Friday, February 3, 2017

When is the same trait not the same?

Different clades and traits yield similar grassland functional responses. 2016. Elisabeth J. Forrestel, Michael J. Donoghue,  Erika J. Edwards,  Walter Jetz,  Justin C. O. du Toite, and Melinda D. Smith. vol. 114 no. 4, 705–710, doi: 10.1073/pnas.1612909114

A potential benefit of trait-centric approaches is that they may provide a path to generality in community ecology. Functional traits affect growth, reproduction, and survival, and so--indirectly--should determine an organism's fitness; differences in functional traits may delineate niche differences. Since fitness is dependent on the environment, it is generally predicted that there should be strong and consistent trait–environment relationships. Species with drought-tolerant traits will be most dominant in low precipitation regions, etc, etc. Since productivity should also relate to fitness, there should be strong and consistent trait–ecosystem functioning relationships.

There are also quite general descriptions of species traits, and the life histories they imbue (e.g. the leaf economic spectrum), implying again that traits can yield general predictions about an organism's ecology. Still, as McIntyre et al. (1999) pointed out, "A significant advance in functional trait analysis could be achieved if individual studies provide explicit descriptions of their evolutionary and ecological context from a global perspective."

A new(ish) paper does a good job of illustrating this need. In Forrestel et al. the authors compare functional trait values across two different grassland systems, which share very similar environmental gradients and grass families present but entirely different geological and evolutionary histories. The North American and South African grasslands share similar growing season temperatures and the same precipitation gradient, hopefully allowing comparison between regions. They differ in grass species richness (62 grass species in SA and 35 in NA) and species identity (no overlapping species), but contain the same major lineages (Figure below).
From Forrestel et a. Phylogenetic turnover for major lineages along a
precipitation gradient differed between the 2 regions.
Mean annual precipitation (MAP) is well-established as an important selective factor and many studies show relationships between community trait values and MAP. The authors measured a long list of relevant traits, and also determined the above ground net primary productivity (ANPP) for sites in each grassland. When they calculated the community weighted mean value (CWM) of traits along the precipitation gradient, for 6 of the 11 traits measured region was a significant covariate (figure below). The context (region) determined the response of those traits to precipitation.
From Forrestel et al.
Further, different sets of traits were the best predictors of ANPP in NA versus SA. In SA, specific leaf area and stomatal pore index were the best predictors of ANPP, while in NA height and leaf area were. The upside was that for both regions, models of ANPP explained reasonable amounts of variation (48% for SA, 60% for NA).

It's an important message: plant traits matter, but how they matter is not necessarily straightforward or general without further context. The authors note, "Instead, even within a single grass clade, there are multiple evolutionary trajectories that can lead to alternative functional syndromes under a given precipitation regime" 


Hans Castorp said...

This is not a provlem of traits, but of the multicasual nature of communities. The soil properties of North America ans South Africa are HUGELY different, with SA much poorer in nutrients. Productivity of course depend both on nitrogen and precipitation, and therefore the difference is exactly what is expected. Climatological mdoel take into little account nitrogen- since they are global - but this is a weakness of such models

Hans Castorp said...

It is a wonderful article, but the abstract is the boring old question: "are communities shaped bu casual (historical) factors or by the nevironment"; I wuppose it will be repeated at least anothe million of times since it is not a scientifici but a phylosophical question which depends on your personal attitude

in the case of the papers the authors think that the differences between SA and NA are historical, when they are at least in part ecological (nitrogen): But they did not test for it and not considereninh the whole set of environmental factors and concluding that case is the ruler is the norm, not the exception

the set is not infinite, we know very well that there are at most 7-10 environemntal factors in plant that must be taken into account, and the belowground factors are among the most important (as any gardner can tell you, it is not enough to wateryour plants, you must also nourish them - but gardners are rarely listened by ecologists)

Caroline Tucker said...

What you suggest is definitely a reasonable explanation for the difference in results the authors saw. But even if there are *only* 7-10 environmental dimension of importance, the complexity of such a system is huge...