Skip to content

Extreme Weather and Wildfire

Available: September 2021

Wildfire visualizations/animations will show CAWFE model simulations for selected fires across California’s different fire weather regions.

For more detail on the visualizations, see this blog post:

Coupled Weather - Wildland Fire Modeling Case Studies

The following animations show coupled numerical weather prediction–wildland fire behavior model simulations of large wildfires created using the CAWFE modeling system.

2020 Creek Fire
Play Video

Ignition: 9/4/20 near Shaver Lake, CA

In first four days after ignition, the Creek Fire expanded 20,000-50,000 acres per day while satellite active fire detection data showed sustained active burning, indicating consumption of large fuel elements.  Below is a simulation of the first day.

2020 Creek Fire (continued)

Attributed to strong upcanyon winds in the San Joaquin River Valley, we find rapid growth driven in the first day by generation of an anticyclonic mesoscale vortex containing updraft cores (red) exceeding 25 m/s over the fire surrounded by a ring of compensating downdraft  (blue) that rapidly pushed winds north up the river valley.

2007 Witch Creek Fire
Play Video

Ignition: 10/21/07 east of Ramona, CA

The Witch Creek Fire, followed by the McCoy and Guejito fire ignitions (included here), were part of a late October southern California outbreak driven by a strong Santa Ana wind event producing exceptional fire growth and damage. Below is a link to a simulation of the wind event and growth from ignition. A vertical cross section of potential temperature through the ignition location suggests strong winds near the ignition were due to pressure driven flow of stable air over low aspect ratio terrain.

2018 Carr Fire Whirl

Period: 7/26/18 Redding, CA

The Carr Fire, having been driven eastward and down toward the central valley by westerlies across the coastal range, created several fire whirls, among which was a large fire whirl with winds exceeding 100 mph when entering west Redding that was detected by radar. CAWFE simulations reproduce fire growth down into Redding, and, when the fire burns across the line where two airflows intersect – the westerlies blowing downslope from the west and dry air up the Sacramento Valley – multiple strong fire whirls spin up and travel through the built areas.

East Troublesome Fire
Play Video
Visualized period: 10/21/20 afternoon: near Granby, CO

Three days after it was reported on Oct. 14, 2020, high winds and low humidity allowed the East Troublesome Fire to spread to over 10,000 acres. From Oct. 20–23, it spread dramatically, with daily increases of 18,000–87,000 acres. This is a simulation of the day of most rapid growth, Oct. 21.

Calwood Fire
Active growth: 10/17/20 afternoon: near Boulder, CO

Following reports of an ignition near the Calwood community center, the Calwood Fire ran nearly 10 km under gusty downslope winds in under 3 hours to Highway 36 along the Front Range, threatening WUI communities, until winds subsided as a weak cold front approached (from upper right).

North Complex (Bear Fire)

Simulation period: 1 a.m. 9/8/20–afternoon 9/9/20 near Quincy and Oroville, CA

On its way to becoming one of the largest fires in California history, the North Complex Fire (composed of multiple fires, most of which were suppressed, but the Claremont/Bear fires merged) grew by over 210,000 acres on 9/8-9/9, driven by strong northeasterly winds. CAWFE simulations show these were concentrated in the river valley of the Middle Fork of the Feather River, with strong pulsing gusts present over the lower foothills above and near Oroville.

Woolsey Fire
Play Video

Simulation period: 11/8/18–11/9/18

On the same day as the Camp Fire, ignitions on the Santa Susanna Field Laboratory site spread south rapidly under strong Santa Ana winds to the Ventura Freeway, which it later crossed overnight, into the Santa Monica Mountains. This simulation shows a river of strong winds coming off the Santa Susanna Mountains to the northeast, generating pulses of winds, particularly over the mountainous bridge connecting them to the Santa Monica Mountains. The fast spread and weaker heat flux reflects the fast-running fires characteristic of sparsely vegetated, chaparral-covered terrain.

Thomas Fire

Figure T1. CAWFE simulation of the first day of the Thomas fire at midnight Dec. 5, 2017. Heat flux from fire (in W/m-2) is colored according to upper color bar. Wind speed arrows (in m s-1) point downstream and are colored according to the lower color bar at right.

