From vol. 106, spring 2014 

By Paul Cordy

DEPOSIT FROM PATH 19.3 IN KOOTENAY PASS. A SIZE 3 AVALANCHE TRIGGERED BY HELI BOMBING // MOTI

IN THE WORLD OF BIG DATA we have become accustomed to interacting with computer models. The search for good snow inevitably begins by consulting the ensemble weather forecast (the consensus weather prediction of five different detailed models of the atmosphere), just as most searches for knowledge these days begin by consulting Google (a complex and dynamic model of the relevance of digital information). So one might expect that any day now Big Data will begin to spread its tentacles into the world of avalanche safety. But are there particular challenges to using computer models for avalanche prediction? Not the least of these may be thecomplexity of geographic and human factors leading to avalanche formation, and also the scarcity of reliable and continuous information about conditions in the start zones.

So how far have avalanche prediction models come, and how might they benefit organizations and individuals? Will they ever be good enough to rely on in Canada? The British Columbia Ministry of Transportation and Infrastructure (MOTI) has a long history of taking the lead in creating digital tools for avalanche practitioners. These efforts have led to the development of one tool that we know and use already: SnowPro. A lesser-known innovation of the MOTI is the computer-based avalanche forecasting system which began more than 15 years ago in Kootenay Pass. Ted Wieck, former information systems manager for the MOTI avalanche and weather branch, spent over a decade developing the MOTI’s first digital highway, weather, and avalanche database. In the beginning, this meant considerable amounts of tedious data entry for technicians, who would have rather spent more time in the hills and on the road. Ted wanted to make all that data useful to the people who were assiduously collecting it for him, and so he became a fervent supporter of computer-based avalanche prediction.

In the mid-nineties, Dr. David McClung and John Tweedy developed and tested software that used manual weather observations (input by the user, of course) to predict the probability of avalanche activity that day. The prediction was based on a statistical model that was created using historical weather data and avalanche occurrence records from the previous ten seasons at Kootenay Pass. As in all computer models (including Google’s search engine), historic data is used to train the model, or in the case of MOTI, determine the relative importance of various weather variables and how to combine them in a way that computes accurate predictions of avalanches.

This is not too dissimilar to the way that we humans learn. Our experience is combined with training to create mental models of how weather creates avalanches. Often we will compare current weather or snowpack structure with previous seasons’ observations to refine our decisions. The original Kootenay Pass model also retrieved the ten most similar instances of weather and presented the data to the human forecaster to further aid in decision making. In the end, both model approaches were 70-80% accurate. Early in the 2000s, James Floyer proved that similar models could be trained on Bear Pass datasets with similar results.

A CROWN ON PATH 19.8, KOOTENAY PASS // MOTI

As a Masters student with McClung at the University of British Columbia, my contribution to this effort was to dynamically integrate numerical weather forecasts and optimize different versions of the model for each of five different highway corridors with active avalanche control programs. In each place, we used ensemble weather forecasts up to 48 hours ahead into each model, thus extending avalanche predictions into the future (all previous avalanche models predicted present probability of avalanches only). As it happens, predicting avalanches in the future mostly depends on the accuracy of weather forecasting, and most avalanche forecasting models achieve similar accuracy irrespective of the type or complexity of the model.

Of course a 70 to 80% prediction rate is horribly inaccurate given the consequences range from traffic hazard to loss of life, and so there always had to be a human forecaster calling the shots. But before dismissing computer models, one must consider the constraints under which they are working.

 
Take weather forecasting as an analogy. European weather prediction is far better than that of western North America because of differences in density of meteorological stations. Weather systems en route to Europe are being broadcast by countless sensors in myriad islands and land masses in the Atlantic, not to mention by the North American sensor network. Reliable data makes for more reliable weather models. By contrast, weather on its way to western North America passes over the Pacific Data Void, a vast stretch of ocean almost uninterrupted by islands and permanent weather stations. So the very same computer models are often inaccurate more than 24 hours in advance.

So too with computer models of avalanche prediction. Greater complexity and precision of avalanche models is unlikely to improve forecast accuracy until we provide such models with more and better information. The data that we provide prediction models couldn’t possibly compete with the human experience. Avalanche technicians explore the terrain, doing hand shears and listening to the snow settling under their skis. They feel temperature changes when fronts come through, just like the sensor networks do, but sensors can’t see the sun hit certain start zones, and they can’t see how snow is loading up there. Really, it’s a miracle that numerical prediction algorithms are accurate at all.

Therefore, the next goal was to integrate information about the snowpack into the model. The MOTI avalanche models had a built-in mechanism for updating the avalanche probabilities based on new information. Previously, this "prior" information was added by the forecaster in response to avalanche control results or other knowledge that was not available to the model. Prior probabilities could just as easily come from a model of snowpack structure and stability such as the red flag method of SnowPro, or the SNOWPACK physical model used in Switzerland. Unfortunately, changing funding priorities and personnel at the MOTI meant that snowpack information was never integrated into predictions, although it is still used in Kootenay Pass. It’s up to the next generation take it to another level.

NORTH FORK AVALANCHE AREA ON THE EAST SIDE OF KOOTENAY PASS // MOTI

Generational change itself was also a major driver of interest in creating the model. During the latest bout of modeling studies, MOTI was facing the near-simultaneous retirement of all of their technicians. MOTI saw that new staff might get up to speed more quickly if they could scan the results for the size, type and spatial distribution of natural or controlled avalanches in the historical records. The idea was to try to decouple the memories of seasons from the people who observe them, and help bridge the loss of team experience when seasoned professionals retire. Furthermore, the benefits of such systems would be more apparent to successive generations of technicians who would be ever more native to the digital environment. Whereas the old ironsides of the avalanche patch are more likely to decry that their Rite in the Rain books have never crashed nor printed error messages, younger generations are more likely to wish they could just use their iPhone and store it in the cloud.

