{"title":"Landslide susceptibility assessment of South Korea using stacking ensemble machine learning","authors":"Seung-Min Lee, Seung-Jae Lee","doi":"10.1186/s40677-024-00271-y","DOIUrl":null,"url":null,"abstract":"Landslide susceptibility assessment (LSA) is a crucial indicator of landslide hazards, and its accuracy is improving with the development of artificial intelligence (AI) technology. However, the AI algorithms are inconsistent across regions and strongly dependent on input variables. Additionally, LSA must include historical data, which often restricts the assessment to the local scale and single landslide events. In this study, we performed an LSA for the entirety of South Korea. A total of 30 input variables were constructed, consisting of 9 variables from past climate model data MK-PRISM, 12 topographical factors, and 9 environmental factors. Sixteen machine learning algorithms were used as basic classifiers, and a stacking ensemble was used on the four algorithms with the highest area under the curve (AUC). Additionally, a separate assessment model was established for areas with a risk of landslides affecting areas larger than 1 ha. The highest-performing classifier was CatBoost, with an AUC of ~ 0.89 for both assessments. Among the input variables, distance of road, daily maximum precipitation, digital elevation model, and soil depth were the most influential. In all landslide events, CatBoost, lightGBM, XGBoost, and Random Forest had the highest AUC in descending order; in large landslide events, the order was CatBoost, XGBoost, Extra Tree, and lightGBM. The stacking ensemble enabled the construction of two landslide susceptibility maps. Our findings provide a statistical method for constructing a high-resolution (30 m) landslide susceptibility map on a country scale using diverse natural factors, including past climate data.","PeriodicalId":37025,"journal":{"name":"Geoenvironmental Disasters","volume":"2 1","pages":""},"PeriodicalIF":3.8000,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Geoenvironmental Disasters","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1186/s40677-024-00271-y","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
Landslide susceptibility assessment (LSA) is a crucial indicator of landslide hazards, and its accuracy is improving with the development of artificial intelligence (AI) technology. However, the AI algorithms are inconsistent across regions and strongly dependent on input variables. Additionally, LSA must include historical data, which often restricts the assessment to the local scale and single landslide events. In this study, we performed an LSA for the entirety of South Korea. A total of 30 input variables were constructed, consisting of 9 variables from past climate model data MK-PRISM, 12 topographical factors, and 9 environmental factors. Sixteen machine learning algorithms were used as basic classifiers, and a stacking ensemble was used on the four algorithms with the highest area under the curve (AUC). Additionally, a separate assessment model was established for areas with a risk of landslides affecting areas larger than 1 ha. The highest-performing classifier was CatBoost, with an AUC of ~ 0.89 for both assessments. Among the input variables, distance of road, daily maximum precipitation, digital elevation model, and soil depth were the most influential. In all landslide events, CatBoost, lightGBM, XGBoost, and Random Forest had the highest AUC in descending order; in large landslide events, the order was CatBoost, XGBoost, Extra Tree, and lightGBM. The stacking ensemble enabled the construction of two landslide susceptibility maps. Our findings provide a statistical method for constructing a high-resolution (30 m) landslide susceptibility map on a country scale using diverse natural factors, including past climate data.
期刊介绍:
Geoenvironmental Disasters is an international journal with a focus on multi-disciplinary applied and fundamental research and the effects and impacts on infrastructure, society and the environment of geoenvironmental disasters triggered by various types of geo-hazards (e.g. earthquakes, volcanic activity, landslides, tsunamis, intensive erosion and hydro-meteorological events).
The integrated study of Geoenvironmental Disasters is an emerging and composite field of research interfacing with areas traditionally within civil engineering, earth sciences, atmospheric sciences and the life sciences. It centers on the interactions within and between the Earth''s ground, air and water environments, all of which are affected by climate, geological, morphological and anthropological processes; and biological and ecological cycles. Disasters are dynamic forces which can change the Earth pervasively, rapidly, or abruptly, and which can generate lasting effects on the natural and built environments.
The journal publishes research papers, case studies and quick reports of recent geoenvironmental disasters, review papers and technical reports of various geoenvironmental disaster-related case studies. The focus on case studies and quick reports of recent geoenvironmental disasters helps to advance the practical understanding of geoenvironmental disasters and to inform future research priorities; they are a major component of the journal. The journal aims for the rapid publication of research papers at a high scientific level. The journal welcomes proposals for special issues reflecting the trends in geoenvironmental disaster reduction and monothematic issues. Researchers and practitioners are encouraged to submit original, unpublished contributions.