Monday, July 6, 2015

Can there be a periodic table of niches?


Are there a limited number of categories or groupings into which all niches can be classified?  I’ll 
admit that my first reaction is skepticism. For those ecologists who think of the similarities and generalities across systems, this may be an easier sell, compared to those who get caught up in the complexities of ecological systems. Classifying niches in this way is apparently a vision that distinguished ecologists have voiced: MacArthur: “I predict there will be erected a two- or three-way classification of organisms and their geometrical and temporal environments, this classification consuming most of the creative energy of ecologists.” 

From Winemiller et al. 2015.
Kirk O. Winemiller, Daniel B. Fitzgerald, Luke M. Bower, and Eric R. Pianka, takes on this rather ambitious goal in a new paper: “Functional traits, convergent evolution, and periodic tables of niches”. The periodic table, of course, is the foundation of chemistry – the predictive, descriptive arrangement of chemical elements based on their atomic number. Ecology may never achieve a similarly simple foundation, but the authors suggest that such a general classification of possible niches (and the species that are within them) is possible. A niche within a table would extend across taxa, habitats, and biomes, and would be seen repeatedly (i.e. periodically) across these.

Perhaps because they (and their reviewers) recognized the ambitious nature of this task, the paper helpfully acknowledges the reasons that a periodic table of niches might be a terrible idea right away. Unlike chemistry, ecology is strongly dependent on context, and stochasticity limits generality. The multi-dimensionality of the modern niche concept limits how few axes such a table could be reduced to. Evolution means that classifying a species’ niche is like trying to hit a moving target.

Examples of convergent evolution are common.
Still, even the chemical periodic table has some fuzzy matching going on – isotopes still group together under a given element, despite variation. “In the same way, elements can have different isotopes,…a niche category could have phenotypic variants but still have ecological properties or functions that are essentially the same.” In particular, the authors argue that convergent evolution has recreated particular suites of traits (niches) in different habitats and distantly related taxa. This has some connection to the idea that, perhaps, much like complex systems, complex arrays of traits may reoccur because they provide stability (e.g. are selected for).

How then to approach this task? Here the periodic table is rooted in a functional trait approach, where observable phenotypes capture niche information. The dimensions of the table are determined based on what must have been the result of long discussions and much difficulty, but the authors restricted themselves to five essential components: abiotic habitat, life history strategy, trophic position, defense mechanisms, and metabolic allocation strategies.
 
From Winemiller et al 2015.

From here, the use of various ordination approaches allow researchers to begin to identify species sharing trait combinations, allowing them to be classified within the table (see paper text for more detail). The combinations of these dimensions observed or unobserved in nature should inform us about the stability of certain niches, and perhaps provide predictions about which species to use for restoration approaches, which species may be invasive in a given system, or to predict shifting distributions.

If you had many different ecologists each develop a ‘periodic table of niches’, each table would be unique, evidence for how difficult drawing general principles and identifying the fundamental ecological dimensions is. Another person might consider dispersal its own dimension, for example, or dismiss defenses. This is especially true because the periodic table presented in this paper is phenomenological, lacking a clear connection with theoretical work, for example. The proof will be in its application and utility – do others adopt it, is it predictive, does it extend our understanding of the niche or improve applications? And I think there is a direction for functional ecology implicit in this work.

Their hearkening to MacArthur makes me wonder what MacArthur would think if he saw ecology today. His prediction that “there will be erected a two- or three-way classification of organisms and their geometrical and temporal environments, this classification consuming most of the creative energy of ecologists” falls short, but not in the ways he might have expected. Here then, is a classification system (and there have been other ideas and versions since his time), but even the 2 or 3 dimensions he generously offers aren't deemed nearly enough to capture ecological diversity. Is the simplicity that MacArthur mentions still considered possible? And I don't think the creative energy of ecologists has been focused on classifying niches in the way he mentions: it is more dispersed amongst topics, and human effects (climate change, fragmentation, habitat loss) have had a dominant role.

Winemiller, Kirk O., Fitzgerald, Daniel B., Bower, Luke M., Pianka, Eric R. 2015.  Functional traits, convergent evolution, and periodic tables of niches. Ecology Letters. DOI: 10.1111/ele.12462

4 comments:

florianhartig said...

I think the problem of ambiguity in approaches such as Winemiller et al. arises because anyone classifying niche dimensions ad hoc will ultimately do so against some implicit model of what's important for an organism, and as people differ in their view on that, so do they differ in their "periodic table".

I'm not so negative about the attempt in general though - in the end, a species' niche must be described by its functional properties. To agree on them, we would need to agree about an explicit (dynamic) model for what is important for organisms of a different class, e.g. plants (by explicit, I mean you could write it down, put it in a computer and make predictions). To work on this does not seem a waste of energy to me. If we have that, we have by definition the important parameters in which species differ, so we just need to fill the data gaps. We can look for trade-offs and maybe eliminate a few dimensions, but very likely we'd end up with far more than 3 axis, but I don't see a problem with that.

Caroline Tucker said...

"I think the problem of ambiguity in approaches such as Winemiller et al. arises because anyone classifying niche dimensions ad hoc will ultimately do so against some implicit model of what's important for an organism, and as people differ in their view on that, so do they differ in their "periodic table"."

I think that's so apt, and one of the underlying problem with attempts to make ecology more general. So either we approach this from a particular system or organism, or we try to make it so general as to be no longer useful.

Dr. Fox said...

What Florian said.

Joachim D. said...

The periodic table of chemical elements is not an arbitrary classification made by humans, in order to pigeonhole elements. It reflects an intrinsic property of matter. All atoms with the same number of protons in the nucleus (and hence electrons in the shell) are chemically identical and do therefore belong to the same element. Okay, isotopes do exist, but that's a matter of differing numbers of neutrons in the nucleus. As most chemistry is concerned with what the valence electrons do, these differences in isotopes can usually be neglected. When isotopes do get important, however, the nuclear physics/chemistry is none the fuzzier for it.

Likening phenotypes to isotopes suggests, to me, that the ultimate periodic table of niches will be one where each niche will be inhabited by exactly one species. Or did I get something wrong there?

P.S.: I remember an early periodic table of niches (from Pianka 1993, I guess) with primary producers, herbivores and carnivores as the main groups and a continuum of r-K-selection instead of discontinuous periods. "What about omnivores and flesh eating plants?", I thought, "and how do you quantify r-K-selection?" IMHO, the new categories seem to suffer from the same ills.