The Avalanche Review, VOL. 11, NO. 2, DECEMBER 1992
Copyright © All Rights Reserved; AAA


by Bruce Jantieson and Colin Johnston

Department of Civil Engineering. University of Calgary, Calgary, Alberta

During the winters of 1990-1992, rutschblock technique and limitations, variability and precision of rutschblock scores, and applications of rutschblocks to slab stability evaluation were studied in the Cariboo and Monashee Mountains of western Canada. The time required for each test was, under many conditions, reduced to 10 minutes or less by using cords, specialized saws or the tails of skis to cut the two sides and the upper wall of the rutschblock. The median rutschblock score was 4 or less on most days when one or more large dry natural slab avalanches were reported by helicopter skiing guides operating within 30 km of the study area. Also, median rutschblock scores were 4 or less near slabs that had been skireleased and individual scores up to 5 were recorded near recently ski-released slabs. In spite of natural variability of rutschblock scores on a particular slope, decreasing the slope angle by l Ox tended to increase the rutschblock score by 1. A tendency for higher and more variable scores was noticed near the top of several slopes.


The rutschblock test was first used by the Swiss army in the 1960's. Its popularity in North America began following Fhn's (1987) calibration of rutschblock scores.

During the winters of 1990-1992, we performed over 1000 standard and non-standard rutschblock tests on dry snow in the Cariboo and Monashee Mountains of western Canada. Field studies included the following topics: variations in techniques; variability and precision of scores; rutschblock scores concurrent with dry slab avalanches; the effect of slope angle on rutschblock scores; and spatial variability of rutschblock scores on particular slopes.


Test sites should be representative of the avalanche terrain under consideration and undisturbed. For example, to gain information about a wind-loaded slope, you need to find a safe part of a similarly loaded slope for the test. The site should not contain buried ski tracks, avalanche deposits, etc. or be within about 5 m of trees where the buried layers might be disturbed by wind action or by clumps of snow that have fallen from the nearby trees.

Rutschblocks done in starting zones provide the best indication of slab stability. However, safety may require that the test be done on less steep slopes with conditions similar to the starting zone.

The following technique is very similar to that described by Fhn (1987). After identifying weak layers and potential slabs in a snow profile, extend the pit wall until it is at least 2 m across the slope (Fig. 1). Mark the width of the block (2 m) and the length of the side cuts (1.5 m) on the surface of the snow with a ski. The lower wall should be a smooth vertical surface cut with a shovel. Dig or cut the side walls and the upper wall deeper than any weak layers that may be active.

Load the rutschblock in the following sequence. The rutschblock score is the loading step produces a clean shear failure:

1. The block slides during digging or cutting.

2. The skier approaches the block from above and gently steps down onto the upper part of the block (within 35 cm of the upper wall).

3. Without lifting the heels, the skier drops from a straight leg to a bent knee position, pushing downwards and compacting surface layers.

4. The skier jumps up and lands in the same compacted spot.

5. The skier jumps again onto the same compacted spot.

6. For hard or deep slabs, remove the skis and jump on the same spot (as recommended by Fhn). For soft slabs or thin slabs where jumping without skis might penetrate through the slab, keep the skis on, step down another 35 cm - almost to mid-block - and push once then jump three times.

7. None of the loading steps produced a smooth slope-parallel failure.


The following interpretation of rutschblock scores (Fhn 1987) applies to tests done in avalanche starting zones. If the slope angle at the test site differs substantially from the slope angle in the starting zone, then the interpretation of the rutschblock score should be adjusted by 1 step for each lox difference. For example, a rutschblock score of 4 on a 25x slope should be interpreted the same as a score of 3 in the adjacent 35x starting zone, provided that snow conditions are believed to be similar.

1, 2 or 3 The block fails before the first jump. The slope is unstable. It is likely that slopes with similar snow conditions can be released by a skier.

4 or 5 The block fails on first or second jump. The stability of the slope is suspect. It is possible for a skier to release slab avalanches on slopes with similar snow conditions. Other observations or tests must be used to assess the slab stability.

