Practice of Science Friday: Mind the Model/Adaptation Gap

Scientist Neil Carter of Michigan State University sets a motion-activated camera with a colleague in Nepal’s Chitwan National Park. Tigers in southern Nepal appear to be changing their habits so they can operate under cover of darkness and avoid coming into contact with humans, scientists said.

There seems to be a disciplinary adaptation gap between some climate scientists and “biodiversity” scientists on the one hand, and the disciplines involved in adaptation…fire science, plants, wildlife biologists and so on, on the other hand.  One of the gaps is that organisms adapt.  Critters and plants adapt, human beings adapt and we jointly adapt to each other.

There is the traditional genetic form of adaptation within species, and there are all kinds of adaptations beyond classical genetics.. behavioral, cultural, epigenetic.  And since these adaptations can’t be modeled (since most of them are unknown) to climate and biodiversity modelers, they don’t exist.  And yet.. in real life, and to certain disciplines, they do exist and are important.

Not to speak of humans.. so we have gaps like reading about crop improvement via new techniques like CRISPR, while at the same time climate modelers are predicting wheat yields in 2070.  GAP! Yet among science institutions, it doesn’t seem to be anyone’s job to notice gaps and attempt to fill them.  I think because while the CRISPR people would easily say “hey we have no clue what’s going to happen by then”, climate scientists seem to spend a great deal of time making predictions and mostly get published if the outcomes are bad… seemingly completely regardless of any characterization of the many uncertainties at the level the CRISPR people and farmers deal with.

Here’s an example of the kind of study I’m talking about..with regard to biodiversity predictions:

A new study by University of Arizona researchers presents detailed estimates of global extinction from climate change by 2070. By combining information on recent extinctions from climate change, rates of species movement and different projections of future climate, they estimate that one in three species of plants and animals may face extinction.

So here are a few papers that talk about wildlife adaptation:

First, mammals are becoming more active at night to avoid us. Here’s a link to an article by Michael Page, and here’s a link to the Science study.

Gaynor and her colleagues noticed animals were becoming more active at night to avoid human disturbances. They have now done a meta-analysis of 76 studies of 62 mammals all around the world. Almost all are shifting to the night to avoid us…
On the other hand, the shift is helping animals survive alongside humans. In Chitwan in Nepal, lots of tigers are able to live near people by being more active at night.

In this sense, the shift to the night may be good. “It’s a way to share space on an increasingly crowded planet,” says Gaynor. “We take the day and they take the night.” Thanks to their nocturnal ancestors, many mammals still have plenty of the characteristics needed to be more active at night, she says. And they are likely to be evolving to be even better at it.

“I would expect that this is an incredibly strong selective force,” says Kate Jones of University College London, who has shown that mammals only became active during the daytime after dinosaurs vanished.

Second, we’re finding out that habitats where critters are currently found might not be the only ones they can live in, maybe not even their preferred. This is in New Scientist by Isabelle Groc. Hopefully, there isn’t a paywall, it’s from 2018.

The story of California’s sea otters is not a one-off. Earlier this year, Silliman and his colleagues revealed a wider trend in a paper aptly titled “Are the ghosts of nature’s past haunting ecology today?“. As a result of conservation efforts, a variety of predators are reappearing in ecosystems they were pushed out of by hunting and development. “It is an exciting time for ecologists,” says Carswell, “because these species are coming back to these ecosystems from which they have been absent for many human generations and they are putting their house back in order.”

Mountain lions are another example. Unsurprisingly, we tend to associate them with mountains. But historical records show that in Patagonia they once lived in open grasslands. As sheep farming became established in South America, they were persecuted – along with their prey, a kind of llama called a guanaco. As a result, mountain lions survived only in the remote Andes away from humans. But in the past 20 years, sheep ranching has declined. “We started to see a change,” says Mark Elbroch from conservation society Panthera. “The mountain lions that had been removed from the open grassland began to come back out of the mountains at the same time as the guanaco was beginning to move back into the grassland.”

Third, critters are moving to places where they didn’t formerly live as far as we know. In this case, apparently without direct human assistance. This story is from Wudan Yan in High Country News (also 2018)

Otters were once unheard of in the Beartooths. In fact, there’s no evidence they’re native to this high alpine environment at all; their arrival appears to be part of the sweeping changes humans have brought to the plateau. In the 1960s, zoologists Donald Pattie and Nicolaas Verbeek spent years surveying the various mammals found in the Beartooths. They found creatures as small as dwarf shrews and as large as grizzly bears and mountain goats, but no otters. Continued but sporadic surveys done by field technicians and researchers at the Yellowstone Ecological Research Center in the 1990s yielded no sign of river otters, either. But for the last decade or so, there have been a few anecdotal reports from Cross, his colleagues, and some of the locals who frequent the plateau.

This of course raises philosophical issues as on this Yellowstone Ecological Research Center website

Are they “invasive species” in this alpine environment, impacting native carnivores like red foxes and American martens, or adaptive survivors seeking a climate refugium (not to mention food bonanza) at higher elevations?

I’m not suggesting we blow through wildlife habitat and ignore their needs. But when we hear predictions about the future, especially the distant future, even by scientists, I think we need to acknowledge that no one actually knows what will happen. And the people working at the interface of people and wildlife are actually the most knowledgeable about them, and how to work toward our continuing coexistence.

10 thoughts on “Practice of Science Friday: Mind the Model/Adaptation Gap”

  1. I would agree there are a lot of variables unknown to modelers trying to provide a view of what might be with species movement and survival, but they could be wrong in both directions – some species may not be quite as adaptable as thought while others may be more so. Plants, of course, may have a harder time adapting compared to generalist predators or herbivores. And animal species that rely on specific plants may have a harder go of it. Modelers are an adaptive lot though, and they will improve on their predictions over time. No matter how you cut it, the rate of extinctions are going to increase and there will be much human suffering due to climate change.

