Modeling of Snow Avalanche Potential in Nordenskiöldland, Central Spitsbergen

Slab avalanche released late April in upper Fardalen, central Spitsbergen, following a brief period of thaw. A well defined fracture line (70-100 cm high) delimits the slide from stable snow above. Photo 2002.04.24.

Ole Humlum, UNIS, Department of Geology, Svalbard, Norway

Click here for list of general avalanche advice

Click here for general information about the physical environment in Svalbard



A dynamic numerical 3D terrain model entitled VirtualWorldForWindows (VW4W) is currently developed and tested at UNIS. This model, which runs on normal PC's, takes topographic data, terrain surface characteristics (geomorphology and vegetation) and meteorology (air temperature at sea level, wind speed, wind direction and cloud cover) as input and yields as output information on various phenomena such as terrain surface net radiation balance, snow cover thickness, snow cover duration, glacier mass balance, ground surface temperatures, growing and freezing degree days, maximum active layer thickness, stable permafrost thickness and the amount of melt water discharge during the summer. Click here for further information on this numerical model. Click here for general information on snow avalanches.

Numerical modeling has its  limitations and the results shown below are research results only, which will be evaluated and tested by field observations. They do not guarantee that all avalanche risk slopes slopes are indicated below. The scale is one significant limitation. Even a slope 5 m high may be the source for avalanches, even though it is not visible in a topographical model such as shown below.



The form of the physical landscape (its topography) is the background for any attempt of modeling geomorphic processes, such as, e.g., snow avalanche activity. The scale and detail represented by the digital terrain model is clearly important for the result of the subsequent modeling. The topographical model used in the present modeling attempt is seen in the diagram below. Clearly this does not address details in the landscape, and provides a general overview only.

Diagram showing topography in Nordenskiöldland.



Most avalanches of dangerous size originate on slope angles between 30 degrees and 45 degrees. They seldom occur below 30 degrees and hardly ever below 25 degrees. Above 45 degrees to 50 degrees sluffs and small avalanches are common, but snow seldom accumulates to sufficient depths to generate large snow avalanches. Slopes steeper than 50-60 degrees usually do not acquire any significant snow layer due to their gradient. One type of output from VW4W is a map (below) showing the spatial distribution of slopes statistical exposed to snow avalanche danger, from a pure slope angle point-of-view.

Map showing distribution of slopes statistical exposed to snow avalanche activity.



Snow avalanche activity requires precipitation in solid form (snow). From this point of view, the mean annual distribution of precipitation across the study area is important. This can be modeled by combining existing meteorological observations from Nordenskiöldland, glacier mass balance investigations, and observations on the equilibrium line altitude on glaciers as mapped from high-quality aerial photos. 

The result of such a modeling attempt is shown in the diagram below. Click here for PDF-file of published paper on how to model precipitation in Nordenskiöldland.


Map showing the modeled mean annual precipitation in Nordenskiöldland (mm w.e.).



Snow blowing into Longyearvalley from adjoining Gruvefjellet mountain plateau during strong SE airflow. 4. April 2005.


The wind significantly redistributes snow during the winter, accumulating the solid precipitation in certain valleys (see photo above), while adjoining mountain ridges and -plateaus are blown almost free of snow.  Presumably, many glaciers in Nordenskiöldland would not exist under present meteorological conditions without this process. The numerical model (VW4W) attempts to model such snow drift. 

Below is a diagram showing where the model indicates that snow is concentrated by the prevailing winter wind (SE) across the mountains. Avalanche potential may be expected to increase with the amount of snow accumulated by wind transport, especially downwind of extensive mountain plateaus acting as source areas for drifting snow.


Diagram showing modeled accumulation of snow by prevailing winter wind (from SE). During periods with wind from other direction, the location of significant snow accumulation will change.


Sky view factor

Snow avalanches, once released,  will continue down across lower parts of the valley side, even though the local slope may be smaller than usually considered dangerous. Some avalanches may even cross the valley bottom and continue some distance op the opposite slope,  depending upon size and speed. The total horizontal distance affected by avalanches is known as the run-out distance. From this point of view a deep and narrow valley clearly is more dangerous than a shallow and broad valley. VW4W are able to calculate the relation between vertical relief and horizontal space in valleys, known as the sky-view factor. A high value (max 1) indicated a horizontal landscape without relief, while a low value indicate a deep and narrow valley. The potential for any snow avalanche crossing even the valley axis is therefore high when the sky-view factor is low, and vice versa. The calculated sky-view factor for Nordenskiöldland is shown in the diagram below. 

Diagram showing the calculated sky-view factor for Nordenskioldland.


Potential for snow avalanches

By combining all the factors mentioned above into a single dimensionless index VW4W attempts to indicate the spatial distribution of zones with high and low potential (risk) for snow avalanches, under typical, modern winter conditions. A diagram showing this potential is seen below. As mentioned above, the distribution of the individual zones will change somewhat during periods with other wind conditions than SE across the mountains. 

A high potential for snow avalanches does not imply that avalanche activity will take place each year, but only indicate that such slopes should be traversed only following proper evaluation of the prevailing local snow and weather conditions. A model can never be a substitute for direct field observations. However, the map below may hopefully assist in planning routes minimizing the snow avalanche potential, even though it should be considered only a first approach addressing a complex modeling issue.


Diagram showing the spatial distribution of modeled snow avalanche potential zones. Higher numbers (darker color) suggest increasing avalanche risk.


Improving the model

Clearly other factors (e.g., crystal geometry, metamorphosis, etc.) will contribute to the potential for snow avalanches, and the present modeling only represents a first attempt towards assessing avalanche danger in Nordenskiöldland. The next important step is verifying or falsifying the output in order to improve the model. 

Avalanche observations are an important factor for operational avalanche warning and the main parameter to carry out an objective verification of avalanche models. Therefore, all field observations (timing, position, size, photos, etc.) on snow avalanche activity in Nordenskiöldland are warmly welcomed (

Avalanche in upper Bødalen, nearly overflowing Little Ice Age moraine in front of Bødalsbreen. Flow direction from right towards left. Pressure ridges were generated by compression during final phase of flow. Avalanche event presumably took place shortly before 19. March 2003. Photo 5. April 2003.  



Latest update by Ole Humlum 19. December 2005.