Scholarlink

<p>Rapid development in Saudi Arabia has made numerous cities vulnerable to floods caused by intense rainstorms due to recent climatic changes. Historical records indicate that Makkah has experienced over 121 flood episodes in 14 decades. Approximately 60% of these floods are severe, resulting in significant loss of human life and property. To develop a flood susceptibility map for Makkah City to mitigate the destructive effects of floods, the performance of four machine learning algorithms were tested: Logistic regression (LR-linear), support vector machine (SVM- margin-based nonlinear), extreme gradient boosting (XGB- tree-based), and an ensemble model (stacking of LR, SVM, and XGB). Geospatial datasets combined geomorphological, geological, environmental, hydrological, meteorological, and anthropogenic aspects to create 12 flood-conditioning factors and flood inventory data (556 flood and 556 non-flood datapoints). These conditioning factors are elevation, distance to stream, slope, aspect, lithology, landforms, relative-slope-position, topographic-wetness-index, slope-length, rainfall, plan curvature, and profile curvature. The flood susceptibility models were trained and validated using 70% and 30% of the flood inventory data. Multicollinearity analysis showed that all 12 conditioning factors were incorporated into the modeling process. Our findings concluded that the XGB algorithm and ensemble model predicted the most among machine learning models, with AUC values of 94.7% and 96.3%, respectively. Variable importance analysis revealed that the distance from streams, slope angle, topographic wetness index, rainfall, and landforms significantly affected flood prediction, suggesting a strong driving process. Anthropogenic activities (urban and infrastructural) were superimposed above the best susceptibility model to produce vulnerability maps. Furthermore, utilizing the ensemble model can enhance the robustness of a flood susceptibility model and incorporate it into the decision makers' and planners' strategies in developing sustainable practices for the area can improve urban and infrastructure resilience to flood hazards.</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

From ancient to recent floods: advances in flood susceptibility modeling and vulnerability, Makkah, Saudi Arabia

  • 期刊论文

  • 发表日期：

    2025年12月22日

查看全文
  • Bosy A. El Haddad,
  • Ahmed M. Youssef,
  • Hamid Reza Pourghasemi

摘要

Rapid development in Saudi Arabia has made numerous cities vulnerable to floods caused by intense rainstorms due to recent climatic changes. Historical records indicate that Makkah has experienced over 121 flood episodes in 14 decades. Approximately 60% of these floods are severe, resulting in significant loss of human life and property. To develop a flood susceptibility map for Makkah City to mitigate the destructive effects of floods, the performance of four machine learning algorithms were tested: Logistic regression (LR-linear), support vector machine (SVM- margin-based nonlinear), extreme gradient boosting (XGB- tree-based), and an ensemble model (stacking of LR, SVM, and XGB). Geospatial datasets combined geomorphological, geological, environmental, hydrological, meteorological, and anthropogenic aspects to create 12 flood-conditioning factors and flood inventory data (556 flood and 556 non-flood datapoints). These conditioning factors are elevation, distance to stream, slope, aspect, lithology, landforms, relative-slope-position, topographic-wetness-index, slope-length, rainfall, plan curvature, and profile curvature. The flood susceptibility models were trained and validated using 70% and 30% of the flood inventory data. Multicollinearity analysis showed that all 12 conditioning factors were incorporated into the modeling process. Our findings concluded that the XGB algorithm and ensemble model predicted the most among machine learning models, with AUC values of 94.7% and 96.3%, respectively. Variable importance analysis revealed that the distance from streams, slope angle, topographic wetness index, rainfall, and landforms significantly affected flood prediction, suggesting a strong driving process. Anthropogenic activities (urban and infrastructural) were superimposed above the best susceptibility model to produce vulnerability maps. Furthermore, utilizing the ensemble model can enhance the robustness of a flood susceptibility model and incorporate it into the decision makers' and planners' strategies in developing sustainable practices for the area can improve urban and infrastructure resilience to flood hazards.