Computers can supplement our memories, help us see broad patterns, and evaluate the importance of various causal factors that govern avalanche formation. Snowpack depths and precipitation intensity can be measured by satellite, and soon we’ll have satellites sensing atmospheric structure and conditions over the Pacific Data Void. With more and more wired backcountry users and the Canadian Avalanche Centre’s geo-referenced recreationist observation database, avalanche information is set to explode. Models can help us to synthesize an oversupply of data into relevant knowledge. That knowledge will always be limited by the data and model that generate it, and may always require a human to make life and death decisions. However with changing personnel and changing climate, it helps to maintain historical perspective on present events. Avalanche prediction models can help to bridge present and past, and to help us tease out the most relevant information that can be used to manage risk.

As the analytical techniques of Big Data inexorably penetrate all aspects of life, I expect that one day they will be as much a part of the furniture of our lives as smartphones. However, research and development in avalanche risk modeling advances through the vision, passion and forward thinking of people like John Tweedy and Ted Weick, who championed the initiative within the MOTI. Although my main research focus has shifted from avalanche models to pollution modeling and mitigation, I maintain a deep interest in the topic. As we approach the critical information density with respect to snow and weather, I look forward to collaborating with the next generation of visionaries and institutional champions that will bring the avalanche world back in step with Big Data.

From vol. 102, winter 2012-13

By Penny Goddard

TREBLE CONE SKI AREA, NEW ZEALAND. THIS BASIN WAS CLOSED DURING THE DAY DUE TO CREEP AND GLIDE CONCERNS. AT 17:00 THE SURFACE WAS STARTING TO REFREEZE, SO THE FORECASTER GAVE THE OK FOR GROOMER OPERATORS TO GO INTO THE BASIN TO WORK. THE AVALANCHE OCCURRED SOME TIME DURING THE NIGHT, FAILING ON DEPTH HOAR AT GROUND. IT DAMAGED THE LIFT BULL WHEEL. // DEAN STAPLES

Avalanches that occur during periods of cooling are important because they can surprise people. The subject first piqued my interest several years ago while watching the sun leave a steep slope on the opposite side of the valley from a ski lodge in New Zealand. When a large slab released from the slope minutes later, it seemed incongruous.

NO OBVIOUS TRIGGER was present: no recent loading by wind, snow or rain, no person and no bomb. The only change I perceived was a rapid drop in temperature as the slope moved from full sunshine to shade and into its associated early evening chill. Days later, the same thing happened on the same slope.

In 2005, I became an avalanche forecaster at Broken River Ski Club in New Zealand. Lingering in the shadows of my mind was an avalanche which had occurred there 13 years prior. The week preceding the avalanche had been stormy, with 142mm of precipitation. Fluctuating freezing levels eventually led to a rain-soaked snowpack. On the day of the avalanche, the weather cleared, temperatures dropped and the snow surface became slick and icy. Staff decided to open the area based on conventional wisdom: cooling and surface re-freezing promote stability.

At lunchtime, a size 4 avalanche failed near the ground on depth hoar, pulling out the entire Broken River basin with a crown up to 2.2m deep. It propagated 800m wide into low-angled terrain, leaving a deposit 20-30m deep. A snow groomer and skiers were in parts of the basin and may have been the trigger, but they were far from the fracture line. Amazingly, because almost all other skiers were inside having lunch only the ski area manager was killed.

A photo of the avalanche hung on the wall in the forecasting office, leaving me chilled and uncertain. Doesn’t an icy, frozen surface mean the snowpack’s locked up? Why did the avalanche fail then and not during the warm storm? Why did it propagate so widely?

So began my investigation. I turned to the books to read up on the phenomenon and learn about the mechanisms behind such events. Beyond some passing references to rapid temperature changes, the standard volley of avalanche reference books left me empty-handed. I tried scientific journals, asked academics and searched online. Very little came to light. So I began to ask my colleagues. A few people had experienced something like
it. Many hadn’t.

A more formal questionnaire followed. In the end, 40 avalanche professionals from around the world responded. The questionnaire focused specifically on ‘re-freeze’ type events (where the snow surface goes from 0°C to below 0°C). I called this a “Cool-Down Avalanche” (CDA). The responses alerted me to the prevalence of surprising, large avalanches during periods of rapid cooling, not just when the snow surface goes from melt to freeze, but also at overall lower temperatures (e.g. a drop from -5°C to -15°C).

This article firstly summarizes the results of the questionnaire, and then highlights a round of cooling-related avalanches in Western Canada during the 2010-11 winter season.

MISTAYA LODGE, WESTERN ROCKIES. OVERNIGHT, JAN. 17/18, 2011 AFTER A STORM WHICH DEPOSITED 1M+ SNOW HAD ENDED. THERE WAS OVERNIGHT AIR TEMPERATURE COOLING FROM -3°C TO -13°C AND WIND (HOWEVER, MANY OF THESE SLOPES WERE NOT LEE TO THE WIND). MORE THAN 20 AVALANCHES RELEASED, SIZE 1 TO 3.5 (MANY 2-2.5) WITH CROWNS 100-150CM, SOME UP TO 200CM DEEP. SEVERAL AVALANCHES WERE OBSERVED IN UNUSUAL LOCATIONS. // DAVE BIRNE

PART 1: THE RESULTS OF THE CDA QUESTIONNAIRE

• In order of descending quantity, observations came from New Zealand, North America, Europe, Asia and Antarctica.
  
• 15 of the 40 respondents had never experienced a CDA. Many more people elected not to answer the questionnaire at all, due to having never experienced a CDA.
  
• About 360 CDA were observed. This number is approximate, as the bulk of observations were poorly recorded, based instead on observers’ memories.
  
• 98% of observed CDA were described as slab avalanches, 2% as loose.
  
• The bulk of the observed avalanches were size 2-3; 14 were size 4 and three were size 5.
  
• 61% were described as ‘glide’ releases.
  
• 20 CDA events occurred within 15-60 minutes of the sun leaving the slope. A further seven occurred less than 15 minutes after the sun left the slope.
  
• 21% of respondents had experienced a close call involving a CDA. These included large avalanches hitting an open highway, burying a ski lift in an area open to staff, and fully burying people in guided groups.
  
• 38% of respondents factor CDA into their decision-making while managing the exposure of people and infrastructure to avalanches. 44% said they do not.
  
• Seven people who had never had a close call involving a CDA factor the possibility of CDA into their decision-making. Interestingly, three people do not factor CDAs into their decision-making, in spite of having had a close call involving a CDA (including involvement in fatal incidents).

