What does a well-founded risk decision look like?


What does a well-founded risk decision look like?

By Mark Smith, Mission Focused Solutions
© Mission-Centered Solutions Version 7 FINAL (June 16, 2017)

[Mark Smith has been presenting Luck Runs Out (What does a well-founded risk decision look like?) at IMT meetings, safety summits at wildfire academies, etc. Where The Big Lie was more a problem statement, Luck Runs Out is meant to focus on actionable, practical steps forward for IMTs and Agency Administrators. The information below started not as an essay but as a handout to go with presentations.]


This document is the result of an ongoing dialogue on the risks that I had within the post-Yarnell Honor the Fallen group. A member asked the rhetorical but critical question: "Should we risk lives for suppression efforts or not?" intelligently accept this risk.


The hard truth is that fighting wildfires is a high risk business. Consider the policy that all firefighters on the line carry fire shelters. It’s a clear recognition that every time firefighters are directly involved in suppressing or managing forest fires, their lives are in danger. In addition, the environment of forest fires exponentially increases in complexity, amplifying the risks. However, what doesn't keep pace is our sophistication to plan, operate and support in this complex environment filled with risk. Our tools are further behind each fire season.

In an attempt to meet this challenge, leaders often make statements such as "No structure is worth a life!" Although true, the statement does not contain any meaningful indications. On the other hand, it is the job of any risk professional to determine exactly the level of risk of the structure. And even though this statement may seem simple, the application of the concept continues to cause confusion among senior forest fire officials.

Where does a well-founded risk decision begin? First of all, it should always start with clearly identified and prioritized values ​​at risk (VaR). It’s the “result” in “Does the result justify the risk?” It is the "gain" in "risk vs gain". This is the "goal" in the "task, goal, end state" of the leader's intention.

If the prioritized risk values ​​are not clearly taken into account, articulated and displayed for everyone to see, all subsequent risk decisions will be profoundly wrong. They should be specifically discussed during the briefing of the agency administrator (AA) and reviewed during all subsequent meetings of objectives, strategy and tactics.

Second, we need to do a meaningful risk assessment for firefighters. The current risk analysis (ICS Form 215a) on the NWCG website reflects a risk approach from the 1970s. One lists the risks and then mitigates those risks. It does not quantify the risk in any way, nor is there any discussion of the level of risk after mitigation and whether this residual risk is acceptable.

Given what we know of the shortcomings of this form, its continued use will be considered negligence and will open agencies and their practitioners to increased liability over time.

In order to see the truth about the levels of risk at which forest firefighters operate and the evolution required to make the best risk decisions, let's first consider the two-axis model, probability / severity. Most incident management teams (IMT) now use a modified risk analysis (215a) incorporating this model.

probability versus gravity

The attraction of this model is the simplicity of its green / amber / red "traffic light" appeal. Unfortunately, it is not sufficiently nuanced for the environment of forest fires. When you really do a solid risk assessment, so many things fall into places like "low-end high" or "medium-high". It also lacks the sophistication of exposure accounting, such as the number of operational periods, the number of fuel cycles, the number of people, etc.

It also does not take into account cumulative cumulative effects such as multiple risks added to each other. For example, Division A presents a high risk due to snags and falling objects, but also due to road conditions and flight operations. It's not just a high risk, it's now HighRisk3.

In the exposure curve on the next page, the mathematical reality of the severity with which the probabilities change when you climb this probability axis is alarming. When you increase your risk or take into account exposure, such as being tired AND being on steep rocky terrain AND being on day 12 of an assignment, the percentages increase exponentially.

gravity versus probability

A key step forward would be to adopt a more sophisticated probability / severity matrix that takes these additional factors into account and more accurately describes the risk spectrum in the wildfire environment.

On the axis of gravity – You can do some things on the side of attenuation that could influence gravity. A kevlar suit from a rider, for example, will help when snapping into a tree, but chances are the rider is still breaking bones. Likewise, a fire shelter can shift the consequences of death to a simple burn, but nevertheless, a significant risk remains.

probability versus gravity

Given the inherent dangers of a forest fire environment, the severity will always be high if things go wrong. Given the likely consequences, most operations on the line of fire will be at medium or high risk. It is extremely rare to find a fire where all operations are in the low risk category.

Most decisions about risk must therefore focus on the axis of probability. For good risk decisions, a template like this should be part of the AA briefing and be posted next to the 215A for discussion. "Should we knowingly risk lives?" – would it be more appropriate to rephrase the text to "Should we place people where the likelihood of something bad happening is high?" You can see on the matrix how quickly the risk escalates when you go from Remote to Improbable to Possible.

