Search

The main purpose of this study was to establish a data-driven approach and assess its potential for shallow landslide susceptibility mapping of the Shennongjia Forestry District, China. For the data processing, Fisher segmenting was used for the classification of topographical factors (gradient, relief amplitude, plan curvature, and normalized difference vegetation index) and the improved Otsu method was used to determine the buffer length thresholds of the locational factors (proximity to roads, rivers and faults). The computations show that these two objective methods help to avoid the uncertainty caused by the subjective judgments and the obtained thresholds are more consistent with actual conditions. Considering the different data types, the locational and topographical factors were evaluated by fuzzy comprehensive evaluation while the geological factors were evaluated by frequency index, and then the three separate evaluations were combined using radar chart analysis. The validation result shows that the susceptibility map based on a data-driven approach has a high accuracy of prediction and this approach will be salutary for risk mitigation and land-use management in mountainous regions. [ABSTRACT FROM AUTHOR]

Landslides are one of the most destructive disasters that cause damage to both property and life every year. The mountainous terrain, heavy and prolonged rainfall and dense development near steep hillsides make Hong Kong as one of the most vulnerable metropolitans to the risk of landslides. Therefore, regional-specific landslide susceptibility assessment in Hong Kong is necessary for hazard management and effective land use planning. Various methodologies have been proposed for landslide susceptibility mappings. By applying the statistically based frequency ratio and logistic regression models in Lantau Island, Hong Kong, this study attempts to add extra value to the literature of evaluating their ‘prediction rate’ (rather than ‘success rate’) for landslide susceptibility mapping in a temporal context. The spatial relationship among landslide occurrence and nine causative factors (elevation, slope, aspect, plan curvature, profile curvature, normalized difference vegetation index (NDVI), distance to river, distance to fault and lithology) were examined, and landslide susceptibility maps were generated by frequency ratio and logistic regression model. Validations of the mapping results were performed by calculating relative operating characteristics (ROC). Virtually all previous studies randomly extracted and reserved a portion of historical landslide records to perform the model evaluation. In contrast, in this study, 1864 (70%) landslide records in the Enhanced Natural Terrain Landslide Inventory from 2000 to 2008 are utilized to train the model, while the subsequent 799 (30%) landslides occurred from 2008 to 2009 are used for model validation. The validation result shows that logistic regression model (84.05%) possesses a better prediction power than frequency ratio model (76.64%) for the study area. This study attempts to offer a new model assessment strategy for landslide susceptibility mapping. The resultant maps also provide a scientific assessment of the risk areas with respect to landslides on Lantau Island and could serve as a basis for decisions or justification of the Lantau development planning. [ABSTRACT FROM PUBLISHER]