Ecoregion-Based Landslide Susceptibility Mapping: A Spatially Partitioned Modeling Strategy for Oregon, USA. Xu, Z., Zuo, P., Zhao, W., Zhou, Z., Shao, X., Yu, J., Yu, H., Wang, W., Gan, J., Duan, J., & Jin, J. Applied Sciences, 15(20):11242, January, 2025. Publisher: Multidisciplinary Digital Publishing Institute![Ecoregion-Based Landslide Susceptibility Mapping: A Spatially Partitioned Modeling Strategy for Oregon, USA [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Conventional non-partitioned Landslide Susceptibility Mapping (LSM), which neglects geospatial heterogeneity, often has limitations in accurately capturing local risk patterns. To address this challenge, this study investigated the effectiveness of localized modeling in the environmentally diverse state of Oregon, USA, by comparing ecoregion-based local models with the non-partitioned model. We partitioned Oregon into seven distinct units using the U.S. Environmental Protection Agency (EPA) Level III Ecoregions and developed one global and seven local models with the eXtreme Gradient Boosting (XGBoost) algorithm. A comprehensive evaluation framework, including the Area Under the Curve (AUC), Landslide Density (LD), and the Total Deviation Index (TDI), was used to compare the models. The results demonstrated the clear superiority of the partitioned strategy. Moreover, different ecoregions were found to have distinct dominant landslide conditioning factors, revealing strong spatial non-stationarity. Although all models generated high AUC values (\textgreater0.93), LD analysis showed that the local models were significantly more efficient at identifying high-risk zones. This advantage was particularly pronounced in critical, landslide-prone western areas; for instance, in the Willamette–Georgia–Puget Lowland, the local model’s LD value in the ‘very high’ susceptibility class was over 3.5 times that of the global model. High TDI values (some \textgreater35%) further confirmed fundamental spatial discrepancies between the risk maps obtained by the two strategies. This research substantiated that, in geographically complex terrains, partitioned modeling is an effective approach for more accurate and reliable LSM, providing a scientific basis for developing targeted regional disaster mitigation policies.
@article{xu_ecoregion-based_2025,
title = {Ecoregion-{Based} {Landslide} {Susceptibility} {Mapping}: {A} {Spatially} {Partitioned} {Modeling} {Strategy} for {Oregon}, {USA}},
volume = {15},
copyright = {http://creativecommons.org/licenses/by/3.0/},
issn = {2076-3417},
shorttitle = {Ecoregion-{Based} {Landslide} {Susceptibility} {Mapping}},
url = {https://www.mdpi.com/2076-3417/15/20/11242},
doi = {10.3390/app152011242},
abstract = {Conventional non-partitioned Landslide Susceptibility Mapping (LSM), which neglects geospatial heterogeneity, often has limitations in accurately capturing local risk patterns. To address this challenge, this study investigated the effectiveness of localized modeling in the environmentally diverse state of Oregon, USA, by comparing ecoregion-based local models with the non-partitioned model. We partitioned Oregon into seven distinct units using the U.S. Environmental Protection Agency (EPA) Level III Ecoregions and developed one global and seven local models with the eXtreme Gradient Boosting (XGBoost) algorithm. A comprehensive evaluation framework, including the Area Under the Curve (AUC), Landslide Density (LD), and the Total Deviation Index (TDI), was used to compare the models. The results demonstrated the clear superiority of the partitioned strategy. Moreover, different ecoregions were found to have distinct dominant landslide conditioning factors, revealing strong spatial non-stationarity. Although all models generated high AUC values ({\textgreater}0.93), LD analysis showed that the local models were significantly more efficient at identifying high-risk zones. This advantage was particularly pronounced in critical, landslide-prone western areas; for instance, in the Willamette–Georgia–Puget Lowland, the local model’s LD value in the ‘very high’ susceptibility class was over 3.5 times that of the global model. High TDI values (some {\textgreater}35\%) further confirmed fundamental spatial discrepancies between the risk maps obtained by the two strategies. This research substantiated that, in geographically complex terrains, partitioned modeling is an effective approach for more accurate and reliable LSM, providing a scientific basis for developing targeted regional disaster mitigation policies.},
