Field observations for estimating the local avalanche danger in the Columbia Mountains of Canada

Snow avalanche danger can vary considerably within the forecast regions, especially large regions. Experienced recreationists routinely use the regional forecast along with local observations to estimate the local avalanche danger. However, some less experienced recreationists are unsure how to interpret the various field observations. To assess a systematic approach, we conducted a field study during the winters of 2006–07 and 2007–08 in the Columbia Mountains of western Canada. Experienced observers rated the local avalanche danger and made 24 observations of weather, avalanche activity and simple manual snowpack tests on approximately 130 location-days. Since the local danger was often rated separately for the elevation bands alpine, treeline, and below treeline, the observations could be applied to 272 individual local danger ratings. Fourteen of the potential predictors yielded significant rank correlations with the local avalanche danger. Reflecting their larger scale, many of the weather variables correlated better with the regional danger forecast than with the local rating. In contrast, some snowpack observations including the hand shear and ski pole test correlated better at the local scale than the regional scale. Classification trees using the regional rating plus three of the local observations exhibited a better agreement with the local danger rating than did the regional rating by itself.