Scholarlink

<p>The rise in coastal flooding has been an issue of global concern, particularly to populations Living in low-lying coastal areas. It has resulted in several casualties, destruction of infrastructure worth billions of dollars, displacement of communities, and ecological damage. This necessitates a review to provide an appraisal of how machine learning can be used for coastal flooding management. The study considered articles published between 2000 and 2025 from academic databases. A systematic literature assessment was carried out based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure an all-inclusive and organized reviewing process. Findings revealed that traditional physical models have a number of weaknesses that include high data needs, calibration, validation, and scaling up. Machine Learning (ML) outperform conventional techniques by effectively handling intricate nonlinear relationships and large dataset burdens. Contemporary challenges to the use of ML in predicting coastal flooding include the requirement for extensive, high-resolution datasets and the risk of overfitting; besides, there are interpretability issues and difficulties associated with incorporating machine learning models into traditional hydrological and hydraulic frameworks. The prospects of ML in coastal flooding are good, as some recent improvements like deep learning architectures, transfer learning, reinforcement learning, and the integration of IoT, hybrid modeling approaches and remote sensing technologies are expected to change entirely how floods are predicted or managed. The study highlights hybrid models as a key future direction, combining the strengths of both ML and traditional methods to improve prediction accuracy and operational efficiency. The study concludes that integrating advanced approaches, such as hybrid models, can help address some of the limitations of conventional models, leading to improved accuracy and efficiency in coastal flood prediction and management. However, while machine learning models show great promise, they should be seen as complementary rather than as complete replacements for traditional methods.</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

Application of machine learning for coastal flooding

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月7日

查看全文
  • Abu-hanifa Babati,
  • Zaharaddeen Isa,
  • Auwal F. Abdussalam,
  • Saadatu Umaru Baba,
  • Bashariya Baba Mustapha,
  • Abubakar Sadiq Musa

摘要

The rise in coastal flooding has been an issue of global concern, particularly to populations Living in low-lying coastal areas. It has resulted in several casualties, destruction of infrastructure worth billions of dollars, displacement of communities, and ecological damage. This necessitates a review to provide an appraisal of how machine learning can be used for coastal flooding management. The study considered articles published between 2000 and 2025 from academic databases. A systematic literature assessment was carried out based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure an all-inclusive and organized reviewing process. Findings revealed that traditional physical models have a number of weaknesses that include high data needs, calibration, validation, and scaling up. Machine Learning (ML) outperform conventional techniques by effectively handling intricate nonlinear relationships and large dataset burdens. Contemporary challenges to the use of ML in predicting coastal flooding include the requirement for extensive, high-resolution datasets and the risk of overfitting; besides, there are interpretability issues and difficulties associated with incorporating machine learning models into traditional hydrological and hydraulic frameworks. The prospects of ML in coastal flooding are good, as some recent improvements like deep learning architectures, transfer learning, reinforcement learning, and the integration of IoT, hybrid modeling approaches and remote sensing technologies are expected to change entirely how floods are predicted or managed. The study highlights hybrid models as a key future direction, combining the strengths of both ML and traditional methods to improve prediction accuracy and operational efficiency. The study concludes that integrating advanced approaches, such as hybrid models, can help address some of the limitations of conventional models, leading to improved accuracy and efficiency in coastal flood prediction and management. However, while machine learning models show great promise, they should be seen as complementary rather than as complete replacements for traditional methods.