Quantitative Risk Analysis of Riding in Avalanche Terrain

[pipedream]

Risk. Our lives revolve around risk but it’s a concept most of us don’t understand well. The human brain is good at many things such as pattern recognition and the nuances of social interaction but our brain is notoriously poor at statistics and probability.

People who are afraid to fly don’t mind driving although, on average, you would have to fly every day for 4,000 years for you to be killed on a commercial aviation flight. Automobile fatalities are so common you hardly see them mentioned in the newspaper anymore (32,000 per year in the U.S.)

We’re afraid to let our kids ride the bus to school yet there are only 3 deaths per year on school busses and 600 per year from parents driving their kids to school.

We’re afraid of shark attacks yet your chance of being killed by lightning is 100 times greater, and being killed in an auto accident is 630 times more likely than being killed by lightning.

So what is the risk of recreating in avalanche terrain compared to other activities? The list below is in units of a “millimort”, which is one death in a million. I have converted all the numbers into millimorts per day so we can directly compare activities using consistent units. A Millimort is a convenient unit because one millimort is the chance of dying per day for an average 20-year-old from all causes.

OK, you say, there’s a big difference in risk between an extreme skier out with their friend with a video camera and a snowshoer walking up the trail. True enough. And those same differences occur in each of the activities I’ve presented here. But each number comes from published data and each is an average of all the risks in that activity. For instance, the number for motorcycles includes both riding a sport bike in a city (an extremely risky activity) and going a few blocks to school on a moped.

At the bottom of the list is the risk of getting out of bed each day. At the top of the list is the activity most mentioned as the riskiest of all activities, Himalayan mountaineering above 8,000 meters where one in 40 people will die.




Here, I've presented the data both as a table and a chart. Notice that the chart displays the data in a logarithmic scale, meaning that each line on the horizontal axis is 10 times greater as you go right. The differences in risk between activities is is so great that it won't fit on a linear scale, I need to change the scale to logarithmic to see it all in one chart.

Backcountry recreation is surprisingly safe but, as usual, it depends. If you use all the risk reduction measures taught in avalanche classes and using modern rescue equipment, the risk is about the same as driving a car one hour to the trailhead (notice driving a car 8 hours is more dangerous than backcountry recreation). This agrees with my experience although I often feel more at risk driving among Utah's famously hyper aggressive drivers in monster vehicles than an ordinary day of backcountry touring.

People often wonder about the risk of randomly center punching a slope without using any risk reduction measures at various danger ratings. These numbers come from Bruce Jamieson’s 2009 ISSW publication where he estimates the risk and compares it with similar estimates from Werner Munter from Switzerland. You can watch a video of Jamieson’s presentation at http://vimeo.com/50900661. You should also look at the rest of their fine video presentations at the University of Calgary Applied Snow and Avalanche Research: http://www.ucalgary.ca/asarc/research/outreach

Someone who randomly center punches 10 slopes per day in terrain rated as Moderate Danger is roughly the same risk as whitewater kayaking or skydiving (but still much less dangerous than riding a motorcycle). Center punching 10 slopes per day in terrain rated as Considerable Danger has roughly the same risk as base jumping, which is considered to be one of the world’s most dangerous sports and one with a notoriously short expected lifespan for regular participants. Doing the same in terrain rated as High Danger is one of the most dangerous activities in the world besides climbing above 8,000 meters.

Bottom Line:
Recreating in avalanche terrain is surprisingly safe as long as it’s done using all the risk reduction measures taught in avalanche classes. It's about the same risk as driving one hour to the trailhead. But randomly center punching 10 avalanche paths per day without any risk reduction measures at Considerable or High Danger can be one of the most dangerous sports in the world.

Source: http://utahavalanchecenter.org/blog-...lanche-terrain

Interesting analysis. Too bad it doesn't include my preferred summer activity of riding a brakeless fixed-gear bicycle in heavy urban traffic. That's gotta be up there with base jumping on sheer risk/reward alone...

[PNWbrit]

Too bad it doesn't include my preferred summer activity of riding a brakeless fixed-gear bicycle in heavy urban traffic.

[obviously]There isn't room on the chart for how tragic that activity is.[/obviously]

[Hugh Conway]

Are there many million user days of US backcountry skiing/snowboarding a year? I know a couple of other people have tried to figure out the "how many people" question, as have I, and it really looks like the answer is "nobody has a decent clue".

