Incident directory

2013 - Yarnell Hill Fire, Arizona

30/06/2013

Country:

USA

  • Open air fires
  • Other

Severity:

Description

Date of event

30th June 2013

Time of event

Fire Started at Approximately 17:40 on June 28th, 2013. (INITIAL CALL)

Name of premises

Yarnell Hill.

Location

Arizona, United States of America (USA)

Service area

The Granite Mountain Interagency Hotshot Crew (IHC) which were hosted by The Prescott Fire Department, City of Prescott, Arizona.

Nature of incident

Fire in open.

Property type

Wildland.

Premises use

Wildland.

Construction type and materials

‘Variable terrain from steep ridges to nearly flat valley bottoms with numerous rock outcrops and boulder piles and fields’. ‘Vegetation consisted of mainly chaparral brush which ranged from 1 to 10 feet in height and in some places was nearly impenetrable’. ‘The last wildfire in the area was in 1966’ (Karels & Dudley et al, 2013).

Occupancy

Wildlife with numerous dwellings and other structures threatened by, and approximately 100 structures destroyed by the fire.

Fire source and location of fire

Accidental – Lightning strike/s

Synopsis

Brief Synopsis

‘Nineteen firefighters died on the Yarnell Hill Fire in central Arizona on June 30, 2013 after deploying fire shelters’. ‘They were members of the Granite Mountain Interagency Hotshot Crew (IHC), hosted by the Prescott Fire Department’. ‘One crewmember was separated from the crew earlier that day and was not at the deployment site’ (Karels & Dudley et al, 2013).

The conditions leading up to the fire were described as very high to extreme fire danger due to the extreme drought that had been present at the time (Karels & Dudley et al, 2013).

‘Late afternoon on June 28, the Yarnell Hill Fire started high on a ridge west of Yarnell, Arizona when lightning ignited multiple fires’. ‘The fire, which was under the jurisdiction of the Arizona State Forestry Division, started in a boulder field in steep terrain with no vehicle access; it was about one-half acre in size’. ‘Responders saw minimal fire activity or spread potential, and they had several safety concerns with putting firefighters on the hill overnight’. ‘In consideration of these and other factors, the Incident Commander prepared for full suppression on the following morning’ (Karels & Dudley et al, 2013).

‘On June 29, resources held the fire in check until around 1600, when winds increased, and the fire spotted outside containment lines’. ‘That evening, the Type 4 Incident Commander ordered a Type 2 Incident Management Team (IMT) and additional resources for the next morning’. ‘The fire grew throughout the night, to an estimated 300 to 500 acres by morning’ (Karels & Dudley et al, 2013).

‘Early on June 30, members of the Type 2 IMT began arriving’. ‘In a briefing at 0700, the incoming Granite Mountain IHC Superintendent accepted the role of Division Alpha Supervisor’. ‘His assignment was to establish an anchor point at the heel of the fire with the Granite Mountain IHC’. ‘The Type 2 IMT assumed command, an action formally announced by radio at 1022’ (Karels & Dudley et al, 2013).

‘For most of the day, the fire spread to the northeast, threatening structures in Model Creek and Peeples Valley’. ‘Around 1550, the wind shifted and the fire started pushing aggressively to the southeast, toward Yarnell’. ‘Fire resources shifted to resident evacuation and structure protection in town’. ‘Only the Granite Mountain IHC remained out on the ridge, on the southwest perimeter of the fire’. ‘Personnel who communicated with the Granite Mountain IHC knew the crew was in the black at that time and assumed they would stay there’. ‘No one realized that the crew left the black and headed southeast, sometime after 1604’. ‘At 1630, thunderstorm outflows reached the southern perimeter of the fire’. ‘Winds increased substantially; the fire turned south and overran the Granite Mountain IHC at about 1642’ (Karels & Dudley et al, 2013).

‘There is a gap of over 30 minutes in the information available for the Granite Mountain IHC’. ‘From 1604 until 1637, the Team cannot verify communications from the crew, and we have almost no direct information for them’. ‘There is much that cannot be known about the crew’s decisions and actions prior to their entrapment and fire shelter deployment at around 1642’ (Karels & Dudley et al, 2013).

 Pic 02

Image from Karels & Dudley et al, 2013.

 Pic 03

Air Attacks reconnaissance photo looking to the Northeast, June 29th at 19:24. Image from Karels & Dudley et al, 2013.

 Pic 04

Granite Mountain IHC on the morning of June 30th, image taken by hikers. Image from Karels & Dudley et al, 2013, courtesy of Joy Collura.

 Pic 05

Image from Karels & Dudley et al, 2013.

 Pic 06

Image from Karels & Dudley et al, 2013.

 Pic 07

Photo taken by Christopher Mackenzie at 15:50 on June 30th. Image from Karels & Dudley et al, 2013.

 Pic 08

Photo texted by Wade Parker at 16:04 on June 30th. Image from Karels & Dudley et al, 2013.

 Pic 09

One of the very large air tankers (VLAT’s), a DC10, dropping fire retardant on the Yarnell Hill Fire on June 30th. Image from Karels & Dudley et al, 2013, courtesy of Rick Tham.

 Pic 10 crop 636649214624009148

Image from Karels & Dudley et al, 2013.

 Pic 11

Image from Karels & Dudley et al, 2013.

 Pic 12

Image from Karels & Dudley et al, 2013.

 Pic 13

Image from Karels & Dudley et al, 2013.

 Pic 14

Image from Karels & Dudley et al, 2013.

 Pic 15

Image from Karels & Dudley et al, 2013.

 Pic 16

Image from Karels & Dudley et al, 2013.

 Pic 17

Image from Karels & Dudley et al, 2013.

 Pic 18

Image from Karels & Dudley et al, 2013.

 Pic 19

Image from Karels & Dudley et al, 2013.

 Pic 20

Image from Karels & Dudley et al, 2013.

 Pic 21

Image from Karels & Dudley et al, 2013.

 Pic 22

Image from Karels & Dudley et al, 2013.

 Pic 23

Image from Karels & Dudley et al, 2013.

 Pic 24

Image from Karels & Dudley et al, 2013.

 Pic 25

Image from Karels & Dudley et al, 2013.

 Pic 26

Investigators inspect on July 3rd inspect the location where the Granite Mountain Hotshots where entrapped by the fire 4 days earlier. Image from Dougherty, D. (2018), original image from Arizona State Forestry Division.

Further information hoping to be identified and still to be located.

Main findings, key lessons & areas for learning

Further information hoping to be identified and still to be located.

City of Prescott Fire Department summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

Arizona State Forestry Division summary of main findings, conclusions, key lessons & recommendations

Taken from; Karels, J. & Dudley, M. et al. (2013). Yarnell Hill fire June 30, 2013 serious accident investigation report September 23rd 2013. [pdf]. Arizona State Forestry Division.

--subaccordion--

Part Two: Learning Discussion

Discussion

The Yarnell Hill Fire was a significant event to those personally affected and to the interagency wildland fire community at large. Because we do not know, and will never know, many of the precise details surrounding the final movements and motivations of the Granite Mountain IHC, we considered many aspects of wildland fire operations in our discussions. This accident prompted us to think about the many unknowns and to explore multiple concepts and perspectives as we tried to reconstruct the incident, to understand it, and to analyse it.

The intent of this Discussion is to inspire readers to think about and to discuss how the wildland fire community can improve at the individual, team, and organizational levels. We do not intend that this Discussion provide readers with finished, concrete answers; instead, we mean it to be a springboard that will prompt readers to think about these issues, to try to understand and learn from the accident, and to find opportunities for improving safety and resilience in their organizations.

Rapidly expanding fires in the urban interface are chaotic by nature. Because lives and property are threatened, emergency response organizations rapidly escalate their actions commensurate with the threat. Members of different organizations attempt to work together under extremely challenging and constantly changing circumstances.

The military term “fog and friction” describes challenges these organizations face as they try to communicate, understand, and respond at the right place and time with the right resources. The Oxford Dictionary defines the fog of war as the “confusion caused by the chaos of war or battle.” “Friction” is the process of coping with ambiguous information, fatigue, and unexpected events. It explains the difference between planned and actual events. “Fog and friction” are as common to wildland fire operations as they are to military operations. The Yarnell Hill Fire had plenty of fog and friction, as does any other fire that quadruples in size in a few hours, threatens people and homes, and requires the integration of many different types of air and ground resources.

This Discussion explores issues and questions that the Yarnell Hill Fire brings to the forefront for all wildland fire operations. Various team members and SMEs raised these issues and questions throughout the investigation. Although we list these issues one after another, they are interdependent and may overlap. We included sample questions to facilitate discussion at different levels of the fire organization, from crews and aviation resources to incident management to agency administration to interagency coordination including research and development.

We present these issues as discussion starters, recognizing that this is not a definitive list of issues raised by this fire, and that these are not the only questions to ask about wildland fire organizations and operations. We challenge every wildland fire organization to identify issues and questions raised in this report that resonate within their organizations, and to initiate and facilitate ongoing discussions.

Sensemaking Frame

Sensemaking refers to how people select what seems important to attend to, and how this influences their actions. According to organizational theorist Karl Weick, who popularized the phrase “sensemaking in organizations,” people cannot possibly cope with all of the raw data and information coming at them at a given moment. Instead, what a person pays attention to is a function of identity, past experience, their understanding of their purpose, and other factors. Sensemaking is a very active process whereby people literally “make sense” of the world around them at each moment.

People engage in sensemaking both individually and collectively. In fire, the term situational awareness describes sensemaking: how comprehensively and how accurately are you making sense of the actual fire environment you are working in? Collective sensemaking is about communication: it is about how crews, IMTs, and host agencies determine potential strategies and tactics, and how they convey and update these during planning meetings, briefings, operations, debriefings, and in after action reviews. Effective risk management communication involves more than simply reporting and transmitting messages. It requires developing effective shared meaning together through dialogue and inquiry. This discussion will frequently return to the concept of collective sensemaking and the role of inquiry in that process.

In a rapidly escalating transition fire, all personnel are simultaneously making sense of two environments at once: the rapidly changing fire environment and the changing organizational environment. At around 1600, an outflow boundary was approaching the Yarnell Hill Fire area with high winds that would hit the fire from a new direction. The sensemaking part comes in terms of the interpretation that people make of that change in the environment, of indicators that change might occur, and of the organization’s changing response.

It is far easier for us to know how we would make sense of the situation in hindsight than it is to know how the Granite Mountain IHC made sense of it. We know that the Granite Mountain IHC was actively making sense of their situation, but we also know that their sensemaking and that of others on the Yarnell Hill Fire did not prevent this tragedy. Because other wildland firefighters have similar training, knowledge, and experience to the Granite Mountain IHC, it is likely that others could “make sense” in a similar manner and suffer a similar outcome. The lessons of the Yarnell Hill Fire are not found in second guessing crew actions in hindsight but in understanding through foresight how things may have made sense at the time. Issues worth discussing for the safety of firefighters on future fires include situational awareness, fire line safety, communications, and incident organization.

