Evaluation of the snow penetrometer Avatech SP2
Hagenmuller, Pascal ; van Herwijnen, Alec ; Pielmeier, Christine ; Marshall, Hans-Peter
Année de publication
Information on the structure of the snow cover is of great importance for operational avalanche forecasting. Manually observed snow profiles, including snow hardness, are therefore widely used to characterize snow stratigraphy. However, such manual measurements are subjective and observer dependent. While the ramsonde resistance profile provides a more objective alternative, it lacks the vertical resolution to identify thin layers and the hardness resolution to identify soft layers, which are essential components of a snow stability assessment. To overcome these limitations, digital cone penetrometers were developed to collect accurate and objective snow cover stratigraphy data. The SnowMicroPen provides highly resolved and accurate hardness profiles but its high price, size and weight, and fragility limit its usage to research purposes. Recently, a new lightweight penetrometer was released, the SP2, intended to fill the gap between the low-resolution ramsonde and the more expensive, accurate, high-resolution SnowMicroPen. We conducted an objective comparison of co-located vertical profiles measured by these three instruments in combination with manual stratigraphy and stability tests, in the French and Swiss Alps and in North America during winter 2015-2016. The SP2 profiles showed stratigraphic features similar to the SMP profiles. However, SP2 measurements were less repeatable with a profile variability generally larger than the spatial variability and influenced by the operator handling such as the penetration speed. The vertical accuracy was relatively low: the total depth was measured with a standard error of 7.5 cm and vertical shifts of the layer position were measured in the range [−10, 22] cm with a standard error of 7.4 cm. Hardness measured by the SP2 showed no significant bias and a standard error of 34 kPa compared to the SMP. The SP2 resolution, as the ramsonde resolution, was too low to detect the weak layer of the considered new snow problem but sufficient to detect the weak layer associated to the considered old snow problems. Nevertheless, on these problems, the accuracy of the SP2 to characterize the slab and weak layer properties was too low to effectively derive stability indices directly from the SP2 profiles.