2015 CSCE Annual Conference Regina - Building on our Growth Opportunities

2015 CSCE Annual Conference Regina - Building on our Growth Opportunities Conference

Improved forecasting of persistent deep slab avalanches with a decision support tool

Dr. Michael Conlan, University of Calgary (Presenter)
Dr. Bruce Jamieson, University of Calgary

Persistent deep slab snow avalanches are generally difficult to forecast because of the relatively large depth to the failure layer.  A decision support tool was developed to help professionals forecast the likelihood of such events.  The tool follows a threshold sum approach, where users answer yes/no questions and all yes answers are summed and compared to threshold values.  The questions were derived from parameters important in the release of persistent deep slab avalanches.  The questions comprise three sections: snowpack conditions, weather conditions, and avalanche observations.  Values within each of the questions were obtained from three independent data sources, including persistent deep slab avalanches that were accessed, a dataset of historical persistent deep slab avalanches, and an expert opinion survey.  A classification tree was used to determine the threshold tool sum that separates days with higher likelihood from days with lower likelihood.  For a dataset of 110 days, the tool correctly classified 89 % of days where naturally triggered persistent deep slab avalanches were observed and 74 % of days where avalanches were not observed.  The tool also indicates whether persistent deep slab avalanches triggered from skiers or snowmobiles are possible, depending on the local snowpack conditions.  The output of the tool only indicates the likelihood of such events within a forecast region and cannot predict when or where they will occur.  The tool may help streamline the decision making process for some avalanche forecasters in mountainous terrain.