BS&M Gets Fancy with Metacommunity Ecology

Guest Blogger: Shauhin Alavi

Advertisements

Last week, I instigated my fellow BS&Mers into veering a little off the beaten path, and convinced everyone to discuss a recent Ecography paper that took a metacommunity approach to studying shape variation in rodents. Metacommunities are sets of communities linked by the dispersal of more than one species. Community scale studies are largely lacking across evolutionary anthropology, and given that many extant primates (and probably fossil hominins) fit nicely in the metacommunity framework, this seemed like a good opportunity to explore the potential applications of this approach to anthropological questions.

The paper was met with mixed reviews within our group. Part of the problem was that not everyone was familiar with some of the metrics and analyses (and jargon) utilized in the metacommunity framework. Namely, community weighted means (CWM), principal coordinates of neighborhood matrices (PCNM), principal coordinates of phylogenetic structure (PCPS), and redundancy analysis (RDA). Admittedly, the paper was not written in the most accessible way. For anyone that might be interested in reading this paper, see the table at the end of this post for my best attempts at explaining these (at least for our group) commonly unfamiliar concepts.

One talking point amongst the BS&Mers was whether we actually learn anything new by zooming out to the metacommunity level. A few people (mostly from the B&S contingent of our group) argued that we don’t necessarily need the added complexity of communities to study how environment influences shape variation, and that this is easily done at the species level. I personally believe that looking at variation in community level indices (like CWM) allows us to look at very large scale evolutionary relationships that we might miss at the species level. By scaling up to metacommunities, we are acknowledging (after accounting for phylogeny) that all members of the ecological community are subject to the same environmental variables, and therefore the same selection pressures. Understanding how the mean trait value for the entire community varies with the environment gives us a starting point when trying to understand function.

Another interesting talking point amongst the BS&M crowd was whether or not the paper should have been rooted in some hypotheses. Some of us expressed dissatisfaction at how function was implied throughout the paper, yet there were no hypotheses given for how shape should vary with respect to environmental variables. I actually think that in this case, not having explicit expectations was a bit refreshing. I certainly wouldn’t have considered some of the traits that the authors found to vary significantly with environment.

One thought I was left with after reading this paper was that I wish the authors had included community weighted variance (CWV) in their analysis. Having a measure of variance would at least tell us if a particular trait is worth considering in the first place. It would have also been nice to have incorporated a model averaged phylogeny (sensu 10ktrees).

BS&M ended with a very productive brainstorming session about how to extend this framework to primatological and anthropological questions. I think it would be interesting to use remotely sensed measures of canopy structure to see how morphology varies across primate metacommunities. I also think it would be important to use this framework to uncover how neutral and niche processes shape primate communities. Others proposed extending this framework to studying how various measures of social organization may affect primate trait evolution at the metacommunity level. And of course, we couldn’t resist discussing how we might extend this framework to studying trait evolution across fossil hominin communities. A fair bit of the hominin discussion was how to surmount the palimpsest nature of the hominin record.

Overall the BS&M crew seemed receptive to the approach presented in the paper and acknowledged the importance of the complexity inherent at the metacommunity level.

Click through for the table of jargon and references! 

Metacommunity2

Sets of communities linked by the dispersal of more than one species
Community weighted means (CWM)3 CWM’s are simple to calculate, and translate to a functional description of the mean strategy in the community. Just calculate the mean trait value for each species in the community, weight said value by the relative abundance of each respective species, and sum across species.
Principal coordinates of neighborhood matrices (PCNM)4 PCNM is a method that explicitly accounts for spatial (or temporal) structure in community assemblages. Spatial structure in communities can be generated by both explanatory variables and autocorrelation in the species assemblage, and can act at global and local scales (distances). In other words, not all species are structured at the same scale, and response variables can act across multiple scales. PCNM seeks to model spatial structure in the community at all scales, identify the scales where species are structured, and decompose them into new orthogonal explanatory spatial variables that account for the effects of the underlying spatial structure.
Principal coordinates of phylogenetic structure (PCPS)5 PCPS is a method for examining how phylogenetic lineages vary across metacommunities. PCPS looks for phylogenetic signals at the metacommunity level, meaning that communities with similar phylogenetic structures also have similar average trait values (i.e. CWM). If there is a phylogenetic signal, then PCPS tests if the traits driving the signal are expressing trait divergence or trait convergence (as a result of ecological filtering), and tests for phylogenetic niche conservatism.
Redundancy analysis (RDA)6 RDA is just an extension of multiple regression, except the explanatory and response variables are data frames or matrices.

References
1) Maestri, R., Monteiro, L. R., Fornel, R., de Freitas, T. R. O., & Patterson, B. D. (2017). Geometric morphometrics meets metacommunity ecology: environment and lineage distribution affects spatial variation in shape. Ecography.

2) Leibold, M. A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., Hoopes, M. F., Holt, R. D., Shurin, J. B., Law, R., Tilman, D., & Loreau, M. (2004). The metacommunity concept: a framework for multi‐scale community ecology. Ecology letters, 7(7), 601-613.

3) Garnier, E., Cortez, J., Billès, G., Navas, M. L., Roumet, C., Debussche, M., Laurent, G., Blanchard, A., Aubry, D., Bellmann, A., & Neill, C. (2004). Plant functional markers capture ecosystem properties during secondary succession. Ecology, 85(9), 2630-2637.

4) Borcard, D., & Legendre, P. (2002). All-scale spatial analysis of ecological data by means of principal coordinates of neighbour matrices. Ecological Modelling, 153(1), 51-68.

5) Pillar, V. D., & Duarte, L. D. S. (2010). A framework for metacommunity analysis of phylogenetic structure. Ecology letters, 13(5), 587-596.

6) Van Den Wollenberg, A. L. (1977). Redundancy analysis an alternative for canonical correlation analysis. Psychometrika, 42(2), 207-219.

Shauhin Alavi is a PhD candidate in the Department of Anthropology at Rutgers University. He studies animal movement and nutritional ecology, and is currently studying cognitive foraging in orangutans.

Author: bonesstonesandmonkeys

The profile of the Bones, Stones, and Monkeys blog.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s