Weather extremes explain community variability better than species richness

Samantha Cady, Samuel D. Fuhlendorf, Craig A. Davis and Scott R. Loss, Natural Resource Ecology and Management, Oklahoma State University, Stillwater, OK, Barney Luttbeg, Integrative Biology, Oklahoma State University, Stillwater, OK, Caleb P. Roberts, Agronomy & Horticulture, University of Nebraska-Lincoln, Lincoln, NE

Background/Question/Methods
The search for a general relationship between biodiversity and ecological stability has continued for decades. Though many studies have shown evidence of a positive diversity-stability relationship (such that diverse communities tend to be more stable), empirical results have largely been inconsistent. We hypothesized that diversity may not be the only—or even the most important—factor regulating community stability. Here, we add context to the diversity-stability debate by not only evaluating the effect of species richness on stability, but also assessing relative contributions of other factors and non-linear relationships. Using 50 years of North American Breeding Bird Survey data, we quantified community change across diverse biomes in North America and investigated how avian community stability was affected by climate (both mean conditions and weather extremes) and species richness (including linear and non-linear relationships). We treated a site variability index as the response variable in a series of 11 linear models, which included climate, geographic, and diversity variables, and ranked models using Akaike information criterion.
Results/Conclusions
The best performing models explaining bird community variability contained minimum precipitation of the driest month and maximum temperature of the hottest month, each of which explained about 5% of the variation in the data. Specifically, very dry and very hot conditions were correlated with highly variable bird communities. We also found that the spatial autocorrelation of community variability is weak, but statistically significant (Moran's I=+0.19, p<0.0001), indicating that local conditions probably play a major role in community stability.