Ignitions: 12/4/18 6:26 PM and 7 PM near Santa Paula, CA

Two ignitions during a strong, long-duration Santa Ana wind event set off the Thomas Fire. During the initial run, the two ignitions merged and the fire raced through Ventura County toward the ocean, driven by winds down the Santa Clarita River Valley and over the southern Sierra Nevada Mountains. This period is characterized as a “wind-driven event.”

Thomas Fire (continued)

The Thomas fire on 12/8/17 at 4 PM PDT.. Heat flux from fire (in W/m-2) is colored according to upper color bar. Wind speed arrows (in m s-1) point downstream and are colored according to the lower color bar at right.

Ignitions: 12/4/18 6:26 PM and 7 PM near Santa Paula, CA

During later periods of the Thomas Fire, the Santa Ana event waxed and waned, varying in where and when strong winds reached the surface. In addition, the fire spread into the mountains north of the river valley, drawing itself up canyons and bowls sheltered from the winds.

Rim Fire
Play Video

The Rim Fire was a massive plume-driven wildfire that experienced days of over 37,000 and 51,000 acres of growth, despite relatively weak ambient winds.

Ignition: 8/17/13 near Yosemite, CA

Figure R1. CAWFE simulation of the 8/20/13-8/21/13 rapid expansion of the Rim Fire, initialized with the NIROPs fire map on late 8/20/13. Heat flux from fire (in W/m-2) is colored according to upper color bar. Wind speed arrows (in m s-1) point downstream and are colored according to the lower color bar at right. The misty field indicates smoke concentration.

Chimney Tops 2 Fire

Figure CT1. CAWFE simulation of the day of rapid growth of the Chimney Tops 2 Fire in Great Smoky Mountains NP down into the town of Gatlinburg, TN (foreground).. Heat flux from fire (in W/m-2) is colored according to upper color bar. Wind speed arrows (in m s-1) point downstream and are colored according to the lower color bar at right.

Ignition: Reported on 11/23/16 in Great Smoky Mountains National Park, TN

The Chimney Tops 2 Fire was one of several fires that burned in North Carolina and Tennessee during a very dry fall in the Appalachian Mountains. While foehns (downslope windstorms) are well known in other regions, conditions created an extremely strong downslope wind event that drove a preexisting fire down through the city of Gatlinburg, TN.

Camp Fire

Figure C1. CAWFE simulation of the Camp Fire. The fire’s heat flux is show in the upper color bar (in W m-2). Wind speed arrows (in m s-1) point downstream and are colored according to the color bar at right. Vectors are shown every 3 grid points.

Ignition: 11/8/18 near Paradise, CA

On November 8, a strong Diablo wind event created strong winds over the Sierra Nevada Mountains, with locally strong winds near the Jarbo Gap and Feather River Canyon. Research simulations using CAWFE showed that extreme winds, created by a shear instability along the top of a shallow, near-surface stable layer (a phenomena that resembles backwards-breaking waves) brought strong winds crashing to the surface, driving the fire rapidly down the slope.

Camp Fire (continued)
Play Video

Figure C2. CAWFE simulation of the Camp Fire. The fire’s heat flux is show in the upper color bar (in W m-2). Wind speed arrows (in m s-1) point downstream and are colored according to the color bar at right. Vectors are shown every 3 grid points.

This vertical cross section down the center of the Camp Fire shows the shear instability as rapidly spreading heavy air (lower potential temperatures) is lifted into slower moving air above it, creating the impression of retroflecting (backwards breaking) waves. As strong winds crashed down to the surface, they spread outward in scallops of strong winds that drove the fire rapidly downslope.

Redwood Valley Fire

Figure R1. CAWFE simulation of the Redwood Valley Fire. The fire’s heat flux is show in the upper color bar (in W m-2). Wind speed arrows (in m s-1) point downstream and are colored according to the color bar at right. Vectors are shown every 3 grid points.

Ignition: 10/8/17 near Potter Valley, CA

On October 8-9, 2017, fourteen large wildfires developed rapidly during a strong Diablo wind event in northern California including the Redwood Valley Fire. Diablo winds travelled south down the Sacramento Valley. The flow was pushed through gaps between ridges, creating narrow rivers of strong winds the peaks of which reached 30-40 m s-1 amidst areas where winds were weak or stagnant.

