Department of Environmental Science, Faculty of Environmental Studies, UPM 43400 Serdang, Selangor, Malaysia
Received 19 September 2009, Accepted 29 October 2009
In tropical areas especially during monsoon seasons intense precipitation is the main caused that trigger the natural shallow landslide phenomena. This phenomenon can be disastrous and widespread in occurrence even in undisturbed forested catchment. In this paper, an attempt has been made to evaluate the susceptibility of natural hill slopes to failure for a popular hill resort area, the Fraser Hill Catchment under different rainfall regimes and soil thickness. A Digital Elevation Model (DEM) was prepared for the 8.2 km2 catchment. A GIS based deterministic model was then applied to predict the spatial landslide occurrence within catchment. Model input parameters include bulk density, friction angle, cohesion and hydraulic conductivity were gathered through in situ and lab analysis as well as from previous soil analysis records. Landslides locations were recorded using GPS as well as previous air photos and satellite imagery to establish landslide source areas inventory. The landslide susceptibility map was produced under different precipitation event’s simulation to see the effects of precipitation to stability of the hill slopes of the catchment. The results were categorized into naturally unstable (Defended, Upper Threshold, Lower Threshold), marginal instability (Quasi Stable) and stable area (Moderately Stable and Stable). Results of the simulation indicated notable change in precipitation effect on Defended area is between 10mm to 40mm range in a single storm event. However, when storm event is exceeded 120mm, the result on Defended area produced by the model tends to be constant further on. For area categorized as naturally unstable (Factor of Safety, SF<1), with 110 mm of precipitation in a single storm event and soil depth at 2 meters and 4 meters could affect 69.51% and 69.88% respectively of the catchment area fall under that class. In addition, the model was able to detect 4% more of the landslide inventory under shallower soil depth of 2 meters.
Keywords: deterministic model; landslides; GIS; slope stability; precipitation threshold.
* corresponding author : Dr. Wan Nor Azmin Bin Sulaiman