利用多源数据对海冰密集度反演的算法验证

发布时间：2018-11-07 17:06

【摘要】：海冰密集度对全球气候变化研究有重要的意义,其反演结果的验证工作也被广泛关注,但结合多源数据反演,同时对两种算法验证的研究较少的现状,该文利用ASPeCt船测海冰密集度数据对Bootstrap算法和NASA Team(NT)算法基于SSM/I数据估算的海冰密集度精度进行验证,并与MODIS影像反演获得的海冰密集度进行对比。研究结果显示两种海冰密集度算法获得的反演结果与ASPeCt船测值偏差分别为2.26%和7.27%,均方根误差分别为11.39%和12.32%。相比之下,MODIS结果与ASPeCt船测海冰密集度比较得到偏差为3%,均方根误差为5.21%。Bootstrap算法、NT算法与ASPeCt船测值比较的偏差和均方根误差显示两种算法精度相近;由于MODIS数据分辨率与ASPeCt船测数据相近,所以其反演精度较优;但因时空分辨率的限制,各种结果都具有一定的不确定性。
[Abstract]:Sea ice intensity is of great significance to the study of global climate change, and the verification of the inversion results has been paid more attention. However, combined with the multi-source data inversion, there are few researches on the verification of the two algorithms at the same time. In this paper, the accuracy of sea ice density estimated by Bootstrap algorithm and NASA Team (NT) algorithm based on SSM/I data is verified by using ASPeCt ship sea ice density data, and compared with the sea ice density obtained by MODIS image inversion. The results show that the inversion results obtained by the two sea ice concentration algorithms are 2.26% and 7.27%, respectively, and the root mean square errors are 11.39% and 12.32 2%, respectively. By contrast, the deviation of the MODIS results compared with the sea ice density of the ASPeCt ship is 3 and the root mean square error is the 5.21%.Bootstrap algorithm. The deviation and the RMS error of the comparison between the NT algorithm and the ASPeCt ship survey value show that the accuracy of the two algorithms is similar. Because the resolution of MODIS data is similar to that of ASPeCt data, the inversion accuracy is better, but due to the limitation of spatial and temporal resolution, all kinds of results are uncertain.
【作者单位】：武汉大学中国南极测绘研究中心;极地测绘科学国家测绘局重点实验室;国家海洋局海洋-大气化学与全球变化重点实验室;
【基金】：测绘地理信息公益性行业科研专项资助项目(201412009) 国家海洋局海洋—大气化学与全球变化重点实验室开放基金课题(GCMAC1405) 国家高技术研究发展计划项目(2013AA12A301)
【分类号】：P731.15;P715.7

【相似文献】