Landslide Hazard Zonation in Sri Lanka: An Assessment of Manual and GIS Based Automated Procedure in Preparation of Geology Weight Map

摘要

Many techniques have been proposed for landslide hazard zonation (LHZ). They can generally be divided into two groups : direct or semi direct hazard mapping in which the degree of hazard is determined by the mapping expert and indirect hazard mapping in which either statistical or deterministic models are used to predict landslide prone areas based on information obtained from the interrelation between terrain factors and the landslide distribution. With the introduction of GIS, in particular indirect methods gained enormously due to its capacity to handle and analyze data with high spatial variability.In the context of Sri Lanka, landslide hazard maps are prepared using a model developed under a research project conducted from mid-1989 to mid-1995. This model is based on the analysis of six major terrain factors with sub factors and factor classes collected from more than thousand landslides occurred in Badulla and Nuwaraeliya districts. For the zonation based on this model, field data is collected according to those factors and the corresponding weight maps are prepared manually. GIS is only used finally as an overlaying and reclassifying tool. In this workflow, very laborious effort is needed for the preparation of geology weight map, especially when complex terrain conditions and large amount of data are involved. One of the reasons is that, unlike all other factors where the basic mapping units are areas, the geology map consists of two major parts : lithological units as areas but structural attitudes as linear or point measurements.In 2009, an approach was discussed how GIS capabilities can be used efficiently to integrate the influence of structural attitudes such as strike or dip directions and dip angles for the preparation of geology weight maps which is an essential part of the LHZ model used in Sri Lanka. Even though the original procedure was based on manual weighting since then, the newly introduced automated procedure has been used by National Building Research Organization to accelerate the mapping procedure.Under this study, a statistical comparison and an assessment were done between the two procedures and necessary modifications to the latter, that is to the automated procedure is proposed to enhance the accuracy of the method.