Studying the interactions between species in a community has been a central focus in community ecology in part because it can influence the community’s stability. Most prior studies have used, correlation-based approaches to understanding how species covary with other species through time within a community. However, such approaches provide an overly simplified statistical description of the structure of how species abundance measures jointly vary. To gain more insight of interspecies dependence structure, we used a “copula” approach which can measure the association between any two random variables in their extreme values (called “tail-dependence”). If smaller (or larger) values of two positively associated variables are more strongly correlated, the variables are said to have stronger lower (or upper) tail dependence. We used a 36 year (1983-2018) percent cover dataset for 15 grassland species commonly found in Jasper Ridge Biological Reserve to measure “tail dependence” in the community. With models and resampling techniques, we considered the implications of tail dependence for the variability through time of total community biomass.
Results/Conclusions
We found many interesting copula features in the data, which showed synchronous as well as compensatory temporal dynamics. For example, when invasive species Bromus hordeaceus was more abundant, the native species Plantago Erecta became rare, but when Bromus was scarce in the system Plantago showed a varied abundance. This aspect of the negative correlation between the species was available from the copula approach but not from standard approaches, and accurately reflected the likely causal relationship between the species: Bromus was the stronger competitor, limiting Plantago’s growth when present, not vice-versa. Secondly, we showed that the distribution through time of total community biomass was influenced by the tail dependence between the individual species. Via a novel resampling approach, we showed that the skewness of the total community biomass time series was influenced by the tail dependence relationships among individual species. Most commonly, the coefficient of variation of the total biomass time series is taken as a measure of community variability, but skewness of the same time series, which we showed is influenced by tail dependence between species, is associated with the “spikiness” of the total biomass time series. Thus skewness is another aspect of community variability. Our results illuminate a new aspect of synchronous versus compensatory dynamics and how they influence community variability.