There are three tables of these metrics in the paper, the “ecological” (I would tend to think “biophysical” since are soils and watershed subsets of ecology? Interesting how terminology and associated thinking changes over time) table is posted at the end.
*****************************************
It seems to me that many wildfire/climate studies suffer from a scientific “streetlight effect”. They correlate area burned with weather conditions or other broadscale data and voila! make conclusions. I mentioned this in my comments to Steve’s post last week.
So I’d like to give a vigorous shout-out to the authors of this study. Vorster et al., Journal of Forestry 2023. It’s called “Metrics and Considerations for Evaluating How Forest Treatments Alter Wildfire Behavior and Effects.” Just logically, wouldn’t you want to have a common idea of exactly what “fuel treatment effectiveness” means.. effective in what sense? So these folks decided to dig in and actually talk about 1) different things it could mean, and 2) what data is available to address those things. This seems incredibly obvious and useful, and it’s only surprising to me that no one did it sooner.. otherwise researchers are basically analyzing past each other (like talking past each other only with research studies).
If we were being logical about scientific funding, we’d ask upfront “why are we asking this question? Whom exactly will this knowledge help?” If we identified a group say, land managers or fire suppression folks, we would ask them what information would help them. A serious problem to me is (and it has been in the past, and continues) researchers often define the problem their own way and then make a strange left turn at the end of the proposal and paper claiming that this information will be useful to people. Nowhere in the chain does there appear to be a reality check, or an advocate for the people whose work is claimed to have been helped.
What I really like about this paper is that this group of researchers decided to dive into all these questions and laid out a framework for a common language. It’s great that this effort was funded (by NIFA, CFRI , NSF and the Forest Service).
Interpretations of forest treatment effects in moderating fire impacts, and whether treatments are deemed effective, can vary widely depending on the audience. Treatment effects are objective measures of the influence on wildfire parameters, whereas effectiveness connotes a human judgment of this effect relative to a value-based goal. News media and the public often ascribe a treatment’s effectiveness to a few metrics: did treatments reduce the number of homes or high-value assets lost? Did treatments contain a fire? In contrast, firefighters may be focused on effectiveness through the lens of their ability to defend structures more efficiently or engage in suppression activities that otherwise would not have occurred (Jain et al. 2021). Meanwhile, land managers might be focused on soil impacts and associated short term watershed risks (i.e., debris flows, flooding, sedimentation, threats to drinking water supplies), as well as longer term ecosystem responses to wildfires, such as forest recovery. Interpretations of effectiveness may also change over time, as different outcomes become more or less important to the management goals of a given group.
Treatments can affect wildfires in a number of ways, including changing fire behavior and intensity, fire size, or footprint, altering impacts to ecological processes, facilitating incident operations, reducing suppression costs, and affecting the number of homes and structures lost (Agee and Skinner 2005; Kalies and Kent 2016; Thompson et al. 2013; Weatherspoon and Skinner 1996). However, quantifying the effect of treatments is complicated by the potentially minor influence relative to numerous other factors driving fire behavior, such as vegetation type, fuel arrangement and load, fire weather, topography, time of day of burning, and fire suppression efforts. In studies that look at these factors combined, the dominant influences on fire severity are often temperature, wind, and vegetation cover type (Birch et al. 2015; Evers et al. 2022; Martinson and Omi 2013; Prichard et al. 2020). Another challenge of quantifying the effect of treatments is the integration of data and processes operating at multiple spatial and temporal scales. Further, the scale of intended treatment effects varies widely. For example, some treatments are designed for local effects (e.g., defensible space around a home) whereas others may be designed for landscape effects (e.g., watershed protection). Fire behavior, typically measured as fire intensity, is commonly reduced following prescribed fire and in areas with previous fuels treatments or basal area reductions (Cansler et al. 2022; Kalies and Kent 2016; Prichard et al. 2020; Ritchie et al. 2007; Symons et al. 2008). Given these interacting factors, treatment effects are hard to quantify yet critical to understand as we are faced with growing costs and losses from wildfires (Bayham et al. 2022; Peterson et al. 2021; Steel et al. 2022; Wang et al. 2021) with increasing size and severity of these fires (Abatzoglou and Williams 2016; Stephens et al. 2014).
Evaluating treatment performance relative to stated or implicit objectives and how landscapes should be managed are topics of active research and discussion (Hessburg et al. 2021; Hood et al. 2022; McKinney et al. 2022; Sánchez et al. 2019). We add to these conversations by identifying a range of metrics to measure treatment effects on wildfire outcomes and considerations, challenges, and recommendations when evaluating and communicating about treatment effects. Here, we (1) present a framework to define metrics of treatment effects on wildfires that can be used to evaluate effectiveness of forest treatments for mitigating wildfire behavior and socioeconomic and ecological outcomes and (2) discuss important considerations and recommendations for evaluating these effects of treatments on fires. We draw on experience and literature primarily from the western United States and use the 2020 Cameron Peak Fire in Colorado, USA, as a case study to illustrate these considerations and evaluate the multiple modalities of treatment effects. Quantifying wildfire outcomes in treated areas provides better data-driven rationale for assessing effectiveness, which can aid in setting realistic expectations for how treatments will fare when confronted by extreme fire behavior and thus inform treatment prioritization and justify costs. This framework and these considerations can guide evaluations of treatment effects and assist managers, researchers, policy makers, and the general public in developing a common language for communicating about treatment effectiveness.
I really like this.. this is the opposite of the “streetlight effect”.. these folks actually are asking for the right info to address the question, including firefighter observations. And noting that data collected for different purposes may not be all that helpful for this purpose.
The treatment effects on wildfire metrics (Tables 1–3) fit into frameworks for characterizing cross-scale cumulative forest treatment impacts, such as the fuel management regime presented by Hood et al. (2022). Conversations about treatment effectiveness are prone to oversimplification and bias by highlighting certain cases to demonstrate a point while ignoring counterfactual evidence. A risk of having so many metrics of effects (Tables 1–3) is that every treatment can be deemed effective or ineffective post hoc by some metric rather than matching postfire metrics with pre-fire intentions for that treatment. We provide the following recommendations for advancing evaluations of treatment effects on wildfire:
Consider multiple treatment effects metrics and consider local context to give a more holistic view of treatment interactions with wildfire and to account for regional differences such as vegetation types, fire regimes, and management practices. Although it is important to align these metrics with the treatment objectives, additional metrics may reveal unintended consequences of treatments and can be just as valuable to adapting treatment methods.
Explore and communicate the range of treatment effect outcomes across burn conditions, treatment characteristics, spatial and temporal scales, and treatment effects metrics.
Improve documentation of suppression activities and firefighter observations, as they are critical for assessing many of the metrics and for attributing the effect of treatment or other drivers of fire behavior.
Improve treatment databases by providing more details and complete attribution of treatment prescriptions, adding historical treatments, providing regular updates, and working towards standardization across agencies so that data can be more readily used during wildfires by incident management teams and firefighters and so effects can be more accurately and efficiently measured.
Advance capabilities to evaluate treatment effects by improving methods for evaluating landscape-scale treatment effects, integrating diverse data streams, and targeting effects that have been difficult to quantify (e.g., watershed impacts, wildlife impacts, fire suppression and postfire recovery costs).
These recommendations can help to better characterize and communicate treatment effects on wildfire, but determining what is effective incorporates additional considerations, such as value systems, management goals, and treatment costs.
Here’s the figure with the ecological indicators..
Not to diminish the importance of this research, but this research seems to diminish the importance of this research:
“However, quantifying the effect of treatments is complicated by the potentially minor influence relative to numerous other factors driving fire behavior, such as vegetation type, fuel arrangement and load, fire weather, topography, time of day of burning, and fire suppression efforts. In studies that look at these factors combined, the dominant influences on fire severity are often temperature, wind, and vegetation cover type.” (Not treatment.)
I also wonder about this:
“Quantifying wildfire outcomes in treated areas provides better data-driven rationale for assessing effectiveness, which can aid in setting realistic expectations for how treatments will fare when confronted by extreme fire behavior and thus inform treatment prioritization and justify costs.”
I didn’t see the emphasis here on identifying extreme fire behavior (which I would expect to be where effectiveness would be lowest). And, as 2nd Law would no doubt point out, “justifying costs” of treatments would have to also consider the likelihood of a fire encountering a treatment.
“…the likelihood of a fire encountering a treatment.”
The likelihood is MUCH higher now, with fires now ranging from 100,000 acres to 1,000,000 acres, happening these days. As far as I know, no one has studied the probability involving these HUGE present-day fires, with fire suppression, firefighting tactics and human-ignitions affecting fire behavior. Otherwise, 2nd Law is just kicking the can down the road. That idea MIGHT have had some minuscule validity, 25 years ago, but not now.
On the other hand, it’s these “megafires” that I gather are least likely to be affected by any treatments. So shouldn’t the key question related to cost-effectiveness be the likelihood of a treatment being encountered by a fire that it could be expected to influence?
Of course, ad nauseum, such treatments aren’t designed to stop wildfires by themselves. There should be tons of examples within the Caldor Fire footprint of places where the fire didn’t crown out, burning with less intensity in areas that were “thinned from below”. When Google Maps becomes updated, I’ll search for some of those areas where I worked in the past. It’s more about fire resilience than prevention, as most Foresters know. People should also remember that most of the Placerville Ranger District had salvage sales happening from 1989 to 1993. People should take that massive fuels reduction into account before comparing it to other big fires, where there was less management, (like on the Dixie Fire, in the Plumas National Forest).
Additionally, the word games seem to continue about “active management” affecting fire behavior. We all know that “Overstory Removal” is considered “active management”, but it does not lead to fire safety. There was a LOT of management within the Caldor Fire footprint (including private timberlands), but some of that was not for fire safety or resilience.
People like to make such assumptions to push their own agendas. We should be focusing on site-specific conditions and treatments, instead of playing word games and pushing agendas.
Additionally, (almost) ALL treatments on the Placerville RD encountered the Caldor Fire. THAT is an important distinction. I could say similar about the other large recent wildfires.
HOWEVER…
I think the Caldor Fire would be a great one to study. The Placerville RD always had an aggressive timber program, being very responsive to getting things done, according to leadership’s ‘orders’. I’m trying to say that whatever the management direction was, P’ville always produced. The 70s and 80s there, clearcutting was King. There was also plenty of fire and insect salvage, as well as Hazard tree projects. Lots of “active management” in the history.
The private timber companies own a very large amount of checkerboarded lands, within the Ranger District’s boundary. Those were obviously heavily managed, with the current owner, SPI, preferring clearcutting. There isn’t many private homes on those private Sections.
What I am getting at is: With all that “active management”, why did the fire steadily march up, and over the mountains? It wasn’t exactly windy. (Yes, we’ve heard about the Trestle Project, but that area was very early in the incident.) I think it’s important to look at past projects and see what worked, and what didn’t. It’s an opportunity to look at a lot of historical projects, and the present-day conditions, post-fire. I would want such a study to be totally objective, seeking out the bad outcomes, along with the good ones.
One conclusion might be that even well-managed forests can burn at high intensity, during a serious drought. I think we have to admit that outcome is probable, and that we should take that into account when managing forests. How? That’s why we should do a study.
One of the key benefits of fuel treatments is increasing the potential for reduced fire severity. “Islands” of burned areas with lower fire severity surrounded by high severity burned areas would provide a variety of benefits for the greater megafire landscape including a source of seeds and microbiota, wildlife habitat preservation, and areas with reduced potential for soil erosion.
Given that:
i) fuel reduction treatments have unavoidable trade-offs such as adverse soil impacts, habitat degradation, carbon emissions, etc., and
ii) only a small fraction of acres treated to reduce fuels actually interact with wildfire.
Therefore, evaluating the overall effect of fuel treatments requires comparing:
a) the adverse trade-off effects on all treated acres, including effects on acres that did not subsequently burn, and
b) the beneficial effects on fire behavior on treated acres that did burn.
This is an important evaluation criterion that does not seem to be addressed by this study.
“i) fuel reduction treatments have unavoidable trade-offs such as adverse soil impacts, habitat degradation, carbon emissions, etc., and”
Modern thinning projects do MUCH less damage to the overall area, with proper mitigation of erosion potential. Foraging habitat for some rare birds, like owls and goshawks is actually improved, while strictly protecting their nesting habitats. Carbon emissions in thinning projects are much less than 80s-style logging. So many fewer log trucks, per day, per week, per month.
“ii) only a small fraction of acres treated to reduce fuels actually interact with wildfire.”
Thinning projects aren’t solely about wildfires. Your old mantra is not valid, in this reality of million acre wildfires. I have seen 5 acre spots that survived the fire, but died the next summer, due to cambium kill and bark beetles. The trees were green, into June, then there was 99% mortality in July.
If done reasonably, logging almost always improves habitat for a much wider and more diverse wildlife population. That is particularly true in the Douglas Fir Region, where closed canopies results in what used to be called a “green desert” — dark, no flowers or songbirds, deer, or elk. Just moss, mushrooms, fungi, and an occasional squirrel; compared to clearcuts with sunny skies, wildflowers, birdsong, butterflies, berries, deer, bear, and people. Based on experience and documentation.
How do you know that “only a small fraction of acres treated actually interact with wildfire”? Are you considering both prescribed burning and mechanical treatment or only mechanical treatment?
There is a number of assumptions here as to what fuel treatments they looked at but only 6.8% were found to have interacted with fire in the lifespan of the treatment: https://www.mdpi.com/1999-4907/7/10/237. I think that is an argument for better design and planning of treatments rather than not doing them at all.
Yes, this whole area of study is interesting..I’ll start a new post and we can add relevant studies there.
Along these lines:
20-year study confirms prescribed burning, forest thinning reduce risks of catastrophic wildfire
https://www.uniondemocrat.com/news/article_49296bda-b029-11ee-95f7-171698d01d03.html
Bob Z: “where closed canopies results in what used to be called a “green desert”
Most unmanaged D fir forests I saw were anything but green deserts, although closed canopy plantations certainly were. Industrial generally opted for CCing these green deserts rather than thinning them. Siuslaw NF started a robust thinning regime in early 1990’s. Cataract Thin, near Minerva OR, was the first replicate conducted under the auspices of a COPE study (Coastal Oregon Productivity and Enhancement). Universal fears were that most residual trees in a closed canopy thinning would blow down in the first heavy wind. Well… the 1996 storm (atmospheric lows comparable to infamous 1962 Columbus Day Storm) arrived almost immediately and caused VERY minimal damage. The results today, 30 yrs later, are quite astonishing. The “control site” (no thinning) remains a green desert. The thinned areas exploded with diversity.
Jim: We are in agreement. Whether thinning or clearcutting a closed canopy (which can also occur through natural seeding and not just plantations), logging quickly results in an “exploding diversity.”
I think you did a great job with the thinning projects on the Siuslaw, but even the thinned stands are not the same as the initial stands that were logged/burned/blew down. So the question now is whether they will be thinned again, clearcut, or allowed to grow old and die in their present condition as a thinned plantation.
For many years I lobbied against plantations established on a grid pattern; which was great for fiber yield and product quality, but only had ephemeral value as diverse wildlife habitat, and often eliminated “natural” firebreaks along streams and ridgelines — “cultural landscape patterns” — and created contiguous canopies capable of carry crown fires. Like the wildfire patterns we’re seeing now.