The following comments made by respondents help address some of the reasons why CDAs are rarely factored into operational forecasting:

• “[This is] much too speculative a theory to apply in an operational forecast.”
• “I see 'cool-down' as the more stable end of the curve.”
• “I don't factor CDAs into management due to a lack of understanding and observations.”
• “I don't factor CDAs in, as it seems a very rare event.”
• “I don't factor CDAs in, as there’s no knowledge base, therefore they are hard to estimate.”
• “The funny thing is, I probably still guide and operate considering cooling down as a good tick for stability.”

CDA CONCLUSIONS

• CDAs (surface refreezing avalanches) were observed around the world.
• Accidents and near-misses have occurred when operators have re-opened previously closed terrain assuming that cooling means dramatically improved stability.
• Some operators actively manage the CDA hazard through closures or explosives control, which are timed to coincide with rapid cooling or surface refreezing.
• They were rarely observed overall–many experienced practitioners have never experienced a CDA.
• There is a feeling that they are too difficult to predict, so there is a tendency to ignore them when making decisions.

LANARK PATH, ROGERS PASS, 8:00, JAN. 18, 2011. THE AVALANCHE WAS SIZE 4.5 AND DAMAGED TEN ACRES OF FOREST. IT FAILED ON FACETS/CRUST AT GROUND. THE AIR TEMPERATURE DROPPED FROM -3°C TO -17°C OVERNIGHT PRIOR TO THE EVENT. IT FAILED AROUND THE TIME THAT SUN FIRST HIT THE SLOPE. // MINISTRY OF TRANSPORTATION AND INFRASTRUCTURE - AVALANCHE AND WEATHER PROGRAMS

PART 2: COOLING EVENTS IN WESTERN CANADA DURING WINTER 2010-11
It is important to distinguish a key difference between Part 1 and Part 2. The questionnaire in Part 1 asked specifically about ‘re-freeze’ CDA events (snow surface going from 0°C to below 0°C). The events listed in Part 2 occurred during periods of rapid cooling within an overall colder temperature regime and did not involve a clear melt-freeze process at the surface.

The included photos show a succession of large avalanches that occurred during periods of rapid cooling in western Canada. Operators described these events as very surprising, eye-opening, historic and unusual.

SOME COMMON FACTORS IN THE EVENTS OF 2010-11:

• Heavy storm loading occurred prior to the event.
• All but one failed on a persistent weak layer.
• Rapid air temperature cooling occurred, often around 7-10°C
  overnight.
• They were mostly very large events with wide propagation.
• In every case, experienced locals were surprised by the
  events.
  

CONCLUSION

What does all this mean? Is there anything more than a sense of vague paranoia to take away? It seems clear that avalanches sometimes occur during periods of rapid cooling, both when the snow surface is going from melt to freeze and at overall colder temperatures. It is unclear whether, or how, cooling itself triggers avalanches—and that is a topic for a whole different study.

What does seem apparent to me is that many near-misses and possibly some serious accidents were caused by faulty decision-making around cooling. The premise that cooling stabilizes the snowpack after a storm (or solar warming) ends is not always correct. Always basing decision making on this premise can lead to premature exposure to avalanche terrain.

The most constant element in these events was that of surprise. In many cases, professionals were just about to (or just had) opened up terrain previously off-limits for public, guests and staff access.

This research is mostly a collection of anecdotes. In order to really understand the mechanisms behind avalanches that occur during periods of cooling (and from there, to be able to forecast them), as a community we need to better document this type of event. I hope that this preliminary investigation will spark some discussion, spawn some more focussed research and perhaps encourage decision-makers to take a ‘second look’ at conditions during periods of rapid cooling.

I presented this topic at conferences in Penticton, and New Zealand last year. On both occasions, numerous audience members revealed that they, too, had experienced surprising avalanches during times of rapid cooling. My feeling is that this phenomenon, while sporadic, is more common than one might expect.

HILDA PEAK, VALKYR RANGE, JAN. 20, 2011. THIS SIZE 4 AVALANCHE WAS AN ISOLATED EVENT WHICH OCCURRED POST-COOLING. THERE WERE NO OTHER LARGE NATURAL AVALANCHES DURING THE CYCLE. IT DESTROYED MATURE FOREST. // RYAN GLASHEEN

ACKNOWLEDGEMENTS

Thanks to the following for their contribution: Bill Atkinson, Mark Austin-Cheval, Reid Bahnson, BC Ministry of Transportation – Avalanche and Weather Programs, Canadian Avalanche Centre, Peter Bilous, Dave Birnie, Stewart Blennerhassett, Kevin Boekholt, Jay Bristow, Wayne Carran, Howard Conway, Rosco Davies, Jef Desbecker, Thomas Exner, Ryan Glasheen, David Hamre, John Hooker, Andy Hoyle, Damian Jackson, Dan Kennedy, Karl Klassen, Mark Klassen, Brett Kobernik, Gary Kuehn, Chris Landry, John Mletschnig, Jane Morris, Shaun Norman, Tom O’Donnell, Christine Pielmeier, Tarn Pilkington, Nicholas Rapaich, Tim Robertson, Davie Robinson, Mike Rubenstein, Mark Sanderson, Mark Sedon, Ron Simenhois, Jim Spencer, Dean Staples, John Stimberis, Frank Techel, Craig Wilbour and Henry Worsp.

From Vol. 101, fall 2012

By Rod Gee

Glide slab explosives control results, 50 mile path. Skeena River corridor, west of Terrace, BC. By Rod Gee

My twenty-five year glide slab education began on an early morning in January 1989. A snowplow operator on Highway 16 west of Terrace reported witnessing a Size 3.5 airborne wet avalanche cross the railroad and highway corridors.

The deposit pushed sections of concrete guardrail into the Skeena River. Fortunately, no one was involved. I arrived at the site shortly after hearing the plow operator’s report. “Argh! It’s the glide slab I’ve been monitoring for the last week," I thought; "Why this morning? It’s not raining, and it’s not warm. Why did it run now? Were there indicators I’d missed?”

