Spatial Analysis
The spatial analysis portion of this project began with the classification of the factor coverages into meaningful information. This process was carried out using a Cartographic model in IDRISI. This allowed data transformation to take place in a logical step by step process. The data preparation occurs in the first half of the cartographic model with the reclassification and standardizing occurring in the second half. Once the factor and constraint information was standardized, a MCE was performed in the Decision wizard module of IDRISI. The following image is the cartographic model used to prepare, transform and standardize the factor and constraint data.
All the Reclass modules in the model represent the classification and standardization step of the spatial analysis. Each factor was standardized to a scale of 0 to 100. The following represents the step by step classification scheme and assignment for each factor.
Aspect
Aspect can influence avalanche release area potential in two ways. It determines which slopes are exposed to direct wind and which slopes are protected on the lee side of mountains. Secondly, aspect determines which slopes are exposed to sunlight and which are shaded, a factor which also influences snow deposition and density. Schmid and Sardemann (2003) determined that leeward snow deposition near ridgelines was the single most important characteristic when determining avalanche release areas. The predominant westerly winds and moderately southerly winds in the study region were therefore used to determine weight values for the reclassification of the aspect coverage. Weights of 0 were classed for areas with no aspect, 30 for 180 to 360 degrees, 50 for 300 to 360 degrees, 70 for 150 to 180 degrees, 80 for 0 to 15 degrees and a weight of 100 was assigned for aspects between 15 and 150 degrees. This produced the following standardized factor image for aspect.
Slope
The slope standardization and classification was based on the conclusions reached by Maggioni & Gruber 2003 in which they concluded that most frequent avalanche release occurs between 30 and 60 degrees. Steeper slopes than this range produce smaller more frequent sluffs and slopes lower than this range tend to remain stable. Therefore the factor weights assigned were; 0 for slopes between 0 and 20 degrees, 55 between 20 and 30, 65 between 60 and 90 and a value of 100 for slopes between 30 and 60 degrees. The following map is the result of the reclassification and standardization.
Terrain Shape
The most susceptible terrain shapes for avalanche release include three of the major terrain classifications produced by the TOPOSHAPE module. The most significant terrain shape is ridgelines, followed closely by convex hillsides and planar sloped hillsides (Schmid & Sardemann 2003). The following weights were assigned to each of the 12 terrain shape classes;Peak- 20, Ridge-100,Saddle- 5, Flat-0, Ravine-75, Pit-0, Convex Hillside-95, Saddle Hillside-20, Slope Hillside-80, Concave Hillside-35, Inflection Hillside-30 and Unknown Hillside-5. These weights were used to produce the final factor map for terrain shape.
Elevation
This factor was incorporated into the MCE because of the nature of snow deposition on the North Shore Mountains. Typically in midwinter, the snowline in the North Shore mountains is between 500 and 900 meters above sea level. The higher ridges are usually covered with snow from November until June. The DEM Project Clip coverage was weighted and standardized as follows producing the elevation factor map.
0 from 0 to 200m
5 from 200 to 300m
10 from 300 to 400m
20 from 400 to 500m
30 from 500 to 600m
40 from 600 to700m
50 from 700 to 800m
60 from 800 to 900m
70 from 900 to 1000m
80 from 1000 to 1100m
90 from 1100 to 1300m
100 from 1300 to 9999m
Terrain Coverage
The terrain coverage layer was quite simple to weight and standardize after all the data preparation was completed to produce the 4 class map. It is well documented that terrain coverage is one of the most influential parameters affecting potential avalanche release areas (McClung 2003). Dense forested areas rarely release because of the stability the forest provides to the snowpack. Open areas with smaller vegetation are much more prone to release activity because there is reduced surface friction to hold snowpacks and grasses and shrubs can produce unstable air pockets underneath the snow surface. Bare rock and scee areas have very little surface roughness to add stability and tend to be the most prone areas for release. The weights chosen for each terrain class were 0 for water and urban areas, 10 for the forested areas, 80 for open areas with smaller vegetation, and 100 for bare rock and scree areas. The factor map for terrain coverage is below.
Boolean Constraint
In addition to the five factors above a Boolean layer was created to mask out all water and urban areas that should not be considered in the Multi-Criteria Evaluation. The layer was initially created in ArcMap by creating a shapefile that digitized all major water and urban areas contained within the study region. The shapefile was rasterized and imported into IDRISI as a Boolean image. All areas within the water and urban areas were given a value of 0 and all other areas to be considered were given a value of 1. The map below is the image layer produced.
Multi-Criteria Evaluation
With the all five factors completed and standardized as well as the constraint layer complete; Decision Wizard was used to perform the MCE. All factors were assigned user defined weights and the fuzzy standardization function was not employed because all five factors were already standardized on the same scale. The factors weights were 0.3 for elevation since it was the predominant factor influencing snow supply and release area locations, 0.25 for Terrain Coverage because of the relative importance that forest cover has on release potential. Slope and Terrain Shape were assigned equal weights of 0.175 because they are both static terrain features that always influence release area potential to similar degrees. Finally a weight of 0.1 was assigned to Aspect, lower than the other four because of the tendency of wind direction to change which can deposit snow in varying locations. The constraint was used to mask out the water and urban areas giving them a final value of 0. The map produced gives suitability for Avalanche Release Potential based on a scale of 0 to 100.
