南极海冰总覆盖范围与面积的不确定性研究

Study on the uncertainty of the total sea-ice cover extent and area in the Antarctic

目前,被动微波遥感数据已经成为提取极区海冰密集度的重要手段,其获取的海冰密集度产品不可避免的存在误差,这样的误差进而会引起南极海冰总覆盖范围和总覆盖面积的不确定性.利用蒙特卡洛方法对海冰密集度产品进行误差的模拟,对随机模拟影像进行统计分析,利用变异系数等指标量化随机误差带给南极海冰总覆盖范围和面积的不确定性.实验发现,当海冰密集度误差较小时,海冰覆盖范围的不确定性相比面积更大;当误差较大时,结论则相反.并且两者的不确定性都与分辨率和季节有关,分辨率越高,不确定性则越小,冬季比夏季的不确定性要小.

Nowadays, the passive microwave remote sensing data has become one of the important means to extract the sea ice concentration （SIC） in polar regions. The SIC products, however, have different degree of errors which will cause the calculated total sea-ice extent and area uncertain. This paper simulated the error of the SIC product by using Monte Carlo method and analyzed the stochastic simulation images. Several indicators such as the coefficient of variation were used to quantify the uncertainty of total- extent and total-area caused by random error. This study revealed： that when the error of SIC product is small, the uncertainty of the total-extent is larger than that of the total-area; whereas, when the error of SIC product is large, the uncertainty of the total- are is larger than that of the total-extent; and the uncertainty of both total-extent and total-area is related to the resolution and season, namely, the higher the resolution is, the smaller the uncertainty would be, and the uncertainty in winter is smaller than that n summer.