Wildfire Magazine :: :: Anatomy of a Blow-Up

  • WUI
  • PREPLAN
  • PUB-ED
  • TACTICS
  • COMMAND
  • TECHNOLOGY
  • FORESTRY
Official Publication
spacer
International Association of Wildland Fire (IAWF)

Browse Back Issues

WILDFIRE MAGAZINE
About Us
E-Newsletter
Media Kit
Subscriptions
Buyers Guide
Job Opportunities
Resources
Fire Chief
IAWF
NIFC
Fire Weather
InciWeb
NICC
Firewise

Recent Comments

Powered by Disqus

Wildfire Magazine > -->

Anatomy of a Blow-Up

By Domingos Xavier Viegas


spacer    spacer    spacer    spacer    spacer

During a wildfire, it's not unusual for a firefront to move quickly with a sudden increase of the rate of spread. This sudden increase of fireline intensity has caused many fatal accidents all over the world.

Because this phenomenon is accompanied by very strong air drafts that surprise all in the vicinity and occurs in canyons or in steep slopes, it's commonly known as the “fire blow-up” or “chimney effect.” However, the designation “blow-up” is hard to translate from English, so “fire eruption” may better describe this phenomenon given its similarity, in qualitative terms, with what happens in a volcano when a very strong convective process is triggered suddenly.

The eruptive behavior of forest fires has been associated in the past with many fatal accidents in Portugal and elsewhere. Among the cases reported in the United States are the 1949 Mann Gulch Fire that killed 13 firefighters and the 1994 South Canyon Fire where 14 firefighters lost their lives. In an accident that occurred in Thirtymile, four people were killed in the sequence of a fire eruption. Given the complexity of the topography in a large part of Portugal that facilitates the development of this phenomenon, the eruptive fire behavior is a relatively common situation in forest fires there.

In spite of this phenomenon's importance in terms of fire behavior and fire safety, there aren't many published studies dedicated to it. For this reason, my research team and I have dedicated particular attention to fire spread in canyons and related phenomena with the objective of improving understanding and increasing personal safety at the firefront. Based on laboratory and field experiments and on the observation of real fires, complemented by theoretical and analytical studies, I've developed a mathematical model that explains and predicts the eruptive behavior of a fire during a blow-up.

But don't worry! Here the processes associated with fire eruption are explained in plain language without complex formulae.

ERUPTIVE FIRE BEHAVIOR

A wrong idea in many minds is the belief that for a given set of topography, vegetation and wind conditions, a firefront propagates with a fixed rate of spread that can be evaluated using some simple or complex model. The studies I carried out at Laboratório de Estudos sobre Incêndios Florestais, the Laboratory for Forest Fire Research, have shown this not to be the case. Rather, for a given set of ambient conditions, the firefront can propagate with a range of spread rates. Fire behavior is dynamic in the sense that its rate of spread changes with time even for nominally permanent boundary conditions.

LEIF has used a testing rig to analyze fire behavior in canyons of arbitrary geometry during the past year. These observations have been checked with full-scale field experiments in the area of Gestosa in central Portugal since 1998. Based on these tests, I developed an original model to explain and predict eruptive fire behavior that was published in Combustion Science and Technology as “A Model for Forest Fire Blow-up.”

In the general case of fires spreading in upslope or with favorable wind, the rate of spread isn't constant. Even if the boundary conditions remain constant, the rate of spread changes continuously due to the convection induced by the fire itself and the related phenomena.

Consider the case of a fire spreading in a slope or a canyon in the absence of wind. If a fire starts as a point at the base of a slope or a canyon, it will have a circular form and propagate very slowly at the beginning. After some time the head of the fire situated on its upper part will become more intense because it is receiving the heat from the remaining area of the fire that is below it. For this reason the head of the fire will develop higher flames and a higher rate of spread. This flame height increase will induce a higher flow of air in its vicinity to feed the intensified combustion reaction. This process feeds itself and will induce an ever-increasing rate of spread.

