Holes in the Yarnell Hill Fire swiss cheese

In another thread there was a discussion about the Yarnell Hill Fire and the fact that when the 19 Granite Mountain Hotshots were killed there was only one aerial supervision aircraft over a very complex fire environment instead of two. Until recently it was more common to have both an Air Tactical Group Supervisor (ATGS) and a Lead Plane or Airtanker Coordinator over a fire that had advanced beyond initial attack.

The ATGS orbits the fire and coordinates, assigns, and evaluates the use of aerial resources, both helicopters and fixed wing. The Lead Plane directly supervises the air tankers, usually flying low and sometimes physically preceding the air tankers before they drop the retardant.

But now we sometimes see those two roles combined into one aircraft, called an Aerial Supervision Module. It can save money, but there is debate about how appropriate it is for a complex fire situation.

One of the recommendations in the first report issued about the fire, by the Arizona State Forestry Division last summer, was for the the State of Arizona to “request that the NWCG develop guidance to identify at what point is it necessary to separate the ASM and Air Attack roles to carry out required responsibilities for each platform”. Other documents released by the state of Arizona last week revealed that members of the Blue Ridge Hotshots said that they witnessed “a near miss” with aircraft, who they described as sounding “overwhelmed” adding that “the air show seemed troublesome.”

The aerial supervision on the Yarnell Hill Fire was only one element, or one slice of James T. Reason’s Swiss Cheese model of accident causation, which is defined in Wikipedia:

In the Swiss Cheese model, an organization’s defenses against failure are modeled as a series of barriers, represented as slices of cheese. The holes in the slices represent weaknesses in individual parts of the system and are continually varying in size and position across the slices. The system produces failures when a hole in each slice momentarily aligns, permitting (in Reason’s words) “a trajectory of accident opportunity”, so that a hazard passes through holes in all of the slices, leading to a failure.

Having only one aerial supervision platform on a very complex fire gives you one slice with some holes. Below we list 18 other holes in the Swiss cheese.

Swiss Cheese modelAnother slice would be supervision of ground personnel. The holes in that slice were:

  • Transitioning that morning from a group of firefighters to only a partial Incident Management Team (all transitions can be tough, but when done hurriedly and to only part of a team, it can be dangerous);
  • Removing Supt. Marsh from the Hotshot crew and making him Division Supervisor. (A reporter who has seen the recently released documents told me that Marsh did not know he would be Div. Sup. until he got out on the fire line that day);
  • No Safety Officer;
  • No Division Supervisors arriving with the IMTeam;
  • No Division Supervisor on the Division adjacent to the accident;
  • An Incident Commander that took over the fire about six hours before the accident;
  • Decision-making was poor, such as failure to designate division breaks, or decide on and communicate a firefighting strategy likely to be successful.
  • Somebody, either Marsh, a Structure Protection Group Supervisor, or an Operations Section Chief (or all of the above), decided for the Granite Mountain Hotshots to leave the safe previously burned black area and walk through unburned brush, where they were entrapped by the fire and killed.

Holes in the planning slice were:

  • No maps given to firefighters that day;
  • No Incident Action Plan that day or the two previous days;
  • There was a poor briefing that morning;
  • Marsh did not attend the briefing because it was given in mid-morning after he and his crew departed the Incident Command Post and headed to their assignment (Marsh did receive some briefing info that morning).
  • There was no complexity analysis completed on day one or day two (the accident occurred on day three of the fire, June 30.) It was completed three hours before the accident.
  • The number of firefighting resources working on the fire for the first three days was inadequate to safely implement the strategy of fully suppressing the fire and protecting the structures and the people in the communities.

Holes in the communication slice:

  • Incorrect radio programming information was given to firefighters that morning which made radio communication difficult;
  • The overhead, such as it was, did not maintain adequate communication with field personnel, which led to inadequate accountability of personnel who were in harms way;
  • Marsh and Granite Mountain did not clearly tell the Operations Section Chief where they were and where they were going when they left the secure black en route to the box canyon;
  • The ASM had difficulty communicating with Marsh and Granite Mountain as they became entrapped, possibly due to him being “overwhelmed” (as described by the Blue Ridge firefighters).

These slices have 19 holes, and when you place the slices next to each other, all it takes is one more hazard that then passes through holes in all of the slices, leading to an unfortunate outcome. Perhaps if one or more of the holes had been plugged by better management of the fire, there would have been a more favorable result.

 

Author: Bill Gabbert

After working full time in wildland fire for 33 years, Bill Gabbert now writes about it from the Black Hills. Google+

13 thoughts on “Holes in the Yarnell Hill Fire swiss cheese”

  1. The more I read about how the fire was managed, the more I am reminded of something that was pounded into me by my EMT instructor during training. You have to know your craft cold BEFORE you respond to a call because when you are in the middle of the action, that is not the time to be wondering what you do next.

  2. Bill that’s a very good summation of what happened, why didn’t the SAIR say that. Address the problems and move forward?

  3. If and when all the facts come out about the Yarnell Fire and the Granite Mountain IHC fatalities, it will become a classic Case Study in Lessons Learned for the next decade – until it happens again! The “holes” in this Swiss Cheese Model were not technology failures as much as they were in the human arena, and that’s where we must place our focus and corrective actions.

  4. The questions left unanswered in the human arena are going to be much more difficult to pin down, but carry the most important lessons. It is our nature to have empathy for deceased and not assign blame. We don’t yet know why they left the black and this is an important question that needs to be answered.

