As rapid climate change brings longer, hotter and often drier wildfire seasons to the western US, the need for large-scale adaptive forest management grows. Along with many regions of the western US, north-central Washington has experienced a steep rise in the incidence, size and severity of wildfires (View ArcGIS StoryMap). This region spans a diversity of forest types, management agencies, and communities that are being threatened by summer wildfires. Lessons learned about the effectiveness of fuel reduction treatments - both in wildfire outcomes and assisting in wildfire response - are therefore highly relevant to many fire-prone forests of the western United States.

Fire excluded vs. fire
Fire excluded forest outside of the 2021 Cedar Creek fire (left) and within the fire perimeter (right).


In this study, we evaluated large wildfires that occurred during the record-setting wildfire seasons of 2014 and 2015 in North Central Washington. During a multi-year drought, the 2014/2015 wildfires burned close to 400,000 ha in total, of which 212,000 ha were forested. Based on remotely sensed fire severity, 25% of affected forests burned at high severity and 37% burned at moderate severity. These burned areas were mostly stand-replacing with high tree mortality.

With warmer and drier growing seasons, burned areas with high stand replacement can pose challenges for tree regeneration and experience long-term conversion of forests to shrub steppe or grassland vegetation. It is therefore important to understand drivers of fire severity and if landscape strategies to increase forest resilience to future drought, forest insects, and wildfires.

Task 1: What were the main drivers of burn severity in large wildfire events of north Central Washington and how was fuel treatment effectiveness influenced by fire weather, biophysical environment, and topography?

In this task, we evaluated how the location and configuration of treatments can influence fire behavior and severity in the context of fire weather, topography and other factors. Our main motivation was to go beyond the classic evaluation of single large wildfire events and focus on a set of regional fires that impacted both the forests and communities of our region. We first started with the 2014 Carlton Complex and then expanded the study area to include all large wildfires in north-central Washington State.

Study Areas
Study area map including classified fire severity of forested areas within the 2014/2015 large wildfire events.
Box Plots
Box plots of the percentage of pixels in treatment units categorized as unburned and low severity by treatment type for early progression dates ranging from 7/15 to 7/18 and later progression dates ranging from 7/19 to 8/10.
Thin and underburn unit that was subsequently burned by the 2014 Carlton Complex fire with low fire severity.

For more information, see:

Prichard, S.J., Povak, N., Kennedy, M.C., and Peterson, D.W. 2020. Fuel treatment effectiveness following the 2014 Carlton Complex Fire in semi-arid forests of north-central Washington State. Ecological Applications. E02104. https://doi.org/10.1002/eap.2104.

Prichard, S.J., Povak, N., Gray, R.W., Griffey, V., Hessburg, P., Kerns, R., Ottmar, R., Salter, R.B. 2021. Landscape fuel treatments and wildland fire management strategies within recent large fire events. JFSP Final Report, Project Number 17-1-01. Joint Fire Science Program, Boise, ID.

For source data, metadata and presentations, see:

TASK 2a: How were past fuel treatments used in safe and effective response?

Throughout our project, we compiled the following lessons learned about how fuel treatments were used to provide safe and effective fire fighting response.

2014 Carlton Complex Fires

Carlton Complex
Burned area reflectance classification, based on Relavized Burn Ratio, around the large, unburned island within the 2014 Carlton Complex. A network of fuel reduction treatments mitigated fire behavior and effects on the N flank of the wildfires and provided defensible space for fighting after early progression days.

A total of 339 recent fuel treatments, including thinning, prescribed underburning, pile burning (including combinations of these treatments within overlapping records) were involved in the Carlton Complex fires. Of these, there were variable effects that were highly dependent on day of burn conditions and landscape position. Specifically, where fuel treatments were aligned with the massive heat column and wind driven fires on July 17, 2014, treatment areas generally had high levels of scorch and severity. However, some treatments were located on leeward slopes and as such were protected from the wind-driven flaming fire front. One of these was a 809-ha Leecher Mountain underburn that was completed in 2001. The treatment unit is located on a steep SW facing slope. As the flaming front burned under sustained 48-km winds from the NW, this older prescribed burn unit supported a flanking fire that backed down the slope toward Benson Creek through reduced fuels. Following the strong winds and extreme fire behavior of July 17 and 18th, firefighting crews were able to use a network of previous fuel treatments along Benson Creek Rd and the 2012 Leecher wildfire to successfully corral the fire. Several factors, including the original flanking fire that backed down a previous fuel treatment combined with subsequent firefighting, resulted in a large swath of unburned forest exists between the north and south portions of the fire.


