Wednesday, June 21, 2017

What do we mean when we talk about the niche?

The niche concept is a good example of an idea in ecology that is continually changing. It is probably the most important idea in ecology that no one has yet nailed down. As most histories of the niche mention, the niche has developed from its first mention by Grinnell (in 1917) to Hutchinson’s multi-dimensional niche space, to mechanistic descriptions of resource usage and R*s (from MacArthur’s warblers to Tilman’s algae). Its most recent incarnation can be found in what has been called modern coexistence theory, as first proposed by Peter Chesson in his seminal 2000 paper.

Chesson’s mathematical framework has come to dominate a lot of discussion amongst community ecologists, with good reason. It provides a clear way to understand stable coexistence amongst local populations in terms of their ability to recover from low densities, and further by noting that those low density growth rates are the outcome of two types of processes: those driven by fitness differences and those driven by stabilizing effects that reduce interspecific competition relative to intraspecific competition. Many of the different specific mechanisms of coexistence can be classified in terms of this framework of equalizing and stabilizing effects. “Niche” differences between species in this framework can be defined as those differences that increase negative intraspecific density dependence compared to interspecific effects. If, as a simplistic example, two plant species have different rooting depths and so access different depths of the water table, then this increases competition for water between similar root-depth conspecifics relative to interspecific competition. Thus, this is a niche difference. Extensions on modern niche theory have offered insights into everything from invasion success, restoration, and eco-phylogenetic analyses.

But it seems as though the rise of 'modern coexistence theory' is changing the language that ecologists use to discuss the niche concept. When Thomas Kuhn talks about paradigm shifts, he notes that it is not only theory that changes but also the worldview organized around a given idea. At least amongst community ecologists, it seems as though this had focused the discussion of the niche to an increasingly local scale, particularly in terms of stabilizing and equalizing terms measured as fixed quantities made under homogenous, local conditions. A recognition of the role of spatial and temporal conditions in altering these variables seems less common, compared to the direction of earlier, Hutchinsonian-type discussions of the niche.

Note that this was not Chesson's original definition, since he is explicit that: “The theoretical literature supports the concept that stable coexistence necessarily requires important ecological differences between species that we may think of as distinguishing their niches and that often involve tradeoffs, as discussed above. For the purpose of this review, niche space is conceived as having four axes: resources, predators (and other natural enemies), time, and space.”

On a recent manuscript, an editor commented that the term 'niche processes' shouldn't be used to refer to environmental filtering since (paraphrased) “when ecologists refer to niche processes, they are usually thinking of processes that constrain species’ abundances locally, confer an advantage on rare species...” But is it fair to say that this is the only thing we mean (or should mean) when we discuss niches? I’ve had discussions with other people who’ve had this kind of response – e.g., reviewers asking for simulations to be reframed from niches defined in terms of environmental tolerances to things that fit more clearly into equalizing and stabilizing terms. That is a good description of a stabilizing process, which is termed a 'niche difference' in the modern coexistence literature. But there is still a lot of grey space we have yet to address in terms of how to integrate (e.g.) the effects of the environment (including over larger scales) into local 'niche processes' or stabilizing effects. It's a subtle argument - that we can use the framework established by Chesson, but we should try to do so without dismissing too-quickly the concepts that don't fit easily within it. In addition, elsewhere the niche is still conceptualized in varying ways from comparative evolutionary biologists who talk about niche conservatism and mean the maintenance of ancestral trait values or environmental tolerances; to functional ecologists who may refer to multidimensional differences in trait space; to species distribution modellers who thinks of large-scale environmental correlates or physiological determinants of species’ distributions. 

The niche is probably the most fundamental, yet vaguely–defined and poorly understood idea in ecology. So, formalizing the definition and constraining it is a necessary idea. And modern coexistence theory has provided great deal of insight into local coexistence and thus has allowed for a better understanding of the niche concept. But there is also a need to be careful in how quickly and how much we restrict our discussion of the niche. It's possible to gain both the strengths of modern coexistence theory as well as appreciate its current limitations. Modern coexistence theory isn’t yet complete or sufficient. It’s currently easier to estimate stabilizing and equalizing terms from experimental data in which conditions are controlled and homogenous, and this can inadvertently focus future research and discussion on those types of conditions. Models which consider larger scale processes and the impacts of changing abiotic conditions through space in time exist, but across different literatures, and these need continued synthesis. There is still a need to understand how to most realistically incorporating and understand the complex interactions between multiple species (e.g. Levine et al. 2017). The application of modern coexistence theory to observational data in particular is still limited, and such data is essential when species are slow lived or experimentally unwieldy. Further, when quantities of interest (particularly traits or phylogenetic differences) contribute to both equalizing and stabilizing effects, its still not clear how to partition their contributions meaningfully.
-->

5 comments:

Brian said...

