Validation of a SeaWiFS dust-correction methodology in the Mediterranean Sea: Identification of an algorithm-switching criterion
Validation of a SeaWiFS dust-correction methodology in the Mediterranean Sea: Identification of an algorithm-switching criterion
复制标题
DOI：
10.1016/j.rse.2009.08.007
复制
发表时间：
2009-12
影响因子：
13.5
通讯作者：
V. Banzon;H. Gordon;C. Kuchinke;D. Antoine;K. Voss;R. Evans
中科院分区：
文献类型：
--
作者：
V. Banzon;H. Gordon;C. Kuchinke;D. Antoine;K. Voss;R. Evans
将一种光谱匹配算法(SMA)应用于地中海西北部的SeaWiFS图像，该算法允许在存在沙尘气溶胶的情况下进行大气校正。目标是找到可用于识别SMA目标像素的标准，并深入了解该方法相对于标准SeaWiFS处理方案(STD)的准确性。这项工作也是首次使用现场数据对SMA进行验证。验证数据集包括从固定浮标站点和2004年8月高级光学特性实验(AOPEX)巡航期间从一艘船上收集的离水辐射。激光雷达船站和沿海AERONET站(法国)和Blida站(阿尔及利亚)提供了补充资料，它们分别记录了浮标附近和尘源附近的气溶胶情况。还提供了AERONET站点的反向气溶胶传输轨迹，从而能够确定潜在的尘埃源，特别是激光雷达观测到的气溶胶层。在研究期间，四次气溶胶事件影响了浮标附近，但事实证明，只有在尘埃占主导地位而不是仅仅存在尘埃的情况下，SMA反演才优于标准处理结果。适用于SMA应用的条件可以使用AERONET参数来定义。它们是高气溶胶光学厚度τa和低ngström指数α(或τa/α>；0.2)的组合。使用STD方法产生的等效SeaWiFS参数也得到了类似的结果，尽管阈值不同。由于选择SMA或STD处理更好的是在大气校正之前以每像元为基础应用该准则，因此提出了使用从SeaWiFS数据获得的与气溶胶模式无关的量的类似试验。因此，在适当的情况下，可以将SMA和STD处理应用于单个图像。
A spectral matching algorithm (SMA) that allows atmospheric correction in the presence of dust aerosols is applied to SeaWiFS imagery in the northwest Mediterranean Sea. The goal is to find criteria that could be used to identify SMA target pixels and to gain insights into the method's accuracy relative to the standard SeaWiFS processing scheme (STD). This work also represents the first validation of SMA using in situ data. The validation dataset includes water-leaving radiances collected from both a fixed buoy site and from a ship during the Advanced Optical Properties Experiment (AOPEX) cruise in August 2004. Supplementary information was provided by the ship LIDAR and coastal AERONET stations in Villefranche (France) and Blida (Algeria) that recorded aerosol conditions near the buoy and proximal to the dust sources, respectively. Backward aerosol transport trajectories were also available for the AERONET sites, allowing identification of potential dust sources, especially for aerosol layers observed by the LIDAR. Over the study period, four aerosol events affected the buoy vicinity, but SMA retrievals proved superior to standard processing results only when dust was dominant, rather than when dust was simply present. The conditions appropriate for an SMA application could be defined using AERONET parameters. They are a combination of high aerosol optical depth τaand low Ångström exponent α (or τa/α>0.2). Similar results are obtained using the equivalent SeaWiFS parameters produced by the STD method although the threshold value is different. Since it is preferable to apply the criterion on a per-pixel basis prior to atmospheric correction to select SMA or STD processing, an analogous test using aerosol model-independent quantities derived from SeaWiFS data is proposed. Thus, SMA and STD processing can be applied to a single image, where appropriate.