I came to the north coast of British Columbia to work in CN Rail’s Skeena avalanche program. I brought seven years of work experience in the Rockies, and ITP training in the Selkirk Mountains and the Coast Ranges. However, I had minimal knowledge of glide slab behaviour.

Glide slab prediction is a challenge, compared to the relative predictability of most maritime snowpack avalanche activity. They are classic poster children for the discussion surrounding why “Hazard Level 2” is perhaps a better descriptor than “Stability Good, with the occasional size 4.” Without start zone instrumentation monitoring glide rates, the CN Skeena program offsets uncertainty to some degree with frequent explosives control, and, where effective, runout zone earthworks.

These are some of the observations on formation and natural initiation I now use to evaluate glide slab stability:

• A low-friction ground surface is important for slab formation, but the degree of support from terrain features immediately below the slab is at least as important for slab failure.
• Rapid, early season snowpack accumulation associated with relatively warm air temperatures increases the likelihood of early- and mid-season glide slab formation.
• Lack of an effective ground freeze prior to snowpack accumulation results in increased mid-winter glide rates.
• Rainfall and meltwater percolation in an isothermal starting zone snowpack may accelerate glide rate by decreasing friction at the slab/ground interface. Free water may also decrease the strength of the supporting snow downslope of the glide slab as well as the slab itself. Rain falling into the glide crack above the slab, likely has a similar net effect. However, rain does not
  guarantee slab failure it is only part of the equation.
• Glide slab failure does not require an isothermal snowpack. Failure may occur before the snowpack becomes isothermal or during the overnight cooling phase of the diurnal cycle,
  and without free water being present at the snow/ground interface.

Explosives Initiation
The ideal condition for explosives control occurs when the slab itself maintains a degree of strength greater than that of the snowpack below the toe and along the flanks of the slab. In an ideal scenario, a combination of terrain and weather factors unbalances the downslope snowpack stress/strength relationship to a greater degree than within the slab itself. The toe and flanks
are now barely able to support the loading of the gliding slab. Explosives applied at this time cause slab initiation by triggering a failure of the snowpack at the toe of the slab.

Technicians Herb Bleuer and Mike Zylicz began experimenting with charge quantity and placement in the Skeena corridor in the early 1980s. They realized that conventional charge quantity was usually insufficient for glide slab initiation, and that charge placement was extremely critical.

They also realized that placing the explosives charge into the glide crack above the slab was ineffective because that was not where the stress/strength relationship was deteriorating. Effective
glide slab control is about “kicking the knees out” from under the slab, and not adding load to the slab itself. Their testing produced reasonable results using 100-150kg ANFO charges placed at the toe of the slab.

The best charge placement is a very specific point where the gliding slab is having the greatest effect on the non-gliding downslope snowpack. Current Skeena corridor glide slab control strategy includes the use of charges of 150 and 500kg on 200-500cm deep slabs. Large charges are used because they increase the likelihood of triggering, which reduces hazard at the runout zone transportation corridor and minimizes the likelihood of natural events disrupting rail operations.

That said, control is not always successful. A complex, ever-changing interplay of factors affects glide slab stability, and the puzzle is not completely understood.

Some factors I consider in evaluating explosives control effectiveness include:

• Control is more likely to be successful on glide slabs poorly supported by the terrain below the slab. For example, a poorly-supported glide slab can be initiated with explosives so it will then trigger a better supported glide slab lower in the starting zone that does not respond to explosives.
• Rain or melt-water at the ground/snow interface is not essential for initiation to occur, but it does increase the likelihood.
• Initiating sections of glide slabs is useful both by reducing the deposit volume of a single occurrence, but also because it exposes the ground surface to solar radiation, which then potentially aids in increasing glide rate by introducing more heat into the slab’s basal layers.
• The strength of the snowpack below and alongside the slab allows the slab to glide a significant distance downslope without failing. Increasing glide rate may indicate decreasing snowpack strength.
• A 300-600cm slab can easily glide 50-100m without initiating if the downslope snowpack and terrain accommodates the glide’s loading effect. Increasing glide rate and/or deteriorating strength of the snowpack supporting the slab are two critical initiation factors.
• Three reliable nearby indicator paths I use to assess an east aspect path prone to glide slab formation have a northwest aspect, but the starting zones are at the same elevation. This
  suggests ambient air temperature affects glide slab behaviour to a lesser, but still relevant, degree.

Prediction and control have improved since the 1980s, but we still include a healthy dose of “art” to the “science” of our craft. Explosives control in January 2012 put a size 4 deposit within 2m of the rail roadbed. Is our understanding of glide slab management improving, or were we just lucky on that mission?

Rod Gee is a CAA and AAA Professional member with 30 years industry experience. His association with Chris Stethem and Associates Ltd. from 1985 to 2011 provided participation in a broad variety of Canadian avalanche programs. Rod is owner of Northwest Avalanche Solutions Ltd., and is based in Terrace, B.C.

From vol. 111, winter 2015

By Colin Zacharias

Conducting an extended column test. Photo by Colin Zacharias

THE WEIGHT OF EVIDENCE
Every winter day we make snowpack observations and extrapolate from observation sites to nearby terrain. Most days, for most avalanche problems, this extrapolation process works and we make key decisions from comparatively few quality bits of information. But it is easy to lose confidence in our abilities when conditions become unfamiliar or our information becomes scarce.

Outside of current avalanching and other alarm signs, and especially during periods of high snowpack variability, experienced observers tend to steer away from drawing quick conclusions from a few snowpack observations. They recognize that one test is just one observation, and to counter possible extrapolation errors they ensure that over the critical timeframe key information is supported and verified.

On the other hand, inexperienced observers may apply too much importance to a persuasive snowpack test result or a single avalanche occurrence and be subject to a confirmation bias. Experienced forecasters, even with a decent amount of information, recognize that at times their best is still in the end just that.

Karl Klassen, Avalanche Canada Public Avalanche Warning Service Manager and mountains guide, recently reminded me with a nice touch of irony that while our data -> information -> knowledge -> wisdom hierarchy (Zeleny 1987) fits into a neat little package, it can also backfire. Depending on the quality and quantity of the data set, its relevancy, and our ability to interpret the info, data isn’t information and information isn’t knowledge, and if one thing is certain, wisdom is a different kettle of fish.