[neufox47]

Hmmm, whole thing is pretty much based on opinion and not data. But I think either the numbers are being heavily skewed by the differences in activity (snowshoeing vs. skiing) to the point that they don't really apply to most bc skiers, or the primary assumptions they started with are just wrong. Maybe its because most of my experience in avy terrain is on Mt. Washington, but the odds of death seem way higher than that study suggests.

I'm curious to see JonathanS's take on this as resident statistician and avy expert / has the attention span to actually run some numbers.

[skifishbum]

like most statistical probability studies the devils in the details.
there isn't any distinction between accessing the terrain you plan on centerpunchin. Seems to me a lot of incidents occur before the skiing part starts
hell peoples definition of center punching differs. how many pople do you know who actually centerpunch anything w/out some form of risk reduction. Wouldn't wearing gear or having a partner be a form of risk reduction?
Isn't the danger rating a subjective guess by humans that spatial varibility and everchanging terrain and weather conditions can make moot
I love the repeated references to risk reduction taught in avvy classes
everybody knows there are no skilled tradesmen who haven't gone to trade school and recieved the peice of paper
or that paying $$$ and attending a class is any real measure of skills
is that piece of paper really much different than the one the chica who puts smilie faces on the foam of my latte has?
I guess she actually had to show that the knowledge was really learnt
where as the avvy cert is given w/out reguardless

[systemoverblow'd]

quant/qual

[icelanticskier]

death, our nations # 1 killer. and will continue to be.........

rog

[hortence]

How can you ascertain "millimorts" without actually knowing how many people are doing the activity. Sure we can get a sense of how many people died in the Himalaya, but without knowing how many people actually went over to climb, the total death is meaningless.

Same with backcountry skiing -- how many go out in a season? Who the heck knows?

It seems like a lot of sketchy stats to conclude that following well established avalanche terrain safety protocols does in fact make you safer.

[Hugh Conway]

How can you ascertain "millimorts" without actually knowing how many people are doing the activity. Sure we can get a sense of how many people died in the Himalaya, but without knowing how many people actually went over to climb, the total death is meaningless.

That's relatively easy to know in the modern era because every 8000m peak requires permits and like all good bureaucracies the numbers are tallied and compiled. Similar data exists for most treks in Nepal, with breakdowns by nationality, month, etc. Well quantified, limited access, negligible means to escape canvasing.

The actual numbers for skiing - who knows?

[Jonathan S.]

Why are we only now in July discussing a blog post from February? (Kind of grim looking back at the marathon figure: at the time I thought of heart attacks, but now for anyone living in Massachusetts...)

Maybe we were all too busy skiing back then -- well, I fly out Saturday morning for my final ski trip of the season (business meeting in Chicago, so while halfway across the continent, might as well keep on going, right?), so briefly:

Are there many million user days of US backcountry skiing/snowboarding a year? I know a couple of other people have tried to figure out the "how many people" question, as have I, and it really looks like the answer is "nobody has a decent clue".

Agreed.

I have put together many quantitative analyses combining various datasets to address issues related to what happens once caught in an avalanche, but taking an epidemiological approach to the chance of being caught (or buried, or killed) based on participation rates -- I can't recall seeing anything like that.

Pretty much the only thing we know is that backcountry skier/boarder days have gone way up in recent years ... yet the number of skier/boarder avy deaths have remained fairly constant. I take that as encouraging evidence of the role of more widespread avy education, the improved focus of current avy course curricula, and more effective companion rescue.

Hmmm, whole thing is pretty much based on opinion and not data. But I think either the numbers are being heavily skewed by the differences in activity (snowshoeing vs. skiing) to the point that they don't really apply to most bc skiers, or the primary assumptions they started with are just wrong. Maybe its because most of my experience in avy terrain is on Mt. Washington, but the odds of death seem way higher than that study suggests.

I find his chart valuable mainly as a heuristic tool, given such shortcomings in its quantitative accuracy.
My version of that (sort of) is a long-term survival calculator (inspired by his actuarial table at the beginning of his book), which you can view and download here (at the fourth page/tab):
https://docs.google.com/file/d/0B6QM...it?usp=sharing

I also think that skier/boarder avy safety is different from many of the other activities in the chart: if you are caught in an avalanche, it is 100% your fault, either b/c of your flawed stability assessment and terrain choice, or because you turned off your own brain and relied on someone else's flawed stability assessment and terrain choice. We just don't have the, say, driving equivalent of getting run over by a drunken driver who crosses the meridian. Therefore, risk is so individualistic, as opposed to facing some common underlying risk.

