Study On Assessment Methodology Of Landslide And Debris Flow Geological Hazards

Landslide and debris flow are common geological hazard. They can induce a series of disasters that may pose a serious threat to lives, property, and even economic development. The comprehensive assessment of landslide and debris flow hazard is a challenging task due to the uncertainties and complexity of various related factors. In this study, based on in-depth study and comprehensive analysis of previous research results, we proposed FAN (Forest Augmented Naive Bayesian) model for landslide susceptibility assessment and BN (Bayesian network) model for debris flow hazard assessment, and developed a GIS-based landslide susceptibility and debris flow hazard assessment systems. The main research work and key developments are as follows:(1) Analyzed the problems of the existing landslide and debris flow assessment model. According to the characteristics of the study area, FAN model was chosen as a model of the landslide hazard susceptibility assessment, and BN model was chosen as a model of the debris flow hazard assessment.(2) Landslide susceptibility is a full of challenge task due to the uncertainties and complexity of multiple related factors. Landslide susceptibility refers to the possibility of landslide occurrence of a terrain unit. Susceptibility does not consider the temporal probability of failure (i.e., when or how frequently landslides occur), nor the magnitude of the expected landslide. Factors associated with landslide susceptibility assessment are selected and a novel methodology for landslide susceptibility assessment based on FAN model and domain knowledge was proposed. Cross-validations of FAN model and ANNs (Artificial Neural Networks) were conducted on two different sample datasets. The landslide susceptibility maps of Pan-Xi district are produced using FAN model and ANNs, respectively. The results indicated that FAN model has excellent anti-interference, good robustness, and proves to be an alternative method for landslide susceptibility assessment.(3) The comprehensive assessment of debris flow hazard risk is a challenging task due to the uncertainties and complexity of various related factors. A reasonable and reliable assessment should be based on sufficient data and assessment approaches reflecting the actual situation. Factors associated with debris flow hazard assessment are selected and a novel methodology for assessing debris flow hazard based on a BN and domain knowledge is presented. Based on authoritative records of debris flow hazards and geomorphologic and environmental data for the Chinese mainland, approaches based on BN, SVM (support vector machine) and ANNs (artificial neural networks) were compared. The results show that BN provides a higher probability (85.66%) of hazard detection, a better precision (89.63%), and a larger AUC (area under the receiver operating characteristic curve) value (0.95) than SVM and ANNs. The BN-based model is useful for mapping and assessing debris flow hazard risk on a national scale.(4) Based on ArcGIS Engine, C#and MATLAB Engine, we developed a GIS-based landslide susceptibility and debris flow hazard assessment system. The assessment system can provide decision-making support and tools for decision makers.All the analysis of both the scientific theory and practical application show that BN model can deal with the uncertainty of landslide susceptibility and debris flow hazard assessment, has great advantages in the hazard assessment and broad application prospects.