SUGGESTIONS FOR SNOW RESEARCH
Apt. 103 - 105 West Kings Road,
North Vancouver B.C. Canada, V7N 2L7
Over the past 38 years I have studied snow; applied results of snow research; and served committees that were concerned with snow control on roads, snow engineering, avalanches, and observations of snow. Many questions which could not be answered were encountered with this activity. I found a few answers by carrying out research, abandoned the search for others owing to difficulties in finding time, working facilities, or equipment for research, and a third category of problems never advanced beyond a list a desired study projects. This paper gives me an opportunity to present a few of the unfilled wishes and to encourage their considereation by future snow researchers and the Center for Snow Science at Alta.
The list of snow problems is long, because numerous activities involve interaction with snow. The principal problem areas are: avalanche safety; loads on structures; traffic on roads; hydrology; locomotion of vehicles on snow; friction of skis; compaction of snow for roads, airfields, and ski runs; blowing snow; and effect of snow on forestry, agriculture, and wildlife. For some applications the economic impact of snow is small, therefore it may be difficult to justify research; often we discover that we can live with the problem without much discomfort. Below, I will concentrate on the few applications which I am most familiar with.
OBJECTIVES FOR SNOW RESEARCH
Snow research should supply information for effective, economical, and safe design and operation of facilities and living conditions in a snowy environment. The information must usually answer two basic questions:
a) What are the maximum design conditions? Examples are: design snow loads on roofs and maximum runout distances of avalanches.
b) What are the conditions now and in the near future, and how will they affect or impact on operations and decisions? Examples are: trafficabilty on a road, snow stability and avalanche hazard, and expected runoff during snowmelt.
The common approach to solving snow problems is:
1. To build a model of relationships between snow properties and the design conditions.
2. To measure index properties of the parameters of the model; they include properties of the snow and the weather.
3. To calibrate the model by determining critical values of the index properties.
Usually, models are based on snow physics and they often require basic research for their development. In order to be useful, a model should:
- simulate the problem,
- be physically based,
- be applicable by users.
Index properties should:
- have a unique relationship to the problem,
- be close to the answer of the problem,
- be readily available,
- be measured reliably and according to accepted standards,
- require a minimum of time for data collection.
Data collection on index properties is a difficult part of solving snow problems. Common concerns are:
- selection of observation sites,
- measuring techniques,
- amount of data that should be collected,
- accuracy and significance of the data,
- application of point measurements to large areas,
- quantifying subjective observations,
This was confirmed when in 1984 a working group distributed a questionnaire to snow researchers in Canada with the request to describe their projects and main impediments. The answers mentioned data collection and instrumentation as the principal problems, and funding and manpower next in importance. Knowledge of the material properties of snow appeared to be a lesser problem.
A common complaint of designers of works in snow environments concerns the lack of readily available data for application of predictive models or even rough estimates of effects. As an example: Where can we find the expected range of density of deposited dry avalanche snow about 1 m deep?
Models and snow measurements are highly probabilistic, primarily because snow has widely varying characteristics and secondarily due to inaccuracies in measurement techniques. When we consider the significance of the variations, it is surprising how few studies have attempted to quantify variabilities and to correlate them with characteristics of terrain, exposure, and weather. I believe too much effort has been spent in the past trying to explain snow metamorphism and not enough attention has been given to measurements of the mechanical properties and their variations.
Some problems can be solved to a large extent with laboratory tests. For example, blowing snow can be simulated in a wind tunnel or a water flume for the solution of snow drift problems on and around structures, on complex terrain, and road cuts. Other problems can be approached by mathematical-probabilistic modelling, for example the accumulation of snow on roofs under variations of snowfall, wind and temperaturs. All laboratory and mathematical models, however, require verifications in the field. For a complex material like snow, research must be supported by field observations.
In conclusion, the priorites of snow research should be:
1. Development of techniques and instrumentation for data collection on index properties.
2. Studies of snow physics to find the best types of models and index properties.
3. Compilation of index properties and development of critical values.
Below, I have suggested a list of problems for future research.
In Situ Tests of Snow Stability
In situ tests can directly test the stability of snow in its natural position. This may be by loading it; or measuring the strength of weak layers, the load on them, and calulating the stability from the combination of strength/load. The tests supply high quality information about snow stability, but they are not applied as extensively as warranted. Ideally, an in situ test for snow stability should meet the following conditions:
- It should stress the weak layer in shear.
- The force applied must stress the snow to failure.