Mt Washington is very different because almost everything you want to ski at treeline is windloaded and at a prime sliding angle. (Truly above treeline is better, but unfortunately weather usually shuts it down, plus the best access is usually through the more dangerous at-treeline terrain.) Any wintertime snowpack is by definition suspect, but on Mt Washington, it's not just suspect: it's been indicted, and standing trail -- and it's not George Zimmerman, it's Michael Dunn.

Now, despite all that, wintertime skiing on Mt Washington has certainly increased in recent years, yet the last skier avy death (as opposed to climbers) was all the way back in 2000, and then before that, the pair of deaths in 1996. Are those the only skiers to ever die in avalanches on Mt Washington? (The chart I compiled for all avy deaths -- skiers or climbers -- goes back only to 1982.) I think we've just been getting lucky lately, given all the close calls. The avy triggered by a professionally guided party using the Obs as a high-priced mountain hut was especially disturbing. I have a bad feeling we're in for more of that...

[icelanticskier]

(Truly above treeline is better, but unfortunately weather usually shuts it down, plus the best access is usually through the more dangerous at-treeline terrain.) Any wintertime snowpack is by definition suspect, but on Mt Washington, it's not just suspect: it's been indicted, and standing trail -- and it's not George Zimmerman, it's Michael Dunn.

coming from someone that spends most of his days up high in winter I will tell you that you really should start spending at least a day or two above treeline in winter to really see what's up. the skiing/weather is a whole lot more forgiving than you or most ec bc users would think.

Now, despite all that, wintertime skiing on Mt Washington has certainly increased in recent years, yet the last skier avy death (as opposed to climbers) was all the way back in 2000, and then before that, the pair of deaths in 1996. Are those the only skiers to ever die in avalanches on Mt Washington? (The chart I compiled for all avy deaths -- skiers or climbers -- goes back only to 1982.) I think we've just been getting lucky lately, given all the close calls.

only 3? really? where's yer chart? I was on scene for those 3 but surely there must be more. 12? or is that the total for all avy related deaths up there?

The avy triggered by a professionally guided party using the Obs as a high-priced mountain hut was especially disturbing. I have a bad feeling we're in for more of that...

that avy was pretty much just a sluff, but the client skied the wrong spot the wrong way in that incident. had the client followed the guides (proper) skiers left line away from the classic loaded skiers right line, all would've been fine. that shoulda been communicated by the guide before hand.

rog

[Jonathan S.]

that avy was pretty much just a sluff, but the client skied the wrong spot the wrong way in that incident. had the client followed the guides (proper) skiers left line away from the classic loaded skiers right line, all would've been fine. that shoulda been communicated by the guide before hand.

I was about to post that the official report had a different account, but now I can't find any officially report, so that might just be my personal recollection (correct or otherwise) of what I was individually told two days later by a Snow Ranger.
This is the only mention I can find on the website:
http://www.mountwashingtonavalanchec...rday-3-2-2013/
I don't anything here:
http://www.mountwashingtonavalanchec...013-summaries/
Any know where to find an official report?

[neufox47]

Rog, he's isolating skier related avys. Also I think he's talking about the wounded warriors guided trip not the skiers in LC.

Sent from my ADR6425LVW using TGR Forums

[Jonathan S.]

only 3? really? where's yer chart? I was on scene for those 3 but surely there must be more. 12? or is that the total for all avy related deaths up there?

Here's the latest version of my chart:




Mount Washington has also had three avalanche fatalities that predate that start of my chart. (I omitted them since they occurred before even Roger Damon started teaching avalanche safety courses.)

Here are two climbing fatalities in Huntington -- note that Tux was officially closed that day, a practice that has long since been discontinued. This is also the latest in the season that a fatal avy incident has ever occurred on Mt Washington.




This appears to be the first avalanche fatality on Mt Washington:




So that looks like only three skier avalanche fatalities ever on Mt Washington.
Divide that by the [unknown] number of wintertime skier visits to Mt Washington and you would derive ... a number that probably has no epidemiological value in estimating the going-forward risk of backcountry skiing there.

[Jonathan S.]

Rog, he's isolating skier related avys. Also I think he's talking about the wounded warriors guided trip not the skiers in LC.

Yes on the first, but no on the second.

More to following (getting this back on the original subject) after mtn biking...

[Hugh Conway]

Pretty much the only thing we know is that backcountry skier/boarder days have gone way up in recent years ... yet the number of skier/boarder avy deaths have remained fairly constant. I take that as encouraging evidence of the role of more widespread avy education, the improved focus of current avy course curricula, and more effective companion rescue.

