FY-3C/VIRR海表温度产品及质量检验

国家卫星气象中心FY-3C/VIRR（visible and infrared radiometer，可见光红外扫描辐射计）海表温度产品在云检测产品的基础上，采用多通道MCSST（multichannel SST）算法进行晴空区海温反演。该文详细介绍了海表温度产品算法、产品设计、质量控制及质量检验方法。FY-3C/VIRR海表温度产品包括5 min段原始投影海温和5 km全球等经纬度投影海温。设计逐像元的海温质量标识，将海温像元分为优、良、差3个等级，用户可根据应用目标选择海温的质量等级。与日最优插值海温OISST（optimum interpolation SST）相比，FY-3C/VIRR 2015年1月—2019年12月的5 min段海温质量检验结果表明:质量等级为优的海温，白天和夜间的偏差分别为-0.18℃和-0.06℃，均方根误差分别为0.85℃和0.8℃；白天海温均方根误差有季节性波动，夏季有的月份均方根误差大于1℃（如2015年7月、2016年7月和2019年7月）；在海温回归系数不变的条件下，夜间海温偏差的季节性波动与星上黑体温度相关显著。从一级数据质量、定位、业务运行状况等方面讨论引起海表温度产品异常的原因，为FY-3C/VIRR历史数据定位、定标和产品重处理及用户应用提供重要的参考信息。

FY-3C/VIRR Sea Surface Temperature Products and Quality Validation

Sea surface temperature (SST) products are generated at National Satellite Meteorological Center (NSMC) of China Meteorological Administration (CMA) from the visible and infrared radiometer (VIRR) on board FY-3C polar orbiting satellite. The production chain is based on FY-3C/VIRR cloud mask products and a classical multichannel SST (MCSST) algorithm is applied to derive SST in cloud-free zones. Operational MCSST procedures and products are described in detail. FY-3C/VIRR SST products are generated in satellite projection at full resolution in 5-minute granule, and in synthetic fields remapped onto a regular world grid at 0.05 degree resolution (5 km).The quality index (QI) information is delivered with each pixel to provide information about the conditions of the processing. They include in particular a quality level in the last two bits of QI (saved in a 8-bit CHAR) for each pixel defined as follows: Excellent, good, bad and unprocessed (cloud, land, no satellite data etc.). Users can select the SST data with certain quality level according to their application purposes (e.g., for climate-related studies, only the SST data with excellent quality level in the time series are used, and for identifying and tracking specific ocean features, users may be more tolerant of lower-quality SST data). The matchup database (MDB) combining FY-3C/VIRR measurements and buoy measurements is built on a routine basis. Validation methods and results are described in detail. The performance of SST retrievals is characterized with bias and root mean square error (RMSE) with respect to Reynolds L4 daily analysis (OISST). The validation bias and RMSE for FY-3C/VIRR operational granule SST with excellent quality level between January 2015 and December 2019 is found to be -0.18℃ and 0.85℃ in day-time, -0.06℃ and 0.8℃ in night-time, respectively. For day-time, the RMSE fluctuates seasonally. Some monthly RMSE is greater than 1℃ in summer. The bias at night is found fluctuating seasonally highly correlated to the black body temperature on board FY-3C since January 2016, and the SST regression coefficient (SST_COEF_V3) is used ever since then. Causes of FY-3C/VIRR SST products anomaly is analyzed, such as L1 data abnormal (e.g., single event upset), navigation error and operational running environmental error. Above all, some important reference information are provided to users for using FY-3C/VIRR SST products and FY-3C/VIRR data re-geolocation, re-calibration and products reprocessing.