Situational Awareness

Wildland fire training emphasizes the importance of situational awareness, or comprehensively and accurately perceiving the environment. It is not possible to “lose” situational awareness except by falling asleep or being knocked unconscious. The important questions are “What are people paying attention to and why?” And, “What are people not paying attention to and why?” For the second question, although it is easy to see in hindsight those things that turned out to be important, it is important not to engage in the counterfactual by assuming a reality that did not exist for the crew. It is better to ask, ”Why might it have made sense to focus on or not to focus on those things at the time?” because others may find themselves in the same situation in the future.

We do not have the benefit of asking questions of the people whose situational awareness we are trying to understand. Nevertheless, using the information available, coupled with the Team’s and the SMEs’ understanding of wildland fire culture, we developed two conclusions that may point to the focus of the Granite Mountain IHC:

  • The Granite Mountain IHC left the lunch spot and travelled southeast on the two-track road near the ridge top. Then, they descended from the two-track road and took the most direct route towards Boulder Springs Ranch. We believe the crew was attempting to reposition so they could reengage.
  • The Granite Mountain IHC had been watching the active fire burn away from their position all day but their observations did not lead them to anticipate the approaching outflow boundary or the accompanying significant fire behaviour changes. These changes included a doubling of fire intensity and flame lengths, a second 90-degree directional change, and a dramatically accelerated rate of spread.

The wildland fire community recognizes that hotshots are capable of handling difficult assignments. One Team member identified hotshots as “engagement experts,” known to be persistent, flexible, and improvisational. This makes them valuable on many types of fires, including transition fires. As the day developed, action moved to the north end of the fire. With the fire reaching trigger points near Yarnell and with evacuations beginning, firefighters probably realized the time it would take to evacuate a town of that size. Although we will never know for sure, we considered how the Granite Mountain IHC might have reasoned: If they stay in the black, they do no good. If they move, they might do some good even if they do not know what that good will look like. They think they can move without it being especially risky.

We have no indication that Operations or anyone else asked the Granite Mountain IHC to move to a new location but we assume they decided this on their own, believing they could reengage and help defend Yarnell. A culture of engagement and a bias for action is part of wildland firefighter identity and a factor in their success, and in this case, a bias for engagement may have prompted them to move.

What were they not focused on? Using available information coupled with understanding of wildland fire culture, we reached the following conclusion:

  • In retrospect, the importance of the 1526 weather update is clear. However, the update appears to have carried less relevance in the crew’s decision-making process, perhaps due to the wind shift (starting at about 1550) that preceded the outflow boundary, or perhaps because of the time it took the outflow boundary to reach the south end of the fire (at 1630). It is possible that they may have interpreted the early wind shift as the anticipated wind event.

The outflow boundary update that the NWS communicated to FBAN at 1526 seems to have been relayed efficiently throughout the incident organization. FBAN radioed the update to Operations, and Operations checked in with the crews to make sure they received the update. The Granite Mountain IHC affirmed and passed along the information to their lookout. We considered why this weather update might have carried less relevance with the Granite Mountain IHC, leading to discussions of desensitization, false alarms, and aging of information.

People in the desert southwest may become desensitized to high temperatures and low relative humidity during certain times of year. As two SMEs figured, crews in the area likely received messages over the preceding two months similarly predicting conditions such as hot weather, dry fuels, and thunderstorms. In other parts of the country, these kinds of predictions are rare; when they do occur, they constitute “strong signals.” Like car alarms in an urban neighbourhood, repetition of strong signals resets the cognitive baseline for what is “normal.” People desensitized after repeated warnings start to rely on other cues to identify new and relevant conditions.

There is also danger that a firefighter may become desensitized to extreme fire behaviour, based on an old mental model that extreme fire behaviour is rare. One SME said, “The unusual is now usual – the scale of fires today is extreme. That’s what’s normal now.” Another said, “This fire went from wildland to WUI (Wildland-Urban Interface) within a burn period. This is part of the new reality. The new normal is extreme fire behaviour.”

Consider the role of false alarms. If weather conditions described in one update do not occur, or occur at a diminished level as happened in this case, what will be the level of confidence in the next weather update? Does that decrease the confidence in future weather updates originating from the same source? Although we will never know the answer to this question, it is worth asking: In the absence of observed cloud and column conditions, to what extent did the Granite Mountain IHC think the weather in the second update was not going to materialize because weather in the first update occurred on a diminished level?

Regarding aging of information, people process information based on a variety of factors including perceived timeliness, reliability, and observations. Weather personnel often issue forecasts covering a set timeframe. If the forecaster considers the update reliable for a finite period, the recipient may draw conclusions about its window of relevance. The older a weather forecast or update is, the lower the receiver’s confidence in the update. If predicted conditions have not materialized in an otherwise dynamic atmosphere, this could further decrease a firefighter’s confidence in aging weather updates.

That the original forecast required updating twice during the afternoon of June 30 indicates how rapidly conditions were changing on this fire. As an update ages, firefighters might base their fire behaviour estimates on their observations more than on other inputs.

While we do not know exactly what time the Granite Mountain IHC headed southeast from their fire line, it is clear that they were in a place where they observed fire behaviour for at least 30 minutes after receiving the last update. We also know the crew would lose awareness of the fire’s location and rate of spread for a short time when they descended from the ridge. We believe it is worth considering that the Granite Mountain IHC might have assumed that the transitioning wind shift that caused their lookout to move was the strong wind shift that the weather update anticipated. It is also possible that they discounted the update because too much time had passed.

While discussing desensitization, false alarms, and aging of information, the SMEs posed the following questions for consideration and discussion:

  • How long does a forecast keep you on the edge of your seat?
  • How long do weather bulletins remain fresh in your mind?
  • How long do you wait before you decide a forecasted condition is not going to materialize?
  • How does this make you think about planning for the worst-case scenario?

The Team and the SMEs also discussed what might improve the signal detection capability of individual resources. Could moving conversations from simple reports to inquiries provide an opportunity for collective sensemaking about the meaning of a weather update? All day, firefighters get information about what might happen. Could a crew discuss what would happen if the weather were to materialize? Or, could Operations ask the crews, “We’ve got this weather coming. What’s your plan? What have you been doing and what do you need?”

Cell phones and iPads are not available to all fire personnel in many firefighting organizations. Some incidents also prohibit personal use of such items except on designated breaks. Should some crews, working under certain conditions, be able to access their own weather intelligence to increase the accuracy and timeliness of information that affects their own safety? Acknowledging that there are trade-offs involved in introducing new technologies, this question merits consideration.

We developed the following additional questions for discussion by various fire resources:

Some Questions for Ground Crews and Aviation Resources

  • As a way to test your engagement and action orientation, how might your unit react to the instructions, “don’t just do something, sit there.”
  • When working in day after day of extremely hot weather with low humidity, when thunderstorms are predicted, what do you do to stay alert to changing conditions?
  • When weather updates come in over the radio, what kinds of conversations typically take place to process the information as a group?
  • What kinds of signals in the fire environment make the hair on the back of your neck stand up, and why?
  • When was the last time you were surprised by fire behaviour? What clues, in hindsight, were there to help you see that things were changing more rapidly than you knew?

Some Questions for Incident Managers

  • How do you as a manager transfer your and your staff’s thoughts about the special concerns on this fire to the incoming resources?
  • How do you ensure that all resources assigned to your fire receive and understand weather forecasts and updates?

Some Questions for Agency Managers/Interagency Coordinators

  • In situations where the system is unavoidably dependent on an individual resource, what could we do differently to give firefighters on the ground better information to make decisions?
  • What are the benefits and drawbacks of equipping firefighters with handheld technology intended to increase their situational awareness?

Some Questions for Researchers in Human Factors, Organizations, Fire Behaviour, etc

  • How might emotion trigger “automatic” movement, and how can we interrupt this circuit to engage in reflection that leads to more purposeful action?
  • How might firefighters stay alert to changing conditions when message repetition might encourage desensitization (e.g., working in day after day of extremely hot weather and low humidity with thunderstorms predicted)?

Fire line Safety

The Yarnell Hill Fire also points to issues of fire line safety and risk management while on the move. As noted in the Conclusions section:

  • We found no indication that the Granite Mountain IHC doubted that the black was a valid safety zone, or that they moved towards the Boulder Springs Ranch because they feared for their safety if they stayed in the black.
  • The Granite Mountain IHC did not perceive excessive risk in repositioning to Boulder Springs Ranch.

The intent of discussing fire line safety is not to second-guess the crew’s actions on this incident. Rather, the intent is to point to issues regarding fire line safety that this fire leads us to contemplate regarding all fires.

Safety Zones

The Yarnell Hill Fire calls attention to how the phrase “good black” conveys a measure of one crew’s safety to other resources on the fire. In generally chaotic conditions when a fire is expanding rapidly and organizational complexity is increasing, if an experienced crew like an IHC reports that they are “good” and “in the black,” this relieves Operations from having to attend to another detail on an already busy fire. Anyone hearing such a report from a crew would automatically consider that they are about as safe as it is possible to be in such a situation.

In terms of collective sensemaking and inquiry, one aspect of the crew’s communication stands out. The crew communicated that they were moving along their escape route to a safety zone, yet others on the fire believed their location was in a safety zone (the black). Personnel in a safety zone do not need an escape route. Others on the fire inquired with the Granite Mountain IHC about their status and location, yet that inquiry did not lead to mutually accurate understanding.

One communication exchange illustrates how inquiry might lead to collective reassessment. At about 1600 after hearing about “a crew in a safety zone,” the ASM asked if they needed to call a time out. Operations replied that it was the Granite Mountain IHC and that they were safe. Then, sometime later, DIVS A followed up and said they were traveling along their escape route to a safety zone. The ASM’s question about pausing operations is a good example of one resource updating situational awareness about another resource’s location and relative safety, and even recommending an action that could have helped update everyone’s collective sense of the crew’s status and location. Nevertheless, we might consider how Operations and the Granite Mountain IHC’s subsequent radio conversation may have led many on the fire to mentally file the crew back in the “safe” category.

The Yarnell Hill Fire also calls attention to firefighter sensemaking about the “green,” including whether there is such a thing as “good green.” Firefighters know that being “in the green” on a fire, surrounded by unburned vegetation, can be unsafe. During indirect attack, there is always unburned fuel between the firefighter and the fire. But consider that prior to 1604 on this incident, all ground personnel assigned to the Yarnell Hill Fire except for the Granite Mountain IHC were in the green. Only the Granite Mountain crew was in the black, normally considered a “safe” location, potentially leading others on the fire to have greater concern for people working in the green than for the Granite Mountain IHC.

