Based On The Geographical Landslide Susceptibility Zoning Of Processing Services

Landslides are the most common and widespread hazards in China. Nowadays, property damage and casualties caused by landslides is increasing. Landslide susceptibility mapping, which is defined as a quantitative or qualitative assessment of landslides proneness in the given area, is a fundamental component of landslide spatial prediction and has been applied widely to land planning and disaster prevention of projects.This paper constructed a model library for landslide susceptibility analysis and implemented the automatic mapping. On this basis, the landslide susceptibility map was combined with weather conditions to develop a WebGIS-based regional landslide early warning system. The main contents are summarized as follows:(1) A model library for landslide susceptibility analysis was constructed invoking GIS components of Geoprocessor programming model in ArcGIS, as well as Python scientific computing package. The model library contains three landslide susceptibility analysis models including logistic regression model, extension model, and weighted linear combination model, as well as the validation method. Meanwhile, given that heavy rainfall is a main trigger of landslides, the model library provides regional early warning function based on a warning model, which couples regional geology setting with weather condition. At last, all the techniques in the model library were published to geoprocessing (GP) service on the platform of ArcGIS Server.(2) Comparative analysis of the three landslide susceptibility map was conducted on the complexity, execution efficiency and reliability. Complexity ranking:weighted linear combination model> extenics model>logistic regression model; execution efficiency ranking:extension model> logistic regression model> weighted linear combination model; and reliability ranking:logistic regression model>weighted linear combination model> extension model.(3) Requirement and function analysis of the regional landslide hazard prediction system were carried out to meet user needs. On this basis, system architecture design and detailed design were conducted to make division of every function module of the system clear.(4) The regional landslide hazard prediction system was implemented based on rich internet applications (RIA) and J2EE programming technique. The system implemented combination of landslide susceptibility mapping function with information system by invoking GP service based on Http and REST network protocols. At last, a sever platform for the system was set up, and the system was published to internet for providing online service including landslide warning and forecasting information and susceptibility analysis.