Overview: On October 26, 2006 at approximately 0700 five firefighters from the USDA Forest Service were entrapped and killed on the Esperanza Incident in Riverside County, California. A Red Flag Warning had been issued for a predicted Santa Ana wind event that was to be accompanied by very low relative humidity due to compressional heating and subsidence of the air mass on the morning of the entrapment.  Eyewitness accounts state that at about 0700 there was area ignition of up to ¼ mile wide flame front and that very strong winds and fire conditions forced personnel of four other engine companies operating in the area to take refuge in their apparatus during the time that Engine 57 was entrapped.

Fire managers may have underestimated the potential for explosive fire behavior on the morning hours of October 26 since at the outset of the fire, the winds had not surfaced to critical levels. However, it should be noted that this fire rapidly grew in size and complexity at approximately 0500 as firefighters were being deployed for structure protection in a narrow canyon in the Twin Pines housing area.

Access into the canyon was a brush lined single lane dirt road that led into a steep chimney. The entrapment site was approximately 2 miles from the nearest paved road. Firefighters had been called to the fire around 0100. First reports on conditions stated the fire to be approximately 2 acres.  It quickly grew to 10 acres with the potential to become a major fire, at which time additional resources were ordered. The fire eventually consumed 40,200 acres, 34 homes and 20 outbuildings. 

Narrative: Research scientists have been using the complex process of coupled weather-fire modeling to examine the fire behavior and atmospheric conditions that were associated with the Esperanza Fire and the entrapment. The Esperanza Fire was selected as a case study since there was sufficient data available for researchers both to input into a fire model (such as a map of fuel properties, time and location of ignition, topography) and to validate its simulations, including airborne thermal imagery of the fire at later times.
 
The official accident report states that alignment of the strong wind with the steep slopes covered in fire climax chaparral and capping from above by an atmospheric temperature inversion allowed the fire to travel with extreme speed and intensity towards Engine 57. This massive thermal pulse of energy quickly over took the firefighters who were not found in fire shelters but were wearing issued PPE. All of the firefighters of engine 57 sustained fatal burns from the entrapment. Temperatures on the ground at the entrapment site are suspected to have been as high as 1,600 degrees Fahrenheit.
 
Dr. Janice Coen, a scientist at the National Center for Atmospheric Research in Boulder, Colorado, has been working with a coupled weather-wildland fire model to understand the combined weather, fire, terrain, and fuel factors that led to the rapid rate of spread and complex fire behavior during the first day of the fire.

The coupled weather-wildland fire model is a weather model to which a fire behavior model has been coupled, so that weather, fuel, and terrain slope determine the fire’s spread rate and direction while the heat released by the fire creates strong up currents of air and wind shifts– i.e. the fire “makes its own weather”. The weather-wildland fire model takes in the fuel type and conditions, initial weather state, terrain, and ignition location, and solves complex mathematical equations and takes into account derivatives and complex data to create the output. The output is a product that accurately depicts weather variables including wind, temperature, humidity, and pressure; fire rates of spread, energy and smoke release; and the effect of the plume on updrafts and downdrafts, and their impact on local weather conditions.
 
The preliminary results from the coupled weather-fire model show that strong atmospheric terrain-induced gravity waves surfaced over the fire and brought down the very strong winds of the Santa Ana wind event. As the wind was driven from Cabazon Peak to the west-southwest across canyons, the uphill flank of the fire made runs up each canyon it passed. The model showed rapid sinking of the air in wave downdrafts over the fire area, strong heat release and winds rising up the slopes from burning in the heavy chaparral. These air motions may have been funneled together and accelerated by progressively narrow canyons - tributaries that lead directly to the promontory point where Engine 57 was over taken by the flame front. At the time of the entrapment, observers report that a plume is estimated to have risen between 18,000-24,000 feet probably due to the rapid rate of combustion in the steep drainages.  Because the atmospheric wind speeds up due to the heat released in extreme burning conditions, there is a very strong interaction/ feedback on the weather from the fire. Photographs show that the smoke was closely hugging topographic features near the ground most of the day. There is ambiguity whether this is due to an inversion layer that is believed to have existed or due to the subsidence of air in the atmospheric gravity waves because there are not weather soundings or other upper air data near the entrapment site.

Dr. Coen states that there are other types of fire models available to fire managers as well, such as BEHAVE and FARSITE. However for the Esperanza Fire researchers have used the coupled weather-wildland fire model because it was crucial to model the weather in the vicinity of the fire, as well as the fire behavior and the fire’s feedback on the weather.
 
Flying above the Esperanza fire on the morning of the entrapment was fire ecologist Dr. Phil Riggan from the USDA Forest Service Riverside Forest Fire Laboratory. In a specially outfitted Piper Navajo fixed wing aircraft, Phil used FireMapper, a special infrared imager to map and monitor the fire. Based on a microbolometer focal-plane array, FireMapper captures infrared radiation and provides information about the physical properties of the fire. The advantage of this type of instrument is that infrared instruments can see the fire line and intensity through dense smoke.

Phil took a series of aerial images of the Esperanza Fire shortly after the entrapment and he presented his findings to other researchers at the Fire Lab. Phil and fellow researcher Bob Lockwood of the Riverside Fire Lab hope that some day that real time images will be able to be sent from the aircraft to incident command staff quickly enough to allow fire managers to make accurate decisions based on real time data. However sending large packets of data from a computer in an aircraft to a specially equipped computer with a radio or cell phone on the ground is not yet routine. You can see examples of their work at www.fireimaging.com

Conclusion: Recently, Francis Fujioka, Janice Coen, Phil Riggan, and Bob Lockwood met at the USDA Forest Service Riverside Forest Fire Laboratory to collaborate on studies such as the Esperanza incident to improve fire science and modeling in the future. After a series of short meetings, the scientists visited the entrapment area first hand. Photographs of the topographical features, fuel type, and fire scars were taken to help aid interpretation in the modeling process. This post fire visit captured a better understanding of fuel types and fire progression that will improve wildfire modeling and help to reveal critical factors about fire behavior. Scientists hope that their findings will enhance firefighter safety.

For more information on fire modeling contact:
Janice Coen, Ph.D.Project ScientistNational Center for Atmospheric ResearchP.O. Box 3000, Boulder, CO 80307-3000(303)497-8986http://www.mmm.ucar.edu/people/coen/