Scholarlink

<p>Extreme weather events, especially heat waves, pose a significant risk to societies by affecting the habitability, well-being, and social life of people. In recent decades, the impact of heat stress on public health has become a global concern. Consequently, human biometeorological indices are widely used to evaluate the relationship between the outdoor environment and human well-being. This study focuses on spatio-temporal modelling of the heat index (HI) for over 33 years in Morocco. It further examines heat waves trends in the country using daily meteorological data from the Global Surface Summary of the Day (GSOD). Heat waves for each station were defined using the 90th percentile approach for the heat index. Subsequently, machine learning methods, such as K-Nearest Neighbors (KNN), Support Vector Regression (SVR), and Neural Networks (NN), were applied to spatialize the results. Our results indicate a significant increase in the occurrence and intensity of heat waves over the study period, with key metrics such as Heat Wave Number (HWN), Frequency (HWF), and Duration (HWD) exhibiting considerable spatial variability. Model comparison revealed that Neural Networks provided the most accurate predictions for HWN, KNN performed best for HWF, and SVR was the most reliable for HWD. This study demonstrates the effectiveness of machine learning tools for high-resolution spatial modeling of heatwave risk, providing actionable maps for climate-health monitoring and adaptation planning. It also fills an important gap by providing nationwide mapping of heat waves in Morocco using a biometeorological index.</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

High-resolution spatiotemporal characteristics of heatwaves across Morocco using machine learning approaches

  • 期刊论文

  • 发表日期：

    2025年10月24日

查看全文
  • Sara Essoussi,
  • Mohamed Oufrad,
  • Zine El Abidine El Morjani,
  • Abderrahmane Sadiq,
  • Arlindo Meque

摘要

Extreme weather events, especially heat waves, pose a significant risk to societies by affecting the habitability, well-being, and social life of people. In recent decades, the impact of heat stress on public health has become a global concern. Consequently, human biometeorological indices are widely used to evaluate the relationship between the outdoor environment and human well-being. This study focuses on spatio-temporal modelling of the heat index (HI) for over 33 years in Morocco. It further examines heat waves trends in the country using daily meteorological data from the Global Surface Summary of the Day (GSOD). Heat waves for each station were defined using the 90th percentile approach for the heat index. Subsequently, machine learning methods, such as K-Nearest Neighbors (KNN), Support Vector Regression (SVR), and Neural Networks (NN), were applied to spatialize the results. Our results indicate a significant increase in the occurrence and intensity of heat waves over the study period, with key metrics such as Heat Wave Number (HWN), Frequency (HWF), and Duration (HWD) exhibiting considerable spatial variability. Model comparison revealed that Neural Networks provided the most accurate predictions for HWN, KNN performed best for HWF, and SVR was the most reliable for HWD. This study demonstrates the effectiveness of machine learning tools for high-resolution spatial modeling of heatwave risk, providing actionable maps for climate-health monitoring and adaptation planning. It also fills an important gap by providing nationwide mapping of heat waves in Morocco using a biometeorological index.