One of the inherent challenges in risk management is that, as humans, we cannot feel when the odds go from 1 in 1,000,000 to 1 in 1,000. But the ; universe perfectly follows all this in real time. This is the discussion that should take place in front of 215A. Not only to list threats and mitigations from rocks, snakes and lightning, but to have dialogue and to know that the real risk lies halfway between the most likely fire and the worst case. It is essential to judge the probability that an operation will meet this next high risk level and to consider the conditions that could precipitate it.

When things go from improbable to possible, it's a big leap, and leaders need to reconsider this risk / reward calculation. Unless there is a cabin in the woods filled with babies, puppies and kittens, then the answer is clear: no, we should not place firefighters in a location where the probability moves to the upper end of Possible and the resulting Extreme risk.

In the wildfire environment, risk levels will regularly be at medium or high risk. There is probably nothing humans can do to prevent this. We should have very high expectations of our AA and IMT decision makers in terms of critical thinking and their sophistication in making acceptable risk decisions, which means that we need tools worthy of the real risks that firefighters take in the current suppression paradigm. According to this current model, we risk lives and, therefore, we have a duty to ensure that it is done intelligently.

In special military operations, a risk must be determined if necessary during the planning process in order to achieve a target. "If we do that, here is the necessary risk that we are going to have to take." At this point, the question arises: "Does the result justify this risk?" if that is the case – it becomes acceptable risk. Otherwise, you are trying to mitigate the risks to the acceptable level. If you can't get there, it's not acceptable and military operators are looking for another way to achieve this goal with less risk. In some cases, the goal must be abandoned all together because the risk is too high in relation to the outcome.

The difference in special operations is that the small unit (i.e. the crew) is heavily involved in mission planning and risk decisions. This is not true in forest fires – an echo of the reason why we still have terms from the Great Depression / era of chain gangs like the crew bosses in job descriptions for forest fires. In forest fires, planning, risk mitigation and decision-making often occur in the absence of those who will be directly exposed to the risk. This places significant responsibility on AA and IMT staff to discuss strategic and operational risks at the Common Operating Picture (COP) meeting, reviewing the expected end state and creating (or validating) targets based on the VaR. These strategic and operational risks need to be further validated and refined during strategy and tactical meetings, where staff specify the necessary risks.

Once articulated on the 215A, it is up to each member of staff to ask the question "Are the residual risks that remain – post-mitigation – justifiable?"

"No tree is worth a life" only tells you what a tree is not worth. But what is it worth? What is the acceptable risk associated with protecting a tree? A structure? A subdivision? Clearly, the responsibility of management is to make and communicate this decision, but in the absence of a meaningful means of making the acceptable decision, operators are often left to interpret this ambiguous intention for themselves.

This gap is not the result of a lack of concern, desire or intent, but rather the absence of the tools necessary for forest firefighters to be able to make all kinds of objective decisions about an acceptable risk. And this is because there is nothing that maps the priority of risk stocks over acceptable risks to protect them. We want to be very clear. Very simple.

Here is an example of what this mapping might look like. As with many examples, there are missing elements and you may not agree with the priority given to VaR, but it is intentional. A completed interagency version should be clear enough so that there is no misunderstanding or disagreement. It is part of the predetermined playbook. Pending a "Red Book" version, it is a AA and IMT responsibility to develop and communicate to all:

EXAMPLE of Acceptable Risk Guidelines

Risk level Type of value at risk
Extreme risk A viable and save human life in imminent danger
High risk Human life in potential danger; the threatened species; essential infrastructure affecting human health; employment centers at regional level; cultural or historical values ​​of national importance
Medium risk Pets; critical watersheds; schools and centers of worship; primary transmission power lines; species at risk; values ​​supporting human livelihoods; neighborhoods; cultural or historical values ​​of regional significance or less
Low risk Recreational values; road networks; electricity distribution lines; non-designated public lands

This matrix provides clear and acceptable risk guidelines for a category of VaR. Incident commanders could make exceptions using the same authority they now have for adjusting work / rest and other directives, but this removes the "What is the value of a tree? " ambiguity.

The challenge remains, however, that without certain measures to be attributed to an identified risk, the assessment is always subjective. "Hmmm … It will take 2.5 hours to bring someone from" C "Division to the hospital in the event of an injury. How do you think it affects the level of risk? "

Some LMIs are starting to assign numerical values ​​to each of the identified hazards and risks. Example: "Less than one hour of medical evacuation to a hospital is low risk, 1 to 1.5 hours + is medium, 1.5 + is high."