- The area under stress should be as large as possible.
- The test should be suitable for application both on a sloping and a level snow cover.
- The preparation of the test site and the repeated observations should require not more than 30 minutes.
- The equipment must be simple and portable.
- A test should give unbiased, quantitative, and objective information.
- The test result should be available immediately without complex analysis.
The following tests - none of them meeting all the conditions - were used successfully but requre additional development.
The Rutschblock is the best available test for direct observation of snow stability, but it has the disadvantage of requiring a site that has an incline and is safe. I wish to suggest as a further development:
a) an analysis of the optimum location where the skier who loads the block should place himself (in the centre, or close to the back cut?);
b) an analysis of the minimum slope incline for which the test supplies useful results.
Loading of a snow column is a stability test similar to the Rutschblock test, but uses a narrow snow column and the load is applied either by pressing with a shovel or by piling up snow blocks. I wish to suggest the development of a calibration for loading and the definition of the limitations of the test. There appears to be a lower limit of slope angle on which the test can be applied, because on level terrain the snow would fail in compression rather than in shear.
The Shear Frame Test is only a strength test, but in association with measurements of the vertical load on the tested snow layer, it has proven to give useful information about snow stability. Numerous studies about the optimum size of the frame and the variability of the tests were published. The following requires further attention:
a) An analysis of the stress distribution with the objective of determining the optimum height and spacing of the cross bars.
b) Development of critical index values for stability in an area.
c) Promotion of its application in operations; I believe, the great value of the test merits its application on a regular basis for operations with study plots close to avalanche starting zones, for example ski areas.
In addition, the development of other tests is encouraged that give quantitative information about snow stability or measure the strength of weak snow layers and boundaries between layers. The field is wide open, but the conditions for useful in situ tests should be kept in mind.
Spatial Variation of Strength and Stability
The variation of snow strength and stability with terrain is a major problem in avalanche hazard evaluation. Variations of depth, density, hardness, and content of free water are concerns in applications such as loads on structures, snow removal, compaction, and trafficability. The variations may be significant over short distances. Several snow researchers - for example, Richard Sommerfeld, Paul Foehn, Bruce Jamieson, Hansueli Gubler, and Karl Birkeland -have made useful studies of the variation of snow properties. I wish to encourage their work. Results should include:
a) Guidelines for estimating likely spots of weaknesses. Experienced ski and mountain guides have learned to suspect weaker spots by judging terrain and the appearance of the snow surface; their experience should be added to the studies.
b) Quantitative information about the variability. The information would assist in risk analysis.
Prediction of Snow Strength
A problem in snow stability evaluation is to estimate for remote slopes how rapidly weak layers may gain strength. Certain types of snow, for example surface hoar, plates, and facetted grains remain weak for long periods of time, but suddenly become strong, sometimes for no obvious reason. I wish to suggest studies about the conditions and the predictability of changes of snow strength under variations of temperatures, loading and other conditions for types of grains that often form weak layers. The studies could include basic analytical studies, laboratory tests, and field observations of shear strength.
Classification of Snow Stability
The definitions of snow stability are not satisfactory as yet. I wish to encourage the development of a uniform classification system with clear definitions. The trend in public avalanche warning definitely is towards issuing information about stability rather than avalanche hazards. The problem is under discussion among countries of Western Europe, and we should co-operate.
Remote sensing of snow depth, settlement, precipitation, wind, and temperatures has become an important part of data collection of operations, but the following problems still need to be addressed:
a) Reliability of instruments. Problems exist with capping of snow collectors, icing and riming of sensors, and failure of transmission equipment due to low temperatures and moisture.
b) Optimum location of the instruments.
c) Cost versus benefits. The high cost of manual observations is one reason for using remote sensing, but often the capital and maintenance cost of the instruments are greater than the savings.
Snow Loads on Buildings
The common approach to estimating design roof loads is to determine ground loads, then with empirical relationships to determine from the ground load the roof load for various shapes and exposures. The ground load is the unit weight of the undisturbed snow on the ground at the buiding site. Most ground loads are estimated from snowpack observations a considerable distance from the building site and adjusted for differences in elevation, exposure, and local climate. Due to lack of better information, many ground loads are estimated from snow depth observations with rainfall added.
The following information need to be developed:
a) Correlations between snow depth and ground load for a variation of climates.
b) Influence of the local meso and micro climate on the ground load.
c) Amount of rain that can fall into and be retained by the snowpack.
