May 23, 2016

Most Weather-Related Wildland Firefigher Deaths Can Be Prevented: Here's How.

Last summer three wildland firefighers died near Twisp, Washington as wind reversed during the afternoon.

In 2013, nineteen firefighters died near Yarnell, Arizona as the gust front emanating from some thunderstorms caused the fire they were working to explode.

Between 1910 and 2014, 1096 wildland firefighters died while they were fulfilling their missions, and a significant number of these tragedies were weather related.

And as I will describe below, most weather-related wildfire deaths can be avoided if fire agencies are willing to apply current generation observing and forecasting technologies, while enhancing critically needed personnel.

As someone who specializes in local weather in the western U.S., regional weather prediction, and the effects of terrain on regional flow patterns, I have always been interested in the weather associated with major wildland fires.   Looking at major recent fires associated with firefighter deaths (such as Yarnell and recent Twisp fires), I have concluded that each was associated with a sudden wind shift and that these wind changes were easily foreseeable by someone possessing meteorological knowledge and state-of-the-art observational resources (click on the links for these two fires to see my blog analysis).

The Yarnell fire deaths were associated with the outflow boundary of cooler air emanating from thunderstorms to the northeast, something evident from weather radar imagery and regional surface observations.  The Twisp disaster was connected with a windshift, associated with cooler air moving across the Casacades, that was predicted by regional weather prediction models and was evident in real time from surface observations.

The convection and gust front associate with the Yarnell Tragedy was evident in weather radar imagery

Recently, I listened to an excellent presentation on "tragedy fires" at the Northwest Weather Workshop that was given by Andy Haner, an experienced National Weather Service meteorologist specializing in wildfire prediction.  In his talk, he went through five fatal fires--Yarnell (2013), Twisp (2015), Thirty Mile (2001), Frog Fire (2015), Beaver Fire (2014)--and found that four of them were associated with clearly predictable weather phenomena.  He had not worked up the Thirty Mile Fire (he is a relatively young man), but I had--it was the same story:  a trained meteorologist could have easily seen the threat coming.  Furthermore, many of the deaths occurred for relatively new and small fires, fires for which no incident meteorologist had been assigned (Incident Meteorologists (IMETs) are forecasters specially trained to work during severe wildfire outbreaks).

Considering the seriousness of this issue, I conferred with some very experienced fire weather researchers in the Forest Service (who preferred anonymity).  They agreed with my and Haner's analysis:  virtually all the major wildland firefighters deaths could have been prevented with better information and guidance using current observations and forecasting technology.

Disturbingly, official reports on these wildland firefighter deaths do not consider or provide superficial coverage of this crucial issue (for example, the initial Twisp report is here).  Very disappointing.

So based on my analysis of several fires and discussions with a number of fire weather experts, let me describe how we could end most wildland firefighter deaths.

(1)  ALL wildland firefighting efforts should have supported by trained meteorologists that are continuously watching the situation.  Importantly, this includes even small or emerging fires (since they are involved in many of the deaths).  This does not require one meteorologists per fire, since a meteorologist can watch over several in an area (think of the meteorologists being similar to flight controllers that supervise several flights at a time).  These meteorologists would be responsible to provide guidance, forecasts, and warnings to fire crews.

(2)  All wildland firefighting crews will have continuous and uninterruptable communications to the meteorologists.   This means that every crew must have satellite phones before they go to a fire.

Weather observation and prediction technology has greatly improved during the past decades, allowing meteorologists to provide radically better support for wildland firefighters.  For example, we have large number of  surface weather observations throughout the U.S., including remote areas.  To illustrate, here is map that shows the latest wind observations over the Washington State (and there are much more available than is shown here).

Weather radar imagery are providing constant coverage over much of the western U.S. (although there are some gaps).  In those areas, high-resolution satellite imagery is available.

New weather prediction technology is now available, such as the High Resolution Rapid Refresh (HRRR) forecasting systems that uses all available observations to create high resolution analyses and short-term forecasts each hour (see below)

And there are many more observing and forecasting resources that I have not mentioned.  Simply put, modern meteorologists have an extraordinary set of tools to determine what the weather is doing now and what it will be doing in the future, even in relatively remote areas.    Importantly, there is no reason that sudden windshifts, the source of many of the wildfire tragedies, can not be diagnosed and forecast.

The deaths of brave young men and women can be greatly reduced if trained meteorologists with the proper tools are assigned to all fires.  With modest resources, such meteorological support could quickly become a reality.   

And there is one more thing, and it is perhaps controversial.   Firefighting teams should never risk their lives to save isolated homes or buildings.   Most of regional forests are meant to burn and have burned for millennia.   Homes and buildings don't belong there and it should be understood by all that the buildings are expendable and will not be protected.  They are not worth risking human lives.

Not worth a human life.