The Redwood Valley Fire had two nearby ignitions on the west side of the Potter Valley (the grass green area in center of Fig. R1). Driven by a shallow river of strong winds, within hours, it crested a ridge and raced down through populated lots into the Redwood Valley.
Animation – click here for mp4 (281 Mb)

Tubbs Fire
Play Video

Figure. CAWFE simulation of the Tubbs Fire at 4:11 AM PDT. The fire’s heat flux is show in the upper color bar (in W m-2). Wind speed arrows (in m s-1) point downstream and are colored according to the color bar at right. Vectors are shown every 3 grid points.

Ignition: 10/8/17 near Calistoga, CA

Coen, J. L., W. Schroeder, and B. Quayle, 2018: The generation and forecast of extreme winds during the origin and progression of the 2017 Tubbs Fire. Atmosphere, 9, 462. PDF

Here, we applied CAWFE to investigate the flow regime and underlying mechanisms associated with the extreme winds and fire behavior during the Tubbs Fire.

As Diablo winds travelled south down the Sacramento Valley and fanned out southwestward over Wine Country, their strength waxed and waned and their direction wavered, creating varying locations near fire origins where wind overrunning topography reached 30-40 m s-1. The flow created peak wind speeds exceeding 40 m s-1 at the crest of some lesser hills. The flow drove the Tubbs Fire over 19 km in 3.25 h into urban areas of Santa Rosa, CA.

King Fire
Play Video
Figure. CAWFE simulation of the King Fire from 9:45 PM Sept. 16 to 4:42 AM Sept. 18. The simulated total heat flux in (colored according to color bar at right, in W m-2) is shown on the surface, along with wind vectors (plotted every 4th grid point) at 21 m above ground level. In the animation, each frame is one minute apart.

Coen, J. L., E. N. Stavros, and J. A. Fites-Kaufman, 2018: Deconstructing the King megafire. Ecological Applications. doi:10.1002/eap.1752.

Ignition: 9/13/14 near Pollock Pines, CA

The King Fire was ignited at 6:37 PM on Sept. 13, 2014, spread during a severe drought in the central Sierra Nevada mountain range in complex terrain covered by mixed conifer forests, which generates complex fuel beds shaped by drought, land management practices (e.g. forest cultivation and harvesting, fire suppression, and fuel mitigation), and burn scars from previous fires. The King fire grew 7 km to the northeast through the evening of Sept. 16 as nearby surface weather stations recorded weak to moderate south-south-westerly winds upon which weak diurnal circulations and gusts of 2-10 m s-1 were superimposed.

From 9:49 PM Sept. 16, when the fire was mapped by the National Infrared Operations (NIROPs) airborne imager, until 1:06 PM Sept. 17, when satellite active fire detection data detected the fire entering the Rubicon Canyon, the fire traveled north over rolling hills. In an afternoon run that was unanticipated in light of weak to moderate ambient winds, the fire grew over 16,200 ha (40,000 acres), racing approximately 25 km to the northeast over the next 11 h � an average spread rate of 2.3 km h-1 � following the canyon to its crest at Hell Hole Reservoir, where growth stalled.

Yarnell Hill Fire

Figure. CAWFE coupled weather-wildland fire model simulation of the Yarnell Hill Fire. The color bar on the right indicates the heat flux (watts per square meter) from the fire, with more intensely burning areas in bright yellow and white, and less intensely burning areas in darker reds. A vector is shown each 4 model grid points.

Ignition: 6/28/13 near Yarnell, Arizona. Simulation of 6/30/13

Coen, J. L. and W. Schroeder, 2017: Coupled Weather-Fire Modeling: from Research to Operational Forecasting. Fire Management Today. 75:39-45.

On June 30, 2013, 19 firefighters were killed during the Yarnell Hill fire, when a gust front from the northeast blew across the fire, changing its direction, and making it spread rapidly across where they were sheltered.

