Application of datamining techniques to the production of landslip susceptibility maps

This initial study has provided proof of concept for a larger project examining landslides and erosion in the Corangamite region in South West Victoria and is jointly funded by the Corangamite Catchment Management Authority, the Colac Otway Shire, the City of Greater Geelong, and the federally funded National Disaster Mitigation Plan (NDMP).

Susceptibility maps show a quantitative or qualitative assessment of the spatial distribution of a hazard (in our case landslides) which exists or may exist in an area, and play a vital role in development approval processes and planning of remedial work. This project examined whether recently obtained high-resolution LiDAR (Light Detection and Ranging) data which provides extremely accurate information about the slope and shape of the land surface can be combined with other factor data using data-mining methods to more accurately identify areas susceptible to landslide.

The susceptibility maps produced by the data-mining algorithms are showing great promise in more accurately defining areas of potential instability. Two recent landslide events within the study area (one on the Bellarine Coast and one along the Barwon River in Geelong) were included in areas predicted as having high landslide susceptibility.

Whilst this pilot study was initially limited to the area of the Bellarine Peninsula, the successful completion of the trial has provided confidence to apply the techniques to a larger selection of regions within the Corangamite area.