We considered what provided for the safety of all the other personnel: possibly distance from the fire, mobility, ability to make sense of what the fire was doing and react appropriately, or a combination of these factors. Members of the Blue Ridge IHC moved through the green all day in areas that subsequently burned. It is conjecture, but possible, that their actions may have assisted in saving the life of GM Lookout as well as preventing the Granite Mountain IHC’s trucks from burning. Obviously, decisions to operate in the green are laden with a variety of assessments of the relative risk of doing so.

We will never know for sure, but we wondered whether the Granite Mountain IHC’s decision to hike through the green might have seemed to them to be a decision to operate in the green just like everyone else. For the crew, this would have meant moving away from a safe location at the time of day when the fire would be most active. Decision makers base such decisions on what they know at the time and their assessment of the risks associated with various courses of action.

Regarding their intended destination, the 0700 briefing on June 30 included identification of the Boulder Springs Ranch as a safety zone. SPGS1 (who had just arrived the evening before) described it as “bomb proof,” a label indicating the Ranch was not only a safety zone, but that it seemed to be an especially good safety zone that could withstand extreme fire behaviour.

Because identifying a safety zone requires judgment and is therefore subjective, firefighters often know the actual effectiveness of a safety zone only in retrospect. If a fire never affects a safety zone with firefighters in it, firefighters may never know whether it would have proven safe. In this case, fire progression maps and aerial photos taken after the fact demonstrate that the large Ranch was well prepared to act as a fuel break in the hills above Yarnell. Inside the perimeter of the Ranch, the preparation and arrangement of the buildings proved effective to minimize damage to the home, livestock, and other values on the property. For these reasons, the label “bomb proof safety zone” seems to have accurately described the Boulder Springs Ranch.

Escape Routes

Continuing from this previous point, the Yarnell Hill Fire also prompts us to think about the connections that firefighters make between escape routes and safety zones. As noted above, we believe the Granite Mountain IHC did not perceive their route as overly risky, or they would not have taken it. Wildland firefighters should consider to what extent a strong vote of confidence about the effectiveness of a safety zone might be interpreted as a strong vote of confidence about potential escape routes for getting there. Conversely, is there some implied measure of the safety along an escape route because it leads to a safety zone?

One might view traveling through an escape route to a safety zone as making educated guesses as to the route and anticipated travel speed while running to a specific point. The educated guess is that the crew can reach the safety zone before the fire reaches them. There are many variables involved in this equation but perhaps the most important one is speed. If the fire can travel at a faster rate than the firefighters, they will lose the race. If they can travel faster than the fire, they will win the race. In order for the educated guess to prove out, the firefighters must predict three things with some degree of accuracy: how fast the fire will travel, which direction the fire will travel, and how fast they will travel. It is possible to misestimate all these factors and suffer no consequences, for example if the firefighters misestimate the fire’s direction of travel but it moves away from their position. However, misestimating any of these variables could cause serious trouble and firefighters misestimating them all may pay the ultimate price.

Unfortunately, all three of these variables are difficult to estimate accurately. Estimating how fast a fire will travel, along with its direction of travel, involves making an accurate fire behaviour prediction while also doing a number of other things. Even in hindsight, knowing what the fire actually did and using the best available hardware and software, fire behaviour analysts can still only approximate how a fire behaved in a blow-up situation. Firefighters in the field must use available information and their own sensemaking to estimate both what the fire is likely to do, and what it could do under a worst-case scenario. Estimates of how fast firefighters can travel can also be problematic. A crew can only move as fast as its slowest member, and unexpected barriers in the route of travel such as rocks, thick vegetation or cliffs, can significantly change the amount of time required to cover a piece of ground.

This points to a key dilemma of wildland firefighting: firefighters on a fire are one educated guess away from potential entrapment. Increased mobility increases the tolerable margin of error; decreased mobility decreases it in terms of how fast a person can travel versus how fast the fire can travel. Firefighters on foot are perhaps the most vulnerable, but many firefighters have become entrapped in vehicles as well.

Wildland firefighters often discuss the need to have multiple safety zones; many firefighters also identify multiple escape routes to the same safety zone, if they exist, although this can require extensive scouting. In hindsight, we know that the Granite Mountain IHC might have arrived at the Boulder Springs Ranch if they had stayed on the two-track road, although it is unclear whether the crew knew that, or how long it might have taken to get there. This highlights another problem posed by limited mobility: because the Granite Mountain IHC was on foot, their ability to scout potential escape routes was limited.

Lookouts

Lookouts, Communications, Escape Routes, and Safety Zones (LCES) is an interconnected system approach to fire line safety, so it is difficult to discuss safety zones and escape routes without also addressing lookouts. The Granite Mountain IHC had a designated lookout for most of the day, until the advancing fire threatened the lookout’s location and forced him to withdraw. This points to one paradox of firefighting: Crews post lookouts to increase safety, but there is no guarantee of the lookout’s own safety. The Granite Mountain IHC never took explicit action to replace this lookout after he was forced to withdraw, but it is likely that DIVS A was serving as a lookout for the crew and that the crew was also exercising their own vigilance. In all the photos of the crew at the lunch spot, they appear focused on the active fire.

Ground resources can also use aviation resources to gather information, but crews are cautioned not to rely on them because aircraft move around, have to land to refuel, and are diverted to other missions. When both air tankers and helicopters are conducting operations in the same airspace, the complexity of the air operations and the communications necessary to accomplish this may preclude routine requests by ground resources for situation updates. While aircraft might have provided situation updates to the Granite Mountain IHC during the time they were hiking and prior to their shelter deployment, there is no evidence that communication occurred. In the current system, ground resources would have to request such an update.

We contemplated a key question arising from the Yarnell Hill Fire that also likely applies to many other fires: Is it necessary to post a lookout when a crew is moving? In some situations, it may not be feasible to post a lookout while a crew is moving due to time and distance considerations. This raises the question of whether a crew can perform its own lookout functions without designating and posting a lookout in a separate location. It is likely that the Granite Mountain IHC was very vigilant to their surrounding environment as they hiked southeast along the two-track road. Because the crew was moving quickly between two safety zones, it was likely not feasible to post a lookout from the crew at another location. Given previous events that day, the crew was likely very aware that posting a lookout in such a dynamic fire environment can pose challenges to the lookout’s own safety and may serve only to separate a crewmember from the rest of the crew.

Whether the crew recognized it or not, their decision to go down the hillside from the Descent Point was a decision to sacrifice some of their effectiveness in serving as their own lookouts. Taking a more direct escape route to minimize exposure in the green generally means traveling a shorter distance and potentially reaching the safety zone more quickly. Moving down the slope into the box canyon meant the Granite Mountain IHC would no longer be able to see the fire. We wondered: Is it possible that they relied on the rock outcropping as a barrier to fire spread? But is it also possible that the outcropping blocked their view of the fire? We will never know if the crew understood that this route of travel required that they sacrifice some of their capacity to serve as their own lookouts. We will also never know if they understood the calculated risk involved in traversing the final distance to the Ranch without the level of situational awareness that a different vantage point might have afforded.

With the help of our SMEs, we developed the following questions for discussion by various fire resources regarding fire line safety.

Some Questions for Ground Crews and Aviation Resources

  • How do you and your crew define “good black”? Firefighters identify and discuss good black, but is there also such a thing as “good green”?
  • How do you assess the risk of hiking into a fire in the green? What differs about your risk assessment when hiking out through the green?
  • When others point out a safety zone to you, what questions do you ask about how they assessed the viability of the site and the safety of the route(s) for getting there?
  • Since all escape routes are necessarily “through the green” or through black that is not very “good,” what characteristics make one escape route better than another?
  • When you identify an escape route, do you also discuss trigger conditions that would prompt reassessment?
  • What assumptions do you make about the safety of other resources? How often do you update that information?
  • Other than at a required annual refresher, how many times per season do you and your crew simulate or play out what you would do during an entrapment, including practicing fire shelter deployments?

Some Questions for Incident Managers

  • When resources leave the black, do you think of it as a move from safety to danger? Should a decision to leave the black, or any other area deemed safe, be considered a decision to accept more risk? Under what circumstances does this require notification to others?
  • What are your own judgments about crews falling back and disengaging? How do you convey this on an incident?

Communications

Communications issues on wildland fires are common. Firefighters usually associate communications problems with technology issues, but communication challenges can occur even when radio systems are working well. Two team conclusions were:

  • Radio communications were challenging throughout the incident. Some radios were not programmed with appropriate tone guards. Crews identified the problem, engaged in troubleshooting, and developed workarounds so they could communicate using their radios. Radio traffic was heavy during critical times on the fire.
  • Although much communication occurred among crews throughout the day, few people understood Granite Mountain’s intentions, movements, and location, once they left the black. The Team believes this is due to brief, informal, and vague radio transmissions and talk rounds that can occur during wildland fire communications. Based on radio conversations, Operations and other resources had concluded that the Granite Mountain IHC was located in the black, near the ridge top where they had started that morning. This resulted in confusion about the crew’s actual location at the time of search and rescue.

The benefit of radio communication is that when the system is working properly everyone can hear the same messages from multiple sources on the fire. Firefighters commonly experience radio issues on fires but they are used to adapting to problems and developing workarounds. There were problems on the Yarnell Hill Fire with tone guards and “dead spots,” or areas where handheld radios could not reach repeaters to transmit messages to other ground resources. To overcome these issues, individuals and crews used time to troubleshoot problems, to reset or reconfigure radios, or to travel to face-to-face meetings. Physical workarounds included lending or sharing radios among crews and using cell phones (text messages and voice calls) as alternatives to radios. Although these alternatives demonstrate initiative and creativity and may permit lengthier, more detailed conversations and interactions, it is worth considering the potential trade-offs.

Moreover, most firefighters constantly monitor more than one radio channel. They might simultaneously scan three or four channels including the tactical channel for their Division or Group, the Command frequency for the incident, the air-to-ground frequency, and their intra-crew channel. They always designate one channel as a “priority” so that traffic on it will pre-empt traffic on other scanned channels. This makes it is easy to miss partial or entire transmissions when all channels are busy, even on a channel that the radio is set to monitor.

Rapidly emerging initial or extended attack actions are hectic, resulting in increased radio communication, competition for radio time, and limited time to convey information. As one Team member said, firefighters have “normalized” overloaded radio traffic when operational tempo increases. During busy times, this can cause a firefighter to feel an increased reluctance to get on the radio. Radio traffic was heavy on the afternoon of June 30 on the Yarnell Hill Fire. The situation compounded in one operational period as the fire moved from wildland fire-focused activity to a fire with concurrent actions in multiple wildland and urban areas. Ground crews, aviation resources, and structure protection resources were all vying for radio airtime. We believe it is worth considering whether early radio problems and heavy radio traffic caused the Granite Mountain IHC to hesitate to add to the radio traffic early in the day, when they might have otherwise relayed periodic updates of their crew’s location.