With Ro as the basic rate of spread for a firefront in the absence of wind and slope (horizontal and flat terrain), the rate of spread of an eruptive fire can be more than 100Ro. In fact, the rate of spread increase can be hundreds of times that value. In such a case the fire can spread along a wide extension of terrain in very few minutes with a destructive capacity that's absolutely out of control. The convection produced by the fire is manifested by the sudden occurrence of very strong winds that produce the characteristic “roar” so well-known by those familiar with forest fires.

This strong wind induces many to think that it or some other strange atmospheric phenomenon produced the fire eruption, but they're completely wrong. The eruptive fire behavior derives from the fire itself and doesn't require any other external trigger. Provided that there's a vegetation-covered slope with an ignition at its base, the fire will increase in intensity until reaching eruption — independent of what happens around it. The steeper the slope and the drier and finer the fuel, the sooner eruption will be reached. If the slope isn't very steep and long, the fire may not accelerate fast enough to reach eruption before the slope is entirely burned.

The result of this process is an extremely dynamic fire behavior that's felt as a continuous increase of the rate of spread with time. A mathematical model based on experimental tests and physical considerations about the convective flow induced by the fire can justify and predict this behavior.

For example, the curve shown in “Eruptive Fire Behavior Model” on page 24 was obtained from laboratory tests with dead needles of pinus pinaster. Despite the facts that the fuel bed was homogeneous, the slope (or canyon) configuration remained constant and there was no wind, the rate of spread increase reached a phase in which it increased suddenly, attaining extremely high values.

The initial rate of spread depends on the slope inclination or of the canyon configuration. For example, in a slope with an inclination of 10° (18%), the initial rate of spread corresponds to point A in the graph, while for slopes of 20° (36%) and of 30° (58%), the initial rates of spread are represented by points B and C, respectively. The time required for an eruption to occur is reduced with the increase of terrain slope. In the case of a 30° slope or in very closed and steep canyons, the fire may reach the eruptive phase practically immediately after ignition.

Something similar happens with other fuel beds, although the time scale changes from one fuel to another. A heavier fuel like shrub or slash will have a longer reaction time, while a lighter fuel like herbaceous vegetation will have a shorter reaction time. However, the overall behavior is entirely the same. Research is being carried out to establish the relationship between time scales for different fuel types and terrain configurations. For example, while the eruptive phase may be reached within five to 10 minutes in herbaceous vegetation, it may need 20 to 30 minutes in shrub vegetation with all other conditions remaining the same.

PRACTICAL CONSEQUENCES

It's not hard to imagine why so many accidents are associated with eruptive fire behavior, especially in the initial stages. When a firefighting team reaches a starting fire at the base of a slope or in a canyon, its slow advance seems to allow for the possibility of controlling it without difficulty using only available resources.

Attacking the fire from its back and advancing along its flanks, the firefighters climb the slope surrounding the fire perimeter. In some cases, team members move along the slope to place themselves above the fire, perhaps in a road or ridge, from where they can make a frontal attack. If this movement isn't performed before the fire reaches the eruptive phase, it may lead to disaster.

The sudden increase of the rate of spread and fireline intensity not only may catch firefighters by surprise, it also may render any fire suppression efforts useless: There's no physical capacity to suppress firefronts with flame heights of 10 to 20 meters and with fireline intensities above 4,000kw/m. For this reason the decision to attack a fire in the base of a slope, even if the fire's relatively small, must be taken with great precaution.

Similarly, one of the basic safety rules is to never fight a fire — or even remain — on a slope above the fire with fuel in between. If there aren't enough resources, the attack shouldn't even be attempted. Firefighting personnel should back off and adopt other tactics that don't involve facing the fire in the middle of the slope. By no means should human and material resources be placed above the fire while it's active.

This is true even for personnel placed on high ground like ridges or plateaus. If an eruption occurs, the hot gases produced by the fire and the radiative and convective heat fluxes that will occur when the fire approaches can endanger the lives of those in such a location. Even a very wide safety zone like a fire break may not be sufficient to guarantee their safety. For this reason, if a blow-up is expected it's better to send all personnel out of that area to avoid unnecessary risks. Even houses and other structures may not remain safe. The placement of watch towers, communication masts, wind turbines and other structures on high points makes them particularly vulnerable to eruptive behavior. In the case of staffed watch towers, the decision to evacuate is of particular importance.

