Risk assessment of karst collapse areas based on the improved fish bone model: An example of the Liuzhou area in Guangxi Province

Taking the karst collapse area in Liuzhou of Guangxi as an example, this study conducted field investigations and analysis to the major controlling and influencing factors of the karst collapse. Considering the relationship of these factors in the fish bone model and the grading influence of fractures, we established the grading buffers of fractures, and thus determined five feature layers of those comprehensive factors. Based on the improved fish bone model, we made spatial overlay analysis on the feature layers of each factor's vector data format, and prepared a zoning map of the risk prediction. It is shown that more than 75% of the discovered collapse groups fall in the areas with a high risk or a moderate risk, which approximately account for half of the total research areas and are mainly located in the central and southern portions of the study area. These areas have thin covers, and frequent karst collapse is common. Less than 25% of the discovered collapse groups are in the areas with a low risk or a very low risk, which are mainly distributed in the north and west of the study area. Urban planning and construction should select those areas with a low risk or a very low risk. This study has effectively predicted the risk of karst collapse in Liuzhou city.