Old snow problems, aka persistent weak layers (PWL) are a fragile layer in the snow pack that takes a very long time to stabilize. The layers are frequently widespread and produce large avalanches and can be susceptible to remote triggering. The most common PWLs are: faceted snow, depth hoar and buried surface hoar. They can produce avalanches days to weeks after the last snowfall, frequently when conditions appear stable. That is, at moderate or even low risk. Sometimes they can be reactivated in the spring by snow creep or meltwater activity.
At altitude there is almost always a layer of depth hoar between the ground and the base of the snowpack but not every year, nor every region, has an old snow problem. This makes the avalanches unexpected, even to experienced skiers including guides.
In 2019 we have already seen some major avalanches along the border region between France and Italy. 2 ski instructors on the Italian side of le Mont Blanc in October. On the 23rd November a group of ski tourers in Valfréjus. The avalanche broke on an old layer of weak, angular grains. Subsequent avalanche control work by the piste services triggered some major slides on this layer. Guide Roberto Ferraris, deputy commander of the Cervinia mountain rescue was killed touring alone on the 15th December. A ski tourer at Alagna on the same day. An avalanche took out a ski lift at Serre Chevalier after a slide was triggered by the piste patrol using a gazex system. Is this the start of another old snow problem in the region?
Avalanche, Rendl, 17th March 2017
2017 was a particularly bad year for old snow problems in the Eastern Alps and French/Italian border. A weak layer formed on the first snows of autumn and was buried by subsequent snowfall in the early winter. The first clue that this layer would be a problem came in the Ferwallthal in the Obergurgl at the end of November. A 10-person group, accompanied by two local guides were climbing the valley. Towards the head of the valley they decided to dig a snowpit and then continued climbing with spacing. Shortly after they heard a settling noise followed by four of the group, including one of the guides, being hit by a large avalanche. The slab avalanche started high above the group then picked up speed on the steep valley walls. One of the group died instantly and another the following day in hospital. The two survivors were also injured.
Investigators dug snow profiles and carried out stability tests and it became obvious that there was an old snow problem. The slide broke on a layer of angular grains on a bed of compact snow that fell in mid September. It was probably triggered from below by the skiers in an area of thin snow cover, the fracture then propagated over 800 meters distance. The avalanche started on extremely steep terrain. The investigators concluded that “the old snow problem, known from early winter in shaded, high alpine areas, where early snow can remain, should be taken extremely seriously.”
This persistent weak layer would return to Austria with serious consequences in the spring. On the 13th of March 2017 a five-person, guided group left the Gaislachkogel to ski the Rettenbach valley. At around 2300m, the group triggered a slab on steep terrain. Four people, including the guide were caught by the slide. Everyone was equipped with beacons and airbags. Three of the group activated their airbags and they were only partially buried. The area had previously been avalanche controlled but clearly not above the weak layer which was close to ground level.
On the 15th an 8-member Swiss group led by a Swiss mountain guide climbed the 2450m high Jochgrubenkopf. During the descent, the group triggered a slab avalanche. Four people were caught and totally buried between 3 and 12 meters. The rescue involved 5 helicopters, dogs and a large number of rescue workers with heavy equipment but none of the buried skiers were recovered alive. The avalanche danger was moderate (⅖) above 2200 meters, low (⅕) below. The slide was skier triggered at 2200 meters on a 40 degree slope on a weak layer of depth hoar (old snow) above a thin snowpack (at that point). The guide had been working since 1974 without incident. Tests by investigators showed that there was a good propensity for the weak layer to propagate in the snowpack and produce large avalanches.
Looking beyond the headline risk the avalanche bulletin noted “Possible danger spots for winter sports enthusiasts can still be found in very steep, little skied, shaded slopes above about 2200m. Where the snowpack is thin, avalanches can still be triggered in weak layers close to the ground, particularly due to high loads.”
On the 17th March a 5 person guided group triggered an avalanche at Rendl on a NNW 40° slope (an avalanche protection wall) at 2030 meters altitude. Three of the group were caught, two were buried and killed under 2 to 3 meters of snow. It seems that moisture penetration down to a ground level weak layer had been a key factor along with a thin snow pack at the trigger point. A buried rain crust increased the stiffness of the slab and aided the propagation of the fracture increasing the size of the slide.
If you fire the gun often enough, you’ll find the bullet
Old snow situations fall into the scope of High Risk, Low Frequency (HR/LF) problems well known beyond the world of snow science. Safety experts classify situations on a risk matrix - Low Risk/Low Frequency, High Risk/Low Frequency etc and it is precisely these HR/LF problems that keep them awake at night. The avalanche bulletin danger level is calculated on a similar matrix: low frequency/small avalanches to high frequency/large slides. Precisely because old snow problems are low frequency, backcountry travellers are unfamiliar with this kind of problem and experienced base intuition doesn’t work or can even be dangerous. Many people will have difficulty even recognizing the situation even where it is mentioned in the avalanche bulletin.
Everyone has a plan until they trigger an avalanche
HR/LF problems can be broken down into two classes: DT and NDT (discretionary time and non discretionary time). DT means we have time to think about the problem and mitigate for it - for example from information in the bulletin or on the ground (whoumps, snow profiles, propagation cracks). Examples of mitigation would be increasing group spacing on PWL slopes or choosing safer, lower angled terrain with low consequence run outs. NDT is when you are actually skiing the slope, if disaster strikes, beyond grabbing your airbag handle, all bets are off.
