基于CFOSAT散射计的海冰识别方法研究

Sea ice identification based on CFOSAT scatterometer

中法海洋卫星散射计(CSCAT)丰富的观测几何信息为极地海冰遥感提供了新的机遇。本文提出一种适用于CSCAT的贝叶斯海冰识别算法,不需要构建海冰地球物理模式函数和计算后向散射系数离海冰地球物理模型函数(GMF)的距离,仅利用海面风场反演伴随的最小残差即可构建CSCAT海冰识别模型。研究结果与欧洲气象卫星组织的海冰边缘线产品进行了比较,表明2021年9月南极和北极区域逐日的海冰覆盖面积估计标准差分别为1%和7%,与其他卫星散射计的海冰识别结果基本一致。这种新的海冰识别方法具有模型参数少、处理速度快、检测结果可靠的优点,对卫星地面系统的业务化处理具有重要的借鉴意义。

The scatterometer onboard China-France Oceanography Satellite(CFOSAT)observes sea surface with abundant viewing geometries,opening up new opportunities for sea ice detection.This paper proposes a Bayesian sea ice detection method for the CFOSAT satellite scatterometer(CSCAT),which only uses the minimal inversion residual derived from the wind inversion procedure,hence it does not need to develop a sea ice geophysical model function(GMF)and to calculate the distance between CSCAT backscatters and sea ice GMF.The results are com-pared with the sea ice edge data from European Organisation for the Exploitation of Meteorological Satellites,which shows that the normalized standard deviation error of CSCAT daily sea ice extent is about 1%and 7%in the Antarctic and the Arctic,respectively,agreeing well with the prior scatterometers.In summary,the proposed meth-od is advanced in terms of model input parameters,processing speed and detection accuracy,so it is of great signi-ficance to the operational ice detection in the satellite ground segment.