Author(s): Jesse Roman. Published on November 2, 2015.

Flame Fundamentals

A conversation with fire scientist Jack Cohen on new research into how wildfire spreads, and how the findings could produce more reliable models to anticipate wildfire behavior. BY JESSE ROMAN

FOR MILLENNIA, WILDFIRES BURNED unchecked, regularly clearing fuel loads in the environment as part of a naturally occurring ecological process. But 200 years of human intervention have transformed many parts of the United States into ticking time bombs, says Jack Cohen, a researcher and fire scientist at the USDA Forest Service’s Missoula Fire Sciences Laboratory in Montana.

What Cohen describes as “our cultural kneejerk approach of eliminating fire at all cost” by suppressing every possible fire has resulted in a fuel-rich landscape. The resulting fires, exacerbated by drought, are larger, more intense, and harder to control. Not every fire will become a monster, but more and more are. This year, for the first time in its history, the U.S. Forest Service spent more than half (52 percent) of its roughly $6.5 billion annual budget on wildfire suppression efforts, compared to just 16 percent in 1995. By several other metrics, 2015 has been one of the worst fire seasons on record. More than 9 million acres have burned this summer in the United States, destroying millions of dollars in property and killing three firefighters and at least seven civilians as of early October.

However, while fire experts like Cohen believe that wildland fire is a critical component to maintaining healthy landscapes, letting every fire burn when there are millions of people living in or at the edge of wildfire-prone areas isn’t viable, either. The trick is to find a safe way to let some wildland fires burn as “an appropriate ecological process,” Cohen says. But science is limited in its understanding of how wildfires spread, so predicting what they will do and where they might go has proven difficult. As Cohen puts it, “we have to understand wildfire better.”

That was the goal when Cohen and others at the Missoula lab, and at partner sites at the University of Maryland and the University of Kentucky, began studying the fundamental processes behind fire spread in wildland fuels. Their results, published in July in the Proceedings of the National Academy of Sciences, have called into question some long-held beliefs about wildfire dynamics and suggest that most models for how wildfire spreads are based on a false narrative.

For decades, fire scientists have believed that wildfire spreads through radiant heat transfer, similar to the sun’s rays, where the fire’s radiant heat ignites fine fuels and vegetation ahead of the fire front. Convective heat transfer, where fuels are ignited when they come into direct contact with flames, was largely dismissed as a primary means of wildfire spread because flames are highly buoyant and were thought only to rise. But as wildland firefighters know, wildfire often spreads horizontally, even in the absence of wind—a phenomenon scientists assumed would be unlikely without radiant heat transfer.

However, fire spread experiments in the Missoula lab have shown that heat transferred by radiation is insufficient to ignite the fine fuels, such as grasses and pine needles, that make up most wildland fuel beds. In fact, according to Cohen’s work, it now appears that convection is the primary way fire is spread in fuel beds such as shrub and tree canopies. This new understanding could lead to more reliable computer models that better anticipate wildfire behavior, which could inform fuel mitigation efforts and improve firefighter safety. It could also give fire managers better information to make decisions about when and where to conduct prescribed fires and when it is possible to let naturally occurring fires burn.

NFPA Journal spoke with Cohen about this new research and what it might mean for the environment and the millions of people living under the threat of wildfire.

It might seem intuitive that flame contact with the surrounding fuel, or convective heat transfer, would be the primary way wildfire spreads. Why hasn’t this been the assumption among scientists?

The reason it was assumed that wildfire is largely spread by radiant heat transfer is because the density of a flame at 1,000 degrees Celsius is about a quarter the density of ambient air at sea level, and that difference produces about 2.5 G-forces of upward buoyant force. So how do you overcome that powerful upward buoyant force and get the flame to go down and flow into the fuels ahead of the flame front? It’s hard to explain—without experiments, this is exactly the argument modelers have used to justify the assumption that radiation is the primary governing mechanism of fire spread.

So physics says that forces within the fire should make the flames shoot up like a hot-air balloon, not down and forward into the fuel bed, making convective heat transfer unlikely. How did you get the notion to question this logical assumption?