Swiss research shows that people tour less when the avalanche risk is higher (Considerable ⅗ or above) but mention of a PWL problem in the bulletin has no effect on overall numbers. Do skiers recognize the problem? Perhaps unconsciously at least. The Valais and Graubünden have more old snow problems due to a continental climate. The avalanche risk is twice that of other areas as a consequence of snowpack construction. Skiers undertake lower risk tours in general.
Swiss researchers have made some other interesting findings. There are three times more ski tours at the week-end compared to Monday to Friday and three times as many avalanche accidents. When the weather is good there are, again, three times as many tours and three times as many incidents. If risk is calculated as the number of tours divided by the number of incidents there is no overall increase in risk taking.
When the avalanche danger level increases there are fewer tours. At level ⅗ (Considerable) there are two-thirds the number of tours compared to level ⅖ (Moderate), at least for ski touring. There are three times the number of serious avalanches at level ⅖ compared to level level ⅕ and five times with level ⅗. Combining these figures the risk of serious avalanche is 2.5x at level ⅖ compared to level ⅕, increasing to 6x at level ⅗. At danger level ⅗ or ⅘ (or bad weather) not only were fewer people touring but they were choosing lower peaks and easier tours and even with this more defensive planning skiers were often turning back before reaching their goal.
What happens where an old snow problem is mentioned in the bulletin? Moderate danger days (⅖ on the avalanche bulletin) are 50% more dangerous when there is a PWL. When the danger is Considerable (⅗), PWL days are twice as dangerous. Perhaps more surprisingly, PWL avalanches are triggered by guided groups in 80% of cases reported in Switzerland. Canadian research on unexpected avalanches echos these findings.
The avalanche doesn't know you're an expert
Why do PWL avalanche catch guided groups? Surely the point of going with a guide is to travel safely? Without firm data about the number of skier days for guided and none guided groups it is difficult to know the extent of the problem. A guide probably skis three times as many days as an experienced ski tourer and maybe ten times that of a “hobby” tourer (100:33:10).
Reports of guided groups getting hit by big avalanches crop up most seasons. Guides operate safely on typical risk 2 and 3 days where the risks are localized so it seems the additional old snow problem is an issue. It may be the nature of guided groups that causes problems. Although, as we saw in the Ferwallthal incident above the guides recognized the danger and spread the group out it is difficult to allow too much spacing and keep control of group members who are maybe less experienced than typical recreational skiers. PWL avalanches often become dangerously large, breaking over the entire slope and to ground level - the High Risk square of the risk matrix. The Foglietta avalanche on the 5th January 2015 illustrates the problem. The group was led by two ski instructors. One acting as a back marker. The avalanche risk was ⅗. They skied the slope one at a time waiting at what they thought was a safe spot but when the avalanche broke it took most of slope and overran the waiting zone. Allowing more space may simply not be feasible for guided groups.
Remote triggering is also a danger with old snow problems. Again it is difficult to leave enough space in a largish group that possibly doesn’t travel quickly due to fitness or experience. The trigger points are often not sensitive to the weight of a single skier but a group traveling relatively closely may be sufficient to provoke the weak layer. Less experienced skiers may also fall (3 times the force of a normal skier turning) more often or not follow the guide’s instructions (group spacing, where to ski) to the letter. The practice of snow farming, closely following tracks may also put more load on the underlying snow pack.
Lies, damn lies and statistics
Applying statistical risk reduction techniques may not be sufficient. We’ve seen that the danger where a PWL is present is 50% to 100% greater. That means that if the PWL is not recognized and not accounted for in the risk reduction technique the danger for a particular slope will be underestimated. Where an old snow problem is mentioned in the bulletin it should be used as an additional risk factor. In the Munther method a Risk Factor of 4: is applied to all critical slopes and altitudes mentioned in the avalanche report, in the case of a PWL a risk factor should be applied to north sector slopes above 35 degrees even if they are not mentioned in the bulletin.
Expect the Unexpected
Guides, operating over the whole season with their own communication channels, are frequently aware of old snow problems but they often describe avalanches incidents as “unexpected”. Given that sectors such as the Graubunden in Switzerland see more incidents than other regions even avalanche forecasters may be underestimating the problem in those sectors (or conversely, overestimating the danger in other regions).
Surveys of backcountry travelers indicate that they think the danger level in the bulletin is too high rather than too low and this trend is more pronounced among guides, ski tour leaders and very experienced recreationalists. Guides also rate the accuracy of the bulletin as slightly below average. These groups use their experience to take on more risk and so are equally likely to be involved in incidents as less experienced users taking into account the number of skier days and some of these incidents, involving PWLs are in the “High Risk” category.
A wide range of information such as snow profiles, stability tests, alarm signals or avalanche observations from different observers flows into the avalanche bulletin. The Swiss bulletin warns of an old snow problem if the main danger is neither fresh or drifting snow. For example for the Graubunden on the 17 December 2019: “Over a wide area various wind slab layers are lying on a weakly bonded old snowpack.” This is a detail worth looking for and taking into account in planning.
Old snow problems are like minefields. You don’t know exactly where the trigger points are on the slope but there are some general rules:
Read the avalanche bulletin carefully - look for words like “buried weak layers”. Where an old snow problem is suspected defensive skiing is required. Slopes should be skied individually or climbing routes changes (skiers are 3x more vulnerable when climbing) but remember the slopes often go big so islands of safety need to be well out of range of a big slide. Study the terrain, PWL frequently trigger on transitions or rocky terrain where there is less snow due to wind erosion. If in doubt, turn back or take another route, particularly if there are obvious clues like weak layer collapse (whoumphs).
A special thanks to Klar on the TGR forum who started an extensive discussion of this problem with respect to statistical methods.