There are times when logistics make it difficult to add weight to the evidence. Poor weather or difficult travel conditions, for example, may prevent access to terrain or study sites. Yet even then assumptions are made and conclusions derived. As Dr. Bruce Jamieson notes in his mountain snowpack presentation for the ITP Level 2 Module 1, “inaccurate assumptions can have serious consequences” when it comes to spatial variability in the mountain snowpack.

Decisions made from a deficit or even partial deficiency of information required to understand the avalanche problem are considered uncertain in light of an applied risk management strategy (as defined by ISO 31000). In the avalanche world we are okay with uncertainty—so long as we know what we don’t know. We understand that as the measure of uncertainty increases so does that long arm of caution when planning to reduce the risk.

In today’s avalanche world in Southern BC and Alberta, professionals rely on a daily information exchange to help manage the complexity of snowpack/terrain variability, to provide a “heads up” early warning system or a nearest neighbor confirmation— “yes, they’re seeing what we’re seeing.” Each day we scan through thousands of bits of data and information on the InfoEx, then go into the field and gather more, aggregate the data into information packets, and analyze and communicate patterns that we refer to as hazard factors.

This article, along with the “How Useful Is the Evidence?” table below, was developed in the fall of 2011 as part of an Avalanche Operations Level 2 Module 3 training course handout to help learners apply the notion of strength and weight to field observations, to use a checklist style verification process, and to encourage quality craftsmanship and a thorough approach when analyzing and discussing snowpack factors. It may help the learner to recognize whether or not their evidence drawn from snowpack tests is helpful to their decisions.

Fig. 1: From CAA Level 2 Module 3 handout

CRAFTSMANSHIP AND CONSISTENCY
“Jeez…. the weather and snowpack vary enough; can’t we all just do the same damn observation the same damn way?”

Regional and operational consistency with technique, application and interpretation ensures the quality of data gathered, recorded and communicated. On professional level avalanche training courses, instructors inform that practice, technique, and a meticulous day-to-day consistency with observations, recording and communication should never be undervalued, nor should the scope of the task be underestimated:

• Ensure that there is an objective for each snowpack test. The early morning safety meeting agenda usually includes assessing the day’s avalanche problem and identifying gaps in knowledge. Know what you’re looking for prior to looking.
  
• Select relevant sites for field test sites using experience and the seasonal observation of how the snow is layered over the terrain. Once sites have proven their worth, they are repeatedly used season to season.
  
• Conduct tests skillfully using standardized, practiced techniques. Observers use established guidelines when conducting, recording, and communicating weather, snowpack and avalanche observations; these come from Observation Guidelines and Recording Standards for Weather, Snowpack and Avalanches (OGRS) and Snow, Weather, and Avalanches: Observation Guidelines for Avalanche Programs in the United States (SWAG).
  
• Ensure consistency within an operation by having employees conduct observations side by side. Discuss technique and compare interpretation during preseason staff training.
  

THE RIGHT TOOL FOR THE RIGHT JOB
The CAA’s OGRS and the AAA’s SWAG provide guidelines for how to conduct and record weather, snowpack, and avalanche observations. Other than a few comments about the observed limitation of certain tests, these guidelines deliberately offer little information on how to apply or interpret the observations as they relate to an avalanche problem or forecast. This knowledge and proficiency is gained through other means, including research articles, professional avalanche training, and on the job training and mentorship.

Of course there isn’t any single test that will reveal exactly what you need to know about snow. Yet every decade or so it seems that guides and forecasters have a new favourite “go to” decision making aid they default to when investigating the current avalanche problem. First it was the Rutschblock test (RB), then the compression test (CT)—or the other way around depending on your region—and now it’s the extended column test (ECT). In a helpful 2010 article “Which Obs for Which Avalanche Type?” Bruce Jamieson and others conducted a field study that did an excellent job of directing attention to those observations that best identify each avalanche concern. The combination of determining the avalanche problem prior to departure (Atkins 2004) and having a good idea about which field observations and tests will best identify the problem is a good start when choosing the right tool for the right problem.

The AIARE Avalanches and Observations Reference included below (published in the AIARE Field Book and instructor materials) was inspired by the aforementioned article and is a useful field reference to help learners target those concerns described in the daily avalanche advisory.

Fig. 2: From AIARE instructor materials and field book, 2012

MANAGING FALSE STABLE AND FALSE UNSTABLE RESULTS
Doug Chabot, forecaster at the Gallatin National Forest Avalanche Center, brings up a good point in a recent blog post: “Snowpit tests are used to show instability, not stability. Never stability. Snow pits (and snowpack tests) do not give the green light to ski; they just give us the red light to not ski. An unstable test result is always critical information. A stable test result does not mean the snow is stable a hundred feet away.”

Chabot’s advice points to the quandary many backcountry recreationists face when analyzing snowpack factors: a test result illustrating unstable snow urges cautious risk reduction, but what does a “no result” mean? Yet estimating where the snow is strong and where the snow is weak is an important skill—particularly for guides committing clients to terrain. Determining stability or the “likelihood that avalanches will not occur” involves a detailed process of gathering evidence, drawing a big picture perspective and not leaping to conclusions from a single observation or test result.

• Knowing the sites that information is coming from, having a systematic or “toolbox approach” to clue gathering (see Fig. 3), and observing the terrain and trends over time are all crucial links in the chain of gathering information and applying it to a hazard analysis. And knowing to what degree those links are missing and then defining the information deficit (whether the uncertainty is weak layer location and distribution, character and sensitivity, or slab characteristic and estimation of destructive potential) is all part of guide and forecaster daily discussion. In addition to the strength, weight and verification checklist provided in Fig. 1, the following points may help when interpreting the day’s investigations.
• A seasonal perspective of where the terrain has historically formed stronger and weaker snow is important. Basal facet development tends to repeat itself in seasonal trends. While near-surface persistent weak layers tend to have a broader distribution, sun or wind effect can result in feature scale variability in weak layer character. For example, DF (decomposed and fragmented snow grain) layers can be unstable locally but may not be problematic on a drainage scale. Expect a higher incident of false stable test results when observing locally unstable layers like DFs, graupel or sun crust/DF interfaces.
  
