A force more deadly than polar bears

When an avalanche engulfed 11 homes on the edge of Longyearbyen on 19 December 2015, people turned out by the dozens to help dig out people who had been buried in the snow. Photo: Courtesy of Trond Olsen/Sysselmannen, all rights reserved.

Top image: Avalanche area in Longyearbyen 19 December 2015. Photo: Courtesy of Trond Olsen/Sysselmannen, all rights reserved.

You might think that polar bears – and the potential for attack – are the biggest danger the Norwegian island archipelago of Svalbard. But avalanches kill far more people on Svalbard than polar bears ever have.

16 February 2017
Text: Nancy Bazilchuk/NTNU

One minute he was warming milk for his coffee. Then his world turned upside down.

Malte Jochmann was in his kitchen in Longyearbyen when an avalanche hit on 19 December 2015. His house was one of 11 that was literally lifted off of its base by a wall of snow. He was tumbled around inside the kitchen along with his partner, their two children and a visiting friend. Snow from the avalanche filled the house, crushing windows and doors, but somehow Jochmann managed to end up on his feet so he could dig himself out, and save the rest of his family and his friend.

Two other people in those houses were not so lucky, and died in the avalanche, buried under two metres of snow or more. The death and destruction from the slide made it the worst natural disaster in this town of 2000 people in recent memory. As of early February, the insurance costs related to the accident, including tearing down the ruins of the houses, and possible new construction, have approached NOK 50 million, or about EUR 5.6 million, according to reports in Svalbardposten, the local newspaper.

But if you look at how much precipitation actually fell, you might not think such a destructive avalanche could be possible, says Alexander Prokop, an avalanche researcher at UNIS, the University Centre in Svalbard.

“They measured 18 millimetres of precipitation at the airport,” Prokop said from his office in Longyearbyen. “But the fracture, the crown height of the avalanche where it broke away from the snow, was up to 4 metres high.”

From immediate action to big picture research
Since the tragedy more than a year ago, Norwegian authorities have taken a series of measures to protect Longyearbyen residents in their Arctic homes. They’ve created a new risk assessment map for the town and instituted an avalanche warning system, currently operated by the Norwegian Water Resources and Energy Directorate. The Norwegian Water Resources and Energy Directorate has also published a case study of Longyearbyen that presented possible methods for protecting buildings and other structures from avalanches and damage from permafrost.

But the climate and risk situation in Longyearbyen, the largest town in the island archipelago, is undergoing enormous change because of global warming. At the same time, there is relatively little long-term information about the frequency of weather conditions that increase avalanche and landslide risks, Prokop says.

“The problem here in Svalbard is that they don’t have measurements that are suited to investigating (these kinds of) hazards. They just have precipitation measurements at the airport, which you can’t really use for snow data,” he said.

Prokop and NTNU researcher Arne Aalberg plan to change that. Prokop is working to install a network of snowfall measuring instruments around the town. This information is vital to improve avalanche forecasting. At the same time, he and his students have taken detailed measurements after the deadly 2015 avalanche, which are being used to fine-tune a computer model that will also help improve forecasting.

Aalberg, who is now leader of the Arctic Technology group at UNIS in addition to his NTNU affiliation, is focused on snow pressures and forces on buildings or other manmade structures. One of his master’s students is investigating the best way to measure these forces in the hills around Longyearbyen. This information could help with the design of avalanche fences or other kinds of structures to protect existing houses or buildings, Aalberg says.

Alexander Prokop in Longyearbyen
Alexander Prokop points to the area where 11 houses were destroyed by an avalanche. Two people died when they were buried in their houses. Strong winds deposited up to 4 metres snow on the slopes behind Prokop. All remains of the ruined houses have been removed. Photo: Nancy Bazilchuk/NTNU.

Arctic and alpine snowpacks are different
Prokop, an Austrian who is also an avid skier, has firsthand knowledge of how communities in the Alps protect their villages from the ravages of avalanches. His latest academic publication, from September last year, is all about a new methodology for planning snow drift fences in alpine terrain.

But while there has been a lot of focus by researchers on avalanche dangers and mitigation in the Alps, much less is known about arctic conditions. And these conditions are quite different.

For one thing, arctic areas like Svalbard are underlain by permafrost, or permanently frozen ground. That matters because the difference in temperature between the top of the snowpack and the bottom, near the ground, is important in helping create conditions that can trigger an avalanche.

Snow scientists like Prokop call this temperature difference a gradient. If the gradient is big, with a large temperature difference between the top of the snowpack and the bottom, the snowpack is more likely to be unstable.

This is because the big temperature differences allow snow crystals to grow. Big crystals that are poorly bonded to the rest of the snow create a weak layer in the snowpack. A weak layer is exactly what it sounds like: a layer in the snowpack where there is a weak or slippery surface or layer that allows the snow on top to slide off.

Weak layers are very common in arctic snow packs, in part because the snowpack is often very shallow. A shallow snowpack can make for a bigger temperature gradient. The changing climate has also made weak layers far more prevalent, because climate change has made it more likely that there are crazy temperature changes. It can be bitter cold in Svalbard on one day, but then rain the next, Prokop says. If a rain crust forms on the snow surface, this provides a slippery surface for new snow that might accumulate on top in subsequent snowstorms.

Knowing where the snow accumulates – and how much
It wasn’t a shallow snowpack that caused tonnes of snow to roar into Longyearbyen on 19 December 2015, however. Instead, high winds swept mountaintops clean, and deposited all this collected snow into protected areas, like the hillside called Lia, which is where the avalanche happened.

That’s because wind-drifted snow is another key factor in determining what will cause an avalanche, Prokop says. “The actual amount of snowfall is not that important in Svalbard, but due to heavy winds and a lot of fetch area, a lot of snow can be drifted into an avalanche release area,” he said. “That was the problem with this avalanche.”