language = {en},
number = {20},
urldate = {2026-01-21},
journal = {Applied Sciences},
author = {Xu, Zhixiang and Zuo, Peng and Zhao, Wen and Zhou, Zeyu and Shao, Xiangyu and Yu, Junpo and Yu, Haize and Wang, Weijie and Gan, Junwei and Duan, Jinshun and Jin, Jiming},
month = jan,
year = {2025},
note = {Publisher: Multidisciplinary Digital Publishing Institute},
keywords = {NALCMS},
pages = {11242},
}
Downloads: 0
{"_id":"ePdz9TaX85nGXoJ3o","bibbaseid":"xu-zuo-zhao-zhou-shao-yu-yu-wang-etal-ecoregionbasedlandslidesusceptibilitymappingaspatiallypartitionedmodelingstrategyfororegonusa-2025","author_short":["Xu, Z.","Zuo, P.","Zhao, W.","Zhou, Z.","Shao, X.","Yu, J.","Yu, H.","Wang, W.","Gan, J.","Duan, J.","Jin, J."],"bibdata":{"bibtype":"article","type":"article","title":"Ecoregion-Based Landslide Susceptibility Mapping: A Spatially Partitioned Modeling Strategy for Oregon, USA","volume":"15","copyright":"http://creativecommons.org/licenses/by/3.0/","issn":"2076-3417","shorttitle":"Ecoregion-Based Landslide Susceptibility Mapping","url":"https://www.mdpi.com/2076-3417/15/20/11242","doi":"10.3390/app152011242","abstract":"Conventional non-partitioned Landslide Susceptibility Mapping (LSM), which neglects geospatial heterogeneity, often has limitations in accurately capturing local risk patterns. To address this challenge, this study investigated the effectiveness of localized modeling in the environmentally diverse state of Oregon, USA, by comparing ecoregion-based local models with the non-partitioned model. We partitioned Oregon into seven distinct units using the U.S. Environmental Protection Agency (EPA) Level III Ecoregions and developed one global and seven local models with the eXtreme Gradient Boosting (XGBoost) algorithm. A comprehensive evaluation framework, including the Area Under the Curve (AUC), Landslide Density (LD), and the Total Deviation Index (TDI), was used to compare the models. The results demonstrated the clear superiority of the partitioned strategy. Moreover, different ecoregions were found to have distinct dominant landslide conditioning factors, revealing strong spatial non-stationarity. Although all models generated high AUC values (\\textgreater0.93), LD analysis showed that the local models were significantly more efficient at identifying high-risk zones. This advantage was particularly pronounced in critical, landslide-prone western areas; for instance, in the Willamette–Georgia–Puget Lowland, the local model’s LD value in the ‘very high’ susceptibility class was over 3.5 times that of the global model. High TDI values (some \\textgreater35%) further confirmed fundamental spatial discrepancies between the risk maps obtained by the two strategies. This research substantiated that, in geographically complex terrains, partitioned modeling is an effective approach for more accurate and reliable LSM, providing a scientific basis for developing targeted regional disaster mitigation policies.","language":"en","number":"20","urldate":"2026-01-21","journal":"Applied Sciences","author":[{"propositions":[],"lastnames":["Xu"],"firstnames":["Zhixiang"],"suffixes":[]},{"propositions":[],"lastnames":["Zuo"],"firstnames":["Peng"],"suffixes":[]},{"propositions":[],"lastnames":["Zhao"],"firstnames":["Wen"],"suffixes":[]},{"propositions":[],"lastnames":["Zhou"],"firstnames":["Zeyu"],"suffixes":[]},{"propositions":[],"lastnames":["Shao"],"firstnames":["Xiangyu"],"suffixes":[]},{"propositions":[],"lastnames":["Yu"],"firstnames":["Junpo"],"suffixes":[]},{"propositions":[],"lastnames":["Yu"],"firstnames":["Haize"],"suffixes":[]},{"propositions":[],"lastnames":["Wang"],"firstnames":["Weijie"],"suffixes":[]},{"propositions":[],"lastnames":["Gan"],"firstnames":["Junwei"],"suffixes":[]},{"propositions":[],"lastnames":["Duan"],"firstnames":["Jinshun"],"suffixes":[]},{"propositions":[],"lastnames":["Jin"],"firstnames":["Jiming"],"suffixes":[]}],"month":"January","year":"2025","note":"Publisher: