基于主被动微波遥感融合的北极海冰密集度反演

Arctic Sea Ice Concentration Retrieval Based on Active and Passive Microwave Data Fusion

  • 摘要: 北极海冰快速变化,通过反照率反馈、热量交换、洋流影响海冰-大气-海水之间的热量、质量和动量交换,加速全球气候变化。利用多源遥感得到的高分辨率海冰密集度产品对监测全球气候变化和航道开发具有重要研究意义和应用价值。针对高时空分辨率海冰密集度产品研制需求,首先利用高时间分辨率的被动微波遥感数据,基于全约束最小二乘方法估算海冰密集度,并将其作为融合密集度的初始化输入;然后基于高空间分辨率的合成孔径雷达数据构建条件随机场模型,获得冰水二分类结果,并将其作为融合模型的类别先验项;最后构建最大后验准则的主被动微波融合模型,实现高精度海冰密集度反演。与现有主流被动微波密集度3.125 km产品(ARTIST(arctic radiation and turbulence interaction study)sea ice algorithm, ASI)目视结果比较发现,主被动微波融合产品包含更丰富的海冰纹理特征,并且能体现冰间水道和海冰边缘区域等低密集度的细节信息。与1 km分辨率的中分辨率成像光谱仪海冰密集度相比,融合产品能获得时间连续的海冰密集度产品,且纹理特征更加清晰。定量分析结果表明,融合产品相对于ASI海冰密集度的平均偏差为(3.6±1.12)%,标准差为(3.5±0.64)%。时间序列分析的结果显示,海冰的增长速率与海冰最小值之间存在负相关关系。在2016—2022年期间(除2020年外),40天内的海冰面积累计增长速率保持相近水平。北极海冰增加的主要区域集中在拉普捷夫海域和东西伯利亚海域,这一现象与这些区域温度的急剧下降密切相关。

     

    Abstract:
    Objectives The rapid changes in the arctic sea ice have profoundly impacted energy exchange processes between the ice, atmosphere, and ocean, further accelerating global climate change through various mechanisms such as albedo feedback, heat exchange, and changes in ocean circulation. High-resolution sea ice concentration (SIC) products generated by multi-source remote sensing are of great significance and application value for monitoring global climate change and ship navigation. To address the need for high temporal and spatial resolution SIC products, passive microwave data with high temporal resolution is used to estimate SIC based on the least squares method, which serves as the initialization input for the proposed fusion model that combines passive and active data.
    Methods A conditional random field model is applied to active microwave synthetic aperture radar data to derive sea ice and water classification labels, which are used as class priors for the fusion model. The fusion model based on maximum a posteriori estimation integrates both active and passive microwave data to achieve high-precision SIC retrieval. Compared to the existing arctic radiation and turbulence interaction study (ARTIST) sea ice (ASI) algorithm for passive microwave concentration products, the fusion SIC product provides finer details of leads and marginal sea ice zones. Compared with moderate-resolution imaging spectroradiometer SIC at 1 km, the fusion product can obtain time-continuous SIC products, and the texture features are clearer.
    Results The results of quantitative analysis showed that the mean deviation of the fusion product relative to the ASI sea ice density was (3.6±1.12)% with a standard deviation of (3.5±0.64)%. The results of the time series analysis showed a negative correlation between the rate of sea ice growth and sea ice minimum. In addition, time series analysis reveals a negative correlation between the sea ice growth rate and the minimum sea ice extent. Except for 2020, the cumulative 40-day sea ice growth rates during 2016—2022 were relatively consistent.
    Conclusions The primary regions of sea ice increase were the Laptev and East Siberian Seas, where the rapid temperature drop contributed significantly to the observed ice dynamic.

     

/

返回文章
返回