Spatial Multi Criteria Evaluation (SMCE) Model for Landslide Hazard Zonation in Tropical Hilly Environment: A Case Study from Kegalle

Perera ENC, Jayawardana DT,


Landslide hazard is one of the most common global hazards. In Sri Lanka, landslides are considered as a disaster, and thus, scientific communities have paid attention to monitoring and prediction of landslide hazards. Landslide hazard zonation (LHZ) is a vital factor for preparedness and mitigation phases in a disaster management cycle. In the Sri Lankan contest, LHZ becomes more significant since 20% of the total lands are vulnerable. Spatial distribution of landslides of Sri Lanka is mainly influenced by geospatial criteria, rainfall distribution, geology, hydrology, geomorphology, land-use, and drainage network. However, the above factors do not equally contribute to determining the landslide susceptibility. This study attempted to map the landslide hazard zones in one of the tropical hilly region: Kegalle District and weighting causative factors rationally using statistical method in GIS environment. In this study, causative factors were weighted and modelled to define hazardous zones by geographical information system (GIS)- based spatial multi criteria evaluation (SMCE). The necessary geospatial data were obtained, processed, and converted into a grid format. The contribution level of each factor for triggering landslides was evaluated by the Analytical Hierarchy Process (AHP) and modelled with the SMCE. The developed SMCE model is at an acceptable level because the acquired consistency ratio value is 0.074 (≤ 0.1). The developed LHZ map shows a 90% level prediction accuracy compared to previous landslides. According to landslide hazard zonation map, 13% (227 km2) of the entire area is a very high landslide susceptible zone, while 37% (634 km2) of the total land area has a high susceptibility to slope failure. Moderate and low susceptible zones were 32% (542 km2) and 12% (203 km2) respectively, and only 6% (96 km2) of the entire study area belonged to the very low landslide susceptible zone.

Relevant Publications in Geoinformatics & Geostatistics: An Overview