Scholarlink

<p>Floods are natural disasters that are intensified by climate variability and land use changes, impacting the social and economic well-being of people worldwide. These floods and vulnerability mapping have explored using conceptual and physical hydrological models, which require large datasets, but often fail to capture the complex interactions between environmental and climate change. This research addresses this gap by applying a machine learning approach, implemented in the Python Programming Language, to map flood susceptibility zones under climate change and land use changes in the Ribb and Gumara Catchments, Lake-Tana Sub-Basin, Ethiopia. Land use maps and flood susceptibility projections for 1991, 2023, and 2050, under SSP 245 &amp; SSP 585 scenarios, were produced using the Random Forest (RF) algorithm, Sentinel-2 imagery, and temporal and geospatial datasets. The model’s performance was evaluated using an accuracy score and area under the curve (AUC), with values exceeding 85% and 0.87, respectively, for most flood classes, verifying its reliability for spatial flood vulnerability assessment. The results revealed a significant transition from low-magnitude to high-intensity flood events, with high and very high flood-vulnerable areas increasing by approximately 72.1% and 57.14%, respectively, from 1991 to 2050. These exacerbations in flood risk were driven by increased rainfall patterns and extensive cropland expansion at the expense of forest, grassland, and shrubland. These findings contribute to achieving sustainable development goals, particularly those related to sustainable communities and cities, and directly support climate action by providing critical insights for decision-makers via high-resolution flood susceptibility maps to enhance resilience against flooding, guide land use planning, and prioritise climate adaptation investments in vulnerable regions.</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

Machine learning approach for assessing flood vulnerability under changing climate and land use and land cover dynamics in ribb and Gumara Catchments, lake Tana Sub-Basin, Ethiopia

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月19日

查看全文
  • Asimamaw Nigusie Asitatikie,
  • Yitbarek Andualem Mekonnen,
  • Dessie Wubetu Melsse

摘要

Floods are natural disasters that are intensified by climate variability and land use changes, impacting the social and economic well-being of people worldwide. These floods and vulnerability mapping have explored using conceptual and physical hydrological models, which require large datasets, but often fail to capture the complex interactions between environmental and climate change. This research addresses this gap by applying a machine learning approach, implemented in the Python Programming Language, to map flood susceptibility zones under climate change and land use changes in the Ribb and Gumara Catchments, Lake-Tana Sub-Basin, Ethiopia. Land use maps and flood susceptibility projections for 1991, 2023, and 2050, under SSP 245 & SSP 585 scenarios, were produced using the Random Forest (RF) algorithm, Sentinel-2 imagery, and temporal and geospatial datasets. The model’s performance was evaluated using an accuracy score and area under the curve (AUC), with values exceeding 85% and 0.87, respectively, for most flood classes, verifying its reliability for spatial flood vulnerability assessment. The results revealed a significant transition from low-magnitude to high-intensity flood events, with high and very high flood-vulnerable areas increasing by approximately 72.1% and 57.14%, respectively, from 1991 to 2050. These exacerbations in flood risk were driven by increased rainfall patterns and extensive cropland expansion at the expense of forest, grassland, and shrubland. These findings contribute to achieving sustainable development goals, particularly those related to sustainable communities and cities, and directly support climate action by providing critical insights for decision-makers via high-resolution flood susceptibility maps to enhance resilience against flooding, guide land use planning, and prioritise climate adaptation investments in vulnerable regions.