The Yarnell Hill Fire

Research background for "Moving Forward With New Information," a PowerPoint presentation on the Yarnell Hill Fire

By Holly Neill

More Info:

  Alan Sinclair report
  PowerPoint in HTML
The PowerPoint presentation included here, a collaborative effort between Alan and me, was presented to the Central Arizona Wildland Response Team (CAWRT) in March, 2016, in Phoenix, Arizona, by Alan Sinclair. The presentation is based in part on the following research.

My interest in the Yarnell Hill Fire and this research project began shortly after the deaths of the 19 Granite Mountain Interagency Hotshot Crew (GMIHC) members on the Yarnell Hill Fire on June 30, 2013.

After reading official investigation reports, many issues remained unresolved in my mind. Three issues in particular prompted me to further action.

Issues #1 and #2:

From the Yarnell Hill Fire Serious Accident Investigation Report (SAIR) Executive Summary pg. 1 -- emphasis added:
    1. Personnel who communicated with the Granite Mountain IHC knew the crew was in the black at that time (1550) and assumed they would stay there. No one realized that the crew left the black and headed southeast, sometime after 1604.
    2. There is a gap of over 30 minutes in the information available for Granite Mountain IHC. From 1604 to 1637 (a 33-minute gap), the team cannot verify communications from the crew, and we have almost no direct information for them.

Three months after the fatalities occurred, in September of 2013, I learned that the U.S. Forest Service had conducted an Aerial Firefighting Use and Effectiveness Study (AFUE) at the Yarnell Hill Fire, during which extensive audio/video recordings were made of aircraft operations. I submitted a public records request for the AFUE data to Arizona State Forestry Division, the agency in charge of the fire. The records arrived 3 months later and I began to review the data on December 3, 2013.

It quickly became apparent that the AFUE recordings had picked up numerous radio transmissions by firefighters on the ground that could be heard in the background, some clearly and some fragmented and difficult to make out. I began to scour the recordings. I'm a former wildland firefighter; we're accustomed to listening to multiple radio frequencies on fire assignments, gaining information and intel on adjoining forces. Several conversations caught my attention.

In background radio traffic, I heard a voice I believed to be that of Eric Marsh, the superintendent of the GMIHC, acting in a division supervisor's role on the fire as Division Alpha, in charge of a specific area including the heel of the fire where GMIHC was assigned and working. In 2014, Marsh's widow Amanda and others who knew him well confirmed it was indeed Marsh's voice. The audio quality of the conversation is poor, so John Maclean and I had the audio professionally enhanced for improved clarity.

During the audio enhancement, the audio specialist pulled out each relevant section of dialog, applied numerous filters, and produced individual files of each section. Included here are the entire continuous original audio and the sections that were separated into individual files or excerpts. They represent a best effort to enhance and clarify the transmissions.

The original audio/video is from: Serious Accident Investigation Records
Video ID: 20130630-16162-O-VLAT-split-1-E-P

This conversation proves two important points:
  • · There was no 33-minute communications gap, as cited in the SAIR. Marsh was communicating at 16:13, during the supposed (16:04 to 16:37) gap.
  • · Further, the audio evidence shows that Marsh gave a status update on his and the GMIHC location and movements, notifying someone that the crew was moving south after 16:04.
It is important to note that additional audio recordings exist to indicate other conversations involving Marsh within the 33-minute gap, including the following audio/video from 16:27. Positive voice identification for Eric Marsh was once again obtained in 2014 from those who knew him well, including Amanda Marsh.

The original audio/video is from: Serious Accident Investigation Records
June 30, 2013

It remains unknown who is speaking to Marsh in either audio.


From the second official investigation of the fire, the Granite Mountain IHC Entrapment and Burnover Investigation prepared for the Arizona Division of Occupational Safety and Health by the Wildland Fire Associates and released in November, 2013.
  • · We could find no evidence that they timed or improved the escape route to Boulder Springs Ranch. (pg.40)
  • · A second escape route was to travel south along the ridge toward the Boulder Springs Ranch and turn east at the descent point. However, the escape route had not been scouted, timed, marked, or improved. (pg.41)
I took my first trip to Yarnell, AZ in July 2013 with my husband Wayne Neill, to attend the GMIHC Memorial Service. I returned again in November of 2013 and viewed the GMIHC fatality site from the ridge above it. The dozer line that was pushed in from the Helms' Ranch to the fatality site to allow access the night of the fire was highly visible in the box canyon below. I began to wonder if anyone had scouted or investigated the area between the ranch and the fatality site for physical evidence along the escape route, before the dozer line was put in. I made numerous inquiries to officials and was able to conclude that this area had not been investigated before. This raised the possibility that evidence was there or may have been destroyed by the dozer. I determined to search the area, but it wound up taking several trips to do a thorough job.

