Accident Route Matrix

Does the Accident Route Matrix (ARM) have a place in understanding avalanche incidents? In this blog we look at systems based incident analysis. These methods look at all the causal factors in incidents far beyond the “who triggered it” strategy of individual decision making.

There are lots of tools for assessing the Human Factors (HF) that contribute to incidents. The Accident Route Matrix (ARM) provides a graphical representation of the route to the incident. This is really useful to highlight Human Factors issues in avalanche incidents, especially for those who just want the headlines.

In this blog I propose that the Accident Route Matrix (ARM) has a place in the context of organisational learning in professional mountaineering associations. By that, I mean the UK member associations that engage in incident reporting.

Simpler systems exist for individual mountain professionals who are operating independently. As they are not subject to the wider influences of an organisation. e.g. a company, bureau or centre.

Avalanche: Does the Accident Route Matrix have a place?
Overhead hazard, an avalanche at an ice climbing venue.

What is the Accident Route Matrix (ARM)?

The Accident Route Matrix (ARM) is a Human Factors investigation tool developed by the Royal Air Force Centre for Aviation Medicine (RAFCAM). Although lots of tools exist that classify the human behaviours and decision making in incidents, they are often difficult to use in combination with traditional professional standards or incident reporting processes.

To create bite sized learning points, incident summaries are often short and superficial. Rarely do mountain professionals want detailed human factors summaries other than the headlines. In military aviation accidents ARM was developed to explain and integrate human factors issues into existing investigation panels.

ARM is based on existing human factor categories, but puts them into a timeline to see where they may have first emerged. These are further categorised at organisation, supervisory, task, equipment, environment or operator level. The proposed benefit of this method is that it helps identify at which point the factor influenced one of four key accident events; hazard entry, recovery, escape and survival. (Harris, 2011).

These levels are fairly generic and I propose that this might be useful in professional avalanche incidents. The hazard entry, recovery, escape and survival phases allow the consideration of other issues such as:

  • Hazard Entry – The recognition of the hazard, narrow margins or safety issues.
  • Recovery – Was there the opportunity to change the course of events, or did further decisions compound the situation.
  • Escape – Is perhaps non applicable in this case.
  • Survival – Choices in selection of equipment such as the use of transceivers, probes and shovels and the influence of those decisions on survivability of the incident may be relevant in this event catagory.

What does ARM look like in an avalanche incident?

To answer the question of what the Accident Route Matrix might look like for an avalanche incident, a simple example appears below. It is based on the narrative account of an avalanche at Aonach Mor (Case Study #4 in Mountain CRM).

Accident Route Matrix - Aonach Mor avalanche (human triggered).
Accident Route Matrix – Human Triggered Avalanche on Aonach Mor

The first day of a week of introductory Scottish winter climbing was a tricky one. The weather was forecast to deteriorate by mid afternoon with significant drifting. Using the gondola at Nevis Range seemed like a good choice to access the snow line and make the most of the days coaching and instructional aims. My two clients were keen and able, so accessing the area around Aonach an Nid was straight forward.

We had a productive morning revising snow belays, and safe travel with intelligent use of the terrain to mitigate the avalanche hazard. Moving up and down some rocky ribs, there was some thin windslab just a few centimetres deep that we could avoid by keeping to the rocks.

The weather wasn’t too bad so we decided to head towards the summit plateau, doing a little navigation on the way. Here we were able to look into ‘Easy Gully’ and notice the tell tale signs of new windslab accumulation on the lee side of the corniced edge. We decided together to finish the day by building some emergency snow shelters, a useful skill for anybody to have in the winter mountains.

To deliver the coaching aims of the day, I needed an area of deposition where the snow had accumulated. Preferably at a reasonable angle so it was easier to dig into the slope. With this in mind we descended back to the rocky ribs bounding ‘G&T Gully’ that we had used before. The area looked subtly different, our footsteps obscured during the hour which we were absent.

I started to descend, my clients just behind sticking to the rocks. There looked to be some good patches of snow lower down and the windslab was shallow, perhaps only 5cm deep at my feet next to the rock rib. Spotting the next group of rocks I stepped onto the other side of the rock rib and was suddenly up to my waist in fresh snow.

Whoommppff – a split second later and the tension crack appeared near my hips propagating across the slope for a hundred metres or so and out of sight. The snow around me seemed suspended for a second despite breaking into big blocks, but then accelerated away taking me with it.

I managed to stop myself but watched with a fixed stare as the snow slowed but covered the flat bench below me. We were all lucky not to be carried down, in part by habitual route choice. The runout distance wasn’t that great and the debris didn’t achieve significant depth. But it was deep enough to bury us easily if we had been carried down.

Case Study #4: Avalanche on Aonach Mor – Mountain CRM

Reflections on Situational Awareness

Failure to correctly perceive the situation:

This incident took me by surprise and affected my confidence for a while. I had failed to correctly perceive the situation. In the first instance I failed to notice the red flags of new windslab accumulation e.g. the slope looked different and our tracks had disappeared. Attentional narrowing was almost certainly present, as I was most focused on finding a suitable location for a coaching exercise, in this case building some snow shelters.

Failure to comprehend the situation:

I failed to truly comprehend that the forecast was one of a deteriorating day, where the potential was for the conditions under foot to change quickly. I was arguably over reliant on my first observations which formed the default values of the day. The situation assessment and mental model was one of stable conditions where I could go anywhere to complete my teaching and coaching aims. When the weather changed quickly this was no longer the case. As a result I was operating with a poor mental model, which was compounded with a desire to fit in one more task before heading back to get the last lift down.

Back to the Accident Route Matrix, these models should always be critically appraised. If you have read this far I should point out that other models are available. These are normally evaluated for variations between analysts. Salmon (2011) provides a comparison of three other methods: Accimap, HFACS and STAMP.

Each was used to analyse the Mangatepopo gorge tragedy in which six students and their teacher drowned while participating in a led gorge walking activity in New Zealand in 2008. The paper is well worth reading but does require academic access via an institution.

Plus points for the Accident Route Matrix

  • Focused on providing a graphical timeline.
  • Considers two transition points (yellow arrows), the first indicating the conditions that were present prior to the event. The second, that were present immediately prior to the incident.
  • Is based on existing categories.

Negative points for the Accident Route Matrix

  • To use ARM in an avalanche context, you do need to be flexible with the existing categories.
  • Organisation and supervision categories may not apply to mountain professionals working independently.

If you have used these models in avalanche incident investigation then please contribute in the comments…


Harris, S., (2011) Human Factors Investigation Methodology, Royal Air Force (RAF) Centre of Aviation Medicine RAF Henlow, United Kingdom [Link]

Salmon, P., et al (2011) Systems-based accident analysis methods: A comparison of Accimap, HFACS, and STAMP, Safety Science.

If you liked this article you may wish to read Attentional Narrowing: Why mountain guides and pilots know it can put you in harms way or the previous blog post Super Puma G-WNSB which examines an aviation incident. Check out Assured Training or E-Learning.

Leave a Reply

Discover more from Mountain Assurance Ltd

Subscribe now to keep reading and get access to the full archive.

Continue Reading