Meh, you are ignoring a far more substantive change in the what and how people are skiing the backcountry that is technique & gear driven, unrelated to any of those.

[Jonathan S.]

Meh, you are ignoring a far more substantive change in the what and how people are skiing the backcountry that is technique & gear driven, unrelated to any of those.

Yes, but if anything, isn't that going in the other direction? In other words, we're not talking about hippies on floppy leather, 3-pins, and nordic-esque skis, but instead skiers on big gear from resort backgrounds who want to ski big steep lines.

So if both the number of backcountry skier/riders days has gone up, and the composition of lines skis has changed, then the annual fatality tally holding steady implies a much lower incident rate.

[Hugh Conway]

I disagree with your causation of the lower accident rate being education (or, more accurately, don't necessarily see much support for said assertion) .

My view is that "information" - in particular on local conditions, in a variety of reporting formats - is much, much, much more accessible and much more frequently accessed than before and much more relevant than education.

[icelanticskier]

Any know where to find an official report?

c'mon man, you know how small that valley is. they took it down to protect the rep of the guide.

rog

[Jonathan S.]

I disagree with your causation of the lower accident rate being education (or, more accurately, don't necessarily see much support for said assertion) .

But we do agree that the avalanche fatality rate per backcountry skier/rider must have gone down dramatically, correct?
And do we agree that the AIARE course curriculum is a vast improvement upon the, well, pretty much lack of any standard approach to avy safety education a couple decades ago (when it was mainly a disjointed combination of snow science and rescue)?
Admittedly, I still lack any direct proof for causal connection between the two, but still, I can hope, can't I?

[icelanticskier]

So that looks like only three skier avalanche fatalities ever on Mt Washington.
Divide that by the [unknown] number of wintertime skier visits to Mt Washington and you would derive ... a number that probably has no epidemiological value in estimating the going-forward risk of backcountry skiing there.

well i'll try not to be at the site for the next one. going 4 for 4 would be trippy

so nice that most of our terrain is so steep that oftentimes the slopes shed the new while it falls in so many areas.

"taking a peek" into some ravines on a high day has often turned an obstacle into an opportunity for an acent when a line naturals wall to wall t2b. love that

rog

[Hugh Conway]

I don't agree that the AIARE course curriculum is any better than the course I took 15 years ago.

What is massively better is the information available on routes and conditions. Particularly for the routes and zones that are now popular to ski. GPS tracks, photo documentation, slide paths, easy to access accounts of previous accidents, weather for the past 30 days, snowfall for the season. All pretty easy to find if you have the time, and much of that can be analyzed and interpreted without "education".

Try going somewhere without information, or with minimal or bad information sometime.

[Jonathan S.]

c'mon man, you know how small that valley is. they took it down to protect the rep of the guide.

Was an official report ever even published at first?
I usually save to my computer anything that is published on-line, but I might have neglected to do so for this one.
All I saved was a passing mention in an article mainly on the solo climber fatality:

The other incident, which involved a guide descending with a client from an overnight at the Mount Washington Observatory, occurred in Lobster Claw, a gully in Tuckerman Ravine. The guide skied a short way down the gully to an island of relative safety where he could watch his client descend, Snow Ranger Frank Carus said on Monday, but when the client began skiing the gully slid. "His slab was probably three to 10 inches [thick] and 50 feet or 60 feet wide," he said. "He was carried about 10 feet."

(And looks like no update to the guide's blog -- would have made for some interesting material!)

[Jonathan S.]

I don't agree that the AIARE course curriculum is any better than the course I took 15 years ago.

Because you don't think much of AIARE, or because you took such a good course way back in 1998 from ... who?
(Not doubtful, just curious. Back in 2001, my wife and I took a very "traditional" two-day course, which looks nothing like current education.)

What is massively better is the information available on routes and conditions. Particularly for the routes and zones that are now popular to ski.

Excellent point (which I had entirely overlooked): mapping tools, recent trip reports, on-line beta requests -- all a huge improvement from years ago, and almost even unimaginable back then what we now have (and for free too).

[Jonathan S.]

Try going somewhere without information, or with minimal or bad information sometime.

I had tours like that early in my backcountry ventures -- quite the contrast with Glacier Peak last month. Although we had never been there, with all the maps, trip reports, photos, advice (thanks again Big Steve!) ... it was still really fun being somewhere for the first time, but sure was better knowing whether we were going and what to expect.
And even Alaska in 2009 in the modern age, that 100k map scale sure lead to some surprises!