Photo: Slide on the Seward Highway, AK about to hit tidewater
Photo by Terry Onslow, www.avalanche.org

Avalanches kill well over 100 people per year.  Approximately 150 deaths are reported per year by the 17 countries that are members of the International Commission for Alpine Rescue (ICAR).  However, many other avalanche accidents occur in other countries that are not a part of ICAR, such as Turkey, Russia, and China.  Avalanches also damage infrastructure, which includes roadways, rail, and pipeline.

There are many attributes to avalanche initiation.  Slope angle, direction the slope is facing, elevation, groundcover type, tree denseness, terrain shape, and snow surface type all play a role in avalanche initiation.  It is expensive and time consuming to monitor all areas extensively by conventional methods, which is done hands on in the field.  Avalanche monitoring is done extensively in popular regions, such as Little Cottonwood Canyon in Utah, reducing avalanche impact on the state's highway system in this area.  Less popular areas that also have high risks of avalanche incidents that could damage infrastructure have low or nonexistent avalanche monitoring.  Cost is an issue.  It is expensive to monitor all areas that are avalanche prone.  However, air photos and satellite images have reduced this cost of getting detailed terrain models that identify and monitor avalanche hazards.  As technology continues to further itself, satellite imagery, digital air photos, digital elevation models, and road network data will greatly assist in observing, predicting, and responding to avalanches.  Digital air photos can be used to manually find and delineate past slides.  Landsatt data can be used to automatically detect avalanche paths on the landscape where vegetation has been disturbed.  Digital terrain models can be used can be used to 3-D models of potential avalanche runout zones.  Remote sensing/GIS-based analyses of avalanche paths are improving.

Clover Point Cartographics Ltd.came up with a model for predicting avalanche potential based on Digital Elevation Model data for their area of interest, British Columbia, Canada.  They wrote a series of text files containing a sequence of commands that can be executed as one command to automate the avalanche modeling process.  These text files are called macros.  The avalanche model predicted the probability of an avalanche occuring based on the terrain and snow accumalation.  Terrain Resource Information Management (TRIM) was their main source of raw data.  TRIM consists of 7027 mapsheets covering British Columbia at a scale of 1:20000 with a coordinate system of Universal Transverse Mercator based on NAD83.  TRIM contains elevation points and breaklines that break up the terrain.  Breaklines show the locations of ridges and valley floors.  A triangulated irregular network (TIN) was made from raw DEM data from TRIM.  TIN is a surface representation derived from irregularly spaced points and breaklines.  From the TIN, polygon coverages are derived to represent slope and curvature.  Breaklines were buffered to form elevation class polygons.  To buffer someting means to form a zone of a specified distance around features in a coverage.  The TIN slope, curvature, and buffer coverages were overlaid and the final coverage was queried using the avalanche prediction model.  An example of a map formed by this process is shown below.

Snow avalanche likelihood based on assessment of terrain and snow accumulation. Ratings are colour-coded. Red indicates that a destructive avalanche is almost certain to occur within a 30 year period. Green indicates that it is almost certain that no avalanche will occur.

Today, Switzerland has the most advanced avalanche monitoring system in the world.  Switzerland's avalanche monitoring system contains high quality remote sensing technology.  Snowpack information from 60 ridgeline sites is combined with remote sensing data is sent to a central computer every 30 seconds.  The central computer calculates the probability of dangerous weak layers of snow near the surface that could cause an avalanche.

Avalanche monitoring will continue to grow as GIS grows.  Avalanche monitoring is expanding to lesser populated areas to prevent casualties and damage to infrastructure.  This will continue to do so as GIS and remote sensing become more advanced and less expensive.

Sources:
Avalanche Monitoring and Mitigation, Thomas Cova, University of Utah
Avalanche Hazard Mapping, Clover Point Cartographics Ltd.
Avalanche!, GEOEurope
Avalanche Hazard Mapping on the North Shore Mountains, Vancouver using Satellite Imagery and GIS, S. McLaren