As a decision maker during my previous military career, I was introduced to a standard risk analysis / risk decision process in the mid 80's. This sample map below has been used only for training events, which is why you don't see categories for enemy strength, enemy cohesion, etc. Imagine it being added to the map as projected incident behavior.

This map is a distillation of the probability / severity matrix, listing the routine variables encountered while training soldiers. This is a "big army" tool, so there are no scuba / night diving categories, just a simple vanilla tool to help quantify the discussion and give managers a common operational picture of risk. In the special forces, we had more sophisticated versions, taking into account obscure factors such as infrared crossing times, the phase and lighting of the moon, the activity of solar flares, etc. Think of fuel humidities of 1000 hours and the Haines index.

Once you have included the fire time and fire behavior variables, and have adapted to other common wilderness variables, you can now have a probability / severity matrix in the briefings or AA tactical meetings and finding actual numerical values ​​to plug into a modified 215A.

(Front – left; rear – right)

This would provide a more quantifiable approach. Inevitably, there will always be situations where, despite well planned mitigation measures, we will still have a high risk residual risk score. Let's say a 25 using the plastic card above. This is where we have to go back to the beginning – the values ​​at risk, the "goal" for the leader. Looking at the example of an acceptable risk table, you will find that the high risk is simply not an acceptable risk for saving three chicken coops and a PC hill.

Now what? If we cannot reduce the risk by other tactics, we will have to revert to the previous C&GS – Strategy meeting. We will have to find another strategy to reduce this risk while achieving the goal. If we cannot create an alternative strategy to reduce the residual risk to acceptable, then we must save even more and revamp the goal that the strategy was supposed to achieve. This process would force decision makers to become much more strategic about suppressing actions – continuing the evolution of engagement thinking that is moving towards the best peak compared to the next peak.

The problem with this kind of thoroughness is that it takes time. If an artificial delay such as the production of IAP due to the availability of the copier or some other factor is at the origin of the quality of our risk analysis, then the tail wags the dog .

The question remains, and should we accept high risk? Let's take our previous example – "Okay, we've planned all of these mitigations and we still have a residual risk score of 25, high risk" – but this time, let's use a different risk value and the same risk guidelines acceptable. "VaR is one of the main power lines to Phoenix and it measures 114 degrees. High risk is acceptable because if this power goes out, some people at risk will die. "This is probably a very appropriate level of risk.

Finally, to have a well-founded risk decision, it is essential to share the risk. Shared risk is a recent buzzword in forest fires, but it's important to really understand what it means. Shared risk means that national leaders create acceptable risk guidelines based on the values ​​at risk, such as the example table. This means that they shared the responsibility and the risk of putting firefighters to protect the power line in the example above. This means that AAs and CIs prioritize risk values ​​as part of the delegation and the dialogue. This increases the quality of fire management interactions with line officers during pre-planning and once the fires start.

The net result of this is that everyone involved – national leaders, agency administrators, incident commanders – shares the responsibility. This means that LMIs perform risk assessments with tools appropriate for the severity of the work, use more objective criteria and create the intention of the leader with the task, goal and end state. By linking the goal to a specific VaR and deciding if the risk is acceptable, they now share the responsibility.

But shared risk means that it is also shared at the operator level. It is much more than a simple refusal protocol. In the current system, the consideration of refusal is completely subjective. "I just don't feel comfortable." When operators receive a division assignment sheet (ICS 204), they have no way of knowing what level of risk they have been asked to accept, so they do not have starting point to go through the risk management process at their tactical level. A 204 grade would include this: Special instructions: Risk level – HIGH – due to the increased density of snags in Division A.

If we regularly included the task, the goal, the end state on a 204, then each DIVS and team leader would also understand the VaR that they are asked to protect. If we included Low, Medium, High, etc. on 204, they would then know the level of risk that IMT deemed acceptable for this VaR. If the table of acceptable risk guidelines were included in the IEP, then they would have all of the ingredients required for their own assessment "Does the result justify the risk?" . Even in a series of mop changes, a low risk yesterday could be high today due to an overnight wind event. Now, crew chiefs and deputy chiefs are also responsible for risk decisions. This is what shared risk really looks like.

Current tools and practices are falling further and further behind the growing complexity of the wildfire environment. The widening gap means that we increasingly rely on luck to succeed. The global game industry’s $ 90 billion in annual revenue is made possible by a universal truth: "Luck Runs Out".

The evolution and use of a few simple tools could have a significant impact on the laudable goal of "significantly increasing the chances of everyone going home" at the end of the next fire season. Let's move on to the risk management process for forest fires from the 1970s and 1990s to the 21st century.


Reproduced with permission from Mark Smith, Mission-Centered Solutions. All opinions are those of the author.