西北太平洋红外辐射计海表温度数据交叉比对分析

COMPARISON IN MULTI-INFRARED PRODUCTS OF SEA SURFACE TEMPERATURE IN NORTHWEST PACIFIC

本文将西北太平洋海域作为研究区域,以2003—2009年的三个海表温度(sea surface temperature,SST)红外产品(AVHRR Pathfinder/NOAA,MODIS/Terra和MODIS/Aqua)为研究对象,分别与Argo浮标数据进行了真实性检验,同时红外产品之间也进行了交叉比对分析。通过评定产品间的差异及使用条件,为融合产品数据源选取和权重分配提供参考依据,用以提高融合产品的数据质量。结果表明,三种红外数据与Argo浮标的平均偏差在±0.2°C之间,均方根误差小于0.8°C,且存在明显的季节性变化,白天的平均偏差均是夏季为正、冬季为负,夜间的平均偏差基本均为负偏差,冬季比夏季的偏差更大,冬季的均方根误差较小;三种红外数据之间的平均偏差在±0.1°C之间,均方根误差小于0.6°C;三个红外产品在空间上均能反映西北太平洋海域的海表温度变化趋势,三个产品之间无明显优劣差异;尽管红外数据的空间覆盖率偏低,但是它提供了高精度和高特征分辨率的数据产品,并弥补了近岸海域缺乏观测数据的不足。

Three infrared radiometer sea surface temperature（SST） products（Pathfinder/NOAA, MODIS/Terra, and MODIS/Aqua） and their sensors were compared and analyzed against Argo counterparts for Northwest Pacific in 2003–2009. Differences among products and their application conditions were revealed to provide information for fusion product selection and weight allocation. The results show that the bias between infrared products and those of Argo is ±0.2°C and root mean square error（RMSE） is within 0.8°C. Meanwhile, infrared data and buoy data displayed obvious seasonal and diurnal cycles. In daytime, bias is positive in summer and negative in winter, while in nighttime, bias is negative in overall, larger RMSE in winter and smaller in winter. Bias among the three infrared datasets is ±0.1°C, RMSE 0.6°C. In general, all the three products could reflect the trend of SST with no obvious difference. Although the spatial coverage of infrared data is smaller, they provide high accuracy and characteristic resolution products, with which data deficiency in coastal waters can be balanced.