6 or 7 The block does not fail on the first or second jump. There is a low (but not negligible) risk of skiers triggering avalanches on slopes with similar snow conditions. Other field observations and tests as well as safety measures remain appropriate.


To study potentially faster techniques, we cut the sides and upper wall with a cord, rail of a ski or 1.3 m long saw, and compared the time requirement with that from adjacent tests using the traditional technique. Average time requirements, excluding site selection and equipment preparation, were only reduced from 10.4 minutes for shovelling the side walls to 9.1 minutes for cutting the side walls with a cord. However, cutting both side walls and the upper wall with a saw or tail of a ski reduced the average time requirement to approximately 5 minutes. These faster techniques have their disadvantages: it is difficult to cut slabs thicker than 0.6 m slabs with the tail of a ski; cords will not cut most slabs containing melt-freeze crusts; and saws are effective under all conditions, but weigh 1.2 to 1.8 kg and are bulky to transport.

To minimize any effect of friction or bonding in the narrow side cuts made by cords or saws, we angled the side walls so that the block was 1.9 m wide at the upper wall and 2.1 m wide at the lower wall. In a concurrent paper recently submitted to the Journal of Glaciology and cited here as JJ, scores from saw- or cord-cut rutschblocks averaged 0.3 more than the scores from shovelled rutschblocks. However, the difference was nor significant at the 90% level or higher based on a two-tailed t-test or Wilcoxon test for matched pairs.


A rutschblock test is effective only for weak layers deeper than ski penetration, and several jumps on a soft slab can result in considerable ski penetration. Since this ski penetration problem can result in erroneously high rutschblock scores and a serious over-estimation of snow stability, we are sceptical of rutschblock results involving weak layers that are within 5 cm of ski penetration. For the soft slabs in our study area, almost all ski penetration problems occurred when the load over the weak layer was less than 400 Pa (4.0 g/cm2). For densities ranging from 100 to 300 kg/m3, this critical load corresponds to slab thicknesses ranging from 0.40 to 0.13 m respectively.


Sets of 36 to 73 rutschblock tests were done on each of 6 slopes that had mean slope angles of 28-33x and varied in slope angle by less than q4x. Median rutschblock scores for the six slopes ranged from 3 to 5. The median score was obtained on 67% of the tests QJ). Scores 1 and 2 steps above the median were obtained on 12% and 2 % of the tests respectively. Scores 1 and 2 steps below the median were obtained on 18% and 1% of the tests respectively. No scores 3 steps above or below the median were obtained.

By assuming the above distributions of deviations from medians are representative of uniform slopes, the probability of a single rutschblock score on a uniform slope being the median is 67%. Similarly, the probability of one score being within one step of the median is approximately 18% + 67 % + 12% = 97%. The probability of the median of two independent tests being within + step of the slope median is approximately 91 % and or being within 1 step of the slope median is approximately 99% (JJ). We suggest that independent tests be 10 m apart.

These estimates of the precision of 1 or 2 tests are appropriate only when the tests are done at sites with 4x of the mean slope angle and on slopes free of trees, rock outcrops or terrain features that might prevent relatively uniform layering of the snowpack. Also, these estimates do not apply to slopes with medians of 1, 2, 6 or 7 for which truncated distributions of rutschblock scores are expected. However, the precision of 1 or 2 rutschblock tests is certainly of practical interest when median scores are in the range of 3-5.


Since rutschblock tests and natural slab avalanches both involve shear failure within a weak snowpack layer, we attempted to correlate natural slab avalanche activity with rutschblock scores. On a total of 80 days during the winters of 1990-92, the rutschblock tests were performed in two study areas that we felt were often representative of widespread snow conditions. These areas are located at 1900 m and 2050 min the Cariboo Mountains, and consist of slopes that are in lee of most storm winds. The avalanche activity was reported by the helicopter skiing guides operating in the nearby areas of the Cariboo and Monashee Mountains. Most of the reported avalanches were within 10-15 km of the study area although some were up to 30 km away.