Firefighters learn in their training to speak in “clear text” (plain English, no codes) during radio conversations and to keep their conversations brief. The reason for clear text is that some users may not understand numeric codes, and asking for clarifications uses airtime. Clear text works well for sharing straightforward information and it works very well when both parties are of a common understanding. However, clear text has limitations when reaching a mutual understanding requires longer interactions. Shortened messages may not fully explain a point, situation, or request; may send an incomplete message; or may lead the recipient to act upon something they thought they heard. Recipients may carry on with a misunderstood message or assumption rather than asking for clarification. While these misunderstandings likely had no direct impact on the Granite Mountain IHC’s fire shelter deployment, confusion over the location of the break between Divisions Alpha and Zulu, and over SEAT drops on the Granite Mountain IHC’s burnout that morning are two examples of communication disconnects that emerged during reliance on radio communications.

Some people knew the Granite Mountain IHC was on the move, but only the Granite Mountain IHC knew their location and intended destination. An early miscommunication caused confusion later about which two-track road the crew was on and which ranch they were heading towards. When DIVS A told BR Supt that they were “picking our way through the black” toward the road at the bottom then to a ranch, BR Supt thought Granite Mountain was going out the two-track road to the northeast. After leaving the lunch spot, the Granite Mountain IHC travelled in a different direction toward the Boulder Springs Ranch. It is clear now, in hindsight, that the message BR Supt perceived was not the message DIVS A believed he relayed. All day, the Granite Mountain IHC and BR Supt were in frequent contact with one another and mutually believed they had a clear understanding between them, so neither noted this miscommunication at the time of occurrence. Unbeknownst to them, they were not communicating in the sense of actually understanding one another’s movements.

The Incident Response Pocket Guide sets out five communication responsibilities for wildland firefighting, including 1) Brief others as needed; 2) Debrief your actions; 3) Communicate hazards to others; 4) Acknowledge messages; and 5) Ask if you don’t know. These responsibilities emphasize the importance of communication content in the overall system, reminding firefighters to share information that others might need and acknowledge information communicated to them. Point 5 reminds firefighters to make an effort to clarify things they do not understand. The guide does not currently capture the need for firefighters to inquire to ensure they have the right understanding. We considered that perhaps point 5 be phrased “Ask questions until you know and you are sure you understand.”

Some Questions for Ground Crews and Aviation Resources

  • When you communicate your assessment of your location, how much detail do you provide to also help others “picture” what you are seeing?
  • Can people who are in other locations help you to assess the risks where you are?

Some Questions for Incident Managers

  • In what ways do WUI fires make already-challenging radio issues more complex?
  • When checking in with your resources, what is the best way to ask for an update to ensure that you receive a report that allows you to visualize what the resources see and know what factors they consider in their assessment?

Incident Organization

Rapidly expanding fires in the WUI are chaotic by nature. Because lives and property are threatened, emergency response organizations escalate their response commensurate with the escalating threat. Our conclusions included:

  • The fire’s complexity increased in a very short time, challenging all firefighting resources to keep pace with the rapidly expanding incident. As complexity dramatically increased starting Saturday evening, fire management went through multiple transitions from a Type 4 through a Type 1 incident in fewer than 20 hours.

Like many high complexity wildland fires, the Yarnell Hill Fire passed through a series of “modes.”12 In this case, we considered Mode 1 to be the initial sizing up and engagement of the fire, characterized first by direct attack at the heel and flanks of the fire. A transition to Mode 2 happened when firefighters largely abandoned efforts to contain the fire through direct perimeter control and shifted to point protection around structures. Mode 2 centred on taking actions to protect values at risk, while anticipating the fire spread and recognizing the inadequacy of attempts at perimeter control. Mode 3 was largely about survival of incident responders and the public, about disengaging and running away from the fire and helping civilians evacuate. Later, firefighters were able to shift back into Modes 1 and 2.

12 The Team numbered these modes for illustration only. Different fires may go through different modes.

The Yarnell Hill Fire incident management organization had a dynamic organizational strategy, scaling up within 20 hours with two fronts of structure protection. In a rapidly escalating fire environment, firefighters simultaneously try to make sense of what the fire is doing and how the incident organization is changing. This prompted us to consider whether some fire personnel might remain focused on one mode while others have moved on to another mode, and we considered the potential implications of this disconnect.

This characterization helps illustrate collective sensemaking because it highlights the importance of all resources mutually understanding which mode the fire is in and whether they are all making the same sense of being in that mode. This ties into the operational strategy, and how different people working on the fire understand and interpret that strategy.

Response to transition fires normally starts with in-briefings and good orientation to individual and collective sensemaking. Mode 1 response is strategic leading to tactical, formulating a plan and initial implementation. In Mode 1, the initial briefings and assignments show that the Yarnell Hill incident started the day well organized. However, as complexity increases and the situation changes, the initial response plan requires adjustment, or actions revert to individual sensemaking. Mode 2 is more of a tactical orientation and on this fire involved a shift from a wildland and WUI focus to primarily a WUI/point protection focus. Mode 3 involves immediate, reactive actions and movement to survival: moving, disengaging, retreating, evacuating.

The Granite Mountain IHC’s initial actions at the heel of the fire could be interpreted as the crew being engaged in Mode 1. Moving toward the Ranch would be a movement toward Mode 2. In the analysis, the Team noted that the Granite Mountain IHC likely thought it was safe to move to the Ranch, first on the two-track road and later by dropping down at the Descent Point. As a consequence, the Granite Mountain IHC would not have transitioned into Mode 3 until after reaching the deployment site, when they had little time and limited options.

We wondered: How might personnel ensure that they are individually and collectively oriented to the same Mode? In discussing this issue, we contemplated several questions:

  • How might formally giving up on “Mode 1” make it more likely for a crew to disengage?
  • Could formally acknowledging a shift in mode on the fire help individual crews reassess their place in the collective effort?
  • But is it also possible that recognizing a shift to “Mode 2” might pull people toward the structures?
  • Signal strength is a potential factor in an individual’s interpretation of mode. During the worst time of day for burn overs to occur, if conditions are all lining up, is it possible to deliver a strong signal about mode that prompts everyone to change their behaviours, including communication behaviours?

We developed additional questions that others should consider.

Some Questions for Incident Managers

  • When incident complexity and operational tempo escalate rapidly, what are some of the things you can do to minimize resultant confusion? Who can help you in managing all the incoming information?
  • Do you think it is important to update resources on your fire as to what strategic and tactical mode the incident organization is in? What ideas do you have on how best to accomplish such an update?
  • What are some ways that you can encourage collective sensemaking among the resources assigned to you?

Some Questions for Agency Managers/Interagency Coordinators

  • Organizational culture on transitioning fires tends to make sense of things individually or in small groups or individual crews. How can we increase our ability to make sense of and share what is happening at the incident scale?

Some Questions for Researchers in Human Factors, Organizations, Fire Behaviour, etc

  • What are some communication strategies to cultivate inquiry that firefighters and incident managers can use to create more effective collective sensemaking?

Improving Resilience

On rapidly escalating transition fires like the Yarnell Hill Fire, complexity can outpace organizational attempts to respond. In other words, collective sensemaking can fall behind the curve, particularly when a fire simultaneously affects multiple fronts. Collective sensemaking often starts strong early in the operational period with face-to-face briefings and agreement on plans. The Yarnell Hill Fire prompts us to consider how, as complexity increases, as resources are adapting to the fire environment, and as operations slide into and out of different modes, it is necessary to develop a new kind of collective plan. This evolution could occur several times in a single day.

Time pressures, radio limitations, crew familiarity, and other issues can hamper efforts to check in and contribute to sensemaking. For example, the need for communication from multiple resources that are focused on direct actions at the same time crowds the airwaves. As a result, the organization designed during initial attack tends to erode and revert to sensemaking by individual resources until announcement of the next operational period’s strategies, tactics, and assignments. This means that collective sensemaking is particularly necessary and difficult at the very moments when autonomous crews most need it: during high tempo situations with escalating complexity.

We considered whether routine use of “inquiry” in communication might increase ground crew capabilities by connecting them to information and expertise available in the larger system. If so, the wildland fire community should consider what it could do to cultivate inquiry, to move wildland fire communications beyond “reporting” and “debriefing,” and to help everyone make sense together, for everyone’s benefit.

Charge to the Wildland Fire Community

This discussion identifies some issues that the Yarnell Hill Fire prompts the wildland fire community to consider. We introduced issues and questions that readers may find useful as they try to understand this accident and learn from it, but this is only a start. Because sensemaking is ongoing and social, we challenge every wildland fire organization to identify and discuss issues and questions raised in this report that resonate within their organizations, to continue the ongoing process of sensemaking.

Conclusions

The Team developed these conclusions through deliberation. The process considered information from a number of sources, including accounts from personnel on the fire, records and logs, physical evidence, knowledge of the firefighting culture, Team observations, and SME sessions.

