Avalanches > Introduction > Risk Reduction Methods > Nivotest
Humans are natural risk takers. If this were not so we would never had come down from the trees and spread out from Africa into the frozen wastes of Europe. There is obviously an evolutionary advantage to taking some risk and one would think that after all this time we would have become quite good at it. However studies in the domains of finance, man-made disasters and avalanche accidents show that we are poor at analyzing just how much risk we are taking.
Prospect theory shows that we weigh the same risk differently depending on whether we believe we are gaining or losing something. Other research has found that individuals with some knowledge are often overconfident, especially in a familiar environment. Safety systems such as Avalungs, ABS and Avalanche Transceivers often mean that we take more risk, Bruce Tremper even postulates that avalanche transceivers have cost more lives than they have saved.
We know a lot about avalanches, both scientifically and statistically. We know that the victims or members of their party trigger 90% of fatal avalanches. We know the weather conditions that lead to avalanches, the different types of avalanches and the slope angles and aspects that are most dangerous. We know a great deal about group dynamics and human factors. And yet around 130 people are still killed annually in the big four alpine countries and North America.
In light of a number of avalanche accidents involving some very experienced backcountry travelers including guided groups, researchers in Europe and North America have looked at the decisions that lead these groups to enter avalanche territory. In Switzerland over the ten year period between 1981 and 1991 there were 52 avalanche fatalities in nine professionally guided ski tours. In France Alain Duclos and Claude Rey noticed that although the overall trend of avalanche deaths was decreasing, that for guided groups had started to increase. Even none guided but experienced backcountry users seem to be taking a lot of risk. Of the 25 deaths last year in France, 6 involved snow professionals and 8 were experienced amateurs. The 2004 season kicked off with the death of a member of the Mountain Police (PGHM) and a near miss for one of France's best known backcountry skiers . Clearly there is a problem with avalanche education. In both Swiss and French training there is now a move away from complicated scientific approaches towards a better understanding of human and terrain factors when assessing risks.
In his excellent articles published recently in The Avalanche Review Ian McCammon suggests that most backcountry travelers short-circuit an extended thinking/analysis process by applying heuristics. Simple rules of thumb used to summarize complex situations. What could be more complex than deciding if a particular slope is going to slide or not? Even the experts say, somewhat ironically, they are only right 50% of the time. Possibly the most basic of these is the avalanche bulletin with its 1 to 5 hazard risk. Rather than understand the details many skiers apply the following rule to summarize the situation: 1 2 go, 3 take care, 4 5 no go. This is perhaps one reason why nearly three quarters of avalanche fatalities occur at risk 2 to 3.
Reducing the many complex factors that make up avalanche risk to a simple go/no go decision is not such a bad idea as long as it is based on a sound analysis of that risk. Avalanche scoring systems do exactly this and they are starting to be used as a standard part of guide and ski leader training in Europe. A scoring system applies a calculation to various aspects of weather and snow-pack, terrain and group experience.
A number of scoring systems have been developed in tandem in various geographic zones: Stop/Go (Austria), SnowCard and Factorcheck (Germany), Red/Amber/Green light (North America), Munter 3x3 and Reduction methods (Switzerland) and NivoTest (Switzerland/France). The fact that three different approaches have originated from Switzerland probably has a lot to do with that country's long history of avalanche research and the fact that it is split into two main linguistic zones.
The most famous of the above are the 3x3 and Reduction method. Developed by Swiss guide and avalanche researcher Werner Munter with the aim to cut avalanche deaths in guided groups by 50%. Munter's methods are currently used in Swiss and French mountain guide training. The Nivotest was developed by Robert Bolognesi in conjunction with the ANENA, the French avalanche research institute and other experts. It is being adopted in avalanche curriculum of French ski group leaders.
Robert Bolognesi is an expert in environmental risk evaluation and has used this knowledge to devise a system for calculating the overall risk presented by an itinerary for the given snow and weather conditions taking into account the strengths and weaknesses of the group.
The NivoTest is based on the classic risk expression:
R = Ht x D
This formula is used to calculate the risks associated with natural hazards. For any Risk (R), Ht is the probability of the event (usually associated with a return period) based on statistical data and D is the potential damage expressed in terms of the vulnerability and value of the objects at risk. Compare this with Munter's Reduction method which defines risk as:
Risk = Danger ÷ Reduction Factors
NivoTest is a catalogue of 25 questions which take into account the weather, the snow-pack, recent avalanche observations, the route and the participants. The classic Avalanche Triangle with the addition of human factors. The NivoTest consists of a card holding a rotary dial a bit like a parking disk and two rules used for calculating 30 degree slopes on Swiss and French 1:25,000 series maps. If the answer to a question is yes, then you add the points indicated on dial. Once the user has completed the test the reverse side gives the risk the user is assuming. The higher the total score, the higher the risk.