An avalanche day is a day in which one or more dry natural slab avalanches large enough to injure or kill a person (class 1.5 or larger according to NRCC/CAA 1989) were reported. For days in which the median rutschblock scores were 2 to 7, the percentage of avalanche days is plotted in Fig. 2. Because some storms restricted helicopter skiing, some avalanches were not observed for several days after they occurred, resulting in estimated dates. The percentage of avalanche days excluding avalanches with estimated dates is plotted separately.

In Fig. 2, the percentage of avalanche days reduces as the median rutschblock score increases. However, even when the median rutschblock score was 5, 6 or 7, there was one or more large dry natural slab avalanches on 8 to 18% of the days. Clearly, rutschblock tests on carefully selected slopes provide only an approximate indication of natural slab stability on surrounding slopes.


For those rutschblocks performed on avalanche slopes, the percentage of those slopes triggered by skiers or people on foot is plotted against median rutschblock score in Fig. 3. Except for two cases when the rutschblock tests were performed one day after the avalanche, the slopes were loaded by people on foot or skiers within 3 hours of the avalanche activity.

The percentage of slopes triggered by people decreases with increasing median rutschblock score as shown in Fig. 3. However, this is a small data set involving only 5 slopes that produced avalanches and 39 that did not. In particular, only twice have we obtained a median rutschblock score of 2 on an avalanche slope. Nevertheless, Fig. 3 like Fig. 2 shows a decrease in avalanche activity with increasing rutschblock score.

Although we have not observed slab avalanches triggered by people on slopes with median rutschblock scores of 5, 6, or 7, this does not mean that all such slopes are safe. Based on a larger data set, Fhn (1987) reports avalanche activity on slopes with rutschblock scores as high as 7 and attributes this result to difficulty with selecting representative sites for rutschblocks. Also, even for rutschblock tests at sites within q4x of the mean slope angle, there is an approximately 14% probability of getting a score one or two steps higher than the median (JJ). Once, when testing a slope that had produced a large slab avalanche, our first rutschblock score was 4 and our second score was a 5, although repeated testing resulted in a median score of 3. Clearly, some slopes that exhibit a single score of 4 or 5 are unstable. This is consistent with Fhn's (1987) interpretation of rutschblock scores.


What does a rutschblock on a 25x slope tell us about a nearby 40x slope? First, unless there is a reason why the layering might be different (e.g. the 40x slope is wind-loaded and the 25x slope is not), we expect the rutschblock to fail on the same layer as a skier might trigger on rutschblock score might be higher on the 25x slope than on the 40x slope because the shear stress caused by the weight of the slab and skier is reduced on the less steep slope.

To study the effect of slope angle on rutschblock scores, we selected 24 sets of 4 or more rutschblocks from data collected during the winters of 1991 and 1992 based on the following criteria: each rutschblock in a set slid on the same surface; each set of tests was completed in 2 to 6 hours; and slope angles within each set varied by at least 8x.

An example of such a set consisting of 42 tests is shown in Fig. 4. The slope angle varied from 23x to 36x and the rutschblock scores varied from 4 to 6. In spite of the variability, there is a general trend for rutschblock scores to increase as slope angle decreases. Based on a straight line fitted to the data in Fig. 4 by least squares, decreasing the slope angle by 12x tended to increase the rutschblock score by 1 step.

The effect of slope angle on rutschblock score was only significant for 10 of the 24 sets of rutschblocks we assessed based on the gamma correlation from nonparametric statistics (JJ). Hence, slope effects are often obscured by natural variability of rutschblock scores. However, for these 10 sets, the decrease in slope angle required to increase rutschblock scores by 1 averaged 10x. This adjustment for slope angle may be useful when steeper sites in avalanche starting zones are judged unsafe.

The minimum slope angle for rutschblocks appears to be approximately 20x since rutschblock scores are inconsistent with a Swiss stability index on slopes of less than 20x(JJ).


