerosol optical characteristics. It is observed 8 fixed times one day by widely distributed meteorological sites in China. It can be a good supplement of satellite retrieved aerosol optical depth data by building the convention relationship between aerosol optical depth data and horizontal meteorological visibility data. In this study, MODIS aerosol op- tical depth product from 2001 to 2009 and horizontal meteorological visibility data from 71 meteorological sites located in Eastern China were chosen as the research data. The research data was used to optimize Peterson mod- el which describes the relationship between aerosol optical depth and horizontal meteorological visibility. Aerosol scale height is the main parameter that influences the conversion accuracy between aerosol optical depth and horizontal meteorological visibility. By regional Gaussian fitting method, the temporal trends of aerosol scale height were fitted. The aerosol optical depth data and horizontal meteorological visibility data in 2010 were used to validate the accuracy and the regional applicability of the optimized model. The result shows the RMSE of aerosol optical depth estimation of optimized model is 0.31, which is lower than that of Peterson model. Compared with the single site optimized model, the accuracy is similar, while the application of regional optimized model is better than the single site optimized model.