  1. Some good ideas - it will take awhile.

  2. Not so sure about the need for satellite phones. Fire crews already carry VHF radios. There's no reason why communications from meteorologists cannot be relayed by radio if there is adequate radio coverage. The key problem here is one of a lack of adequate forecast information, not a lack of communication.

  3. Your photo of the house buried in the forest is sobering. The builders & owners are asking for it. First, they should have a defensible (nonflammable) perimeter around their buildings. Second, in an area prone to wildfire, structures should be designed and built that do not burn easily.

    Even in western WA we all need to be fire safe. Our lush environment can support devastating wildfires. We personally live in an forested area. Our (new) house is built with fairly nonflammable materials (steel roof, concrete & stone walls, but some wood fascia). We did leave some big hundred-year-old maples fairly close to the house. Hard to kill ancients.

    Agencies (city, county, state, federal) that issue permits should share in the responsibility to protect people (including firefighters).

  4. I especially like your comment regarding private inholdings within state or national forest boundaries. I could not agree more.

  5. Good suggestions Cliff.

    The fact that the way we manage fires has changed (completely) in the last 30 years should also be addressed, or at least discussed openly so the public understands it.

    I have fought forest fires. There is very little that is comparable to the objective 30 years ago (put the fire out) to the objective today (manage the fire - and the actual policy is still evolving). Comparing acres burned annually today vs. that of the past is a false comparison due to the amount of forest land that is intentionally burned now during these events. There are some valid reasons for this (safety, protection of property, fuel build-up due to the 20th century policy), but what is happening today, and the level it is happening, is something the public should be better made aware.

  6. Hey Cliff, I am a wildland firefighter and I was up at the Twisp fire this summer. You are right about the wind shifting and cutting off their escape route down woods canyon rd, but we do have meteorologist that monitor our fires. We have fire fighters that sit at the boundry of the fire that reports wind speed, direction, rh, tempurature and fire activity. We also have dispatch give us weather updates on all of the fires that we are on throughout the day. Also, we start everyday with a briefing that includes what the weather is going to be like. Every fire fighter with a radio will hear these weather updates throughout the day.

  7. At some point during the 90's, it was decided that the National Forest Service should stop performing routine culling of deadwood within their areas of purview - it was speculated that the primary reason why this change occurred in their MO was because of wealthy individuals building vacation homes within fire - prone areas. If individuals want to build homes in areas prone to fire and flooding (i.e. shorelines in hurricane - prone regions), they should be required to buy private insurance to cover their damages, and not rely on government subsidized policies that only encourage more risky building because of their low cost and extensive coverage. If you can't afford it, then don't build it.

  8. Alec,
    I really respect what you do. To give you proper warning, it takes more than someone at boundary of the then it is generally too late. You need someone watching the regional observations who has a good knowledge of the local meteorology. Most incipient small fires do not have a meteorologist watching. That is the problem. And certainly, your colleagues that died were not given proper warning..cliff

  9. You say it will take only "modest funding," but you must realize FS is vastly underfunded. Until they can get legislation through congress that funds major fires from a disaster fund like other natural disasters instead of the current system where FS has to raid fire prevention (e.g. controlled burns and thinning), trails, and other budgets to fight fires, the funding situation is not going to improve. Senators Wyden, Cantwell and others have been trying to get this legislation through congress for several years to no avail. Fat chance of getting it in an election year.

  10. I would argue your last suggestion (never risking human lives to save structures) is the most important first step in protecting wildland firefighters.

  11. It would be a simple matter to assign a IMET to the dispatch center to monitor weather conditions.

    Be aware that most dispatch centers cover a HUGE area. I will leave it to others to see if an IMET could monitor conditions over such an area effectively.

  12. I was the communication leader at the nearby Chelan Complex/Wolverine fire when the Twisp River burn over occurred. Even though the Twisp River "start" was not our fire, we still supported it. We had set up radio links to the forest service radio system as a back up and through out the day, a fire weather meteorologist from the Anchorage NWS would use our "crossband" link repeater to access radio frequencies being used to give moment by moment updates with the blessing of CWICC or Central Washington Inter Agency Communications Center. She knew how critical fire weather was and was doing her best to ensure that firefighters knew what was going on.
    Weather alerts were going out on multiple frequencies constantly in a radio foot print in excess of 2000 square miles. Satellite phones were junk in the area because of the deep valleys and high mountain peeks made them paper weights because they just didn't work well with a mountain between the user and the satellite. We had our coffee pot in the radio shack set up on cases of them because they didn't work in the local topography. I am confident that our fire's weather alerts were likely being heard because my radio operators were some of the first to know when the medical helicopters were cancelled due to confirmation of fatalities.
    There are a lot of issues that go on with people getting burned over. Some of these include "warning fatigue" where weather warnings are coming out constantly and frequently, and its normal for fire fighters to operate in "red flag" conditions. "Trigger points" operationally become subjective. Issues also include meteorologists being "UTF" or unable to fill, because there are not enough to go around to all of these fires. Many meteorologists with the NWS either are not interested in fire weather or their home offices won't let them be part of the program (staffing issues). Other problems (some of which have been mentioned on this blog before) is when weather interacts with topography especially wind at a local/micro level, and the lack of effective weather radar in the area.
    There are also other human issues too. Can you expect a meteorologist fresh off a plane from DFW (for example) to understand local weather and topography? I bring this up because said meteorologist said an 'atmospheric scientist is not a meteorologist', and that he went to a true forecasting school at Texas A&M. "You just need a hard science degree to apply to the NWS". This guy was not at the fire in Washington, BTW, but my interaction with him has forever left me concerned about consistency in the "IMET" or incident meteorologist position. Most of the IMET's I have met or work with at fires are bright, talented, and extremely hard working. They are there because the care.

