Aided by raster-based Geographical Information System （GIS） function

one method for terrain based landslide risk assessment for Lantau Island of Hong Kong was given in the paper. In this method

slope

aspect

vegetation covered

geology

elevation and distance to drainage line were considered as landslide related elements

and each element as given different weight in landslide risk assessment matrix. In raster based landslide risk assessment methodology

the grid size （pixel） selection for dataset concerned is a key task and must be solved. To find the most appropriate grid size for each elements

5m

10m

20m

40m and 80m five kinds of grid size were tested in this study. Following the state that landslide risk assessment result is the key to evaluate the appropriate grid size

landslide risk assessment parameters characteristic

risk assessment results were analyzed and compared thoroughly

20m grid size was found the appropriate grid size for the pilot area according to most features described in the paper such as area frequency for each kind of grid size

landslide frequency under each grid size

area frequency for each class landslide susceptibility under different grid size. Generally

the appropriate grid size selection for certain cases are connected to area of pilot region

user requirement

gromorphology and topography characteristics of pilot region

and other case sensitive elements. Based on the results obtained in the pilot area

one experimental formula was given for grid size selection of landslide risk assessment.