Guizhou Province Based On Gis Landslide-prone Multi-model Evaluation

Landslide is the most widespread geological hazard in the world. A lot of landslides were occurred in the recent years. Some of these landslides seriously threat personal and property safety because of their large scale. As a valuable tool for assessing current and potential landslide risks, landslide susceptibility mapping based on GIS was used for land-use planning and infrastructures layout, and developing the early-warning systems. With combination of landslides inventory and 10 causal factors such as elevation, slope, lithology and so on based on GIS, the paper established the landslide susceptibility map in study area of Guizhou province through using several mathematical models. The results of the study were used by relative departments to decrease the loss of lives and properties of the local people.As intensive theoretical models, weighted linear combination model (WLC), logistic regression (LR) and self-organizing map neural network (SOM) have been introduced to evaluate landslides susceptibility in this paper. The achievements of the research work are as followings:(1) Entropy-based weighting method was used to calculate objective weights of causal factors according to the area density of occurred landslides. Trapezoidal fuzzy number weighting (TFNW) approach was used to assess the importance of each subclasses of a causal factor. Finally, the landslides susceptibility map of Guizhou province in China was created by using WLC model.(2) Selected the key factors of landslide according to Certainty Factor (CF) which was deduced from the area of landslides inventory and the area of subclasses of factors. With the helping of GIS spatial analysis tools and SPSS, LR model, which described the relationship of hazards (dependent) and the key factors, was used to create landslide susceptibility map.(3) In order to get reasonable parameters of SOM, vectors which were composited by landslide area densities had been trained 100 times. Then, the landslide susceptibility map was established with SOM model. (4) Three models and results of above were compared under the same conditions to get the distinguishing features and applicable conditions of each model.(5) In order to guide land utilization and plan infrastructure, the landslide susceptibility map was published on web based on ArcGIS Server.