首页 > , Vol. , Issue () : -
MODIS 3km DT(Dark Target)卫星气溶胶光学厚度(Aerosol Optical Depth, AOD)数据产品已广泛应用于大气污染监测，但受反演方法限制，该数据产品像元缺失严重、时空覆盖度低、精度偏低。相比，MODIS 10km DT_DB_Combined AOD数据产品因融合DT和DB(Deep Blue)两种反演算法，一定程度上可弥补MODIS 3km DT AOD数据产品在时空覆盖度与精度方面的缺陷，但分辨率偏低。此外，受气溶胶组分来源的季节变化与地表反射率估算的季节性误差影响，MODIS AOD数据产品精度同时也存在季节性特征。 本文由此以京津冀为试验区，顾及AOD季节变化特性，开展MODIS 10km DT_DB_Combined AOD数据产品偏差纠正下的地统计反演模拟(Bias-corrected Geostatistical Inverse Model, BGIM)降尺度算法研究。试验同时引入AERONET地基观测数据和MODIS 3km DT AOD数据产品作为降尺度结果的绝对与相对验证标准。 结果表明：季节偏差系数纠正下生成的MODIS 3km DT_DB_Combined AOD与10km DT_DB_Combined AOD、3km DT AOD数据产品的绝对精度相当，验证R2分别为0.79、0.70、0.71；且相比MODIS 3km DT AOD数据产品，季节偏差系数纠正下的MODIS 3km DT_DB_Combined AOD数据与其相关系数可达0.93；在时间覆盖度和空间覆盖度方面可分别提升11.21%和11.44%，其中春、冬两季空间覆盖度提升效果尤为显著。 研究结果证实BGIM降尺度算法可有效估算MODIS 3km DT_DB_Combined AOD数据，提高MODIS 3km AOD产品的时空覆盖度，并同时抑制原有MODIS 10km DT_DB_Combined AOD数据产品的季节性高估现象。
MODIS 3km DT(Dark Target) satellite aerosol optical depth(AOD) data products have been widely used in air pollution monitoring. However, due to the limitations of inversion method, MODIS 3km DT AOD products have a serious shortage of pixel, low temporal and spatial coverage and limited accuracy. By contrast, MODIS 10km DT_DB_Combined AOD products integrate two inversion algorithms, Dark Target and Deep Blue(DB), to some extent, which can make up for the shortcomings of MODIS 3km DT AOD data products in terms of temporal and spatial coverage and accuracy, but the resolution is low. In addition, affected by the seasonal variation of aerosol component sources and the seasonal error of surface reflectance estimation, the accuracy of MODIS AOD data products also have seasonal characteristics. This paper takes Beijing-Tianjin-Hebei region as the experimental area and takes into account the seasonal variation characteristics of AOD to conduct a Bias-corrected Geostatistical Inversion Model (BGIM) downscaling algorithm research using MODIS 10km DT_DB_Combined AOD products. AERONET ground-based observation data and MODIS 3km DT AOD data products are introduced as absolute and relative verification standards for downscaling data. The results show that: the absolute accuracy of the seasonal corrected MODIS 3km DT_DB_Combined AOD data, 10km DT_DB_Combined AOD and 3km DT AOD data products are pretty close, verification R2 are 0.79, 0.70, 0.71, respectively. Compared with MODIS 3km DT AOD data products, the correlation coefficient of the seasonal corrected MODIS 3km DT_DB_Combined AOD data is as high as 0.93, and the temporal coverage and spatial coverage are increased by 11.21% and 11.44%, respectively, among which the spatial coverage in spring and winter is particularly significant. The results confirm that the BGIM downscaling algorithm can effectively estimate MODIS 3km DT_DB_Combined AOD data, improve the spatial and temporal coverage of MODIS 3km AOD products, and suppress the seasonal overestimation of the original MODIS 10km DT_DB_Combined AOD products.