Avalanche Research: Scientists Behaving Badly
Marriage Peril

by Douglas PageŠ 1999

In the past three years avalanches have killed 482 people, including three researchers. Any research that leads to better avalanche prediction is therefore valuable. Obtaining avalanche data, however, is always difficult and frequently dangerous. Douglas Page profiles a team of scientists who have devised a novel means of data collection - they place themselves inside a flowing avalanche.

At approximately 2:30 on a Saturday afternoon last May, a group of Memorial Day weekend climbers descending the standard South Side climbing route of Oregon's Mt. Hood triggered a 150 m long, 10 m wide avalanche. As avalanches go, this one was puny. Nevertheless, four climbers below the trigger-point were caught in the slide, which fortunately stopped before sweeping the party into the abyss of a burgshrund crevasse, the highest and often deepest fissure on any glacier. These climbers escaped unhurt.

Unfortunately, avalanches frequently produce far more tragic results. For instance, one person was killed and five injured June 12 when 12 climbers were caught in a springtime avalanche on the upper reaches of Mt. Rainier in Washington. In France, at least 12 people died and five more were missing after a January 23 avalanche swept away a group of teenagers and their guides hiking near the ski resort of Les Orres in the southern French Alps. In March, 1997, over 100 people waiting for buses in northern Afghanistan died when a massive slide covered the road.

Avalanches, one of nature's most abrupt and unforgiving calamities, result from unstable slabs of snow, from the addition of heavy snowfall to an insecure mass of ice and snow, from the removal of part of the base of the mass (caused by melting and eroding) or from sudden shocks from explosions or earth tremors.

Violent Shiver

Some of the more furious avalanches are capable of shaving a mountainside at speeds approaching 500 kph. There is no escape for those below when mountains shed their sheets of snow with so violent a shiver. In the first nine months of this year, worldwide avalanches killed 218 skiers, snowboarders, hikers, villagers or motorists. Three of those victims were engaged in research activities.

Obviously, any research that might lead to better avalanche prediction, or eventually even prevention, is valuable.

Ed Adams and Jim Dent, associate professors of engineering mechanics at Montana State University-Bozeman, contribute to this research. They're part of a group of researchers at Montana State University that has 20 years experience studying avalanches.

"We're interested in what type of environmental and meteorological parameters influence snow metamorphism and how that may lead to subsequent failure and thus to an avalanche," says Adams. "Jim Dent is looking at what's happening dynamically as avalanches flow, and I'm trying to see if we can measure some of the structural properties of snow when it's in motion."

This Montana State crew conducts their studies a little differently than other avalanche researchers, however. Avalanches are usually studied as potential avalanches or after the event, often from a distance. Some avalanche areas are even being studied using data from satellite images. Adams and Dent prefer a more personal approach. They collect data from inside an avalanche, as it sweeps around and over them.

Gathering Information at the Source

"Many avalanche models do not work well because they contain parameters that are not easily measured," says Dent. "To use models, they must first be 'calibrated' by modeling known avalanches." Currently, model parameters are back-calculated by matching speeds and final runout positions, which are correlated with avalanche type, size and terrain. In order to build better models of avalanche motion, more information about that motion needs to be gathered. The Dent and Adams team gathers the information at the source.

Avalanche research is never easy. By nature, avalanche areas exist where the angle of repose is treacherously steep and often inaccessible. This group has looked at avalanche initiation, stability and impact pressures. To understand avalanches, one must understand snow. Tampering with snow in a slide area is risky even for experts. Recently, Adams and Dent have focused their interest on avalanche velocities and the microstructure of snow, which is how they employ their novel means of collecting some of this information. Two or three researchers occupy a 2.5 m square by 2 m high shed placed in the avalanche path while above, colleagues tickle the snow crest, inducing an avalanche, whereupon the instruments, researchers and hut get run over by the slide - the avalanche research equivalent of playing in the street.

The researchers will tell you this is done with safety in mind. The shed is positioned below rocks the size of four-door sedan. "There's a big window. You can see the avalanche go by," says Dent, who makes it sound as secure as pulling over to wait for a patch of ground fog to roll past. "No one has gotten hurt so far."

Hope that the Numbers were Right

The avalanche doesn't just "go by". It flows over and sometimes buries the wood-frame shed and those in it. The researchers will deny it, but inside the shed during these exercises there is sometimes just a fleck of fear - when the shaking stops and the avalanche swallows all sound, when the entire world has suddenly been compressed to the dimension of an outhouse. In these eternal seconds the face tightens around a hope that the numbers were right and something regrettable hasn't surged down the funnel.

"The slope we're working is fairly small, so we can control things," says Dent, unconcerned that something bigger than they expect will get loose on the 32 degree concave slope. "The avalanche path is 10 m wide and maybe 200 m long."

Still, snow weighs a minimum of 20 pounds per cubic foot. Even the modest avalanches that these researchers spar with carry hundreds of tons of snow. It can sometimes take several anxious moments for the surface team to exhume the shed occupants. According to an avalanche reporting system used in Canada, a 100 m avalanche, with a mass of 100 tons and an impact pressure of 10 kPa - smaller than these running at Adams and Dent - can bury, injure or kill a person. "Most of the mass transfer and destructive momentum of an avalanche is in the layer closest to the ground where most of the snow is," says Utah avalanche expert and mountain climbing guide Jim Frankenfield. "The dust cloud, although sometimes spectacular, has a much lower potential for destruction."