The animation depicts a CAWFE simulation of June 30 (370 m horizontal grid spacing), from 2 am – 8:15 PM local time. The fire is initialized in the model using the 3 AM VIIRS active fire detection map. Each frame is 1 minute apart.
In the simulation, solar heating stirs up the boundary layer circulations throughout the day. Convection occurs in outer domains (not shown) to the northeast (upper right), creating high-based convective clouds as upper level air flows southeast over the Mogollon Rim. Rain falls into a very dry boundary layer, creating a broad gust front that reaches the south edge of the fire at frame 936. The fatality occurred around 4:45 PM. Comparison with the fatality report suggest the simulated timing for the gust front reaching the northeast edge of the fire and the fatality site are within 15 minutes of occurrence.

Little Bear Fire

Ignition: 6/4/12 near Ruidoso, NM

Coen, J. L. and W. Schroeder, 2013: Use of spatially refined remote sensing fire detection data to initialize and evaluate coupled weather-wildfire growth model simulations. Geophys. Res. Lett. 40:5536-5541.

The Little Bear Fire was ignited by lightning strike in the Sierra Blanco Mountains of New Mexico. It burned 17,939 ha (44,330 ac) and 254 buildings and was the most destructive fire in New Mexico State history.

High Park Fire

Ignition: 6/7/12 near Fort Collins, CO

Coen, J. L. and W. Schroeder, 2015: The High Park Fire: Coupled weather-wildland fire model simulation of a windstorm-driven wildfire in Colorado’s Front Range. J. Geophys. Res. Atmos. 120:131-146

The High Park fire is reported to have been ignited by lightning strike. It grew rapidly during a Colorado Front Range downslope windstorm, destroying 259 homes and burning 87,284 acres.

View: towards the north. Fort Collins lies to the right, and Poudre Canyon is the northernmost boundary of this domain. Each frame is a minute apart, the total animation covers the local time from 5:45 am 6/9/12 to 3:00 am 6/10/12. The misty field represents smoke, colored by concentration – higher concentrations are more opaque (linearly with concentration) and darker. The colors identifying the burning parts of the fire are the sensible heat fluxes released by the fire (see color bar to right, in W/m^-2). Darker browns are lower fluxes. The surface appears dark brown where fire has passed. Red arrows represent the near-surface horizontal wind speed (the length of the arrow) and direction (the arrows point downwind).

Esperanza Fire

Ignition: 10/26/07 near the Banning Pass, Cabazon, CA

Coen, J. L. and P. J. Riggan, 2014: Simulation and thermal imaging of the 2006 Esperanza wildfire in southern California: Application of a coupled weather-wildland fire model. International Journal of Wildland Fire 23, 755-770.

The Esperanza fire was ignited on the upwind edge of the San Jacinto mountains during dry, windy Santa Ana conditions. The following simulations show the winds near the surface and fire growth. This work was the first to simulate simultaneously the evolving meteorological flow, fire behavior, and fire-induced flow for a landscape-scale naturally evolving fire. It captures the rapid spread of the fire to the west-southwest driven by both Santa Ana winds and topographic effects, recreating the splitting of the fire, feathering at the leading edge, and other distinctive features. For infrared imagery of this fire from research aircraft, see The simulation was visualized using VAPOR producing the animations below.

Big Elk Fire

Ignition: 7/17/02 near Pinewood Springs, CO

Coen, J. L., 2005: Simulation of the Big Elk Fire using coupled atmosphere-fire modeling. International Journal of Wildland Fire 14, 49-59.

This simulation shows several hours in the early period of the Big Elk Fire, a 4400 acre wildfire ignited by a tailpipe. Fire behavior was extreme reflecting the extremely dry conditions throughout Colorado (including the lowest fuel moistures ever recorded in the area). Initial spread was rapid, moving up a south slope of ponderosa pine mixed with Douglas fir with crowning and torching into high density thin-stemmed lodgepole pine at upper elevations. The red field shows where the air was warmed at least 10 degrees by the heat released from the fire. The misty white field represent smoke, with denser areas representing higher concentrations. The arrows show the wind speed (shown by the length of the arrows – longer arrows being stronger winds) and direction near the surface. This case represents a relatively simple scenario, with no large-scale weather features – winds were driven primarily by solar heating of mountain slopes, producing weak afternoon upslope conditions during the active fire periods.

Extreme Weather Team
Wildfire Simulations

The Extreme Weather Team is using weather modeling to identify locations where combinations of factors produce extreme winds that can drive fast-spreading fires.

Through this research, we can help point out particularly vulnerable fire weather locations in California.