Yesterday I attended the Northern Rockies Avalanche Safety Workshop; and was thoroughly impressed.  Avalanches are unique in the world of objectives hazards.  There are no books which use hundreds of pages to discuss rockfall.  There are scientific conferences on Grizzly behavior, but none which are targeted for passionate lay hikers and focus on attack avoidance.  Whitewater safety classes, and even wilderness first aid classes, discuss human factors as they contribute to safety but much, much less than modern avalanche education.  I left the workshop convinced that this is because avalanches prey upon the greatest weaknesses in human decision making in a way which is, in the world of outdoor recreation, unequaled.  Nothing else comes especially close.  Irresolvable ambiguity intersects with lust long labored for in a way designed to encourage stepping over a line which can never be actually seen.  The analogy is imperfect, but there is an interesting paper to be written about the correlations in behavior between experts caught in avalanches and people with proper education who nonetheless catch STIs.

The workshop ran all day, with well scheduled break and good free coffee.  The weather was not especially inspiring, which helped pack the house with well over 200 people.  It even, and this is truly shocking for NW Montana, almost started on time.

In addition to the thought-provoking picture painted by the speakers, as detailed below, I got to inspect first hand airbag packs from BCA, WRAY, and Mammut. I was impressed with how poorly all of them where designed. The BCA Float 32 doesn’t have enough structure to support it’s own (considerable) empty weight against torso collapse. The market is ripe for the picking.

The first speaker was Scott Savage, a professional forecaster who did a secondary analysis of avalanche incidents among professionals in SW Montana (he was working at Big Sky Resort).  The most revealing piece of data was that out of 3 organizations he only had 33 official incidents to work with.  It could be that these places are just that safe over a length of time, but Savage seemed to suggest that the desire to sweep such things under the rug was to blame.  In any case, you can’t do any interesting correlations with an N of 33.  He also coded that causes for each incident, but seemed to not have addresses inter-rater reliability.  A similar study with a bigger sample size and more exacting analysis would be very interesting.  Savage’s general conclusions were noteworthy but unsurprising: that while expert decision making typically operates beyond mundane heuristics, it can fall victim to comparable traps with similar consequences.  Savage was an excellent speaker, engaging and organized.

The second speaker was LeeAnn Allegretto from the National Weather Services, who shared some things about the weather.gov site which I did not know.  In particular, the scrolling map box in the right margin on the normal forecast page, form which you can get a predictive graph custom tuned to a given 2.5k square (it predicts for the elevation noted at the top).

Third speaker was Karl Birkeland, an eminent figure in avalanche science.  Another great presentation with lots of geeky details well put forth, with two major take-aways: the ECT seems to have the lowest rate of false-safes amongst widely used snowpit tests, and when controlled for other variables the results for the ECT generalize remarkably well across slope angles.  In other words, a test on a 15 degree slope will reliably tell you what’s going on with a 40 degree slope, assuming aspect, snow depth, etc are comparable.

The fourth speaker was pro skier Elyse Saugstad, who was caught in and yesterday spoke about an avalanche (one of four that day) this past February in Washington.  You can read more about it here and on Saugstad’s blog.  Her candor was admirable, and she stated multiple times that under comparable conditions, but perhaps without that particular experience in her pocket, she would ski the line again.

The fifth speaker was Chris Robinson, a former Navy SEAL who gave a rambling yet compelling talk about decision making under fire.  It rotated around the salient idea that given how little time avalanche victims often have, and how good beacons have become, the biggest variable is not the beacon search itself but how the searchers react and use their resources.  He had plenty of anecdotes involving guns and video footage of SEALs shooting at pirates of the coast of Somalia.

The sixth speaker, Dale Atkins, is a long-time avalanche professional and unfortunately by far the worst speaker of the day.  His last slide said that we should avoid avalanche decision-making which hinges on probability, which sounds like a good idea.  He did not coherently suggest how this might be accomplished.

The best part of the workshop was not only the speakers themselves, but the way in which their various ideas interlaced and presented a picture which focused intensely on decision making.  The overall take-away was that there is no objective standard for evaluating hazard.  The snowpack may be able to be read, like a book, by those with enough knowledge.  Yet just like a book the variety of possible interpretations is exactly the same as the number of readers.  Our interpretations are created by who we happen to be on that particular day, and thus uncertainty evaluation (another good yet detached idea from Mr. Atkins) is merely an externalized version of self-examination.  In my limited experience, this is something which the avalanche community is only just beginning to look at fully.  To be blunt, Ms. Saugstad’s repeatedly insistence that she “could ski the slope [which avalanched] blind folded and backwards” as if that was at all relevant is as good a testimony as any here.  My hope is that we’re seeing the terms of discussion change, and drastically.  It is needed.

I believe it was Warren Miller who described modern freeskiing style as akin to “a bag of cats headed for the river as fast as possible”. I think that’s appropriate, and that more broadly a focus on gnar at all expense will need to be reigned back in if the above is to be taken seriously.