中国地区MODIS Terra/Aqua MAIAC气溶胶光学厚度(AOD)产品反演误差对比分析  

Comparative analysis of retrieval errors of MODIS Terra/Aqua MAIAC aerosol optical depth(AOD)products in China

作　　者：陈翔[1,2] 汪洋 周佩 张丽坤 徐玲琳 CHEN Xiang;WANG Yang;ZHOU Pei;ZHANG Likun;XU Linglin(Institute of Geography,Fujian Normal University,Fuzhou 350117;School of Geographical Sciences,Fujian Normal University,Fuzhou 350117)

机构地区：[1]福建师范大学地理研究所,福州350117 [2]福建师范大学地理科学学院,福州350117

出　　处：《环境科学学报》2023年第7期220-232,共13页Acta Scientiae Circumstantiae

基　　金：国家自然科学基金(No.42001295);福建省公益类科研院所专项(No.2020R11010009-1);福建省自然科学基金(No.2022J01132098)。

摘　　要：最新的多角度大气校正(MAIAC)气溶胶光学厚度(AOD)反演算法已应用于Terra和Aqua两颗卫星上的中分辨率成像光谱仪(MODIS),在优化反演精度的同时将AOD产品的空间分辨率提高到1 km级.本研究将2003—2020年重叠期间Terra和Aqua中国地区MAIAC AOD产品与全球气溶胶自动观测网络(AERONET)AOD数据进行对比验证,分析该产品在中国地区的适用性和两颗卫星产品间的差异性,并探讨邻近云像元与观测几何对MAIAC AOD产品的影响.结果表明:MAIAC AOD产品在中国地区有较好的适用性,Terra和Aqua两颗卫星性能表现基本相近,整体验证结果有超过70%的数据落在期望误差(EE)范围内(EET=73.12%,EEA=70.83%).研究发现邻近云像元是影响AOD产品精度的重要因素之一,邻近云像元会造成AOD反演高估,且这种现象随着邻近云像元的数量而增加.经过质量控制剔除邻近云像元的影响,发现观测几何也会引起产品间的偏差,过小或过大的太阳天顶角会造成更大的平均偏差.总体而言,中国地区MODIS Terra/Aqua MAIAC AOD产品间差异较小,Aqua AOD数据平均偏差更小(MBT=0.0111,MBA=0.0048),年际间变化也更加稳定,更适合应用于长时序分析研究,但产品使用过程中应考虑临近云像元和观测几何对产品精度的影响.The latest Multi-Angle Atmospheric Correction(MAIAC)aerosol optical depth(AOD)retrieval algorithm has been applied to the Moderate Resolution Imaging Spectroradiometer(MODIS)on the Terra and Aqua satellites,optimizing the retrieval accuracy while the spatial resolution is optimized to 1 km level.To verify the applicability of this product in China and the difference between the two satellite products,this study compared the MAIAC AOD products in China with Aerosol Robotic Network(AERONET)data during the overlapping time series of Terra and Aqua from 2003 to 2020.The comparison and verification were carried out to discuss the influence of adjacent cloud pixels and observation geometry on the MAIAC AOD product.The results show that the MAIAC AOD product possesses good applicability in China,the performance of the two satellites is basically similar,and the overall verification results are more than 70%of the data falls within the expected error(EE)range(EET=73.12%,EEA=70.83%).It is also found that adjacent cloud pixels are one of the important factors affecting the accuracy of AOD products.The adjacent cloud pixels will cause overestimation of AOD retrieval,and this phenomenon increases with the number of adjacent cloud pixels.After quality control to eliminate the influence of clouds,it is found that the observation geometry can also cause errors between products,as well as a too-small or too-large solar zenith angle will trigger a larger average bias.In general,the differences between MODIS Terra/Aqua MAIAC AOD products in China are small.The mean bias of Aqua AOD data is smaller(MBT=0.0111,MBA=0.0048),and the interannual variation is more stable,which is more suitable for long-term time series analysis and research.However,the influence of adjacent cloud pixels and observation geometry on the product accuracy should be considered during the application of the product.

关 键 词：气溶胶光学厚度(AOD) 多角度大气校正(MAIAC) 中分辨率成像光谱仪(MODIS) 邻近云像元 气溶胶自动观测网(AERONET) 中国地区 

分 类 号：X513[环境科学与工程—环境工程]