  • The Granite Mountain IHC was a fully qualified, staffed, and trained hotshot crew. They were current with the required training and met work/rest guidelines. The crew followed all standards and guidelines as stated in the Standards for Interagency Hotshot Crew Operations and the Arizona State Forestry Division’s Standard Operational Guideline 804.
  • The Yarnell Hill area had not experienced wildfire in over 45 years. It was primed to burn because of extreme drought, decadent chaparral, and above average cured grass loadings.
  • Although Yavapai County had a Community Wildfire Protection Plan, many structures were not defendable by firefighters responding to the Yarnell Hill Fire. The fire destroyed over one hundred structures.
  • Radio communications were challenging throughout the incident. Some radios were not programmed with appropriate tone guards. Crews identified the problem, engaged in troubleshooting, and developed workarounds so they could communicate using their radios. Radio traffic was heavy during critical times on the fire.
  • The fire’s complexity increased in a very short time, challenging all firefighting resources to keep pace with the rapidly expanding incident. As complexity dramatically increased starting Saturday evening, fire management went through multiple transitions from a Type 4 through a Type 1 incident in fewer than 20 hours.
  • The Granite Mountain IHC had been watching the active fire burn away from their position all day but their observations did not lead them to anticipate the approaching outflow boundary or the accompanying significant fire behaviour changes. These changes included a doubling of fire intensity and flame lengths, a second 90-degree directional change, and a dramatically accelerated rate of spread.
  • The Granite Mountain IHC left the lunch spot and travelled southeast on the two-track road near the ridge top. Then, they descended from the two-track road and took the most direct route towards Boulder Springs Ranch. The Team believes the crew was attempting to reposition so they could reengage.
  • The Granite Mountain IHC did not perceive excessive risk in repositioning to Boulder Springs Ranch.
  • The Team found no indication that the Granite Mountain IHC doubted the black was a valid safety zone, or that they moved towards the Boulder Springs Ranch because they feared for their safety if they stayed in the black.
  • Although much communication occurred among crews throughout the day, few people understood Granite Mountain’s intentions, movements, and location, once they left the black. The Team believes this is due to brief, informal, and vague radio transmissions and talk rounds that can occur during wildland fire communications. Based on radio conversations, Operations and other resources had concluded the Granite Mountain IHC was located in the black, near the ridge top where they had started that morning. This resulted in confusion about the crew’s actual location at the time of search and rescue.
  • In retrospect, the importance of the 1526 weather update is clear. However, the update appears to have carried less relevance in the crew’s decision-making process, perhaps due to the wind shift (starting at about 1550) that preceded the outflow boundary, or perhaps because of the time it took the outflow boundary to reach the south end of the fire (at 1630). It is possible they may have interpreted the early wind shift as the anticipated wind event.
  • An Air Attack and/or an ASM provided aerial supervision coverage throughout the day including at the time of the accident.
  • The ASM working the fire was very busy fulfilling lead plane duties, which limited their ability to perform full Air Attack responsibilities over the fire at the same time.
  • During some limited times, aircraft were not available due to adverse weather and refuelling needs.
  • At the time of the shelter deployment, a VLAT was on station over the fire waiting to drop retardant as soon as the crew’s location was determined.
  • The judgments and decisions of the incident management organizations managing this fire were reasonable. Firefighters performed within their scope of duty, as defined by their respective organizations. The Team found no indication of negligence, reckless actions, or violations of policy or protocol.

Recommendations

1. The Team recommends that the State of Arizona review and possibly update its approach to mitigating wildfire threat to Arizona life and property. This could be modelled after communities such as Prescott, AZ; Santa Fe, NM; or communities in other states. This process could be a cooperative effort to reduce hazardous fuels and improve overall suppression efforts for communities that are at a high risk from wildfire.

2. The Team recommends that the State of Arizona review the state-wide wildfire communications plan and program, as well as similar programs and plans in other states, for possible improvements.

3. The Team recommends that the State of Arizona work cooperatively with its fire co-operators to develop a wildland fire staff ride for the Yarnell Hill Fire incident. The staff ride is a process of conveying the lessons learned from this incident for future fire leaders.

4. The Team recommends that the State of Arizona request the National Wildfire Coordination Group (NWCG) to review current technology that could increase resource tracking, communications, real time weather, etc. For example, this may include GPS units and weather applications.

5. The Team recommends that the State of Arizona request the NWCG and/or Wildland Fire Leadership Council (WFLC) to charter a team of interagency wildland fire and human factors experts to conduct further analysis of this event and the wildland fire communications environment.

6. The Team recommends that the State of Arizona request the NWCG to 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.

7. The Team recommends that the State of Arizona request the WFLC/NWCG to develop a brief technical tip for fire supervisors/agency administrators on the effective use of VLATs. These are new, emerging fire suppression tools that the ground-based fire supervisors may be utilizing regularly in the future.

--end--

Further information hoping to be identified and still to be located.

United States Fire Department trade union/representative body/s summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

National Institute for occupational Safety and Health (NIOSH) summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

Underwriters Laboratory (UL) summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

National Institute for Standards and Technology (NIST) summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

Fire Protection Association (FPA) summary of main findings, conclusions, key lessons & recommendations

No information identified to date and/or still to be located.

Wildland Fire Associates report summary of main discussion and conclusions

Taken from; O’Brien, D., Anderson, E., Hicks, B., Larsen, D., & Schulte, D. (2013). Granite Mountain IHC entrapment and burn over investigation; Yarnell Hill Fire – June 30 2013, prepared for Arizona Division of Occupational Safety & Health. [pdf]. Wildland Fire Associates.

--subaccordion--

DISCUSSION

The following discussion identifies areas of concern where difficulties existed on the Yarnell Hill Fire. Our conclusions are based upon professional experience and the industry standards for wildland fire that include the 10 Standard Firefighting Orders, 18 Watch Out Situations, Lookouts Communication Escape Routes and Safety Zones (LCES), 2013 Interagency Standards for Fire and Aviation Operations (Red Book), Incident Response Pocket Guide (IRPG), and the Wildland Fire Incident Management Field Guide (PMS 210).

The analysis of the decision-making process has led to the identification of four primary areas of concern:

· ARIZONA STATE FORESTRY DIVISION

· YARNELL HILL FIRE INCIDENT MANAGEMENT

· DEPARTURE FROM STANDARD PRACTICES

· FATIGUE

Below we highlight the specific conditions and events that support why these four areas of concern resulted in the burn over.

PRE-EXISTING CONDITIONS IN YARNELL

Yarnell, Arizona is a classic example of the wildland urban interface (WUI) situation. The structures within town are located in chaparral scrubland that had not burned in at least 40 years. The Yavapai Communities Wildfire Protection Plan approved in 2004 provides direction for hazard fuel removal on lands including the Yarnell area. The Bureau of Land Management has been working to reduce hazard fuels in Peeples Valley and Yarnell. Between 2005 and 2011, $169,000 was spent in the Yarnell area to clear 375 acres, and $27,500 was spent in 2007 to clear 40 acres near Peeples Valley17.

Based upon the Arizona State Forestry Division (ASFD) 2013 Season Outlook released in March 28, 2013, the state of Arizona was in a drought situation. Fine fuel moistures were approaching single digits as early as late March and the Energy Release Component was above normal and trending upward. Yavapai County had live green fuel moistures in chaparral that were below normal. Yarnell was mentioned as having chaparral with below average live fuel moisture and older chaparral stands with high dead-to-live ratio that may “prove resistant to control efforts due to the low live fuel moistures.” The dry winter and late spring precipitation had led to a delay in new seasonal fine fuel growth. The conclusion of the report is that the Yarnell area had high fire potential18.

17 AZCentral.com, Brush Clearing Saved Homes, July 17, 2013.

18 2013 Arizona Fire Season Outlook, page 10.

Arizona State Forestry Division

The ASFD is responsible for fire suppression operations on 22 million acres of State Trust land and private property located outside of incorporated communities19. ASFD has employees that work on three districts and in the state office. Each District has a District Forester who, in the case of the Phoenix District, also fills the role of FMO. The Phoenix District Office maintains field offices in Prescott and Yuma, and has three fire crew coordinators throughout the District. Yarnell is part of the Phoenix District.

ASFD responds to an average of 476 wildfires annually (based upon a 10 year average, with 2006 having substantially more fires than the other nine years), which burn an average of 24,000 acres per year. ASFD supports twelve 20-person Arizona DOC fire crews, supplemental summer preparedness resources, and has cooperative agreements with 250 fire departments and federal agencies20.

ASFD had the authority for the suppression of Yarnell Hill Fire.

INCIDENT ACTION PLAN

The Arizona Revised Statute 37-623 Section on Wildfire Suppression Strategies states that wildfire suppression operations shall be conducted to “minimize both suppression costs and resource losses, consistent with resource values to be protected and shall consider fire behaviour, the availability of suppression resources, the values of the natural resources and property at risk, and potential cost of suppression.”

In order to meet the intent of this statute, a comprehensive and coherent Incident Action Plan (IAP) should have been articulated. An IAP “contains objectives reflecting the overall incident strategy and specific tactical actions and supporting information for the next operational period. The plan may be oral or written.”21

Based upon incident documents and interviews, we believe that the ICT4 worked hard to develop and convey the incident strategy and tactics to resources as they arrived on the fire. However, given the complexity of the rapidly evolving fire situation, the ICT4 did not adequately brief incoming resources on June 29 or provide a written IAP for the incoming IMT2 on June 30.

Based upon our interview with ICT2, we have concluded that when ICT2 arrived at the Incident Command Post (ICP) he observed an obviously fatigued ICT4. Realizing that the fire situation was very dynamic and intensifying, ICT2 took over the fire despite the fact that certain key members of the team had not yet arrived. ICT2 provided the 0930 briefing to resources that had arrived at the ICP. Some resources were not at the 0930 briefing because they had already been assigned and working on the fire line. Based upon incident documents and interviews, ICT2 was working in a diligent and professional manner, although the situation was deteriorating.

19 Arizona State Forestry Division website (http://www.azsf.az.gov/, accessed on November 3, 2013).

20 Arizona House of Representatives, Committee on Agriculture and Water, Minutes of Meeting, March 3, 2011.

21 NWCG Glossary of Wildland Fire Terminology.

The ultimate result was that ICT4 and ICT2 failed to convey a coherent strategic plan for suppressing the fire that was uniformly understood by ground and air resources from initial attack through the entrapment and burn over. An IAP with formalized strategies and tactics known to all resources assigned to the Yarnell Hill Fire, starting with initial attack, would have decreased the amount of confusion and miscommunication that occurred.

EFSA AND AGENCY ADMINISTRATOR BRIEFING

The ASFD failed to give clear management direction to the incoming IMT2 because they had not completed the Escaped Fire Situation Analysis (EFSA) required by their own policy for fires escaping initial attack22. A Complexity Analysis was not completed until June 30, after the IMT2 had taken over the fire. ASFD published their Wildland Fire Situation Analysis (WFSA) decision on July 4. The ICT4, acting as Agency Administrator, provided the briefing for IMT2.

Yarnell Hill Fire Incident Management

The following is a discussion of the decision points starting with the ignition of the Yarnell Hill Fire and ending with the entrapment and burn over. We examined the decisions that were made through the lens of the outcome and, where appropriate, suggest where different decisions could have been made based upon current policy and guidelines.

QUALIFICATIONS

During our investigation, we reviewed the Incident Qualifications for significant personnel assigned to the Yarnell Hill Fire. We found that everyone was qualified for the positions in which they were serving. We also examined the Type 1 Certification for the Granite Mountain Interagency Hotshot Crew (GMIHC), along with the training records for each firefighter. We have determined that GMIHC met the Type 1 Crew qualification. Each of the firefighters met the basic qualifications for the positions they held on GMIHC.

INITIAL ATTACK

At approximately 1700 on June 28, the Yarnell Hill Fire was started by lightning. The Air Attack Group Supervisor (ATGS) for the Doce Fire was requested to fly over the Yarnell area and size up the fire. The ATGS said that the fire was in a boulder field with no vehicle access. The size-up also included that the fire was less than half an acre, only burning on one side, and did not appear to pose a threat to structures or people. A conference with the Yarnell Fire Department (YFD) personnel confirmed that the fire was inaccessible by road. The ICT4 decided not to initially attack the fire until the following morning based upon the ATGS size-up, input from YFD, and concerns about firefighter safety. The decision to defer initial attack was made without input from the District Office because the AFMO was the duty officer for the day. The ICT4 planned for initial attack the following morning.