• One of the best tools for determining the nature of snowpack variability is to simply observe and memorize how the current snow surface or near surface condition changes over the terrain. Knowing the extent of surface hoar, facet, crust, or graupel formation and the distribution of storm snow and wind redistribution of snow helps to form a baseline when later estimating snowpack strength. Imagine yourself a heli ski guide with the opportunity to travel over 10s or 100s of kilometres of terrain on any given day. Using your eyes, your skis, and a few quick penetration and hand tests provides insight into what to expect when the snow surface becomes a buried weak layer. “Quick tests,” while not subject to the same formal research as standard tests, still provide helpful information to an approximate depth of 45cm (Schweizer and Jamieson 2010).
  
• A checklist sum of snowpack structural properties (a.k.a. “yellow flags” or the Snowprofile Checklist (Jamieson and Schweizer 2005)) provide valuable clues about which layer interface is most likely to result in a localized failure/fracture. However, as the checklist sum has a tendency to overestimate instability (false unstable=false alarm), further tests are conducted to determine propensity for propagation (Winkler and Schweitzer 2008). The combination of CTs (with fracture character) and the profile checklist sums provide an excellent tool to determine which layer is worth testing prior to a propagation saw test (PST) or ECT propagation propensity test.
  
• The large column snowpack tests that employ taps or jumps to apply a load to the slab (e.g., the ECT and RB) may still indicate a “no result” when a significant weak layer is buried approximately 1m or deeper—and/or when stiffer snowpack layer characteristics (e.g., a crust) reduce the likelihood that surface taps are affecting the deeper weak layer. The cautionary note is that skier triggering of a layer of this depth may still occur from shallower or weaker area (see case history below). In this scenario, one would not use the ECT or RB as the sole observation tool. It may be more prudent to identify the deeper weak layer with a CT or deep tap test (DT) and if a sudden fracture is observed choose to conduct a PST (or choose a shallower location for an ECT) to observe propensity for crack propagation in the layer. The combination of the small column test (which may err on false unstable but identifies fracture character) combined with a large column test (testing for propagation propensity) both reduces the likelihood of a missed observation and provides more information with a verified result. This “toolbox approach” may help interpret a potential “no result” or a false stable result.

The 2010 Schweizer and Jamieson article “Snowpack Tests for Assessing Snow-Slope Instability” provides an updated, excellent perspective directed at a general audience on snowpack test use and limitations. The following summary points have been paraphrased from the article:

• A good test method should predict stable and unstable scenarios equally well.
  
• Column tests are particularly helpful for assessing persistent slab conditions.
  
• Small column tests (CT and DT test) are useful for identifying weak layers and likelihood of initiation but have a tendency to overestimate instability (false unstable) conditions. Observing fracture character improves, to a degree, the interpretation of the test results. These tests are a better indicator of layer character than instability.
  
• Large column tests are better at predicting propensity for fracture propagation than small column tests, particularly when used in combination with other large column tests. Comparative studies suggest that the RB, ECT, and PST have comparable accuracy.
  
• With large column tests, repeated test results in the same location are useful but the tests repeated on similar, nearby slopes add value.
  
• Each test has a margin of error. Even with very experienced observers an error rate of 5-10% is to be expected. Site selection and interpretation require experience.
  

A TOOLBOX APPROACH TO INVESTIGATING LAYERS OF CONCERN

The Toolbox Approach in Fig. 3 may help students avoid the relatively high number of false predictions that occur due to a combination of several factors, such as extrapolation from single tests and high snowpack variability. The diagram supports a dialogue encouraging students to take a step-by-step approach and observe clues from a combination of tests and observation methods. For example, the combination of both the “yellow flags” checklist sums and fracture character in compression tests provide clues, not confirmation about whether or not a “propagation likely” scenario exists, which is then verified with a large column test that is suitable for testing within limitations posed by the particular snowpack structural properties. Understanding test limitations, matching the test to observed structural properties and verifying observations with complementary tests may improve the ability to interpret the test results and reduce false stable or false unstable predictions. I created this diagram and instructional method five years ago and have included it on the L2M3 and American professional level courses.

FIG. 3 The toolbox approach, ver. 6, Zacharias, 2015 

A CAUTIONARY TALE
Backcountry winter travelers are always encouraged to make weather and snowpack observations in the field, and when possible identify on a drainage and slope scale what the public avalanche advisory describes for the region or range. For the most part, this is an effective risk management strategy. However, there have been a number of close calls, incidents and avalanche accidents with backcountry users increase their risk by not managing exposure when gathering information or misinterpreting the observations they collect. In December 2007, a fatal avalanche accident occurred on Tent Ridge in Kananaskis Country when two backcountry skiers were killed conducting a snow profile in the start zone of an avalanche path. Older examples of riders conducting tests on or very near the slope and being subsequently killed include Wawa Bowl, AB, and Mt. Neptune, BC in 1984, Thunder River, BC in 1987, and White Creek, BC in 1993. More recent incidents include Ningunsaw Pass, BC in 1999 and in Twin Lakes, CO in 2014, where a group of seven dug a profile and conducted eight CTs on a slope before choosing to ski it (see CAIC Incident Report for more information).

There are also several examples of “close calls” where test results gathered and extrapolated to chosen terrain illustrated one problem but not the primary concern. A recent example occurred in December 2013 in Hope Creek, BC involving two backcountry skiers. This is an unfortunate example where a combination of well-intentioned observations formulated a confirmation bias and decision making trap. The rider’s observations prior to descending the slope included three existing ski tracks on the slope, 15cm recent snow, light winds, -3°C and no recent avalanches. The group conducted several tests with the following results: CTM16 (SC), ECTP 23, and “numerous ski cuts in the start zone,” all revealing a significant surface hoar layer (size 7-10mm) 40cm deep but nothing deeper. A DT also revealed no results on deeper layers. The group decided that the surface hoar layer was manageable and to ski the slope one at a time. Rider 1 skied the slope with no problems and stopped 400m below, adjacent to the path trim line. Unfortunately, Rider 2 triggered the slope after landing an air low down in the start zone. The resulting D3 avalanche fractured 100m wide on basal facets 80-120cm deep and well below the surface hoar layer. The fast moving avalanche debris caught Rider 1 on the path’s edge before he could scramble to safety. Both involved were carried approximately 700m downslope. Both were buried and badly injured but were able to self extricate, call for help, and were successfully rescued
(Editor’s Note: read a first-hand account of this avalanche by Billy Neilson in The Avalanche Journal Volume 106).