In about 2003, myself and my colleague Mark Finney at the USDA Forest Service started discussing this topic, and we each made the observation to the other that we didn’t see ignitions happening without flame contact. So over a period of time we began doing exploratory experiments that continually reinforced this notion that flame contact has to occur in order for the ignition to happen. We decided specifically to set aside the question of how fire spreads and instead focus strictly on the fundamental question of how the fuel particles are heating up leading to ignition. We are now asking fundamental physics questions from the premise that fire spread is the result of continued ignition, so if you don’t understand ignition, you don’t understand fire spread.

A key part of your research was to observe fire very closely to see what patterns were revealed. What did you see?

In lab tests we saw that flames tend to produce a very regular kind of pulsing behavior and we started noticing these peaks in the flame front. Three of the four of us who worked on this project are former firefighters, and we have all seen these flame peaks from small grass fires to huge crown fires. We always assumed without question, even in the lab, that it was because of heterogeneities in the fuel—a cluster of fuels producing a higher intensity at this one location than some other location—and we passed it off.

But then Mark came up with the idea of creating highly reproducible fuel beds—absolutely perfect in consistency, in spacing, in size and character of the fuel bed—so that we could see exactly what was happening at the flame front without these fuel variations. We unencumbered ourselves from the notion that we wanted a rate-of-spread model and therefore had to use real fuel in our experiments. The bottom line is the physics doesn’t change. The context may change, but the physics of the heat transfer and flame dynamics remains the same, regardless of the fuel particles used.

What did you learn from observing the fires spreading in the engineered fuel beds?

We set up these fuel beds, made of laser-cut cardboard, in a wind tunnel, and lo and behold we were still seeing these same peaks and troughs in the flame front. We took a closer look with high-speed videography and by sampling temperatures 500 times per second, and we started seeing structure in the flame front. We started seeing the rotating structure and the outflow coming from the trough areas at fuel level.

The fluid dynamics get relatively complex and abstract, but the bottom line is you start generating pairs of counter rotating vortices within the flaming front. Where the rotation is up in both cases, you get a peak in the flame and where it’s down you get down drafts and it flows out the front side. That explains how a highly buoyant gas like a flame will end up getting pushed down and forward into the fuel. So the action is where the troughs are, not where the peaks are. The fuel heating to ignition is at the troughs.

Do you believe you have found some fundamental ‘truth’ of wildfire? Do you see this phenomenon in all wildfires?

You see it everywhere. And the extremely interesting thing is that we see it regardless of size. We have blown cold air across a hot plate in the wind tunnel and we produce the same kinds of vortex pairs. We have also observed these vortices in large crown fires. When it scales up from very slow laminar flow to a turbulent crown fire, it gives us great hope that we will be able to actually characterize the pulse frequency and eddy size that is responsible for the flame contact on the unburned fuel adjacent to the flame front.

If the existing models based on radiant heat transfer are wrong, and wildfires do spread through convection, how will this understanding enable you and others to better predict how wildfires will spread and move?

Whether we can diagnose the conditions that would lead a wildfire toward one behavior or another is still up for question. Our hope is that by better understanding the fundamental processes of how fuel ignites, we can, if not make predictions on where the fire will go, certainly diagnose the conditions under which it would spread or not spread.

What are you personally hoping will come of this research?

When Mark and I started working on this, our primary motivation had to do with allowing for an appropriate ecological occurrence of fire. We look at fire suppression as an increasing cost for diminishing return. In our cultural kneejerk approach of eliminating fire at all cost, we are creating a huge negative impact. We’ve actually changed the landscape—the composition and continuity of the fuel—by putting out fires until we can’t. We have changed the ecology of wildland fire; now wildfires are more inclined to burn more severely and extensively than they otherwise would. That is definitely influencing the human benefits that we get from our landscape, such as water quality, wildlife, commodity production, and other natural resources. We are not going to end up with the same vegetation composition and structure on the landscape that we have come to expect.

Could your research eventually help enable forest managers make better decisions as to which fires to allow to burn and which to put out?

We have 300 million-plus people living in the contiguous 48 states, along with all of our infrastructure, so we can’t deal with free-burning fire the same way we did 200 years ago. That means we have to understand it in order to proactively manage it in a more appropriate ecological manner. The more we understand something, the more comfortable we get with it.

I think the fundamental message here is that wildfire is inevitable and that wildfire under extreme conditions is inevitable. We need to be able to deal with this in some way other than a highly reactive and expensive failing process.

Interview conducted and edited by JESSE ROMAN, NFPA Journal staff writer. Top Photograph: Richard Barnes/OTTO