Plant communities in the tropical forests of Central and South America are famously diverse, and by some estimates are home to more than a third of all tree species on earth.
Two processes are thought to be particularly important for maintaining the high diversity in these regions: dispersal limitation, which reduces the range that species or their offspring can travel and allows unique ecological communities to develop in different regions; and negative density dependence, which reduces species survival when they are disproportionately common and helps prevent common species from driving rarer species extinct.
In order to better understand the effects of these two processes, we developed new analytical tools based on analysis of spatial moments as detailed in our recent paper.
Spatial moments describe variability in the structure of spatial patterns, much as classic variance is used to describe variability around a mean value in many statistical tests. A major advantage of these new tools is that they can be linked directly to analytical “spatial moment models” of population dynamics. This allowed us to generate quantitative estimates of the spatial scale of dispersal, and the strength and scale of density dependent interactions, based solely on survey data describing the location of individuals at a single moment in time.
We applied these methods to tree and liana (woody vine) species in the 50 hectare Forest Dynamics Plot on Barro Colorado Island, Panama. We first summarized the spatial distribution of each species using spatial moments, and then tested for significant associations between species spatial distributions and seven commonly measured functional traits: growth form, dispersal syndrome, tree canopy layer, adult stature, seed mass, wood density, and shade tolerance. Lastly, we related these significant associations to the parameters in our spatial moment models.
We found that these quantitative measurements of dispersal and density dependence were strongly associated with many commonly-measured species traits. For example, our results confirm that increases in tree size tend to be associated with decreased aggregation and better dispersal, and show that increases in seed mass, wood density, and shade tolerance are associated with less intense negative density dependence. Our results, therefore, suggest that simple trait measurements can provide insight into the processes that govern spatial structure and diversity in ecological communities.
We are optimistic that these methods will facilitate the process of scaling up current theoretical and empirical understanding, and hope that our study can serve as a guide for others who wish to incorporate these tools into their analyses.
Adam T. Clark, University of Minnesota, USA