Those involved generously provided the CAA occurrence report with snowpack observations and insight into what gave them confidence to venture onto this particular slope. This event is a helpful wake-up call as we can all place ourselves in their decision making shoes. In hindsight, it is revealing to examine the Kicking Horse Mountain Resort local forecaster’s public video statement issued on Vimeo on December 13, 2013 for the nearby backcountry terrain. The forecaster warned there is a “basal weakness at the bottom of the snowpack that is still reactive,” and “skier triggered size 3 avalanches have occurred,” and “avalanches had triggered larger slopes sympathetically,” and that “now is the time to be very mindful of slope history.” He went on to emphasize “without that degree of confidence that an avalanche has happened [on your slope of interest], you are really rolling the dice hopping onto big terrain.” This incident—though occurring over one week after the video statement—illustrates that when it comes to managing deeper persistent slabs, the careful observations and good well-learned techniques of the backcountry travelers were not sufficient to protect them from the lingering hazard. It also reveals the big-picture perspective of the forecaster, who clearly warned of the more serious basal concern.

Experience with this type of problem, experience monitoring unstable snow in a shallower snow climate, experience matching specific tests to specific problems, and experience managing false stable results and prioritizing the key concerns are all factors that may have given the experienced forecaster a different perspective than the backcountry riders. In this case, the knowledge of how the snowpack lay over the terrain held more weight than even a series of test results, all of which drew attention to a secondary problem that, while significant, was less so than what lurked below.
The bottom line is snowpack tests used to predict instability, while valuable when employed appropriately, are not foolproof. As Schweizer and Jamieson state obviously and importantly in the aforementioned 2010 article, “decisions about traveling in terrain should not be based solely on stability (snowpack) test results.”

REFERENCES
American Avalanche Association and USDA Forest Service National Avalanche Center. 2010. Snow, Weather and Avalanches: Observation Guidelines for Avalanche Programs in the United States. 2010. Pasoga Springs, CO: AAA. http://www.avalanche.org/research/guidelines/pdf/Introduction.pdf

Atkins, Roger. 2004. “An Avalanche Characterization Checklist for Backcountry Travel Decisions." Proceedings of the 2004 International Snow Science Workshop. Jackson Hole, WY. http://arc.lib.montana.edu/snow-science/objects/issw-2004-462-468.pdf.

Avalanche Canada Incident Report Database. Golden, Hope Creek Draining, Privateer Mountain. December 29, 2013. http://old.avalanche.ca/cac/library/incident-reportdatabase/view.

Campbell, Cam. 2008. “Testing for Initiation and Propagation Propensity.” Canadian Avalanche Association Level 2 Module 3 Lecture.

Canadian Avalanche Association. 2014. Observation Guidelines and Recording Standards for Weather, Snowpack and Avalanches. Revelstoke: CAA. www.avalancheassociation.ca/resource/resmgr/Standards_Docs/OGRS2014web.pdf

Chabot, Doug. “Another viewpoint on a Backcountry Magazine article.” Gallatin National Forest Avalanche Center Blog, November 6, 2015. http://www.mtavalanche.com/blog/another-viewpoint-backcountry-magazine-article

Colorado Avalanche Information Center. Colorado, Star Mountain. February 15, 2015. https://avalanche.state.co.us/caic/acc/acc_report.php?acc_id=526&accfm=inv

Gauthier, Dave, Cameron Ross, and Bruce Jamieson. 2008. “How To: The Propagation Saw Test.” University of Calgary Applied Snow and Avalanche Research. October 2008. http://www.ucalgary.ca/asarc/files/asarc/PstHowTo_Ross_Oct08.pdf.

Jamieson, Bruce. 2009. “Mountain Snowpack and Spatial Variability.” Canadian Avalanche Association Level 2 Module 1 Lecture.

Jamieson, Bruce. 2003. “Risk Management for the Spatial Variable Snowpack.” Avalanche News 66 (Fall 2003): 30-31.

Jamieson, Bruce and Jürg Schweizer. 2005. “Using a Checklist to Assess Manual Snow Profiles.”
Avalanche News 72 (Spring 2005): 57-61.

Jamieson, Bruce and Torsten Geldsetzer. 1996. Avalanche Accidents in Canada Volume 4: 1984-1996. Revelstoke: Canadian Avalanche Association.

Jamieson, Bruce, Jürg Schweizer, Grant Statham and Pascal Haegeli. 2010. “Which Obs for Which Avalanche Type?” Proceedings of the 2010 International Snow Science Workshop. Squaw Valley, CA. http://arc.lib.montana.edu/snowscience/objects/ISSW_O-029.pdf

Klassen, Karl. 2002. American Institute for Avalanche Research and Education Level 2 Student Manual. 2002 Version.

Ross, Cameron, and Bruce Jamieson. 2008. “Comparing Fracture Propagation Tests and Relating Test Results to Snowpack Characteristics.” Proceedings of the 2008 International Snow Science Workshop. Whistler, BC. http://arc.lib.montana.edu/snow-science/objects/P__8177.pdf.

Schweizer, Jürg and J. Bruce Jamieson. 2010. “Snowpack Tests for Assessing Snow-Slope
Instability.” Annals of Glaciology 51(54): 187-194.

Simenhois, Ron and Karl Birkeland. 2007. “An update on the Extended Column Test: New Recording Standards and Additional Data Analyses.” The Avalanche Review 26(2).

Zeleny, Milan. 1987. “‘Management Support Systems: Towards Integrated Knowledge Management.” Human Systems Management 7: 59–70.

“Skiers Died Testing for Avalanche.” The Calgary Herald, December 11, 2007. Accessed December 10, 2015. http://www.canada.com/story.html?id=75d72a3c-c743-4c65-a0c2-a18728813217.

