Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake
Seismic landslides are characterized by wide distribution and strong destructiveness. On July 22, 2013, the Min-Zhang earthquake occurred and a large number of casualties and building burying were caused by the geological disasters induced by seismic motion. The present research aims to generate sei...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Maximum Academic Press
2023-01-01
|
| Series: | Emergency Management Science and Technology |
| Subjects: | |
| Online Access: | https://www.maxapress.com/article/doi/10.48130/EMST-2023-0005 |
| _version_ | 1797655452426698752 |
|---|---|
| author | Hanxu Zhou Ailan Che |
| author_facet | Hanxu Zhou Ailan Che |
| author_sort | Hanxu Zhou |
| collection | DOAJ |
| description | Seismic landslides are characterized by wide distribution and strong destructiveness. On July 22, 2013, the Min-Zhang earthquake occurred and a large number of casualties and building burying were caused by the geological disasters induced by seismic motion. The present research aims to generate seismic landslides susceptibility prediction maps of Min-Zhang earthquake using different machine learning algorithms, providing reference for disaster prevention and reduction in earthquake-affected areas. Five machine learning algorithms including K Nearest Neighbors (KNN), Naive Bayes (NB), Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM) are implemented and the sample dataset was prepared based on the landslide inventory map from open data repository. A total number of 4660 samples containing seismic landslides and non-landslide were collected. The influencing factors of seismic landslide include peak ground acceleration (PGA), epicenter distance, elevation, slope, aspect, plan curvature, profile curvature, fault distance, river distance, and normalized difference vegetation index (NDVI). The performance of five target machine learning algorithms is evaluated and compared using determination coefficient R2 and AUC value of ROC curve. The results indicate that the RF and SVM model have more accurate prediction ability with higher AUC value reaching 0.999 and 0.998, respectively, and the NR model has relatively poor performance resulting from the potential correlation of various influencing factors. Finally, the seismic landslide susceptibility of the Min-Zhang earthquake was mapped using the five trained models and it could offer useful information for seismic hazard management in the future. |
| first_indexed | 2024-03-11T17:14:38Z |
| format | Article |
| id | doaj.art-83b32636fe68455fa2dbf37c68de3056 |
| institution | Directory Open Access Journal |
| issn | 2832-448X |
| language | English |
| last_indexed | 2024-03-11T17:14:38Z |
| publishDate | 2023-01-01 |
| publisher | Maximum Academic Press |
| record_format | Article |
| series | Emergency Management Science and Technology |
| spelling | doaj.art-83b32636fe68455fa2dbf37c68de30562023-10-20T03:42:57ZengMaximum Academic PressEmergency Management Science and Technology2832-448X2023-01-013111110.48130/EMST-2023-0005EMST-2023-0005Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquakeHanxu Zhou0Ailan Che1School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, ChinaSchool of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, ChinaSeismic landslides are characterized by wide distribution and strong destructiveness. On July 22, 2013, the Min-Zhang earthquake occurred and a large number of casualties and building burying were caused by the geological disasters induced by seismic motion. The present research aims to generate seismic landslides susceptibility prediction maps of Min-Zhang earthquake using different machine learning algorithms, providing reference for disaster prevention and reduction in earthquake-affected areas. Five machine learning algorithms including K Nearest Neighbors (KNN), Naive Bayes (NB), Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM) are implemented and the sample dataset was prepared based on the landslide inventory map from open data repository. A total number of 4660 samples containing seismic landslides and non-landslide were collected. The influencing factors of seismic landslide include peak ground acceleration (PGA), epicenter distance, elevation, slope, aspect, plan curvature, profile curvature, fault distance, river distance, and normalized difference vegetation index (NDVI). The performance of five target machine learning algorithms is evaluated and compared using determination coefficient R2 and AUC value of ROC curve. The results indicate that the RF and SVM model have more accurate prediction ability with higher AUC value reaching 0.999 and 0.998, respectively, and the NR model has relatively poor performance resulting from the potential correlation of various influencing factors. Finally, the seismic landslide susceptibility of the Min-Zhang earthquake was mapped using the five trained models and it could offer useful information for seismic hazard management in the future.https://www.maxapress.com/article/doi/10.48130/EMST-2023-0005seismic landslidesusceptibilitymachine learninggismin-zhang earthquake. |
| spellingShingle | Hanxu Zhou Ailan Che Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake Emergency Management Science and Technology seismic landslide susceptibility machine learning gis min-zhang earthquake. |
| title | Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake |
| title_full | Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake |
| title_fullStr | Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake |
| title_full_unstemmed | Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake |
| title_short | Seismic landslide susceptibility mapping using machine learning methods: A case study of the 2013 Ms6.6 Min-Zhang earthquake |
| title_sort | seismic landslide susceptibility mapping using machine learning methods a case study of the 2013 ms6 6 min zhang earthquake |
| topic | seismic landslide susceptibility machine learning gis min-zhang earthquake. |
| url | https://www.maxapress.com/article/doi/10.48130/EMST-2023-0005 |
| work_keys_str_mv | AT hanxuzhou seismiclandslidesusceptibilitymappingusingmachinelearningmethodsacasestudyofthe2013ms66minzhangearthquake AT ailanche seismiclandslidesusceptibilitymappingusingmachinelearningmethodsacasestudyofthe2013ms66minzhangearthquake |