2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 1
Presentation Time: 1:30 PM-5:30 PM

KLANAWA LANDSLIDE HAZARD MAPPING PILOT PROJECT


ROLLERSON, Terry, Golder Associates Ltd, 1462 Broadview Road, Gabriola, BC V0R 1X5, Canada, MAYNARD, Denny, Denny Maynard and Associates Ltd, 222 West 22nd Sreet, North Vancouver, BC V7M 2A1, Canada and HIGMAN, Shelley, Weyerhaeuser Co Ltd, 65 Front Street, Nanaimo, V9R 5H9, Canada, trollerson@golder.com

The Klanawa Landside Hazard Mapping Pilot Project utilizes categorical and scale data from terrain and landslide inventories to produce semi-quantitative landside hazard maps for forest management purposes. The term “semi-quantitative” is used because although statistical methods are used to generate landslide density values, the terrain and landslide inventories used to generate the basic data are qualitative products based on air photo interpretation, limited field validation and professional judgement. This project is part of a larger study that includes watersheds in the Vancouver Island Ranges, Coast Mountains, and Queen Charlotte Islands.

The approach depends on statistical inference based on landslide densities determined over relatively large areas and involves four steps:

1. Map and database creation, involving terrain and landslide inventories in a series of logged areas representative of a larger area.

2. Intersection of terrain, landslide and other data (e.g., bedrock geology and slope) using GIS.

3. Analysis of the data using univariate and multi-variate statistics to identify relationships between predictor variables and landslide rates.

4. Developing simple algorithms to generate landslide hazard maps that portray expected hazard as landslide densities for forest road construction and harvesting.

We assume that most landslides related to logging and road construction occur in the first 5 to 15 years after harvest. We also make the assumption that relationships developed in logged areas can be extrapolated to similar terrain in forested areas within the same geographic area. What is not known is if the landslides rates documented after the first-cycle harvest will reflect those that will occur after second-growth harvesting.

The landslide density values developed by this approach are considered an “index” of landslide hazard. They provide an objective way to rank landslide hazards. The landslide density values represent mean values for different categories of terrain. The landslide rate that occurs within a specific map unit can vary from the mean value reported for that category of terrain.