Scholarlink

<p>Changes in sea level can affect coastal erosion and inundation. Sea level change can be understood from sea surface height (SSH), which is the instantaneous sea surface height measured using satellite altimetry. Accurate SSH projections help in designing coastal infrastructure and disaster preparedness strategies. The primary objective of this study is to identify the climatic variables that significantly influence the sea surface height (SSH), perform statistical downscaling through different scaling approaches, and select the best-suited general circulation model (GCM) for future projection of SSH under climate change scenarios. Correlation analysis and variance inflation factor were used to examine the relationship between SSH and predictors. In this study, SSH is downscaled using long short-term memory (LSTM) and support vector machine (SVM) techniques for three locations, Jaitapur, Kochi, and Mumbai, situated along the western coast of India. The variables that substantially impacted SSH in Kochi were sea surface temperature, sea surface pressure and salinity, while in Jaitapur and Mumbai, sea surface pressure and wind speed were the influencing variables. These variables have been scaled using different feature scaling techniques, such as standardization, Min–Max scaling, Max-Abs scaling, and robust scaling before downscaling. The findings show that the LSTM outperforms SVM for the Min–Max scaling approach. Future projections up to 2100 were made by selecting suitable GCMs for each specific site for two climate change scenarios, SSP1-2.6 and SSP5-8.5.</p>

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

Projections of sea surface height under climate change scenarios using machine learning approaches

  • 期刊论文

  • 发表日期：

    2026年3月12日

查看全文
  • Akhila Unni,
  • S. K. Pramada

摘要

Changes in sea level can affect coastal erosion and inundation. Sea level change can be understood from sea surface height (SSH), which is the instantaneous sea surface height measured using satellite altimetry. Accurate SSH projections help in designing coastal infrastructure and disaster preparedness strategies. The primary objective of this study is to identify the climatic variables that significantly influence the sea surface height (SSH), perform statistical downscaling through different scaling approaches, and select the best-suited general circulation model (GCM) for future projection of SSH under climate change scenarios. Correlation analysis and variance inflation factor were used to examine the relationship between SSH and predictors. In this study, SSH is downscaled using long short-term memory (LSTM) and support vector machine (SVM) techniques for three locations, Jaitapur, Kochi, and Mumbai, situated along the western coast of India. The variables that substantially impacted SSH in Kochi were sea surface temperature, sea surface pressure and salinity, while in Jaitapur and Mumbai, sea surface pressure and wind speed were the influencing variables. These variables have been scaled using different feature scaling techniques, such as standardization, Min–Max scaling, Max-Abs scaling, and robust scaling before downscaling. The findings show that the LSTM outperforms SVM for the Min–Max scaling approach. Future projections up to 2100 were made by selecting suitable GCMs for each specific site for two climate change scenarios, SSP1-2.6 and SSP5-8.5.