2015 North Star – Tonasket Ranger District

Primed by regional drought and hot, windy conditions, The 2015 North Star Fire had explosive fire behavior with 64-80 kph winds that contribute to 19-km run through mixed conifer forests in a single day. A total of 67,000 ha burned on the Colville Reservation, representing the largest fire on Trust Lands in recent history. Tod Camm, the FMO for neighboring Tonasket RD to the north, started his first day of his new job on the day before the North Star started. He and local fire crews called for firefighting resources but ended up needing to conduct some proactive burnouts before the North Star burned onto USFS lands.

North Star
Left photo: example of a unit that was thinned and recently underburned (Lost 108) that was used in burnout operations in advance of the 2015 North Star Fire. Right photo: example of a thinned unit that was a planned prescribed burn that was successfully burned as a summer prescribed burn in advance of the 2015 North Star Fire.

Overall, the total area of the burnout in advance of the North Star Fire was around 1600 ha. Fire severity mapping shows low severity throughout. Shawn Plank visited one area and witnessed the burnout as a mild backing fire with 0.6-m flame lengths. Even though the crews initially wanted to jump on the spots, they strategically used the roads as containment lines and practiced patience as they hand-ignited units for light backing fire effects rather than letting burnouts rip uphill from the road. The network of roads along with several past fuel treatments were used strategically as containment lines to corral the North Star to the east.

Fuel treatments
Network of fuel treatments near the Aeneas Valley, Washington that were used in firefighting operations and helped to corral the wildfires to the east.

Just south of these treatments, an untreated area designated as a wildlife corridor had high post-fire mortality. The understory is dominated by fireweed and is already regenerating with lodgepole pine seedlings. Without a subsequent fire, this ponderosa pine site likely will transition to a lodgepole pine-dominated forest. As part of the site visit, we discussed the implications of this management trajectory of dense, mixed conifer forest that historically had a low to mixed severity fire regime burning severely in a wildfire and then regenerating as a lodgepole pine forest with a high severity fire regime.

This panorama was taken to capture the transition of a high-intensity wildfire moving into the treatment area, left to right. A large network of roads and previously treated areas provided more decision space for local crews to reinforce contingency lines with strategic burnout operations.
Post-fire tree mortality
Post-fire tree mortality in an untreated forest within the 2015 North Star Fire area, Tonasket Ranger District.

2018 Boyds Fire

Several WA DNR and WDFW managers highlighted this case of fuel treatment effectiveness, and details are also provided in a January 2019 “Ear to the Ground” media story by the WA DNR. The 2018 Boyds Fire started in the Sherman Creek Wildlife Area, located across the Columbia River from the town of Kettle Falls, WA. As it quickly spread, the wildfire threatened some homes, power lines and a nearby mill. Previous forest restoration treatments by the Washington Department of Fish and Wildlife were used to assist firefighting operations. Recent thinning and prescribed underburning treatments allowed firefighters to construct firelines that successfully contained the fire during hot, dry and windy burning conditions.

Sherman Creek Wildlife Area treatments
Location of the Sherman Creek Wildlife Area treatments (pink) that assisted in safe and effective response to the 2018 Boyd Fire. Source: https://washingtondnr.wordpress.com/author/washingtondnr

2018 McLeod Fire

McLeod Fire
Past fuel treatments on the eastern edge of the 2018 McLeod Fire were used as the basis of burnout operations along the Eightmile Rd near Winthrop, WA.

During the summer of 2018, firefighting teams on the McLeod Fire posted on social media that firefighting options along the Eightmile Road were aided by past fuel reduction treatments and contrasted the situation along Twisp River Rd near the 2018 Crescent Mountain Fire, which was being fought concurrently and was lined with fire-excluded, untreated forests. Treatments such as those shown below allowed for road access and burnout opportunities on the east Flank of the McLeod Fire.


