Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach
Despite the fact that precipitation, topography, and geology can control the landslide magnitude scaling law, threshold equations for the landslide scale distribution are still difficult to accurately determine. In this study, the controlling precipitation, topographic, and geologic factors for land...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-12-01
|
| Series: | Geomatics, Natural Hazards & Risk |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/19475705.2025.2487805 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849762361502597120 |
|---|---|
| author | Qi Li Zidan Liu Ziyu Tao |
| author_facet | Qi Li Zidan Liu Ziyu Tao |
| author_sort | Qi Li |
| collection | DOAJ |
| description | Despite the fact that precipitation, topography, and geology can control the landslide magnitude scaling law, threshold equations for the landslide scale distribution are still difficult to accurately determine. In this study, the controlling precipitation, topographic, and geologic factors for landslide scales of both argillaceous clastic rock landslides and shale landslides are examined based on the red-bed soft rock landslide (R-SRL) inventory in Guangdong Province, southern China. The dataset includes the location, magnitudes, and lithology of 1,187 R-SRL records between 1994 and 2015. Our investigation uses spatial and temporal scales to attempt to estimate the R-SRL activities and employs a statistical approach to determine ‘where’, ‘when’, and ‘how’ large landslides will occur in the study area. Correlations are derived by combining the number and volume of year-wise shale and argillaceous clastic rock landslides as well as the maximum rolling rainfall (MRR) intensity. The magnitudes of both types of landslide occurrences are closely related to the MRR intensity. The landslide scales of both argillaceous clastic rock and shale landslides are divided into five levels. Thresholds for the volume level of the R-SRL scale are derived by identifying the MRR intensity that triggers a landslide event using a log-log regression plot approach. For each level, the threshold curves of argillaceous clastic rock landslides are higher than those of shale landslides. These thresholds are validated with a dataset of 237 significant rainfall events (SREs) for the 1994–2017 period. The fortuities and skill scores of receiver operating characteristic (ROC) analysis are created and calculated to evaluate the performance of the volume-level thresholds. The threshold curves for the giant-sized (GS) landslide type exert the first-order prediction with a 98.73% and 99.58% accuracy for shale and argillaceous clastic rock landslides, respectively. Using a Bayesian modeling framework for predicting the probability occurrence of landslides triggered by a rainfall event above the defined rainfall threshold, we found that high intensity rainfall events play a more important role in triggering R-SRLs than their long duration. |
| format | Article |
| id | doaj-art-e732994d2bcf45ef9c95801a0d76705e |
| institution | DOAJ |
| issn | 1947-5705 1947-5713 |
| language | English |
| publishDate | 2025-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | Geomatics, Natural Hazards & Risk |
| spelling | doaj-art-e732994d2bcf45ef9c95801a0d76705e2025-08-20T03:05:45ZengTaylor & Francis GroupGeomatics, Natural Hazards & Risk1947-57051947-57132025-12-0116110.1080/19475705.2025.2487805Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approachQi Li0Zidan Liu1Ziyu Tao2School of Civil Engineering and Transportation, Guangzhou University, Guangzhou, ChinaGuangzhou City Polytechnic, Guangzhou, ChinaSchool of Civil Engineering and Transportation, South China University of Technology, Guangzhou, ChinaDespite the fact that precipitation, topography, and geology can control the landslide magnitude scaling law, threshold equations for the landslide scale distribution are still difficult to accurately determine. In this study, the controlling precipitation, topographic, and geologic factors for landslide scales of both argillaceous clastic rock landslides and shale landslides are examined based on the red-bed soft rock landslide (R-SRL) inventory in Guangdong Province, southern China. The dataset includes the location, magnitudes, and lithology of 1,187 R-SRL records between 1994 and 2015. Our investigation uses spatial and temporal scales to attempt to estimate the R-SRL activities and employs a statistical approach to determine ‘where’, ‘when’, and ‘how’ large landslides will occur in the study area. Correlations are derived by combining the number and volume of year-wise shale and argillaceous clastic rock landslides as well as the maximum rolling rainfall (MRR) intensity. The magnitudes of both types of landslide occurrences are closely related to the MRR intensity. The landslide scales of both argillaceous clastic rock and shale landslides are divided into five levels. Thresholds for the volume level of the R-SRL scale are derived by identifying the MRR intensity that triggers a landslide event using a log-log regression plot approach. For each level, the threshold curves of argillaceous clastic rock landslides are higher than those of shale landslides. These thresholds are validated with a dataset of 237 significant rainfall events (SREs) for the 1994–2017 period. The fortuities and skill scores of receiver operating characteristic (ROC) analysis are created and calculated to evaluate the performance of the volume-level thresholds. The threshold curves for the giant-sized (GS) landslide type exert the first-order prediction with a 98.73% and 99.58% accuracy for shale and argillaceous clastic rock landslides, respectively. Using a Bayesian modeling framework for predicting the probability occurrence of landslides triggered by a rainfall event above the defined rainfall threshold, we found that high intensity rainfall events play a more important role in triggering R-SRLs than their long duration.https://www.tandfonline.com/doi/10.1080/19475705.2025.2487805Rainfall thresholdred-bed soft rock landslidevolume-level thresholdsROC analysisBayesian modeling |
| spellingShingle | Qi Li Zidan Liu Ziyu Tao Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach Geomatics, Natural Hazards & Risk Rainfall threshold red-bed soft rock landslide volume-level thresholds ROC analysis Bayesian modeling |
| title | Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach |
| title_full | Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach |
| title_fullStr | Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach |
| title_full_unstemmed | Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach |
| title_short | Enhancing landslide-scale rainfall threshold predictive modeling for rainfall-induced red-bed soft rock landslide occurrence using a stock-taking approach |
| title_sort | enhancing landslide scale rainfall threshold predictive modeling for rainfall induced red bed soft rock landslide occurrence using a stock taking approach |
| topic | Rainfall threshold red-bed soft rock landslide volume-level thresholds ROC analysis Bayesian modeling |
| url | https://www.tandfonline.com/doi/10.1080/19475705.2025.2487805 |
| work_keys_str_mv | AT qili enhancinglandslidescalerainfallthresholdpredictivemodelingforrainfallinducedredbedsoftrocklandslideoccurrenceusingastocktakingapproach AT zidanliu enhancinglandslidescalerainfallthresholdpredictivemodelingforrainfallinducedredbedsoftrocklandslideoccurrenceusingastocktakingapproach AT ziyutao enhancinglandslidescalerainfallthresholdpredictivemodelingforrainfallinducedredbedsoftrocklandslideoccurrenceusingastocktakingapproach |