Nice post! I agree with you that niche is such a central concept in ecology yet so vague (and variously) defined. A couple of years ago several faculty at UMaine had a reading group on niche theory for a semester. We came to the same conclusions (important, not fully developed).

I for one am going to vote that coexistence theory, while important in its own right, is not going to be our last best description of niche theory itself.

I also continue to be amazed at how poorly the notion that niche theory doesn't have to be built on the idea of each species preferring a distinct part of the niche axes. Colwell & Fuentes 1975, Rosenzweig in several of his 1990s habitat selection papers, Keddy in several papers, Wisheu 1998 in a very decisive meta-analysis paper all point to the idea that our core concept of niche is probably off (briefly also reviewed in the 2006 TREE paper I had with Enquist, Weiher & Westoby). Namely they all suggest that what inclusive niches, or shared preference niches are far more common. Species don't prefer worse environments. Everybody wants to live in the nicest environment. But they end up actually being found in the best environment that they don't get outcompeted in. Once people grasp this idea they see it everywhere. But most niche theory still assumes distinct preferences for niches. We're going to have to recognize the shared/inclusive niche idea before we get very far in niche theory.

Caroline Tucker said...

Hi Brian - thanks, this is a helpful comment (I obviously need to re-read some of the references you mention). I agree with the idea that species ends up where they don't get outcompeted, and coexistence theory gets only part of the way to describing the dynamics involved.
I find myself wanting us to do a way better job of defining what exactly we mean by the niche, but also getting a bit annoyed when little subfields decide on their own what they mean, so that you get fragmented approaches all using the same word, and sometimes being quite certain that they are 'correct'... A glaring example is that both functional ecologists and evolutionary biologists talk about the niche, and both fields should have obvious connections since traits reflect evolutionary processes and yet the two groups don't mean the same thing at all.

Jeremy Fox said...

A quick technical comment:

"Further, when quantities of interest particularly traits or phylogenetic differences) contribute to both equalizing and stabilizing effects, its still not clear how to partition their contributions meaningfully."

This is a generic issue with any partition. A partition of A and B and a partition of the *underlying causes* of A and B are two different things. For instance, in evolutionary biology the Price equation partitions contributions of selection and transmission bias to directional evolutionary change in mean phenotype. But the same underlying cause (say, an environmental change) can generate both selection and transmission bias.

The deep issue here is that most partitions are partitions of effects, not causes. For instance, in modern coexistence theory, "equalizing mechanism" and "stabilizing mechanism" are sort of misnomers; they aren't "mechanisms" in any everyday sense. Rather, "equalizing mechanism" is "anything that has the effect of reducing density-independent interspecific variation in low-density growth rate". And "stabilizing mechanism" is "anything that makes low-density growth rates negatively frequency-dependent". If you have a partition of effects, it's hard, maybe impossible, to then turn around and say you also want a partition of the causes. I have an old post on this: https://dynamicecology.wordpress.com/2011/07/18/should-we-classify-mechanisms-by-their-causes-or-their-effects/.

In the context of the Price equation, multiple regression is the only generic tool I'm aware of for doing something like a partition of the causes underlying each term in the Price equation. See Fox & Harpole 2008 Ecology. But I don't know that that generalizes to modern coexistence theory.

Mark Vellend said...

My preference is to stop hoping the term niche will settle into a widely agreed upon meaning. I predict that will not happen in our lifetimes, or those of any of our students. So, I say we drop it and just say what we actually mean in any given situation. I try to use it only if a clear link with an existing use of the term is needed to communicate something. Curmudgeonly yours...

Caroline Tucker said...

That's how I feel about the term 'functional diversity'...
I don't disagree with that approach, though from a purely historical perspective, I think it would be sad to see the term 'niche' be discarded.