Scale-dependent trait filtering of woody diversity in North America

Background/Question/Methods

One of the key factors shaping biodiversity is environmental filtering that select species from a species pool according their environmental tolerance. We explored the role of environmental filtering across spatial scales, from community scale to continental scale. Geographical patterns of the significance in trait filtering, their environmental drivers, and their relation to species richness patterns were explored using USGS tree range maps and SALVIAS forest plots located in North America. We used four essential tree functional traits (maximum height, specific leaf area, seed mass, and wood density) and four climatic variables (temperature, precipitation, and their seasonality), and tested the significance in trait filtering using null models based on the random selection of species from species pool.

Results/Conclusions

We found strong non-random geographical patterns in means of trait values, but almost no evidence for the significant change in the variances of trait values compared to the species pool. This pattern was consistent across spatial scales. Moreover, neither environmental harshness nor lower species richness was associated with the relative reduction of the variance of trait values for most traits. Plot-scale correlations between climate and the mean or the variance of trait values often largely differed from the correlations observed at large-scale. These findings indicate that climate is a key factor selecting for mean trait values of woody plant communities, but its effect on the variance of trait values is much weaker. Moreover, the effect of environmental filters is scale-dependent, likely due to the biotic interactions acting at small spatial scales