Relating meteorological variables to the natural slab avalanche regime in High Arctic Svalbard

Traffic is affected by natural slab avalanche activity in the mountainous landscape around Svalbardʹs main settlement Longyearbyen, at 78° N in the High Arctic. Without any avalanche forecasting system in place, the analysis of meteorological variables to identify the best suitable discriminators for avalanche and non-avalanche days is timely and important. Thus we have studied the 4 major meteorological variables air temperature, wind speed, precipitation and snowdrift for 0, 24, 48 and 72 h prior to an avalanche and a non-avalanche day, deriving 32 combinations to test differences. Combined with an avalanche dataset including 156 natural slab avalanches that released on 20 avalanche days between 2007 and 2010, we carried out a Wilcoxon rank sum test and linear regressions to differentiate the best avalanche potential discriminating variables. Nine meteorological variables differed significantly. Precipitation and snowdrift 24, 48 and 72 h prior to an avalanche and non-avalanche day were the best predictors. Minimum, maximum and average wind speeds could also be used as good indicators of avalanche activity, whereas air temperature was not of any significance. Major shortcomings of our analysis were the infrequent field observations, mainly during the polar night and the small number of avalanche days compared to non-avalanche days, unsuitable for significantly discriminating avalanche from non-avalanche days. The latter is however a specific characteristic for the Svalbard landscape, as well as the large amount of cornice fall triggered slab avalanches that are not included in this study. Measuring precipitation in the High Arctic barren and windswept landscape is difficult, causing uncertainties in precipitation amounts, therefore caution must be used when interpreting the results.