From volume 104, fall 2013

By Troy Leahey

The start zone. Photo by Jim Bay.

This case study details a skier-triggered avalanche that occurred in the backcountry accessed from Revelstoke Mountain Resort (RMR) on February 22, 2013. The party included five young men, all new to Revelstoke. Three members of the party were involved in the avalanche. Two of these were partially buried; the deceased (an RMR staff member) was fully buried despite wearing an airbag. It was a size 2 slab avalanche with an approximately 40cm fracture line, which failed on an early February surface hoar layer.

Analyzing accidents has long been an important part of the CAA ITP program and in other risk  reduction industries. It is easy to be the armchair quarterback after an incident and identify the mistakes people made, but I believe it is also important to identify what was done well in the rescue effort to use as a good learning tool.

My involvement in this accident was on a number of different levels. As a member of Revelstoke Search and Rescue, I assisted Buck Corrigan and Ryan Buhler with the body recovery the day after the accident. I watched and helped interpret a Contour headcam video worn by one of the rescuers during the rescue effort for local members of the RCMP and a representative of the BC Coroners Service. This was a sobering but interesting piece of footage; although the actions of those involved on the scene would be apparent to most readers, the footage required some explanation and interpretation for non-skiing investigators. I interviewed some surviving members of the group for clarification on the location of those involved and their actions for the Coroners Service.

I also led a debrief at RMR for friends and co-workers of the deceased, including the other young men involved in the accident. This debrief was the most challenging public speaking exercise I have ever undertaken. Our company president asked if I could speak about the accident, as staff members and friends in the community had unanswered questions, and answers can help lead to understanding and closure. I wanted to be quite frank about the obvious mistakes made, as they were an opportunity for learning. I also wanted to commend the survivors on their effort and help them recover some confidence and dignity. I focused on the following points in the debrief.

This photo shows the position of the skiers when the avalanche occurred and where they ended up. Their uptrack is in red, the avalanche is outlined in black, and the blue dot is where the victim ended up. Photo by Troy Leahey.

MISTAKES:
• The most obvious mistake was the aggressive terrain choice—the danger rating at alpine and treeline was high. The five males were young, aggressive skiers with limited backcountry experience. This is the demographic we may expect to see on the avalanche fatality list. The deceased had skied in the terrain the day before and felt confident in his decision to lead the rest of the group to that zone. He had spent many days in this area in what had been a mostly stable season to that point. It was also the first major surface hoar cycle we had experienced in the season, and this type of avalanche failure may not have been familiar to a young German in Canada for his first winter. The CAC’s avalanche forecast was bang on—as it usually is—but youth, overconfidence and a desire to ski steep powder caused this glaring information to be ignored.

• Poor group management. The first three skiers on the uptrack were all involved in the avalanche; they were obviously not well spaced out enough and were all engulfed by the slab. Luckily the last two in the group were slower and not affected by the slab failure, and were able to execute a fairly quick rescue. Complacency on uptracks is a common problem for less experienced backcountry travellers. The group of five had been split into the first group of three up front and the slower pair at the back, which leads to a lack of communication and no consensus in the go or no-go decision. The most educated member of the group had been a student of mine on a CAA Avalanche Operations Level 1 course the previous year in Whistler. I spent a lot of time speaking with him about their decisions that day. He admitted to having reservations about the decisions being made, but did not speak up and deferred to the deceased as the leader, since he was most familiar with the terrain.

• Improper use of an avalanche balloon pack. Wear the crotch strap if you are wearing an airbag. The deceased was near the surface with his airbag inflated and clearly visible from 100m away. He had not attached the crotch strap of his airbag. As the overburden of the slab he triggered from mid-slope overran his position in the toe of the debris, the airbag was lifted away from his back and above his head. This caused two serious problems. As the balloons and pack were pushed forward and downhill it lifted the pack, causing the chest strap to catch on his chin and impede his airway. Secondly, the buoyant airbag also pulled the victim’s arms above his head, restricting movement and the ability to use hands to clear his own airway. The Contour video showed quite clearly the victim’s lifeless arms well above his head. The space between the balloon pack and the victim’s back was approximately 40-50cm.

The avalanche path where the victim was caught is on the right. Photo by Jim Bay.

WHAT THE RESCUERS DID WELL:

• The two not involved took a safe route to the toe of the debris and did not expose themselves to any additional hazard. They quickly and efficiently went into rescue mode when they saw the airbag on the surface, and began the excavation by first removing the pack and digging to the head to clear the airway.

• The two partial burials ended up mid-path and were able to self-rescue and do a transceiver search of the slide path down to the victim.

• Group members performed good first aid on the victim. Once they uncovered his head, they immediately cleared an ice chunk from his mouth and pulled him out to a prone position where they started CPR. Considering the environment, they performed excellent CPR with quality air movement as exhibited by the face of the victim in the video.

• The survivor with the most training took control of the rescue effort. He dispatched one of the members of the party to start moving back to the ski area boundary to report the accident as there was no cellular reception on the accident scene. This individual really led the first aid efforts as well.

• After at least a half hour of CPR, they made the decision to try to move the body with an improvised toboggan. They did not get very far as the conditions were very deep, but were able to move the body to a safe location out of the avalanche path. They then made the very difficult decision to leave the body and return to the ski area, as the weather conditions were deteriorating quickly. They left the body in a seated position under a tree with the inflated airbag and flagging tape arranged to mark its position, making our recovery very easy the next morning.

This was an unfortunate accident involving a group of nice young people from around the world enjoying the mountains in Revelstoke. There is nothing ground breaking about this accident other than the question of whether the airbag crotch strap could have made a difference. So why did I write this for The Avalanche Journal on a 30° July day when I’d rather be fishing? On a personal level, it brings back a bunch of vivid images and unpleasant emotions that make me sad. However, on the big-picture level this is an opportunity for others to learn and avoid mistakes in the future. A case study is really a story; this story may be repeated in Whistler, Banff or beyond. Hopefully the mistakes made and the triumphs that occurred in this accident will resonate and help others make better decisions in the mountains. That makes me happy.

Have a fun, safe winter.