22 Arizona State Forestry Division, Policies and Procedures, September 23, 2008. Note: The EFSA has been replaced by the Wildland Fire Decision Support System (WFDSS).

Initial attack on the morning of June 29 was delayed because a helicopter large enough to move the six-person crew safely to the site and evacuate them if needed was not available. People with local area expertise did not tell him about local trails and roads that could be used to hike in to the fire.

By mid-afternoon on June 29, the fire jumped over the two-track trail. ICT4 started ordering additional resources. The initial attack forces had clearly failed to “stop the fire and put it out in a manner consistent with firefighter and public safety and values to be protected.”23 ASFD did not declare that the fire had escaped initial attack. Had they made that declaration, the decisions from that moment forward would have been proactive, rather than reactive. Based upon the Wildland Fire Incident Management Guide (PMS 210), the ICT4 would have completed a complexity analysis, implemented risk management protocols from Incident Response Pocket Guide (IRPG), determined and documented incident objectives, and reviewed the Extended Attack Safety Checklist. Based upon interviews and incident documents, we have found no evidence that this occurred.

STRATEGIES & TACTICS

The initial strategy for the Yarnell Hill Fire was full suppression. Direct attack was made on June 29 with Single Engine Air Tankers (SEAT) dropping fire retardant by mid-morning. A Hotshot Crew arrived at the fire around 1100. The Hotshot Crew continued direct attack throughout the day. Direct attack by the SEATs was discontinued in the early afternoon. The full suppression strategy was modified to include point protection for Peeples Valley and Yarnell by the evening of June 29.

By the morning of June 30, tactics still included direct attack on the ridge at the south end of the fire by GMIHC, with aviation resources being used to slow the spread of the fire on the north and east flanks. Structure protection tactics and trigger points were established in Peeples Valley and Yarnell.

Planning OSC and DIV A decided that GMIHC would establish an anchor point using the burned area and flank using direct attack whenever possible. GMIHC was to join their line with the dozer line at the bottom of the hill and then work to the north in an effort to keep the fire out of Yarnell and Peeples Valley.

The S-336 Tactical Decision Making in Wildland Fire Course textbook contains a section on appropriate tactics in the Southwest which says that direct attack works well on small fires. However, when planning for larger fires “a number of items need to be considered before deciding on strategy; topography, fire behaviour and intensity, rate of spread, availability of needed resources, logistics in moving and supplying firefighters and of course, probability of success…Indirect attack is also used, especially in lower elevation fuel types. Acreage is often sacrificed for lower suppression costs and higher probability of success. Direct attack on a fast moving desert or brush fire is seldom successful. Using natural barriers and roads when burning out is very common...”

Based upon incident documents and interviews, we found no evidence that a risk assessment for the strategies and tactics were examined. We also could not find evidence that the probability of success for the chosen strategy and tactics was examined.

23 NWCG Glossary Definition of Initial Attack.

An alternative to the implemented tactics could have been to establish the anchor point as they did, burn out along the two-track trail that existed at the top of the ridge, and then burn out along the jeep trail that they used to hiked in, ending at the old grader. This tactic would entail indirect attack with burnout, and would have provided a secure line from the ridgetop to the valley floor. This tactic would have supported the strategy of point protection in Peeples Valley and the town of Yarnell. This concept is displayed in Figure 15.

 Pic 27

Figure 15. Yarnell Hill Fire’s final outline perimeter (thick, orange outline). Topographic map showing proposed burnout and dozer line. Deployment site marked with a green circle.

Based upon interviews and incident photos, our opinion is that GMIHC was trying to ignite a test fire in anticipation of burning out indirect line along the two-track that they used to walk in when SEAT drops extinguished their burn out. GMIHC was then told by ATGS to go back to direct attack. From the interviews and incident documents, it is unclear whether the DIVS A spoke with Field OCS about the SEAT extinguishing the test fire. Also unclear is whether DIVS A spoke to Field OSC about GMIHC retreating to the burned area because the tactic of building direct line was not feasible.

RISK MANAGEMENT

Safe implementation of the strategy and tactics requires constant re-evaluation due the continual change in predicted and observed fire behaviour. One of the main requirements of risk management is to identify trigger points for re-evaluating strategies and tactics being applied on the fire. ICT4 initially selected a tactic of direct attack for the Yarnell Hill Fire, including the establishment of an anchor point and flanking the fire as it headed to the north and east. This tactic was not fully implemented on June 29. As a result, early in the morning of June 30, a decision to have GMIHC establish the anchor point and make a direct attack on the fire was made between Planning OSC and DIVS A during a conversation at the Yarnell Fire Station. Neither Planning OSC nor GMIHC had actually seen the fire.

The chosen tactics had the following limitations:

· THE SUSTAINED LINE PRODUCTION RATE OF A 20-PERSON CREW IN THE CHAPARRAL BRUSH TYPE IS ONLY 436 FEET AN HOUR, THE SLOWEST PRODUCTION RATE OF ANY FUEL MODEL24.

· FIREFIGHTERS USING HAND TOOLS ARE EFFECTIVE ONLY AGAINST FLAME-LENGTHS OF ABOUT 3.5 FEET BECAUSE OF THE HEAT GIVEN OFF BY THE FIRE25.

· THE RATE OF SPREAD ESTIMATES FOR THE YARNELL HILL FIRE EXCEEDED 436 FEET PER HOUR WITH FLAME LENGTHS GREATER THAN 3.5 FEET.

At the June 30 noon meeting between IMT2 Command and General Staff, they discussed current perimeter control efforts and decided to continue with existing strategy and tactics.

On June 30, the fire moved to the northeast, then to the east, and eventually to the south. The SPGS2 reported flame lengths of 40 feet with rates of speed up to 16 miles per hour occurred, yet no one seemed to recognize these signs as trigger points that should have led to a change in tactics and relocation of GMIHC. The probability of success for the chosen tactic of establishing an anchor point and flanking the fire diminished greatly each time the fire changed direction. We found no evidence that:

· GMIHC OR DIVS A SUGGESTED TO PLANNING OSC THAT THE TACTIC WOULD NOT WORK,

· THAT PLANNING OSC FOLLOWED UP WITH DIVS A OR GMIHC TO GET THEIR IMPRESSION OF THE CHANCE FOR SUCCESS,

· OR THAT IMT2 REEVALUATED THE TACTICS OR DISCUSSED MOVING GMIHC SO THAT THEY COULD REENGAGE WHERE THEY WOULD BE EFFECTIVE.

24 Wildland Fire Incident Management Field Guide, 2013, PMS-210, page 109.

25 2010 Incident Response Pocket Guide (IRPG), page 79.

ARIZONA INCIDENT MANAGEMENT TEAM

According to the 2013 Southwest Area Mobilization Guide,

“The Arizona Divisions of Forestry and Emergency Management jointly sponsor the Arizona Incident Management Team (Arizona IMT). The team consists of employees of the Forestry Division and other areas of State government and from fire departments throughout the State. While the majority of the team’s experience is wildland fire, the team is organized with “multi-hazard” intent and is used to manage a wide range of incidents and events at the local, State, and Federal levels within Arizona as well as other areas of the country. The Arizona IMT can be configured and will respond to meet the requirements of almost any all-risk incident up to and including a Type 2 level. The team will respond as a Type 3 IMT, as a Type 2 Short IMT, as a Type 2 Long IMT, or as ordered depending on the needs and desires of the Agency Administrators responsible for the incident.”26

The Arizona Incident Management Team, referred to as IMT2, was ordered as a “short team” and arrived without some of its key members. The following discussion outlines what happened based upon dispatch logs, interviews and daily logs. The importance of the discussion is that the team that arrived was already short-handed and had to fill in with people that were not part of the initial team order. This led to an organization that lacked the initial cohesion needed to successfully take over a complex fire.

The initial team order did not include a Safety Officer (SOF) as suggested by the Southwest Area Mobilization Guide for ordering a short Type 2 Team27. The SOF regularly assigned to the team was unavailable due to injury. Orders for two SOFs were placed on the evening of June 29 that went unfilled. They were re-ordered on the next day at 1230 with a request that they arrive by 1700. The responding SOFs arrived separately at 1455 and 1530 on the afternoon of June 30. The SOFs were assigned on the fire shortly before the entrapment and burn over. The SOF duties include “monitoring and assessing hazardous and unsafe situations and developing measures for assessing personnel safety.”28

The individuals who filled the Operations Section Chief (OSC) positions were not originally ordered as OSCs, although they were both qualified. One individual was originally ordered as an Incident Commander Type 3, but was reassigned as Planning OSC when he arrived on the fire. Field OSC was originally ordered as a Division Supervisor, but reassigned when he arrived on the fire. The Air Support Group Supervisor (ASGS) and Air Operations Branch Director (AOBD) were ordered as part of the original team order. The rest of the Operations Section was filled with individuals who were ordered as individual resources, not as part of the formal team. This includes the individuals assigned as Structure Protection Specialists (STPS).

The Planning Section Chief (PSC) was ordered as part of the team, but did not arrive at the ICP until late afternoon. The Fire Behaviour Analyst (FBAN) was also part of the initial team order and arrived in time for the 1000 briefing. A GIS Specialist was ordered as part of the team; however this position is not listed in the final organization chart of those assigned to the fire.

26 2013 Southwest Area Mobilization Guide, page 60-10.

27 2013 Southwest Area Mobilization Guide, page 60-7.

28 NWCG Glossary of Terms.

The initial team order did not request a Logistics Section Chief (LSC), but the person who filled that role was ordered as a Based Camp Manager (BCMG) in the initial team order. The initial team order did request an Ordering Manager (ORDM), a Supply Unit Leader (SPUL) and a Communications Unit Leader (COML). The COML did not arrive until after the 1000 briefing.

The Finance Section Chief (FSC) was originally ordered as the Cost Unit Leader (COST). A Time Unit Leader (TIME) was on the original order.

The results of IMT2 initially missing key people or having them arrive after the morning briefing led to the following deficiencies as the team took the fire over from IMT4.

· THE PSC WOULD HAVE BEEN ABLE TO BRING FOCUS AND COHERENCE TO THE 1000 BRIEFING AND DISTRIBUTE MAPS TO ALL RESOURCES.

· AN SOF ON-SITE THE MORNING OF JUNE 30 WOULD HAVE VIEWED THE FIRE AND FIRELINE ASSIGNMENTS STRICTLY FROM A SAFETY VIEWPOINT, NOT THE TASKORIENTED VIEWPOINT OF AN OSC.

· COML WOULD HAVE BEEN ABLE TO HELP THE TEAM TO ESTABLISH CONSISTENT COMMUNICATIONS WITH GROUND AND AIR RESOURCES BEFORE THEY LEFT THE ICP.

