Propagation Tests in SnowPilot

The Extended Column Test (ECT) and the Propagation Saw Test (PST) show crack propagation, a fundamental part of the avalanche process. Many studies have examined the accuracy of these tests at predicting stability, but only a few compare the tests side-by-side. Side-by-side tests are the only way to fully control for the many factors that affect crack propagation. Moreover, most of the comparisons have been from research data. We have limited knowledge of how these tests are being used by avalanche professionals and backcountry travelers. SnowPilot is the largest public database of stability tests in the world. In this study, we examine 256 snow pits from the SnowPilot database with 513 ECTs and 345 PSTs conducted side-by-side. Because results of the ECT and PST cannot be directly compared, we simplify test results by classifying them as unstable or stable. We classify a test result as unstable if it is 'ECTP/ECTPV' or 'PST End' with a cut length ≤ 50 cm; otherwise we classify a test result as stable. We find that: 1) PSTs showed unstable results more often than ECTs, 2) the subjective stability rating ('stability on similar slopes') was correlated with ECT stability, but not with PST stability, 3) PSTs were used on deeper slabs than ECTs, and 4) PST use increased with a decreasing stability rating, relative to ECT use. Result (1) is supported by 2 of 3 other studies with side-by-side tests. We suggest a potentially larger 'cracked' area in the PST as one cause, resulting in a larger crack nucleus and increased edge effects that promote propagation. Result (1) contradicts previous work that shows the PST has a higher false-stable rate than the ECT. One would expect Result (1) to cause a lower false-stable rate. Result (2) shows that either the ECT is an accurate test or that users are basing their stability assessment on ECT stability results. This correlation is problematic for studies that use SnowPilot's stability field to infer test accuracy. Result (3) supports studies and guidelines that suggest the PST is more effective for deeper slabs, but may also be an artifact of the guidelines for the test. Result (4) may arise because users are more likely to add a PST, in addition to an ECT, as stability worsens. We advise going in the other direction, adding a PST while stability is increasing since this is often more difficult to forecast. Percent of unstable results in side-by-side tests. Sample size shown in vertical type.