Team Depth Hoar Homepage

Methodology

The search for avalanche fatality data:

We attempted to find all avalanche fatalities that occurred on the western side of North America.

Data we found:

Avalanche fatality data was obtained from Dale Atkins of the Colorado Avalanche Association as well as Snow Torrents (USA accidents) and Avalanche Accidents in Canada. Canadian fatalities were plotted from 1955 to the spring of 2004 and American fatalities were plotte from 1983 to 2004 due to data availablity.

Problems with the data:

Some of the fatalities had ambiguous locations, for example, Monashee Mountains. Therefore we were unable to pinpoint the exact location of some deaths. The data obtained containing American fatalities contained greater inconsistencies pertaining to the locations of the fatalities. For example, some locations are listed as entire mountain ranges whereas other locations are as specific as the name of a chute at a ski area. Reasons for this inconsistency are unclear. However, in our data set Canada has greater accuracy when specifying where fatalities occurred. One reason why the U.S. may be less consistent is because predicting and forecasting avalanches is comprised of independent avalanche stations located throughout mountainous states. This could allow for some inconsistencies when reporting avalanche fatalities. Even still, we had to find the geographic coordinates of the fatalities which was a challenge in itself. In the GIS world there is a saying that states: "output gained is only as great as input used." In other words, the conclusions that are made are based upon the data used in the study. If faults are apparent in the source data then they may compound themselves into the final analysis, skewing the results. Fortunately it was the number of data points held greater importance than the actual location for our analyes.

Map data:

Basemaps for our location of interest were easy to obtain as they were located on the UBC Geography server. However, obtaining elevation data for all of North America turned out to be a difficult task. After what seemed to be days of searching for free data on the internet with no luck, Jeff Phillips presented us with exactly what we needed, a DEM of North America. Team Depth Hoar was now ready to plot the coordinates on a basemap and DEM with NAD83 projection. In this project, DEM’s of British Columbia and Alberta of a resolution of 769m was used. A DEM of North America was also used with a resolution of 1000m.

Building our own Database:

Perhaps the most challenging and time consuming part of our project was organizing and simply obtaining all of our data. Between the 4 members of our team, we had to put together a database consisting of over 5000 fields all in the same format. This took some incredible teamwork and organization. Our data base consists of the following information:

year (1955-2004)

group number

fatalities

lat/long coordinate (decimal degrees)

type of activity

ski area (all of western North America)

ski area coordinates

The following is the list of activities used for the classification:

Ø Skier

Ø Climber

Ø Snowmobile

Ø Building

Ø Road

Ø Railway

Ø Ice Climber

Ø Unknown

Ø Other

Building the maps:

Plotting:

For each incident, the majority of latitude and longitude coordinates were acquired from http://www.bivouac.com/ (Canadian Mountain Encyclopaedia) and http://www.lat-long.com/; coordinates were then entered into Arcmap and plotted on a base-map. Because the coordinate were obtained from an external source, the accuracy of the data points is only as good as the source itself. We deem these sources to be credible and reliable. For the purposes of the study, the trend of the data points rather than the exact location was what we were after. Therefore, when a location was specified such as the Chugach Range for example, a coordinate was chosen for approximately the middle of the range. Other locations such as the name of a specific ridge in the backcountry were more accurately mapped.

Data Processing:

There were extensive problems with the data processing. Due to limited storage on the UBC Geography server, and high usage rates, the GIS program used (ArcMap) crashed on three occasions, during which all the completed and formatted maps were lost. Near the end of the project when most of the maps were ready for export to the website, the ArcMap file became corrupted for an unknown reason, and it was not able to be restored. These problems resulted in less analysis than Team Depth Hoar had anticipated, as there were many days of just re-doing the same procedures that had failed earlier.

Using ArcMap, Team Depth Hoar used the spatial analyist to create slope and hillshade maps from the DEMs. This processing allowed a spatial projection showing that wherever there are steep slopes, there are avalanche deaths. In addition, population densities were plotted along with the avalanche fatality data, and this showed that there is almost no relation at all between population density and avalanche fatalities. This shows that in pretty much all backcountry locations, regardless of access, avalanche deaths occur.

Metadata:

Source: Our data for our basemaps was collected from UBC Geography in 2005

Our data for avalanche fatalities were from

USA-“Snowy Torrents” Text- on avalanche accidents in the U.S.

Also, Dale Atkins from the Colorado Avalanche Centre provided us with additional fatality data.

Canada- “Avalanche Accidents in Canada” Text-1955-2005 fatality accidents.

Datum: We used NAD83 Albers projection for

Resolution: Resolution varies between US and Canadian maps:

- 769 meters for British Columbia

- 1000 meters for North America

Classifications:

Skier: refers to fatalities associated with skiing and snowboarding, inbound, out of bounds or back country

Climb: refers to fatalities associated with snowshoeing, hiking, mountaineering or any

other like activity

Ice Climb: refers to any technical ice climbing activity. This is a fairly new sport, with spatial restrictions.

Snowmobile: refers to fatalities involving snowmobile use

Building: refers to any fatalities occurring within or around a building, of whatever structure type

Other: refers to any other fatality causing activity. This includes hunting, working, play, etc.

Railway: refers to fatalities occurring on near railways

Unknown: refers to unknown activity type

Avalanche Fatalities in Canada and USA

INFORMATION

MAPS