Landslide hazard mapping of Ibrahim River Basin, Lebanon
详细信息 查看全文
- 作者：C. Abdallah ; G. Faour
- 关键词：Remote sensing ; GIS ; Visual interpretation ; Statistical correlation ; Modeling
- 刊名：Natural Hazards
- 出版年：2017
- 出版时间：January 2017
- 年：2017
- 卷：85
- 期：1
- 页码：237-266
- 全文大小：
- 刊物类别：Earth and Environmental Science
- 刊物主题：Natural Hazards; Hydrogeology; Geophysics/Geodesy; Geotechnical Engineering & Applied Earth Sciences; Civil Engineering; Environmental Management;
- 出版者：Springer Netherlands
- ISSN：1573-0840
- 卷排序：85
文摘
Landslide susceptibility and hazard maps were established using an adapted Value Analytical Bi-Univariate (VABU) method for Nahr Ibrahim watershed located in Mount Lebanon. The site covered an area of 312 km2 compromising 3 % of Lebanon. Pan-sharpened IKONOS imageries were used to delineate landslides. Morphological, geological soil, hydrological and anthropic parametric factors affecting landslides were extracted under GIS environment. Statistical correlation between existing landslides and related factors using both univariate and bivariate analysis was also conducted. The latter along with landslide and landslide preconditioning and triggering factor databases enabled establishing the landslide susceptibility and hazard maps for the area. The visual interpretation of the satellite imageries has allowed the delineation of 229 landslides covering an area of 7.6 km2. The study indicates, depending on bivariate remote sensing and GIS statistical correlations (Kendall tau-b correlation), that lithology is the most influencing on landslide occurrence, having the highest correlation with other parameters (i.e., 7 times correlated at 1 % level of significance and 3 times at 5 %). It also shows that statistical correlations with landslide exist best between other parameters but with lower levels of significance. The proposed mathematical decision-making method (VABU) that considered two-level weights for mapping MM susceptibility/hazard showed that 19 and 7 % of the watershed are identified as very high susceptibility and hazard areas, respectively. The accuracy of the VABU model was checked and proved to be approximately 74 % in both maps.