Wayne and I returned in March of 2014. During this trip Wayne found a cut stob just east of the fatality site. It could have been cut by GMIHC as they were clearing a deployment site, which they tried to do, or in an effort to clear a P or personnel line. Time did not permit further scouting, but we vowed to return for additional search efforts.

I was able to return again for the first-year anniversary in June 2014. On this occasion I spent two nights at the Helms' Ranch, and on June 29 I asked Alan Sinclair to join me in scouting the area again. Alan found a chainsaw-cut stob much closer to the ranch than the one previously found, and located south of the dozer line.

I returned again with a larger group in November 2014, which included Wayne, John Maclean, and another friend. During this trip with more people available we conducted an extensive grid search of the area between the fatality site and the Helms' Ranch. Wayne and I found additional cut stobs, located closer to the ranch than to the fatality site, and in the same vicinity of Alan Sinclair's find. These stobs were also found on the south side of the dozer line.

I returned by myself in June 2015 and found another cut stob in the same vicinity.

Wayne and I also hiked the steep slope above the fatality site that GMIHC descended. We walked in a switchback pattern looking for any indication of cutting or thinning above the deployment site. Numerous small-diameter branches with burned and angled tips were scattered across the slope, indicating a widespread and noticeable burn pattern. We concluded that the same burn pattern was frequent and scattered, without being limited to any specific location. We did not find any other suggestive stobs on the steep slope.

I photographed and took GPS coordinates for each cut stob found between the fatality site and the ranch, with the following exception: Alan Sinclair photographed the stob he found, after he found it and removed it, and before I could photograph it in its original location. I did take a photo of the location where he found the stob, the location was marked with a rock and an upright stick, and I took GPS coordinates at the location.

I then plotted the GPS data onto a Google Earth map. It was interesting to note that even though collection had occurred over 15 months, from March 2014 to June 2015, the cut stobs were all located in relative alignment, and south of the dozer line.

I repeatedly asked Lee and Diane Helm, the ranch owners, if they were aware of any pre- or post-fire chainsaw cutting in this area, the box canyon just to the west of their property, or had any cut wood been dumped there by them or anyone else?

The Helms did have brush clean-up work done around their property several weeks before the fire, but all slash was removed by truck and taken off the property. They also had a wood fence installed earlier, and all scrap material was hauled off the property as well. They did not cut any trail through the box canyon themselves, nor were they aware of anyone else doing so. They neither hiked the area nor did they use the area for recreation.

I also questioned several firefighters who were on scene after the fatalities occurred, from the evening of June 30 to the following morning and into the next day. Was any chainsaw work done in this area, either along the dozer line or for mop-up purposes? All answers were "no". I was repeatedly told that there was no reason to run a chainsaw to support the dozer line operation. I was able to determine that there were some mop-up efforts around the immediate ranch perimeter by firefighters with hand tools, on July 1, 2013, but chainsaws were not used, nor necessary.

The PowerPoint presentation makes it clear that these findings are inconclusive and only suggestive. It is unfortunate that the dozer line was put in before this area could be scouted for indications that an escape route may or may not have been improved. It is also unfortunate that official investigators drew absolute conclusions without a thorough investigation of the area.

The presentation also makes it clear that the SAIR encourages additional research: 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. (SAIR pg. 46)

The question remains: Did GMIHC mark, time, scout, or improve their escape route to the ranch?

It is important to keep in mind that research and investigative efforts are ongoing, and the information presented here is in no way all-inclusive.

But the information presented here, in summary, suggests (without concluding) that GMIHC may have made efforts to improve their escape route to the ranch, and it further suggests, even more, that the box canyon should have been more thoroughly examined by official investigators after the fire.

Whatever it adds to our understanding of this tragic event, these findings can and should inspire us to continue to ask more questions and seek more answers.

We owe the fallen an accurate legacy; one that will stand the test of time.