Satellite-derived sea surface temperature (SST) presents discrepancies in both magnitude and temporal variability in the wind-driven coastal upwelling region off the Cape Frio (23°S) in Southeastern Brazil. Level 2 (L2) data from the Moderate Resolution Imaging Spectroradiometer (MODIS) Terra and Aqua, and Level 4 (L4) SST products are compared with in situ measurements. Strong positive biases up to 4°C are systematically observed in L4 products during upwelling events and decrease to less than 1°C in non-upwelling days. Lower biases were found in L2 MODIS SST, and the statistical performance of L4 SST products did not present significant improvements in days when high-resolution clear-sky observations from MODIS were available. This indicates that spatio-temporal interpolation and ingestion of microwave-based data sets in L4 SST analyses are likely the main cause of observed discrepancies during upwelling events. Improved L2 SST retrieval algorithms and L4 SST analyses that take into account upwelling-related regional characteristics are required for more accurate satellite SST products over coastal regions.

来自卫星的海面温度（SST）在巴西东南部的弗里奥角（23°S）附近的风力驱动的沿海上升流区中，在大小和时间变化方面均存在差异。将中等分辨率成像光谱仪（MODIS）Terra和Aqua的2级（L2）数据和4级（L4）SST产品进行了现场比较测量。在升流过程中，在L4产品中系统地观察到高达4°C的强正偏差，而在非升流期间则降至低于1°C。在L2 MODIS SST中发现较低的偏差，并且从MODIS获得高分辨率的晴空观测数据后，L4 SST产品的统计性能在几天内并未表现出显着改善。这表明L4 SST分析中基于时空的插值和基于微波的数据集的摄取可能是上升流事件期间观测到差异的主要原因。需要改进的L2 SST检索算法和L4 SST分析，并考虑到与上升流有关的区域特征，才能在沿海地区获得更准确的卫星SST产品。