| 长序列北极海冰覆盖数据集对比分析 |
| |
| 引用本文： | 武胜利,刘健.长序列北极海冰覆盖数据集对比分析[J].海洋学报,2018,40(11):64-72. |
| |
| 作者姓名： | 武胜利 刘健 |
| |
| 作者单位： | 国家卫星气象中心, 北京 100081 |
| |
| 基金项目： | 国家重点研发计划（2018YFC1407204）；公益性行业（气象）科研专项（GYHY201306049） |
| |
| 摘 要： | 国家卫星气象中心使用2011年至今的风云三号卫星数据开发了一套基于Nasa Team2（NT2）算法的北极海冰密集度数据集，并可实时业务更新。将该数据集与其他国家不同机构业务运行并实时更新多种同类型数据集进行横向对比分析，其中包括:（1）美国冰雪中心基于Nasa Team（NT）算法以及SSM/I、SSMIS数据制作的1978年至今25 km分辨率全球极区海冰覆盖数据集；（2）美国冰雪中心基于Boot Strap（BS）算法以及SSM/I、SSMIS数据制作的1978年至今25 km分辨率全球极区海冰覆盖数据集；（3）美国NOAA基于多种卫星资料、地面观测数据以及海冰模型制作的2004年至今4 km分辨率北半球海冰覆盖数据集（IMS）。对比表明，上述数据集在北极地区不同的时空范围内存在一定的偏差。以分辨率较高的IMS数据集为基准，对其他3种长序列数据集进行初步评价，总体最大偏差超过100×104 km2，其中，NT2数据集过估较明显。经过与IMS数据集多年各月监测最大值的对比订正，NT2数据集过估情况得到改善。在此基础上的分析结果表明，NT、BS、NT2等3种数据集与IMS数据集相比，过估区域主要分布在海岸线附近，夏季过估比冬季更加明显，少估区域与算法、月份相关性明显，夏季少估面积也较冬季更大。NT、BS、NT2等3种数据集之中，NT2数据集与IMS数据集偏差最小，NT数据集次之，BS数据集与IMS数据集偏差最大。结果表明使用风云三号卫星数据的北极海冰覆盖数据集精度与国外3种同类型数据集相当。
|
| 关 键 词： | 海冰覆盖 长序列 数据集 |
| 收稿时间： | 2018/1/19 0:00:00 |
| 修稿时间： | 2018/5/9 0:00:00 |
Comparison of Arctic sea ice concentration datasets |
| |
| Wu Shengli and Liu Jian.Comparison of Arctic sea ice concentration datasets[J].Acta Oceanologica Sinica (in Chinese),2018,40(11):64-72. |
| |
| Authors: | Wu Shengli and Liu Jian |
| |
| Institution: | National Satellite Meteorological Center, Beijing 100081, China |
| |
| Abstract: | An arctic sea ice concentration dataset was operationally produced in National Satellite Meteorological Center (NSMC) using FY-3 data and Nasa Team2 algorithm. Currently, there are several long time series sea ice concentration data sets are operational produced in different organizations. Several typical operational data sets were compared with the FY-3 sea ice concentration data set include:(1) 1978-present global SIC data set using SSM/I, SSMIS data and Nasa Team (NT) algorithm from National Snow and Ice Data Center (NSIDC); (2) 1978-present global SIC data set using SSM/I, SSMIS data and Boot Strap (BS) algorithm from NSIDC; (3) 2004-present northern hemisphere sea ice coverage data (IMS) set using multi-remote sensing data, ship observation and model result from National Oceanic and Atmospheric Administration. Analysis results show that there are deviation between different data set in north polar area. Based on the higher resolution data set (IMS), we compared the other 3 kinds of data sets, the total biases are more than 1 million km2. For NT2 data set, there is a significant overestimate. After the bias correction, the overestimate of NT2 was improved. The further bias analysis found that most of the overestimated area of NT, BS, NT2 compared with IMS was located in land/sea border area. For both overestimated area and underestimated, summer error is larger than winter error. For the total bias, NT2 have the lowest value while BS have the highest value separately. Results show that the FY-3 Arctic sea ice concentration dataset have a similar accuracy with other operational sea ice concentration datasets. |
| |
| Keywords: | sea ice concentration long time series dataset |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋学报》浏览原始摘要信息 |
| 点击此处可从《海洋学报》下载免费的PDF全文 |
|