Depending on the weather, the Montana State group normally can expect to instigate five or six avalanches on the same southwestern Montana slope between January and April, most of them nudged down the mountain with the help of a two-pound cast primer bomb. The researchers call the area the Revolving Door because they use the same avalanche track again and again after each fresh snow.

There is a dose of danger in nearly everything this group does, including getting to work. A helicopter flies most of the group's equipment to the site north of Bridger Bowl Ski Area in the Bridger Mountains above Bozeman sometime during the fall. From then on, the researchers must ski in with the remaining supplies and instruments on their backs. After riding to the top of Bridger ski lift, the group takes a rope tow to the ridge, then hikes along the ridge until they're well out of the ski area, before dropping down to their Revolving Door avalanche area.

"I've been knocked down and carried a distance in avalanches, but nothing dramatic," says Adams, with professional detachment. "I've been involved in some recovery of bodies, but fortunately I've been on the other side of the rescue." Some years ago, Adams became so interested in avalanches while cross-country skiing in Utah that he abandoned English, his original field of study, to pursue engineering mechanics - "the math and physics side of engineering". He now teaches about the thermodynamics of snow in the civil engineering department at Montana State. His interest in the microstructure of snow takes him occasionally to Antarctica.

One of the goals of the Dent and Adams research is to attempt to build avalanche velocity profiles. "It's a little questionable how they flow, how they act given their granular material," says Adams. "Avalanches are often modeled just like a fluid. We're trying to get a better perspective on snow as it's moving by." Adams also wants to start piecing together snow microstructure during a flowing avalanche - how it breaks apart, whether they can see blocking in the material, how cohesive it is and how fractured it is prior to stopping.

Working Inside a Billowing Avalanche

"If you look at stopped snow, it tends to settle very quickly and get quite hard. It's difficult to get a lot of information out of it," Adams says, explaining why it's necessary to work inside a billowing avalanche.

The shed at Revolving Door has evolved to serve this purpose, where instrumentation has been developed and installed to measure velocity, density, depth, shear and normal stress of moving avalanches.

According to Dent, the researchers first remove the snow from the slope beside the 'sedan' rock and outside the wall of the instrument shed, exposing the top 60 cm of the shed-wall parallel to the direction of the flow. When the avalanche is triggered, the snow flows down the chute, over the rock, over the shed, rubbing along the exposed shed wall. Instrumentation mounted on this wall is used to measure flow properties as the avalanche goes by. The large window in the shed allows the avalanche to be observed and photographed by the scientists inside.

Mounted in the wall of the instrument shed are an array of rugged, industrial US$2.00 photoelectric sensors, 7.3 mm in diameter, comprised of unfocused infrared light emitting diodes (LED) and an infrared sensitive phototransistor. "Light from the LED is back-scattered by the avalanche to the phototransistor," says Dent. "The amount of infrared light reflected from the snow surface is a function of the structure and density of the snow. Since that structure and density changes from point to point in the avalanche, the amount of light reflected from the avalanche varies with time as the avalanche passes the optical sensor." When the data is reduced, density property of the snow can be deduced.

A second sensor watches the same snow from a short distance downstream. The output of the two phototransistors produce similar time-varying signals. Finding the time lag between both signals enables the velocity of the snow to be calculated.

Density and velocity measurements are made at several heights in the avalanche, with particular attention directed to the running surface.

Shear and normal stresses are measured using a 23 by 28 cm roughened aluminum shear plate mounted in a sturdy box on two cantilevered arms to the side of the shed below the running surface so the plate is flush and parallel with the surface of the ground. As the avalanche slides over the plate, the cantilevered arms deflect. The amount of deflection from the resting position is converted using a calibration curve produced by known weights, thus yielding avalanche stress and pressure data.

The group has made several discoveries. "The main one is that an avalanche pretty much slides as a whole instead of flowing like water," says Dent. "It's more like a block sliding than a flowing liquid." Other results indicate that avalanche deformation (folding and faulting) is concentrated near the running surface where the snow density is found to be greatest. Upward from the surface, the gradient fall off greatly while the density also declines.

What remains is to collect data on avalanches in various types of conditions to further refine the construction of models of avalanche motion. Temperature measurements in the flow are planned as well as the use of load cells to measure pressure distribution against various obstacles, the team says. Additional instruments are also being developed for the shed to measure air pressure and velocity of the powder or air blast that accompanies the avalanche.

None of this information may be much help to the unfortunate or reckless snowboarder or skier who find themselves swallowed by a gush of "white death", but it contributes to the body of knowledge that helps researchers predict various aspects of an avalanche. By obtaining measurements of the snow in motion, as well as prior to and subsequent to an avalanche event, the scientists are learning to build better avalanche models - which may in time lead to better predictive parameters, which in turn could save lives. Including their own.


Suggested Reading

Dent, J.D., et. al. "Density, Velocity and Friction Measurement in Dry Snow Avalanche," Annals of Glaciology, in press.

Dent, J.D., et. al. "Density and Friction Measurements in a Flowing Dry Snow Avalanche," in Proceedings International Snow Science Workshop, Banff, Alberta, Canada (October 1996).


This article appeared in Science Spectra (No. 16, 1999).

Comments? Questions? Assignments? douglaspage@earthlink.net
Back to Top. Return to Home Page.