Incorporating Landslide Probability into Operations Planning

Current landslide hazard analysis does a poor job of informing the planning process for forest operations in that landslide models provide maps of landslide hazard instead of quantitative measure of landslide impact. Other aspects of forest management planning (such as road surface erosion) provide quantifiable estimates of impacts such as tons of sediment delivered to the stream. This paper outlines an approach by which landslide probability is used to provide quantitative evaluation of landslide impacts of alternate forest management plans. At the local scale, landslide probability can be used to estimate the probability of landslide impact of a road crossing or a stream habitat. The number of landslides that might be expected across a landscape during a harvest rotation can also be estimated. This probability approach requires estimation of landslide probability for different forest management activities on different hillslopes. The simplest way to estimate the probability of a landslide following a given management on a given slope is to observe many similar activities on similar hillslopes and calculate the fraction that subsequently slid. A more flexible approach would be to use logistic regression, which allow empirical modeling landslide probability as a function of any landslide producing feature.