A Methodology for Verification of Nearest Neighbours Avalanche Forecasts Based on Qualitative Expert Assessments (PSAM-0322)


Snow avalanche related decision-making includes winter road closure, ski track safety and warning of backcountry avalanche hazard issued to the public. For many years, statistically-based forecast techniques such as Nearest Neighbours have been used in avalanche prevention and decision-making, thus creating a sizeable database from which to learn. The benefits realised by statistical models depend on how well the requirements of decision-making are matched and on the value of the information delivered by the models. In order to set up a knowledge base for the verification of these techniques, we collected daily assessments on the characteristic, local avalanche situation made by local avalanche experts with permanent presence in the terrain. The information was collected by use of a web-blog in verbal form including descriptions of the weather situation, snowpack build-up, snow instability and mitigation measures when available. After the first winter of blogging, we matched the collected information with the results of a Nearest Neighbour implementation adapted to and used for local avalanche forecast. Bearing in mind the small size of the 1st-year sample, the comparison indicates that the textual expert descriptions give a comprehensive summary of the daily snow conditions, that the tested model tends to well describe the weather conditions but displays little skill in selecting snowpack conditions. Regarding snow stability the results of the comparison are ambiguous and display no clear tendency as of yet. The method described thus seems to have considerable potential, given larger samples, to study the strengths and weaknesses of model outputs for avalanche forecasting in detail.

  • Summary/Abstract
  • Introduction
  • Data - The Blog Knowledge Base
  • Method - Verification of the Similarity of Nearest Neighbors
  • Results
  • Discussion
  • Conclusions
  • Acknowledgments
  • References

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In