    Just as important is to understand why they perished in the spot that they did. Why was the decision made to hunker down when there was still time to run. Why did they choose to backfire around their position in highly volatile fuels. Has anyone interviewed the spotter to find out what they were taught concerning fire shelters? Were they taught that burning out and using a shelter was a lifesaving method?
    I understand that to run they were only left with two choices: back the way they came, or to their right. Both directions were uphill through heavy brush with scattered large boulders. Running uphill through the brush and boulders would have been slow going.

    But has anyone really looked into this? There was certainly “some” time between when they felt they had a problem and the time they got into their shelters, perhaps six minutes or longer. Consider that most hotshot guys can run a six minute mile, it is not unreasonable that they could have run a quarter mile back the way they came, even through the boulders. The two hikers out ran it, why is it so unreasonable that if the hotshots would have dropped their packs and saws and ran like hell, that they couldn’t have added valuable minutes?

    Was there a mistaken belief that fire shelters were a viable alternative and by not running, but rather spending precious moments cutting and setting backfires before deploying the shelters? Did the backfire in volatile fuels accelerate the entrapment, rather than lengthened it? Was the decision not to run, rather than the leaving of the black or the lack of tactical response, that ultimately caused their deaths? Should the culture of “not running” be looked at more closely?

    Does the use of fire shelters need to be examined more closely, especially their placement in different terrains and vegetation? The video mostly deals with a forested situation, with little specific direction about the type of terrain that the Granite crew was in. Why has Canada abandoned the use of fire shelters? Should the use of fire shelters be limited to locations were deploying them would actually prolong life, rather than give a false sense of security?

    The Swiss cheese modeling is very insightful and is most effective when as many slices as possible are scrutinized. Perhaps a slice should be devoted to the questions of the human factors, why they left the black and also why they chose to do a burnout and use the fire shelters. Were there any other options that might have had a different outcome? If so, then it is this lesson that needs to be shared with the firefighting community.

  5. Big question for me is why not run like hell. Seems like it would be human nature to drop packs and accelerate. Difficult terrain, uphill, but adrenalin alone might have got them to the boulder field. So many unanswered questions.

  6. With due respect, my impression was that GMH didn’t have that much time to prepare the site, and that the terrain would likely have been too difficult to manage a fast enough pace to escape flame front accelerating uphill …a lot of those boulders are humongous and many of the spaces in between had a lot of choking chaparral. Admittedly, that’s just from looking from the highway, but the terrain might have been even nastier than it looks. Has anyone timed themselves on comparably steep uphill terrain with chaparral of the same density it was there? Of course, if the time line information that had been deduced previously was wrong, then we may not know just how much time GMH had to prepare for the flame front.

    I think Joy and Tex got out of the fire’s path before there was a question of having to outrun the fire, in addition to their having expert local knowledge of the terrain and perhaps being more aware than the fire bosses of the entrapment potential. They also weren’t carrying heavy equipment and hadn’t been working fire line for weeks almost nonstop.

    1. Pat Byrnes on December 16, 2013 at 9:52 pm said:

      Joy:

      EXACTLY- well said. perfect comment. That area was chokingly with chaparral and boulders

      Sonny:
      however if they left the area to the south or the north of the deployment area they could have gotten into huge boulder areas with very little vegetation/fuel. Some boulders even were cave-like with flat ground area to deploy with very little vegetation. In fact, the fellow that said why not run like hell was correct. You could not do that in the maze of manzanita/cat claw/scrub oak/brush that was there but they could have had they stayed on the 2 track road. I choose to go to the Congress side (less fuel and many boulders) where we dropped off to another gravel road yet once we got to that road we had very little time as the smoke and fire was right on our heels in the distance as you can see in one of Joy’s photos. What you do not understand or can not understand is once the fire took off about 12:24pm that they did not take off- I did at 12:38 and Joy rejoined me 40 minutes later when I went back for her.

  7. Firefighters don’t want to drop their gear. Its kinda driven into our heads. Also, they wouldn’t have deployed if there was any other option. The idea of pulling out the shelter is a scenario that we don’t even want to think about. It absolutely had to be the last option.

  8. When you see that their lookout came very close to a deployment, and that Willis had also identified a deployment site (but no safety zone) in a different area that also likely would have been problematic, it does seem there may have been an overreliance on the ability to survive burnovers by deploying, particularly when they were faced with a certainty of direct flame contact. That’s three cases in three different physical locations.

    Since GM’s physical circumstance where they deployed should not have suggested that deployment was survivable, certainly one valid option was to drop gear and, if not run, seek better terrain. As an active trail runner who can easily run 6 minute mile splits, I personally still have to slow to a walk when going steeply uphill on rugged terrain for any distance, even if in more open terrain and not choked chaparral. So, it is not clear that GM had any options you would call good. But, identifying that where they were was non-survivable and telling people to hoof it would have given a chance to some of the crew.

  9. Drop gear and run might have been their only survivable option . Best and last chance. Maybe too little time. Maybe black smoke. But it takes incredible discipline to get 19 men to stick together till the hellish end. The fact that they deployed rather than scramble is a question for further learning. No judgement here. To be human is to make mistakes.

  10. Why are there no holes in the swiss cheese for the first 18 hours of the fire, June 28th and early morning of the 29th. No mention of that in the theory. The ratio of “Risk to Exposure” and it’s duration is way far less during that period. A quick, aggressive and proper initial attack has acceptable risk with most all conditions in the favor of a successful suppression effort compared to days or weeks of defensive suppression tactics.

Comments are closed.