CASE STUDY

On Aug. 5, 2003, a couple who owned a local forest died in the north of Portugal because of eruptive fire behavior.

The topography of the area where the accident occurred is shown in the map on page 21. It is a slope above the Douro River, not far from Freixo-de-Espada-à-Cinta and near the Spanish border. At about 14:30 hours, a forest fire began at point A, and local firefighters responded immediately. In spite of the extreme fire danger conditions, firefighters managed to control and practically surround the fire. Their intent was to push the fire to the bottom of the slope against the river so that it would extinguish itself. At about 17:00 hours the fire was practically surrounded and considered to be suppressed, with the exception of a small firefront less than 30 meters long near the base of a double canyon at point B.

The two forest owners had left their house to watch the fire and see if it was endangering their forest stand near point C. They were more than 2 kilometers away from the firefront, which they couldn't even see due to the terrain. They would have felt safe as the main fire seemed to be extinguished. Having checked on their trees, they were climbing back toward their car to return home.

It was then that the fire eruption occurred. The small firefront that had been at the bottom of the slope for about 20 minutes suddenly became a colossal wall of flames that swept the entire slope to its top and even beyond. The fire caught the forest owners, and only by a miracle did it not kill some of the dozens of firefighters and civilians who were still around the fire.

On the top of the slope at point D, an automatic weather station was engulfed by the heat wave produced by the fire eruption. The data recorded by the station were used to reconstruct the fire eruption.

The air temperature during the day of the accident is shown on page 20. The indicated values are 10-minute averages recorded automatically by the station data logger. At about 18:30 hours, the air temperature rose suddenly over 20 minutes until it reached a value of 55°C, or 131°F. Anyone exposed to gases of that temperature would not escape easily without serious injuries, if at all.

The average value of wind velocity on the same day is shown on page 20, too. The wind velocity was around 10 to 15km/h, but it increased suddenly to reach average values of 65km/h and a maximum value of 96 km/h.

Based on the analysis of field experiments on shrub vegetation and on the real cases of South Canyon and Thirty-mile, the velocity of the wind induced by the fire was computed and is shown as a full line curve in “Wind Velocity Comparison” on page 21. The origin of the time scale was taken as the beginning of the fire eruption. The other two curves correspond to the average and maximum (10-minute) values recorded by the meteorological station in the same period of time. The model predicts very well the sudden increase in the wind velocity that was observed during the fire eruption and depicts its order of magnitude with reasonable accuracy.

These data illustrate much better than many words the power of the fire and its destructive potential. In this case two people died. It's hoped that no more victims are produced by accidents such as this one.

TERRAIN IS KEY

Among the situations of extreme forest fire behavior, the fire eruption is one that has the highest potential risk. Its occurrence depends essentially on the terrain configuration and has little to do with meteorological conditions or even vegetation. During the South Canyon Fire, the shrub vegetation that covered the slope where the accident occurred had a high moisture content, and the down-slope fire was propagating very slowly for a couple of days. During the blow-up the situation was completely different.

All that's necessary for a fire eruption to develop is an ignition and a slope with sufficient extension. Based on this study, the eruptive phenomenon is one of the more predictive situations of extreme fire behavior. However, it continues to surprise and overtake even experienced and knowledgeable people. Perhaps this description of eruptive behavior and the physical and mathematical model may encourage firefighters to avoid unnecessary risks so that such accidents don't happen again.

Domingos Xavier Viegas is a professor and head of the department of mechanical engineering at the University of Coimbra, Portugal. He has a degree in mechanical engineering from the Instituto Superior Técnico and a Ph.D. in aerodynamics from the University of Coimbra. He is the coordinator for the Centre for Studies on Forest Fires, a member of the board of directors of the International Association of Wildland Fire and is on the editorial advisory board of the International Journal of Wildland Fire.

Want to use this article? Click here for options!
© 2011 Penton Media, Inc.

spacer
gipoco.com is neither affiliated with the authors of this page nor responsible for its contents. This is a safe-cache copy of the original web site.