Fhn (1987) notes that rutschblock sites near ridge crests are seldom suitable. Our studies of rutschblocks indicate that, compared to the lower part of a slope, scores may increase and become more variable near the top of a slope even if chat upper part is steeper.

A set of 44 rutschblocks from a 27x to 35x slope is shown in Figure 5. In the lower six rows, most scores range from 4 to 6, the median score is a 5, and there is only one score of 7. In the top three rows which are almost as steep, scores range from 4 to 7, the median is a 6, and there is at least one score of 7 in each row. The weak layer of graupel was less evident in these upper rows, possibly because the wind had removed much of the graupel from the upper part of the slope.

A set of 20 rutschblocks on a 19x to 36x slope are shown in Figure 6. In the bottom six rows, scores range from 3 to 5 and the median is 4. In the top two rows which are steeper and near the top of the slope, the median is a 4 but scores range from 3 to 7.

Figures 5 and 6 show examples of higher and more variable rutschblock scores on the upper part of a slope even though the active weak layer varies from graupel to surface hoar. This suggests that single rutschblock tests, and probably other slope tests such as ski cuts, done near the top of slopes may be less indicative of slope stability than tests done farther down the slope.


1.Cutting the side walls of rutschblocks with a specialized saw or the tail of a ski can reduce the time requirement by approximately half. Cutting the sides and upper wall with a cord extended around poles at the top corners of a rutschblock can reduce the time requirement slightly. These faster techniques do not appear to affect the score significantly.

2.The rutschblock technique is only suitable for weak layers deeper than ski penetration. For soft slabs, problems with skis penetrating too close to weak layers are rare when the weak layer is buried by a slab weighing more than 4.0 g/cm2.

3.On a uniform slope that varies in slope angle by q4x or less, one test has an approximately 67% probability of being the slope median and an approximately 97% probability of being within 1 step of the slope median. The median of two tests has an approximately 91 % probability of being within + step of the slope median and an approximately 99% probability of being with 1 step of the slope median.

4.As the median rutschblock score obtained at a representative location increased from 2 to 7, the percentage of days on which large dry natural slab avalanches were reported (most within 10-15 km) was reduced from 57% to 14%. However, large dry natural avalanches were reported on 8-I8% of the days when median rutschblock scores were 5, 6 or 7. Hence, rutschblock tests provide only an approximate indication of natural slab stability for slopes several km away.

5.As the median rutschblock score obtained in avalanche starting zones increased from 2 to 5, the percentage of those slopes that were released by a person on skis or foot decreased from 50% to 0% . No slab avalanches occurred when the median rutschblock score was 5, 6 or 7. However, individual rutschblock scores ranged as high as 5 on avalanche slopes that were triggered by people. More tests are needed to clarify the relationship between rutschblock scores and slab stability for human triggers.

6.Decreasing the slope angle by l Ox tended to increase rutschblock scores by 1 although the effect of slope angle on rutschblock score was obscured by the natural variability of rutschblock scores on 14 of 24 slopes.

7.A single rutschblock test, and probably other slope tests, done near the top of a slope may be less indicative of slope stability than tests done farther down the slope.


Fhn, P.M.B. 1 987. "The rutschblock as a practical tool for slope stability evaluation." In: Avalanche Formation, Movement and Effects, IAHS Publ. 162, 223-228.

J.B. Jamieson and C.D. Johnston. "Rutschblock precision, variations on technique, and limitations", submitted to the journal of Glaciology.

NRCC/CAA. 1989. "Guidelines for Weather, Snowpack and Avalanche Observations", National Research Council of Canada and Canadian Avalanche Association, NRCC Technical Memorandum 132.


We are grateful to Mike Wiegele Helicopter Skiing and the Natural Sciences and Engineering Research Council of Canada for financial support of this collaborative research and development project funded through the Council's University/Industry program. Mike Wiegele Helicopter Skiing also provided logistical support and a productive working environment. Many thanks to Mark Shubin and Jill Hughes for their dedication and careful field measurements.

The Avalanche Review, VOL. 11, NO. 2, DECEMBER 1992
Copyright © All Rights Reserved; AAA