· ICT2 HAD TO ASSUME ALL THE MISSING FUNCTIONAL DUTIES WHICH WAS A SIGNIFICANT WORKLOAD.

· THE TEAM THAT WAS ASSIGNED TO THE FIRE LACKED THE COHESIVENESS THAT IS EXPECTED WHEN A TEAM IS ORDERED.

As soon as the ICT2 saw the scope and potential of the fire, he started seeking the closest qualified resources, including a SOF, through every channel available to him. The ICT2 called resources directly and bypassed the Resource Ordering Status System (ROSS) to ensure that people with the necessary skills were in place as soon as possible to assist fighting the fast moving chaparral fire. The ICT2 had little choice but to accept the fire on the morning of June 30, however the job was made more difficult based upon the way the team was ordered. They did not arrive as a cohesive and functioning unit and spent the day trying to bring order to a very chaotic situation.

Communications on the Yarnell Hill Fire were inadequate from the time IMT2 arrived because the COML arrived late. COML was not available to clone radios at the morning briefing. Tone guards were also a problem. Lack of communication is a significant safety problem.

An additional problem with the way the team arrived is that without a PSC, maps are not readily available to resources going to the fire line. GMIHC was not provided with a map or aerial photo by ICT4 when they arrived on the fire. A map would have helped the crew estimate how far the Boulder Springs ranch site was away from the lunch spot and evaluate alternative escape routes including the two-track road to Boulder Springs Ranch. Visually, the ranch looks close from the top of the ridge where GMIHC initiated their descent into the canyon. The heavy brush in the canyon, combined with the rocky nature of the area, made travel difficult and slow. They may have underestimated the speed with which the fire was moving.

BRIEFINGS

On June 30, the IMT2 morning briefing at 0930 lacked necessary effectiveness because many Command and General Staff members had not arrived at the fire. The ICT2 had to present many parts of the briefing that should have been presented by other Command and General Staff positions because the PSC, SOF, COML and one of their usual OSC’s had not arrived at the ICP by the time of the briefing. ICT4 was present to assist with the briefing as needed.

Based upon incident documents and interviews, it appears that most of the information flow for GMIHC occurred through informal conversations prior to departing to the fire line at 0800. GMIHC and DIVS A were not at the 0930 briefing.

A good example of the lack of communication from the briefing that occurred on June 30 was the test fire that GMIHC was igniting. During the morning at the top of the ridge, GMIHC was planning to burnout a small section of line to create the fire perimeter down to a two-track road. While they were igniting the test fire, two SEAT retardant drops extinguished their test fire between 1130 and 1145. The Air Tactical Group Supervisor (ATGS) did not know the purpose of the burning and dropped retardant on it, forcing GMIHC to go to direct attack on the fires perimeter. Planning OSC failed to inform ATGS of the tactics for the fire. During our interview with the ATGS on ASM1, he stated that they did not have a firm understanding of where the division breaks were. He stated that they could see the dozer, but did not understand its mission. The dozer was constructing a contingency line from west to east, which when complete, would allow a firing operation to be conducted when conditions were favourable. If ATGS had known this, they could have reinforced the dozer line with retardant instead of picking a location just to the north.

UNCLEAR DIVISION BOUNDARIES

Neither DIVS A nor DIVS Z was at the 0930 briefing at the ICP. At different times, Field OSC instructed DIVS A and DIVS Z to establish the division boundary. DIVS A thought that DIVS Z wanted to establish the Division break somewhere near the top of the hill, which would leave Division A with only the top of the ridge heading north. DIVS Z could not figure out how to establish the Division break with DIVS A and travelled back to ICP to seek clarification. DIVS Z also was unclear as to how to initiate effective suppression actions. During an interview, DIVS Z stated that “there was just no good ways to connect any dots at that time. It was just really difficult to see anyway to connect a piece of line.”

ATGS of ASM1 believed that Division A was on the western edge of the fire along the ridgeline heading north from the heel of the fire. Planning OSC thought that Division A was working from the anchor point to the East, down-slope to the valley floor. Planning OSC also thought that the western edge of the fire perimeter was not a major concern because the fire could not move to the west over the ridge without encountering a change in aspect, vegetation types and prevailing wind direction.

OPERATIONAL OVERSIGHT

The Blue Ridge Interagency Hotshot Crew (BRIHC) spent part of their time waiting for an assignment and part of their time working on improving the dozer line. The dozer encountered a locked gate and “No Trespassing” sign which stopped line production. The objective for the dozer was to construct line from the old grader to the east to anchor into a physical feature that would allow for a burnout to protect Yarnell. This objective was established by STPS1, and it is unclear as to whether Field OSC and Planning OSC were aware of this plan.

DIVS A did not feel that GMIHC needed the help from BRIHC. However, GMIHC made little progress in establishing an anchor point. With limited resources available, Planning OSC could have used both GMIHC and BRIHC in Division A to establish the anchor point and connect with the dozer line. Assigning both GMIHC and BRIHC to work together would have used the concept of mass action. The crews would have been able to burnout from the top of the ridge to the dozer line fairly quickly. Giving the crews separate assignments appears to be a result of poor communication, which led to poor coordination.

COMMUNICATION BETWEEN DIVS A AND OPERATIONS

The fire situation deteriorated throughout the day of June 30. The incoming IMT2 personnel had to orient themselves to the current fire behaviour, the resources currently on the fire and plan for the impending wind shift. There had to be a coherent flow of information between Field OSC and ATGS, and between Field OSC and DIVS A. The Field OSC was limited by the lack of a PSC who would translate the conceptual plan into an actual plan to be implemented on the ground. Based upon our interviews, we believe that none of the following items were intentional, but a function of an overwhelming and understaffed situation.

Planning OSC stated that “since we had not developed a plan…as we got…things going we would just assign them out.”

The decision to establish the anchor point and flank the fire was made between Planning OSC and DIVS A during a conversation that occurred while standing in the Yarnell Fire Station early in the morning of June 30. Neither OCS Plans nor DIVS A had actually seen the on-the-ground fire situation. Once GMIHC had arrived at their work area, Planning OSC should have asked if the burned area was an adequate safety zone in the event of extreme fire behaviour. If GMIHC did not consider the safety zone adequate for an extreme fire behaviour situation, then their assignment would have been changed to one that would be in a safer area.

Once the SEAT drops had extinguished the test fire that GMIHC was igniting, the crew tried to build direct handline, which subsequently failed. We found no evidence that DIVS A notified Planning OSC that the tactic of going direct had failed. Such a notification should trigger a reassessment of both strategy and tactics.

Planning OSC did make efforts to check on the crew with helicopter flyovers and radio calls in the morning and early afternoon. During an interview, Planning OSC said that he believed that GMIHC was safe because they were located in over 200 acres of previously burned landscape, some of which had been cold for 24-36 hours. By 1540, the first trigger point in Yarnell was breached and STPS1 called for an evacuation of the town. A short time later, the fire reached the second established trigger point and was breached which called for all structure protection personnel to retreat to their pre-identified safety zones. Planning OSC was standing with STPS1 when these commands were made.

GMIHC did not stay in the burned area safety zone. They moved toward a previously identified safety zone at Boulder Springs Ranch. Planning OSC did overhear a radio transmission from GMIHC saying to someone that they were using their predetermined route to the structures. Although Planning OSC believed as long as the crew had one foot in the burned area they would be safe, earlier in the day, he should have discussed with GMIHC whether the burned area was an acceptable safety zone.

Some Incident Management Teams require Division or Group Supervisors to call back to the ICP by a specified time to confirm that the requirements of LCES have been put in place. This protocol was not in place on the Yarnell Hill Fire.

Departure from Standard Practices

In determining the standards that guide professionals in the field of wildland fire management, we identified the 2013 Interagency Standards for Fire and Fire Aviation Operations (Red Book) and Wildland Fire Incident Management Field Guide (PMS 210) as established industry standards. We also referred to Arizona Revised Statues and City of Prescott guiding documents as needed. In addition to these resources, we also used the 10 Standard Firefighting Orders and LCES. Through our interview process, a clear picture emerged that ground-level firefighters treat the 10 Standard Firefighting Orders and LCES as rules and upper level managers tend to treat the Orders as guidelines. As a result of our observations, we have chosen to treat the 10 Standard Firefighting Orders and LCES as rules because they should have guided the actions of GMIHC on June 30.

10 STANDARD FIREFIGHTING ORDERS

We have applied the 10 Standard Firefighting Orders to the Yarnell Hill Fire:

1. Keep informed on fire-weather conditions and forecasts

Planning OSC briefed GMIHC on fire weather conditions and forecasts at the Yarnell Fire Department during the morning of June 30. The crew was later informed twice over the radio about weather warnings from the National Weather Service concerning approaching thunderstorms with associated strong winds.

2. Know what your fire is doing at all times

GMIHC was positioned on a ridgeline that had an unobstructed view of the fire movement and intensity. The crew had a lookout posted for much of the day. Their lookout eventually had to move because the fire reached pre-established trigger points that meant that he was in danger from the fire. GMIHC no longer had a lookout after their lookout evacuated his position. ATGS was in the air above the fire when GMIHC decided to change locations; however the crew did not ask ATGS to serve as their lookout.

3. Base all actions on current and expected behaviour of the fire

GMIHC based their actions on the fire behaviour they had observed for several hours.

4. Identify escape routes and safety zones, and make them known

GMIHC had identified their vehicles and the Boulder Springs Ranch as good safety zones. The Ranch site was large and well-constructed, with wildfire in mind. The site withstood the flames of the Yarnell Hill Fire as it burned around the Ranch. The buildings sustained very little damage and the owners stayed in the main house as the flaming front passed. Granite Mountain had several escape routes to select from. We could find no evidence that they timed or improved the escape route to Boulder Springs Ranch.

5. Post lookouts whenever there is possible danger

GMIHC posted a lookout when they were building direct handline. However, GMIHC did not have a lookout posted during their descent to the safety zone. The lookout had left his post because trigger points used to ensure his safety had been breached. During the critical period when GMIHC was traveling to the safety zone, the lookout was moving the crew vehicles to a safer location as requested by his supervisor. Based on interviews, we found no evidence that GMIHC requested that ATGS or anyone else in a position to see the crew’s location, watch the fire for them as they travelled to Boulder Springs Ranch.

6. Be alert. Keep calm. Think clearly. Act decisively

Evidence shows that even up to and including their last radio transmission, DIVS A and GMIHC were alert, unimaginably calm, thinking clearly, and taking decisive actions.

7. Maintain prompt communications with your forces, your supervisor, and adjoining forces

GMIHC maintained communications with everyone on their crew and division. DIVS A had some difficulty maintaining communication with Planning OSC. GMIHC did not notify their supervisor that they planned to move to an alternate safety zone.