However to fully understand the questions that make up the NivoTest it is necessary to read an accompanying book called Attention avalanche! This is a short, 100 page booklet that explains each of the questions in a couple of paragraphs, giving the probability of an avalanche and the potential damage. The book can easily be slipped inside a pocket for reference on the terrain. Each question has its own risk factor, ranging from 1 (lowest) to 5 (highest). By reading the explanations that Bolognesi gives it is possible to deduce the associated risk factor.
Both the NivoTest and Munter methods are heavily based on statistical observations about avalanches in the European Alps. NivoTest has been evaluated by a large number of snow professionals, for example Snow Dynamics, based in the Chugach range in Alaska used the system over two seasons and fed their results back to Mr Bolognesi. However the weightings attached to various terrain and snowpack risks may not translate precisely to other latitudes or to the dryer climate of the American and Canadian Rockies. Research by Schweizer and Jamieson shows that risks on northern slope aspects are roughly similar for Canada and Switzerland. Given good statistical measurements for aspect and slope angles it should be possible to adapt both systems for American use.
We used the NivoTest on a ski tour we made during the late winter in the Belledonne mountains in the French Alps (see photo) then cross checked the results with the Reduction method.
Answering the NivoTest questions takes about five minutes; it is a good idea to involve all members of the group in this process. We answered yes to the following questions out of the 25 with our running total (shown by the disk) in brackets:
The NivoTest says that a score of 1-7 is a Go, 8-25 is doubtful and greater than 26 is a No Go. These are indicated by a series of 'smilies' on the reverse of the disk: :-), :-|, :-).
The risk figure given in the avalanche bulletin is the basis of the Reduction method. In Europe avalanche bulletins are issued daily and cover fairly small geographical areas. The idea of the Reduction method is to take this headline risk and see if it can be reduced to an acceptable level, less than 1. A risk of 1 is statistically equivalent to 1 death per 100,000 person/trips. During the 1980s Switzerland saw 1 death per 50,000 person/trips so this equates to a 50% reduction in fatalities.
For our trip to the Selle du Puy Gris the avalanche hazard was considerable. For each level on the avalanche scale Munter doubles the risk, 2 4 - 8, so this gives a danger value 8. The steepest slope angle was 33 degrees (reduction 4). The slope is north facing (no reduction 1), the group is large (no reduction 1) but the slope is regularly skied so is compacted (reduction 2). Plugging this into the calculation we get:
8 ÷ (4 x 1 x 1 x 2) = 1
With this risk value we should try to look at other reduction factors such as improving group spacing. The calculation took less than a minute.
It is worth examining the questions asked by the NivoTest in a little more detail. These are divided into five areas: weather, snow conditions, avalanche activity, route choice and group. These equate roughly to the three filters of: conditions, terrain and human used in Munter's 3x3 method. For example, under group factors we are asked if one of the group members has poor ski technique. In the corresponding book we learn that this has several consequences, the person may be slower leading to a greater exposure to danger, the person may fall more often which puts a greater trigger load on any instabilities, the person may be less attentive to danger signs and will probably be less able to organize a rescue in the event of an avalanche. This clearly increases the chance of triggering an avalanche and increases the consequences of any slide. Bolognesi sees the NivoTest as a learning tool that will instill good backcountry practices in its users. It also enables direct comparisons to be made between route choices. This is one reason it has been adopted in training ski leaders in France.
When we got to our proposed route, rather than our compact group of 4 we found there were another 17 people ahead of us on the slope, this is a frequent problem on popular routes in the crowded European Alps. Bolognesi says that because NivoTest consists of a large number of questions, none of them with a significant weighting, it is tolerant of these factors and this won't change the overall conclusion. In this case after applying the Reduction method and not being able to find a further reduction factor such as good group spacing we headed for a less crowded climb.
It is obvious that both NivoTest and the Reduction method cut down skiing on steep, shady slopes. Statistically these are the most dangerous but they are often the very slopes we would like to ski. There is a danger that users will consider this conservatism unrealistic in the real world. With the NivoTest it can be seen that it only takes a few risk factors to get a doubtful prognosis. In their tests Snow Dynamics considered the risk to be moderate at the lower end of the doubtful scale. Bolognesi says that his test should be seen more as a pause for reflection by the user who should either consider an alternative route or think carefully about the factors increasing the risk. His book has a complete section on risk reduction covering group management and search and rescue.
Despite the introduction of the NivoTest into ski leader training there still seems to be some reluctance to explicitly adopt scoring systems. In his work on heuristic traps McCammon found that 73% of accidents occurred when there were 3 or more obvious indicators of hazard. So clearly there are some serious errors being made in risk assessment, at least in the groups caught by avalanches. Scoring systems such as the NivoTest can be viewed in a similar way to pre-flight checklists used by pilots. They can ensure that obvious hazard indicators are not overlooked by a group. In the future a more formal adoption of these techniques may need to be mandated for certain backcountry users. For example club ski leaders could be required to keep a log-book showing their decision making process.
<< | Risk Reduction Methods | >>