  13. Jason,
    I appreciate your detailed comments. Clearly, there was a failure. If satellite phones don't work in a certain area (I have used them in the mountains), then we need reliable VHF or other communications. You need experienced meteorologists who know the area. But to be honest, what happened that days was not complex and it was well forecast. Firefighters should have been pulled back as the wind shift approached, just as the Yarnell folks should have been pulled off the fire when the gust front got close. Your point about there not being enough meteorologists available is a good one. That is exactly my point...more need to be assigned to assist firefighters in the field. ..cliff

  14. VHF and frequency propagation also can have issues that could be a blog unto itself. However, there has been recent research that radio propagation (including VHF) in the presence of fire reduces radio signal strength. Research at the University of Adelaide in Australia by Jonathan Boan ( discusses this.
    In my experience, heavy fire activity does seem impact radio function in the VHF spectrum. So when you need it most, that's when it doesn't work. Its the atmosphere again; it would be really neat if more research could go into this including maybe a local study. It would appear that a lot of local factors can impact radio during a fire. I wish I had a suggestion for research funding. Thank you.

  15. Of course firefighters should not risk their lives to save someone's cabin in the woods. Those property owners should also be paying full cost for insurance, if they have any. I'd be surprised if any insurance company would take that risk.

  16. Interesting no one has brought up the standard firefighting orders, something I can remember being trained on as a young guy wen I was fighting fires. Here they are:

    See number 1? Also 4 and 5? With our 20/20 hindsight, we can see where the orders were tragically not followed.

    My crew boss was a hard ass about these orders. I thought he was a little over the top at the time but am now grateful for the professional he was. We were on some hairy fires and never in danger.

  17. Patrick... my experience in having a house in the woods is that fire insurance is not hard to come by, but depending where you live, it can be more expensive. This is especially true in areas of California where wildfires are frequent. I don't think it's nearly as big a deal here in Washington, especially Western Washington. I know in my case, my insurance company does not care that my property is extensively forested compared to neighbors who have no trees.

    Fire departments are obligated to fight fires threatening homes. That's why we pay them our tax money. If counties want to change that, they should stop issuing building permits for dangerous locations or require some sort of 'opt out" clause on fire protection.

    I could just as easily argue that homeowners in cities should be paying higher rates, or have no insurance available, when their house is close enough to another house to have a fire spread. If houses are too close together, we should not ask firefighters to risk their lives to fighting a fire that is worsened by the known, dangerous practice of locating close to neighboring houses.

    Or we could both agree these arguments are silly.

  18. Fire fighter safety would enhanced in my opinion by reasonable timber harvest in areas where choked, overgrown forests have been mismanaged National Forest Service. The "let it burn" policy in the Okanogan and Wenatchee forest in 2015 was the cause of loss of life and homes.
    The best option is to transfer all National Forest to the Washington State DNR. Manage forests, fund schools, open access for recreation... win win win

  19. ***News Flash***

    Don't build your dream home in the middle of a damn forest.

  20. Breaking news: The Seattle PI is reporting that the Save KPLU campaign reached their fundraising goal of $7 million!

  21. A good read just published.

  22. I'm the Fire Chief in a district known occasionally for challenging interface fires in high wind.

    My legal marching orders are to prepare for and respond to those fires, usually at the incipient stage prior to the arrival of State and Federal wildland resources, and often in defense of properties which Cliff might include in his gallery of poster children write-offs.

    Time permitting, it would be nice to have real-time weather guidance 24/7, but failing that, for my money, one of the most useful firefighter-saving tools is FLAME, or fireline assessment method (, which develops skills in foreseeing change affecting fire behavior.

    -- Jim Appleton, Mosier Fire Chief

  23. Defensible space around forested homes is important both in forests and at the forest edge.

    In many cases people living in a forest provide stewardship services that are valuable.

    I do not think that lives should be lost to protect structures but I also do not believe that we can just abrogate private property rights. But if you build in a high risk place you are on your own.


Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

Undergraduate Scholarship Fund in Honor of Steve Pool

Steve Pool was a leading television meteorologist in Seattle for nearly 40 years..... but he was so much more. In addition to providing mete...