Planning OSC ineffectively communicated the tactics to be used for the day with all of his forces. There is evidence that the aerial resources did not understand tactics being used by forces on the ground. There is also evidence that DIVS A and DIVS Z could not agree where the division break should be placed.

8. Give clear instructions and insure that they are understood

ASFD failed to:

· PROVIDE A WFSA OR WFDSS DOCUMENT AND RATIONALE FOR SELECTING ITS SUPPRESSION ALTERNATIVE TO THE IMT2;

· PROVIDE THE IMT2 WITH CLEAR WRITTEN DIRECTION IN THE FORM OF A DELEGATION OF AUTHORITY LETTER, WHICH IS CONSIDERED TO BE MARCHING ORDERS BY INCIDENT COMMANDERS;

· THE PLANNING OSC DID NOT GET AVIATION RESOURCES AND GROUND RESOURCES ON THE SAME TACTICAL PLAN. GMIHC WAS ATTEMPTING TO BURN OUT FIRELINE AND ATGS ORDERS TWO RETARDANT DROPS ON THEIR BURNOUT. SIMILARLY, THE STRUCTURE PROTECTION GROUP WAS USING A DOZER TO CONSTRUCT CONTINGENCY LINE NEAR YARNELL, BUT THE AVIATION RESOURCES CHOOSE TO DROP RETARDANT ON A SIMILARVECTOR CLOSE TO THE DOZER LINE. AIR RESOURCES MISSED THE OPPORTUNITY TO REINFORCE THE DOZER LINE WITH RETARDANT BECAUSE THEY WERE NOT PROPERLY COORDINATED WITH THE STRUCTURE PROTECTION GROUP.

9. Maintain control of your forces at all times

GMIHC died together in a very small space. No one ran. This is a testament to the exceptional leadership abilities of GMIHC Superintendent and Captain.

10. Fight Fire Aggressively, having provided for Safety First

ASFD had a strategy of full suppression using the tactic of direct attack. When the tactic failed, the managers of the fire did not reassess the strategy or tactics. A reassessment should have resulted in GMIHC moving to an area of the fire where they would have been safe and could be used effectively.

Although GMIHC successfully followed most of the 10 Standard Firefighting Orders and LCES, this section discusses the errors that were made by the crew.

The LCES checklist suggests that more than one escape route be available and that escape time and safety zone size requirements will change as fire behaviour changes. GMIHC initially had multiple escape routes, including the ability to walk back to their vehicles (an option that was closed off when the vehicles were moved to safety). A second escape route was to travel south along the ridge towards the Boulder Springs Ranch and turn east at the descent point. However, this escape route had not been scouted, timed, marked or improved. At the descent point, they had a third option of turning to the west, escaping over the ridge and down to Highway 89 on the Congress side of the mountain. A fourth option would have been to continue along the two-track road to the south and east to the Boulder Springs Ranch. There is no evidence that GMIHC had scouted and timed alternative escape routes or checked the escape route they used for loose soils, rocks or excessive vegetation. There is also no evidence that the crew had evaluated the escape time versus the potential rate of spread based upon the afternoon weather forecast.

LCES was designed to be a simple way to ensure that fire line resources have identified their Lookout, Communication, Escape Route and Safety Zone. LCES reinforces the 10 Standard Firefighting Orders and 18 Watch Out Situations.

A second error made by GMIHC is that they did not have a lookout when they made the descent to Boulder Springs Ranch. GMIHC did a very good job of having a lookout posted while they established the anchor point and constructed line. Based upon interviews and incident documents, we could find no evidence that they requested a lookout as they travelled towards Boulder Springs Ranch.

Finally, GMIHC had an obligation to notify their supervisor where they were moving and what route they would be traveling. The confusion that surrounded the search for the crew after the entrapment and burn over illustrates the importance of notifying the supervisor.

The 10 Standard Firefighting Orders were developed by a Task Force in 1957. The expectation was that firefighters would memorize them and use them to keep themselves safe. The project was undertaken after the Inaja Fire in California where 11 firefighters died. The effort to develop the 10 Standard Firefighting Orders also led to the development of the fire shelter and Nomex clothing for wildland firefighters, and a call for improved training especially in the area of fire behaviour. The Inaja Fire of 1956 was viewed by some as the final straw in a series of fire fatalities:

1943 Hauser Creek, CA, 11 fatalities

1949 Mann Gulch, MT, 13 fatalities

1953 Rattlesnake, CA, 15 fatalities

1956 Inaja, CA, 11 fatalities

The 10 Standard Firefighting Orders were written in an attempt to provide rules that would be able to stop the increasing number of fire fatalities.

Fatigue

Based upon interviews, fatigue appears to be a factor in the decisions that were made by ICT4 during the Yarnell Hill Fire. Timesheet records indicate that he had worked 28 days straight as of June 28. ICT4 sized the fire up on the evening of June 28 before returning home for the night. The following morning, he arrived in Yarnell and started a shift that would last for over 30 hours. The Incidence Response Pocket Guide states that going 24 hours without sleep affects your decision-making ability the same way a blood-alcohol-content of 0.10 would29. This level of exhaustion could impair decision making ability and situational awareness.

Fatigue may have been a factor for GMIHC as well. Their work records indicate that they had worked 28 out of 30 days during the month of June. The crew had worked 13 of a 14-day tour. Although technically not a violation of the work-rest guidelines, cumulative fatigue resulting from working 28 out of 30 days may have been a factor in their decision making process.

29 IRPG, PMS 461, January 2010.

CONCLUSIONS

We have determined that the following factors directly contributed to the entrapment and burn over:

· FIRE BEHAVIOR WAS EXTREME AND EXACERBATED BY THE OUTFLOW BOUNDARY ASSOCIATED WITH THE THUNDERSTORM. THE YARNELL HILL FIRE CONTINUALLY EXCEEDED THE EXPECTATIONS OF FIRE AND INCIDENT MANAGERS, AS WELL AS THE FIREFIGHTERS.

· ARIZONA STATE FORESTRY DIVISION FAILED TO IMPLEMENT THEIR OWN EXTENDED ATTACK GUIDELINES AND PROCEDURES INCLUDING AN EXTENDED ATTACK SAFETY CHECKLIST AND WILDLAND FIRE DECISION SUPPORT SYSTEM WITH A COMPLEXITY ANALYSIS.

· THE INCIDENT MANAGEMENT DECISION PROCESS FAILED TO RECOGNIZE THAT THE AVAILABLE RESOURCES AND CHOSEN ADMINISTRATIVE STRATEGY OF FULL SUPPRESSION AND ASSOCIATED OPERATIONAL TACTICS COULD NOT SUCCEED. THIS ALSO REMAINED THE CASE WHEN THE STRATEGY CHANGED FROM FULL SUPPRESSION TO A COMBINATION OF POINT PROTECTION AND FULL SUPPRESSION.

· RISK MANAGEMENT WEIGHS THE RISK ASSOCIATED WITH SUCCESS AGAINST THE PROBABILITY AND SEVERITY OF FAILURE. ASFD FAILED TO ADEQUATELY UPDATE THEIR RISK ASSESSMENT WHEN THE FIRE ESCAPED INITIAL ATTACK LEADING TO THE FAILURE OF THEIR STRATEGIES AND TACTICS THAT RESULTED IN A LIFE-THREATENING EVENT.

--end--

Further information hoping to be identified and still to be located.

IFE Commentary & lessons if applicable

None produced at this time.

Known available source documents

Further information hoping to be identified and still to be located.

Local Fire Department Incident Report/s

No information identified to date and/or still to be located.

United States Fire Service trade union/representative body/s Incident Report/s

No information identified to date and/or still to be located.

National Institute for occupational Safety and Health (NIOSH) Incident Report/s and/or improvement notices

No information identified to date and/or still to be located.

Local Police Department Incident Report/s

No information identified to date and/or still to be located.

EMT/Ambulance Department or Service Incident Report

No information identified to date and/or still to be located.

Underwriters Laboratory (UL) Reports/investigations/research

No information identified to date and/or still to be located.

National Institute for Standards and Technology (NIST) Reports/investigations/research

No information identified to date and/or still to be located.

Fire Protection Association (FPA) Reports/investigations/research

No information identified to date and/or still to be located.

Coroner’s report/s and/or improvement notices

No information identified to date and/or still to be located.

Related Government Correspondence

No information identified to date and/or still to be located.

Other information sources

Karels, J. & Dudley, M. et al. (2013). Yarnell Hill fire June 30, 2013 serious accident investigation report September 23rd 2013. [pdf]. Arizona State Forestry Division. Also available at http://www.iawfonline.org/Yarnell_Hill_Fire_report.pdf

O’Brien, D., Anderson, E., Hicks, B., Larsen, D., & Schulte, D. (2013). Granite Mountain IHC entrapment and burn over investigation; Yarnell Hill Fire – June 30 2013, prepared for Arizona Division of Occupational Safety & Health. [pdf]. Wildland Fire Associates.

Dougherty, J. (2018). Twelve Granite Mountain Hotshot Families File Wrongful Death Claims Seeking $237.5 Million. [online]. Available at; http://www.investigativemedia.com/twelve-granite-mountain-hotshot-families-file-wrongful-death-claims-seeking-237-5-million/ [Accessed 9th March 2018]. Investigative Media.

Numerous resources including PowerPoints, documents and video available at;

Unknown author. (Unknown date). Yarnell Hill fire entrapment fatalities (2013). [online]. Available at https://www.wildfirelessons.net/orphans/viewincident?DocumentKey=1a2dac92-1d79-420f-be0e-1aa616a40a70 [Accessed 22nd March 2018]. Wild Fire Lessons Learned Centre.

Further information hoping to be identified and still to be located.

Service learning material

No information identified to date and/or still to be located.

Videos available

ABC15 Arizona. (2013). New video from the fallen Granite Mountain Hotshots. [Online]. Available at; https://www.youtube.com/watch?v=gfmPgAA2AMU [Accessed 9th March 2018].

Storyful News. (2014). New videos released from deadly Yarnell Hill Fire (1). [Online]. Available at; https://www.youtube.com/watch?v=7UVL8pxSBJc [Accessed 9th March 2018].

Storyful News. (2014). New videos released from deadly Yarnell Hill Fire (2). [Online]. Available at; https://www.youtube.com/watch?v=KqRn1Ro9J8E [Accessed 9th March 2018].

Unknown author. (Unknown date). Yarnell Hill fire entrapment fatalities (2013). [online]. Available at https://www.wildfirelessons.net/orphans/viewincident?DocumentKey=1a2dac92-1d79-420f-be0e-1aa616a40a70 [Accessed 22nd March 2018]. Wild Fire Lessons Learned Centre.

Further information hoping to be identified and still to be located.

Go back