2021 Cedar Creek Fire

The 2021 Cedar Creek Fire started along the Hwy 20 corridor near Mazama, WA. Over the next week, the fire burned along a long set of interconnected ridges south toward Winthrop, WA. The Virginia Ridge Timber Sale was commercially thinned in 2019. In VR Unit 1, slash removal with machine piling and some hand piling was conducted along Wolf Creek Rd. In VR Unit 2, logging slash was left dispersed on site in addition to large landing piles. The WA DNR planned to burn piles and was looking into options for underburning. However, the Cedar Creek Fire burned the units first with variable impacts on firefighting crews. Many large landing piles were dismantled by crews and moved to a safer location. Burnout operations were successfully conducted along Wolf Creek Rd, and the recent thinning and slash removal provided crews with options for conducting a light underburn that resembles a spring or fall prescribed burn within that unit. However, dispersed logging slash on VR Unit 2 presented a different situation in which heavy slash contributed to high severity effects and 100% tree mortality. A combination of burnout and wildfire burned through that unit and resulted in relatively high tree mortality. However, firefighting crews were able to corral the fire away from the highly valued Sun Mountain Lodge and into more remote locations, using the two Virginia Ridge units as part of their firefighting operations.

Pre/post Cedar Creek fire
Pre- and post-wildfire photos taken within Virginia Ridge Unit 2. This unit had particularly high dispersed logging slash prior to the 2021 Cedar Creek fire and had 100% post-fire tree mortality.

TASK 2b: How did firelines perform in the 2014/2015 wildfire seasons, and what are best methods for calculating metrics of fireline effectiveness?

As wildfires increase in size and severity in many regions of the western US, the economic and environmental costs of fighting fires are becoming major questions to guide future wildfire management. In this task, we evaluated a method proposed by Gannon et al. (2020) to calculate metrics of fireline effectiveness. The four metrics proposed by Gannon et al. (2020) offer a way to summarize how much investment was made in fireline construction, how much fireline was actually engaged in the fire (i.e., burned over or held) and of the line that was engaged, how much held. An overall ratio of the length of held line to the total fireline constructed provides an overall assessment of fireline effectiveness, considering how much fireline was both engaged and held.

Overall, a total of 2205 km of fireline was evaluated in this analysis, of which 71% was bulldozer line (either as new dozer line or bulldozed old roads), 21% was existing road used as completed line, and 8% was hand line. Not counting road as completed lines, at total of 1742 km of dozer lines or hand lines were constructed in these wildfires. Fireline effectiveness was strongly associated with fire severity with the highest effectiveness in unburned and low severity areas vs. lower effectiveness in areas that burned at moderate to high severity.

We are currently drafting a manuscript on project findings. Preliminary results are presented in our JFSP Final Report:

Prichard, S.J., Povak, N., Gray, R.W., Griffey, V., Hessburg, P., Kerns, R., Ottmar, R., Salter, R.B. 2021. Landscape fuel treatments and wildland fire management strategies within recent large fire events. JFSP Final Report, Project Number 17-1-01. Joint Fire Science Program, Boise, ID.

TASK 3: Can strategic configurations of fuel treatments across fire-prone forested landscapes be effective at reducing future burn probability and predicted flame lengths?

In our final task, we used the Interagency Fuel Treatment Decision Support System (IFTDSS) to evaluate fuel treatment scenarios that can guide future implementation of the Cohesive Strategy in north central Washington State. We focused on two priority landscapes identified within the Washington Department of Natural Resource's 20-year Forest Health Strategic Plan for eastern Washington State, including Methow Valley and Republic. Strategic fuel treatment layers were created by classifying the LANDFIRE Biophysical Settings (BPS) layer into units with the same vegetation type, topographic position, and aspect classes. Because our study areas are dominated by low elevation mixed conifer forests, we applied a consistent fuel reduction treatment (forest thinning followed by a combination of pile burning and prescribed underburning) to randomly selected treatment units across scenario landscapes, ranging in treatment intensity from 10% to 60%.

The new IFTDSS makes running fire simulations very straightforward. We developed an overview of our approach with step-by-step instructions.

Comparisons of treated landscapes are illustrated below:

Burn Probabilty by Percent Treated

Burn Probabilty by Percent Treated

Conditional Flame Length by Percent Treated

Conditional Flame Length by Percent Treated

Here are those simulations in video form:

For source data and fireline effective maps, see: