Modeling for Decisions IV. In Practice – Climate Change and the Rio Grande Cutthroat Trout (and Forest Planning)

It’s fortuitous that we have this recent example of how a court viewed a population model for an at-risk wildlife species that addresses climate change. The court included the usual caveat that, “Deference to the agency “is especially strong where the challenged decisions involve technical or scientific matters within the agency’s area of expertise.”

It is undisputed that the Service attempted to estimate the effects of climate change by using both “moderate” and “severe” predictions of expected effects, and that for the severe model, it “increased the risk function over time by 20 percent for the 2040 forecast and 40 percent for the 2080 forecast.” 79 Fed. Reg. at 59,147–48. The Plaintiffs take issue with the Service’s observation that the differences in results from the moderate and severe climate change models were “not particularly large.” Disbelieving that this could be a correct conclusion, the Plaintiffs thus suggest that the models “are driven by the Service’s assumption that climate change will have relatively little influence on the threats to individual Trout populations.” (# 76 at 26.)

But the Plaintiffs’ argument begs its own question, assuming that the Service’s models are infected by false preexisting assumptions that climate change effects with be minimal. It is essential to note that the Plaintiffs have not gotten “under the hood” of the Service’s models and pointed out any methodological, programming, or data entry flaws with them. Rather, the Plaintiffs simply argue that the models must be flawed because they produced results with which the Plaintiffs disagree. It may be that the models are flawed, but it may also be that that Plaintiffs’ (and the Service’s as of 2008) expectations about climate change effects are misplaced. Ultimately, it is the Plaintiffs’ burden to demonstrate an error in the Service’s actions, and simply pointing out that two different methodological approaches to calculating the effects of climate change in the far future produced two different results, one of which the Plaintiffs disagree with, does not suffice to carry that burden.

The threatened inquiry takes a longer-term view, asking whether the species might become endangered in a more distant future. But the threatened inquiry is necessarily closed-ended; once the Court has reached the endpoint of the “foreseeable future” (a term found in ESA, and defined recently by regulation) — which the parties here agree is 2080 — the Court’s ability to prognosticate must also come to an end. After 2080, nothing can be foreseen, all is simply speculation. So it is meaningless to ask whether a species will be threatened as of 2080, because it is impossible in 2080 to engage in the long-term future examination that the threatened analysis requires. By 2080, a species must have either reached the level of endangered and be at immediate risk of extinction, or it never will.

As the Plaintiffs observe, it appears that the Trout is on a “slide towards extinction.” (# 76 at 35.) But if the Service’s models are correct — and in the absence of a challenge, the Court must assume that they are — that slide will not be completed as of or immediately following 2080. At that time, there will still be 50 populations of Trout remaining, a number that the Service believes (and the Plaintiffs have not disputed) is enough to ensure the species’ survival through some indeterminate point in the future. What might become of those 50 populations after 2080 is beyond our ability to foresee; the curtain has come down and the movie has ended. We could attempt to speculate about what might happen thereafter — the 50 populations could persist, they could perish, new populations could be discovered, old habitats could become viable again — but speculation is all it would be. Our ability to predict what might happen has come to an end.

This analysis and decision actually has some important implications for forest planning (from the 2013 Rio Grande Cutthroat Trout Conservation Strategy).

Of the total 1,110 km (690 mi) of occupied habitat, 698 km (434 mi) (63 percent) are under Federal jurisdiction, with the majority (59 percent) occurring within National Forests (Alves et al. 2008).

Range-wide, a large proportion of the watershed conditions within the forests that have Rio Grande cutthroat trout are rated as “functioning at risk,” which means that they exhibit moderate geomorphic, hydrologic, and biotic integrity relative to their natural potential condition (USFS 2011)

Land management activities are currently practiced according to the Carson, Santa Fe, and Rio Grande National Forest Land and Resource Management Plans, and BLM Resource Management Plans. During scheduled revisions, the forests and BLM field offices will evaluate the current Land and Resource Management Plans and update as necessary to provide adequate protection for Rio Grande cutthroat trout with current best management practices. Land management activities that would result in the loss of habitat or cause a reduction in long-term habitat quality will be avoided.


If the trout is a warranted for listing (even if precluded by higher priorities), it is a “candidate” species under ESA.   The Planning Rule requires that forest plan components conserve candidate species (which under ESA means the same thing as recover). Since this decision that listing is no longer warranted was reversed, that should mean the species is again a candidate species.

Of course, national forests where the trout is found have been revising their forest plans after 2014, when listing was no longer considered warranted. Consequently, the Rio Grande cutthroat was considered for inclusion as a species of conservation concern. The Rio Grande National Forest has identified the species as an SCC in its final plan (currently in the objection period). The requirement for forest plans for SCC is for plan components to maintain a viable population.

Logically, a species that is warranted for listing should warrant greater protection than one that is not. So it’s possible that the Rio Grande will need to reconsider plan components in areas that are important to this species, or to at least document why this change doesn’t make a difference.

Which could bring us back to the modeling question – how does the Forest Service show that it is meeting the NFMA requirement to provide ecological conditions necessary for this species?   If there is a working population model for a species, then those factors that may be influenced by national forest management should be examined to determine how they could change as a result of forest plan decisions, and whether or how that could affect the model results.


7 thoughts on “Modeling for Decisions IV. In Practice – Climate Change and the Rio Grande Cutthroat Trout (and Forest Planning)”

  1. Jon – that the crux of the matter, isn’t it? Being able to identify which components of the population model are influenced by land management actions. I do not know the intricate workings of population models, but I wonder how easy identifying these components can be. If the task is relatively easy, then yes, your suggestion makes sense because the questions you pose also make sense.

  2. My observation is that the trend is towards less meaningful analysis of forest plans. Yet, they can’t legally ignore the best available scientific information, and that should include relevant state-of-the-art models. At a minimum, I think §219.3 requires them to explain why they did not use this methodology and chose another.

  3. Jon, I really appreciate you diving into this and reading the cases.

    I would only add that while I understand the judge’s point of view from the legal perspective,in the real world, there is no reason to think that models will predict any better for 2060 than for 2080. So to me that’s the difference between biological conservation and court conservation- people figure out what needs to be done, and move in the appropriate direction until they find out it’s not working and then pick another direction.

    Conservation by court case, on the other hand, in my experience, has led to keeping people from doing management activities of the type that that plaintiffs have a history of being against. Which may not be necessary, and is usually not sufficient, for recovery.

  4. I second Sharon’s comment. The way biologists (and other resource specialists) manage resources is based on their experience and the best available science. As science evolves, it is expressed in models, which can then be used in formal analyses that may be subject to court review. It doesn’t go the other way, with courts blessing certain models and then -ologists managing resources to fit those models. Maybe it does sometimes, but I think there would be the danger of trying to game the model rather than doing what is best for the species/resource. The models are always imperfect representations of the complicated interplay of site-specific factors that are known to affect a species or resource. If you don’t know anything about fish biology, it might be helpful to look at the variables in a model, but all of those variables (and more!) are already incorporated in fisheries management.

    That being said, I can imagine that reporting that directly references model outputs could be useful. For example, LRMP-required monitoring reports are often idiosyncratic and hard to compare between forests. Some kind of standardized monitoring (maybe based on USFWS models) of species of concern might be helpful.

    Thanks for posting this, Jon. The Court’s opinion is interesting and well-written.

    • Conor, coordinated long term monitoring is the Holy Grail of FS work. I remember when Chief Thomas tried to get this hands around this one.. it wiggled out of his grasp and continues to pose challenges. It’s probably worth another discussion on its own.

  5. It would be interesting to hear about examples of where environmental litigation has prevented actions that would contribute to species recovery (Sharon’s allegation). There are cases where there may be a benefit to the species, but the contribution to recovery is debatable because plaintiffs believe that that harm (usually short-term) outweighs any benefit (usually long-term). Or there may be competing interests of multiple at-risk species that limit what can be done. But otherwise?

    To Conor’s point (I think), as this case reminds us, the only thing that matters to a court is what has been documented. And models are actually a good way of documenting scientific facts and relationships. (But the users also have to document why and how they apply their models.)


    Serendipitously, I rediscovered this discussion of the Paiute cutthroat trout restoration in California. There the arguments against rotenone included harm from chemicals (including to another at-risk species), and it looks like it was a NEPA issue about effects in wilderness that decided the case of Californians for Alternatives to Toxics v. USFWS in 2011 (quoting from a later iteration of the case):

    “Essentially, the court held that, while the decision to reinstate the PCT to its historical range would contribute to the overall wilderness character, the manner in which Defendants chose to implement the project—use of a gasoline-powered Auger—failed to take into account the effect the project would have on other native species, specifically invertebrates, such as stone flies.”


Leave a Comment