Scholarlink

Accurate estimation of melt pond depth is pivotal for comprehending the dynamics of sea ice melting. In this study, we employed a Multilayer Perceptron (MLP) model based on Landsat-8 and ICESat-2 imagery to estimate the depth of melt ponds. Compared to currently existing methods, the MLP model exhibits higher accuracy in estimating melt pond depth. This offers a new perspective for accurately assessing the depth of melt ponds and provides a viable reference for estimating the melting rate of sea ice in the future. Additionally, the study found that the primary distribution area of deep melt ponds in the Canadian Arctic Archipelago is shifting southward, and their depth is decreasing year by year.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimation of Melt Pond Depth Based on Deep Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zhihao Wang,
  • Shuaishuai Lyu,
  • Xiaoyi Shen

摘要

Accurate estimation of melt pond depth is pivotal for comprehending the dynamics of sea ice melting. In this study, we employed a Multilayer Perceptron (MLP) model based on Landsat-8 and ICESat-2 imagery to estimate the depth of melt ponds. Compared to currently existing methods, the MLP model exhibits higher accuracy in estimating melt pond depth. This offers a new perspective for accurately assessing the depth of melt ponds and provides a viable reference for estimating the melting rate of sea ice in the future. Additionally, the study found that the primary distribution area of deep melt ponds in the Canadian Arctic Archipelago is shifting southward, and their depth is decreasing year by year.