Multidisciplinary Digital Publishing Institute","keywords":"NALCMS","pages":"11242","bibtex":"@article{xu_ecoregion-based_2025,\n\ttitle = {Ecoregion-{Based} {Landslide} {Susceptibility} {Mapping}: {A} {Spatially} {Partitioned} {Modeling} {Strategy} for {Oregon}, {USA}},\n\tvolume = {15},\n\tcopyright = {http://creativecommons.org/licenses/by/3.0/},\n\tissn = {2076-3417},\n\tshorttitle = {Ecoregion-{Based} {Landslide} {Susceptibility} {Mapping}},\n\turl = {https://www.mdpi.com/2076-3417/15/20/11242},\n\tdoi = {10.3390/app152011242},\n\tabstract = {Conventional non-partitioned Landslide Susceptibility Mapping (LSM), which neglects geospatial heterogeneity, often has limitations in accurately capturing local risk patterns. To address this challenge, this study investigated the effectiveness of localized modeling in the environmentally diverse state of Oregon, USA, by comparing ecoregion-based local models with the non-partitioned model. We partitioned Oregon into seven distinct units using the U.S. Environmental Protection Agency (EPA) Level III Ecoregions and developed one global and seven local models with the eXtreme Gradient Boosting (XGBoost) algorithm. A comprehensive evaluation framework, including the Area Under the Curve (AUC), Landslide Density (LD), and the Total Deviation Index (TDI), was used to compare the models. The results demonstrated the clear superiority of the partitioned strategy. Moreover, different ecoregions were found to have distinct dominant landslide conditioning factors, revealing strong spatial non-stationarity. Although all models generated high AUC values ({\\textgreater}0.93), LD analysis showed that the local models were significantly more efficient at identifying high-risk zones. This advantage was particularly pronounced in critical, landslide-prone western areas; for instance, in the Willamette–Georgia–Puget Lowland, the local model’s LD value in the ‘very high’ susceptibility class was over 3.5 times that of the global model. High TDI values (some {\\textgreater}35\\%) further confirmed fundamental spatial discrepancies between the risk maps obtained by the two strategies. This research substantiated that, in geographically complex terrains, partitioned modeling is an effective approach for more accurate and reliable LSM, providing a scientific basis for developing targeted regional disaster mitigation policies.},\n\tlanguage = {en},\n\tnumber = {20},\n\turldate = {2026-01-21},\n\tjournal = {Applied Sciences},\n\tauthor = {Xu, Zhixiang and Zuo, Peng and Zhao, Wen and Zhou, Zeyu and Shao, Xiangyu and Yu, Junpo and Yu, Haize and Wang, Weijie and Gan, Junwei and Duan, Jinshun and Jin, Jiming},\n\tmonth = jan,\n\tyear = {2025},\n\tnote = {Publisher: Multidisciplinary Digital Publishing Institute},\n\tkeywords = {NALCMS},\n\tpages = {11242},\n}\n\n\n\n\n\n\n\n\n\n\n\n","author_short":["Xu, Z.","Zuo, P.","Zhao, W.","Zhou, Z.","Shao, X.","Yu, J.","Yu, H.","Wang, W.","Gan, J.","Duan, J.","Jin, J."],"key":"xu_ecoregion-based_2025","id":"xu_ecoregion-based_2025","bibbaseid":"xu-zuo-zhao-zhou-shao-yu-yu-wang-etal-ecoregionbasedlandslidesusceptibilitymappingaspatiallypartitionedmodelingstrategyfororegonusa-2025","role":"author","urls":{"Paper":"https://www.mdpi.com/2076-3417/15/20/11242"},"keyword":["NALCMS"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"article","biburl":"https://bibbase.org/zotero/NAAtlas2024","dataSources":["qLjf8q88GSLZ5dAmC"],"keywords":["nalcms"],"search_terms":["ecoregion","based","landslide","susceptibility","mapping","spatially","partitioned","modeling","strategy","oregon","usa","xu","zuo","zhao","zhou","shao","yu","yu","wang","gan","duan","jin"],"title":"Ecoregion-Based Landslide Susceptibility Mapping: A Spatially Partitioned Modeling Strategy for Oregon, USA","year":2025}