The methods for estimating roof loads from ground loads and from climate information have received considerable attention in recent years. Analytical models of snow accumulation, physical modelling in wind tunnels, and full scale observations on roofs were applied.
Risk Analysis, Risk Acceptance, and Cost-Benefit Studies
Many activities concerned with snow contain a risk. Risk analysis involves an identification of possible damage (measured in terms of potential fatalities or property damage) and an estimate of the probability of it to occur. In snow engineering, risk assessment has been applied for land-use planning in avalanche paths and snow loads on buildings. There is a need to apply it also for:
- Traffic on roads with respect to avalanches;
- Users of ski areas with respect to avalanches;
- Avalanche hazard evaluation for back country travellers;
- Problems of deep snow and blowing snow for road traffic.
For all applications, tolerable risk levels should be defined by making comparisons with other hazards such as fire, floods, flying, boating, and injury in skiing. The tolerable risk levels should be clear and understandable by professionals, management, and politicians.
Cost-benefit studies are closely related to risk analysis. Studies should be carried out as examples of whether or not an avalanche control is worthwhile for specific applications.
Present models for estimating avalanche speeds, runout distances, and impact pressures serve many engineering applications but the following problems still require basic and analytical studies and calibration of models:
a) Effect of the pulsating nature of dry flowing and powder avalanches on structures.
b) The nature of impact pressures as functions of size and shape of the loading surface.
c) Prediction of the flow depth of avalanches, particularly of the dense core of a mixed dry flowing and powder avalanche. Information about the flow height is significant for the design of deflectors, and identification of surface areas exposed to avalanche forces.
d) Density of the snow in motion.
e) Speeds and runout distances for avalanches on short slopes (typically with differences in elevation less than 100 m).
f) Field observations of the speeds of all types and sizes of avalanches.
About 25 years ago snow removal from roads, railways, and runways of airports dominated the list of snow problems. In the meantime, efficient snowplows have been developed and effective methods of snow removal and ice control be identified. Snow removal problems now are concerned with operations: planning for the winter, levels of service, efficient use of equipment, training of manpower, timing of operations, and prediction of ice formation. At this time there seems to be little demand for knowledge about snow properties.
The results of the research on predictive models and data about snow properties must be made available to users. Two problems arise in selling the information:
Often studies were carried out many years ago, but the results became buried in files or were lost. Users are not aware of them, and new researchers often re-invent old techniques. There is a need for agencies that ensure the continuity of data collection and availability of information. The Swiss Federal Institute for Snow and Avalanche Research at Davos is a good example of an agency where data were maintained for more than 50 years and experience is transferred from one generation of researchers to the next. It is hoped that the Centre of Snow Science at Alta will be able to assume a similar role in the U.S.A.
The second problem concerns a mental block which many practicing snow technicians have towards applying research results. In order to assist application of results, snow researchers should not be satisfied by describing their experiments in scientific journals; they should be translated into a language which managers and technical personnel understand. Meetings such as the International Snow Science Workshop, and publications such as the Avalanche Review assist by making research available to practicioners.
Difficulties in Snow Research
Additional difficulties in carrying out snow research are:
In snow, numerous parameters usually appear in combination; they cannot be studied in isolation.
Most field observations need to be made in the winter when university staff and students are occupied with courses and are unable to spend sufficient time in the field.
Access to study sites, timing, and logistical support for field work is often difficult. Experience has shown that a person should be in the study area on most days of the winter and be prepared to make observations on short notice when the conditions are right.
Operational personnel (ski areas, roads) usually cannot be relied on for data collection. In most cases, operational staff is too busy with its own work at the critical times or not motivated enough.
Research projects require much time because of the complexity of the material and the limited annual time available for field work. The long time often does not allow researchers to meet the demand for results and publications that is necessary for maintaining their scientific reputation.
Several factors contribute to difficulties in obtaining support for snow research:
Usually there is a need only to solve a problem which snow is associated with, but not to obtain information about snow. In many applications snow is a nuisance, rather than a problem and the commercial benefits of snow research are not obvious.
Generally, the problems have a short duration and disappear when the snow melts. Because the human mind tends to blot out bad experiences, it is difficult to stimulate support in the summer for a problem that has "disappeared".
Most modern researchers are burdened by administration, including finding funds. Almost all of them feel that the hassles invoved make it not worth the work.
Snow research has numerous applications, researchers have a great selection of problems, but carrying out snow research requires much dedication, time, and patience.