Land use is expected to affect species and vegetation and those effects should leave a signature in functional trait distributions. Or as we put it elsewhere, land use change causes changes to the ecological filters determining what traits are viable under particular conditions. Possession of such viable trait values is what specifies which species may be viable under such land use change (ignoring effects of biotic filters and immigration subsidies). Yet we know relatively little about the outcomes of the interaction between land use change and species traits on vegetation.
So our broad question was, do equivalent land use changes in different parts of the world (different abiotic and biotic environments) result in similar changes to trait distributions?
Commonly, one might ask whether the mean of some trait distribution changes, e.g. does the mean of potential heights of species decrease? You could also ask whether the trait diversity changes, e.g., whether the distribution becomes narrower or broader. Diversity changes might reflect patterns of convergence or divergence in traits. But another way to think of it is to consider the trait distributions, and to ask whether species are lost from particular parts of the distribution, or added at particular parts of the distribution. For instance, you might expect additions to tails of the distribution if a particular land use change causes the shift of a biotic or abiotic filter.
In this work we analysed 15 datasets (from five continents) of comparisons between intact forest and logged, or logged and grazed, or logged, grazed and abandoned forest. We measured changes to trait distributions, in location, dispersion and at the tails. And we assessed whether those changes are consistent between the tree canopy and understory, and between regions. In particular, the abiotic conditions resulting in gradients of productivity might be expected to affect the response of traits.
There is a wealth of individual results in the paper, but at the broad scale we found that:
- Species richness and trait diversity did not show coordinated reductions with land use change (as we’d predicted before).
- Interestingly, understorey trait changes were more consistent than canopy responses, reminding us to measure both components of vegetation.
- Pastures had more diverse leaf sizes and dispersal syndromes and lower diversity in heights, seed mass and pollination mechanisms compared with forest understories.
- When we found evidence of changes in trait variation with land use change, it could often be ascribed to additions or deletions from one end of the trait distribution, e.g., loss of the tallest heights or additions of really small seeds.
- Trait changes were relatively inconsistent across the regions, and those changes could not be explained by NPP, except for dispersal type.
What does this mean? Well, context probably plays a strong role. Broad classification of land use change may be insufficient to capture the processes involved in vegetation change. We still have a way to go towards generalized understanding of plant functional trait responses to land use change.