    • And so my question would be… extinctions are going to increase and there will be great human suffering… what is the point of writing papers that say 20% or70% of species will go extinct.. when we really don’t know? Perhaps the funding would be better spent actually helping species not go extinct.

  2. Tigers changing their habits to avoid humans is not evidence of adaptability to climate change. The return of a large predator that was extirpated due to human persecution, that is then relaxed, is not evidence of adaptability to climate change.

    Some animals will certainly be able to adapt to climate change. North American river otter is a species that is likely to adapt (unless you are aware of a model that suggests otherwise?). I don’t understand how raising a few anecdotal examples of species not threatened by climate change and adapting to climate change, or even 1,000 anecdotal examples, is somehow evidence that we shouldn’t trust the models of climate change effects on biodiversity.

    As the saying goes, all models are wrong, but some are useful. Sure, we don’t know exactly what will happen, but we know there is an extremely high probability that there are a lot of animals and plants that will not adapt to climate change. The models all point in the same direction. The only way to prove/refute which species will or will not survive is to administer the treatment (i.e., climate change) and see what happens. Without allowing them to go extinct, we need to use the best data available to make decisions. When it comes to federally threatened and endangered species, the rule is to err on the side of the species.

    • Ben, I didn’t say that they were adaptations to climate change. But they are adaptations. Animals and plants (and even viruses) do adapt. So far in expected and unexpected ways. I agree that we need to use the best data available and that ESA errs on the side of the species.

  3. There you go again: “… I think we need to acknowledge that no one actually knows what will happen. And the people working at the interface of people and wildlife are actually the most knowledgeable about them, and how to work toward our continuing coexistence.” I don’t understand the need to separate land management from modeling. Wouldn’t any of the “disciplines involved in adaptation” want to base their decisions – especially long-term planning decisions – on their best guess at “what will happen?” And why wouldn’t they want the information about that that models/modelers would provide (accompanied by relevant confidence intervals and caveats)? And why wouldn’t modelers attempt to account for potential “disruptive” change to the best of their ability? I don’t think the alternative of ignoring the future has a lot going for it.

    • Jon, I don’t see it as “ignoring the future”. And I don’t think we need to “separate land management from modeling”.. but I think at some point modelers cross the line of utility, and it is wise for land managers to hold these projections lightly.

      I think we can take model results and acknowledge their uncertainty (although this isn’t usually described well in climate models). But I think there’s a point at which they are not particularly useful we might be setting the wrong course by thinking they are more correct than they really are.

      There are other ways to deal with uncertain futures… scenarios with no regrets policies and so on. Our friends the economists and other decision scientists have been dealing with this for some time.
      Also important is the waste of time, talent and treasure (often government funded) for projections that are not particularly useful .. nowhere in our scientific institutions is there a discussion that says “there are so many uncertainties that this paper exceeds the threshold of believability.”

      • I don’t think this statement is true at all: “…nowhere in our scientific institutions is there a discussion that says “there are so many uncertainties that this paper exceeds the threshold of believability.”” That’s what the scientific method, statistics, peer review, top tier journals, and repeatability do.

        There is a considerable body of science finding that climate change will have significant adverse effects on biodiversity, and very little support otherwise. The fact that some species have a high adaptive capacity is not profound and does not in anyway suggest that the findings that biodiversity is at risk from climate change is incorrect. There are many species that have a low adaptive capacity to climate change and those are the ones the models are trying to hone in on so that we don’t waste time, talent and treasure worrying about them all.

        • Well, no, Ben, in my experience that’s not the way it always works. For example, for obvious reasons modelers review other modelers’ papers. What if there was open peer review where say, plant breeders, got to weigh in on how realistic it is to predict yields in 2070?

          I’m fine with “biodiversity (well not so fine because that’s an abstraction)” climate change could harm plants, animals, invertebrates, viruses and so on. I’d even think it might be useful to say… based on modeling we think “these classes of organisms might be more at risk in these places than in others.” (this often turns out pretty common-sensical, but not always, and this hypothesis can be verified through time).

          Every species is not at risk. Some are. Which ones? Can we know in advance? What characteristics do they share? What adaptations or lack thereof have surprised us? Do we see any patterns? To me those are all interesting and worthwhile questions that benefit from both modeling and the observations of the species experts (including geneticists).

          Claiming 1/3 will go extinct… not so much.

          • I wouldn’t argue against improving scientific review processes to get a more cross-disciplinary perspective. I don’t think the possible lack of that necessarily undermines the value of model projections to the degree you suggest.

            Which species? ESA says those at risk in the “foreseeable future,” based on the “best available scientific and commercial data available.” Meeting this requirement has to address what is foreseeable, and the effect of climate on habitat and populations must be a part of that. The “best available” climate data has been argued about in court, and models have been accepted based on the best available scientific opinions. (I’m glad to see you agree above that “we need to use the best data available” – even if it was not given cross-disciplinary review.)

          • Sharon, As the FS Chief with three names once wrote, land management decisions are made based on science, values, laws and politics. Just as modeling predicting global temperature increase over time is important for policy makers, so is modeling predicting global species decline. Among other things, this kind of information motivates voters to vote for politicians who align with their values concerning climate change. Regional/local modeling is exceptionally important too, as you say. Modeling of climate change impacts on individual tree species was conducted on the Rio Grande National Forest a few years back. That information helps foresters make decisions on which tree species to plant where. There is already a lot of general information about species characteristics that make them more or less vulnerable to disturbance, but as you say, there is still much to be done.


Leave a Comment