Installing snow depth measuring instruments around Longyearbyen will allow Prokop and others to know not only how much snow falls, but also how much snow has been redeposited by the wind. Across the world, avalanche experts consider a daily new snow depth of 30 cm – whether wind deposited or freshly fallen – to be a critical amount that can increase avalanche risk.

“That is a key amount where you have to take action,” he said.

Measuring snow depth with laser scanner
Alexander Prokop out in the mountains outside of Longyearbyen, measuring spatial snow depths. Photo: Alexander Prokop/UNIS.

Building a predictive model
Another way to improve the ability to predict avalanches is by creating a computer model, where you enter different snow conditions for different locations and see what the model predicts.

Prokop and his students took action right after the deadly 2015 avalanche to collect vital information to build exactly this kind of predictive model.

“What we did was try to document the event, what actually happened,” he said.

“What did the snowpack look like? How much snow had drifted in? What about the wind? We had pictures and laser scanning, and we created a high-resolution snow surface model so we can actually know how much snow was released” during the avalanche.

The proximity of the avalanche slope to the town, and the fact that people took pictures of the slope as the wind piled up snow on the slope and around buildings also helped provide very useful information, Prokop said.

“This is very important and is often missing, because you don’t see where the avalanche was released or no one can go up there,” he said. “Here in Longyearbyen, it is easy to go there, and we have pictures prior to the avalanche.”

Prokop has a master’s student who is working with information from other avalanches that occurred around Longyearbyen to fine-tune the model. Then, he says, you can just change the snow height and see how far the avalanche would travel, or how much snow it might deposit.

Graphic of avalanches in Longyearbyen Janaury 2017.
An example of how terrestrial laser scanning can show spatial snow depth changes; the illustration shows avalanches that occurred in January 2017 in Longyearbyen. Graphic: Alexander Prokop/UNIS.

Snow loads and protective fences
While Prokop is studying avalanches, Aalberg, as a building engineer, is interested in understanding what kinds of forces the snow exerts before it slides away in an avalanche. This kind of information is important if the town were to decide to build some kind of protective structure to prevent avalanches from happening.

Elsewhere in Norway and in other avalanche-prone areas, for example, towns erect a series of snow fences on slopes where enough snow might build up and tumble down as an avalanche.

There are at least two types of fences, Aalberg says. The first kind are composed of open fences (which are essentially wind screens, half of the area of which is open). These structures slow the wind which causes the snow to be deposited on the leeward side of the fence.   “It’s a kind of engineered re-location of snow accumulation,” Aalberg said. “It piles up in suitable, ‘safe’ areas, instead of on the road or on a railway or whatever you are trying to protect.” The challenge with this kind of snow fence is that it requires a fairly large area for the snow to accumulate, he says.

The second kind of protection that can be built includes fences, walls, or soil dams that are placed in series along the slope. These structures do not necessarily prevent snow accumulation, but help anchor the snow so it can’t begin sliding and cause an avalanche. Snow will still accumulate, but is “anchored” safely if the fences are well anchored in the ground.

However, “if you are going to take any safety actions, you need to know how much force a structure will experience,” Aalberg said. Although there has been a fair amount of research on these kinds of forces on mainland Norway, Svalbard experiences far different conditions than on the mainland, he said.

“No one can guarantee that the design manuals from the mainland will work,” Aalberg said. “But if you are thinking about building something that will hold the snow back, you have to know how the snow behaves.”

Arne Aalberg
Arne Aalberg in his office at UNIS in Longyearbyen. His computer shows an early mock-up of a design for a snow pressure-plate measuring device. Photo: Nancy Bazilchuk/NTNU.

Aalberg hopes to collect data over several years using a pressure plate measuring system that has been developed as part of one of his master’s student’s research. Then, he says, if the town decides to build some kind of protective structures, engineers will have the information they need.

Whether the town will choose to build snow fences or other protective structures remains an open question, however.

“Svalbard is a place where there are so many historic structures,” he said, which are protected by Norwegian law. Then there’s the actual physical effect of building big fences on the mountainsides around the town.

“It would be ugly,” he said. “It would be a bit of an attack on the environment. It would take 50 years for the vegetation (that would have to be disturbed to build the structures) to come back.”

More dangerous than polar bears
While the Norwegian Water Resources and Energy Directorate has now started providing avalanche forecasts for Svalbard, Prokop says he wants to create a sustainable avalanche safety system that is based on the island archipelago itself.

“We want to help the town to know when it is dangerous and when it is not, and how to behave,” Prokop said. Such an automated system also needs to be independent of local, knowledgeable individuals, he said, even though local knowledge is very important.

“People stay in Svalbard an average 4 years and then they leave,” he said. “The problem is that then the knowledge also leaves.”

As one example of this, Prokop pointed out that there had been a hazard map created for the Longyearbyen area in 1992, which showed that Lia, the area where the avalanche occurred, was at risk. All in all, the tragedy was an important, if distressing reminder, of the power of the environment.

“It might be interesting for people to know that in the last 15 years, 7 people died from avalanches,” he said. “But in the last 35 years, only 3 people have died from polar bear attacks. It is more likely that you will die from an avalanche in Svalbard than from a polar bear. ”

Alexander Prokop, Emily S. Procter, A new methodology for planning snow drift fences in alpine terrain, Cold Regions Science and Technology, Volume 132, December 2016, Pages 33-43, ISSN 0165-232X, https://dx.doi.org/10.1016/j.coldregions.2016.09.010.


This article first appeared in Gemini Research News
(Science news from NTNU and Sintef).

Arctic Geology Arctic Technology Research