One of the greatest uncertainties surrounding climate change predictions is the role of species interactions. This is because changes in any climate variable could alter the abundance of species directly or indirectly via interactions with neighbors. Notably, the effects of species interactions could both amplify or negate the direct effects of climate; for instance, the negative effects of warming on plant growth might be offset if warming has stronger negative effects on competitors or herbivores.
Here, we estimate the effects of two drivers on interaction coefficients in perennial plant communities at two elevations in Colorado, USA. Multiple plots within these alpine systems were experimentally exposed to increased temperatures and removal of the dominant species over several years. We adapted an existing theoretical framework to infer species-specific growth rates and pairwise interaction coefficients as a function of experimental treatments. By examining communities with and without the dominant species, we estimated the effect of dominant species on per capita interaction strengths of other species pairs---an example of a higher-order, multispecies interaction. Temperature treatments allowed us to estimate both the extent to which pairwise interaction strengths at each elevation vary with increasing temperature and how the relative importance of higher-order effects changes across this same gradient.
Results/Conclusions
In general, the responses of plants to the treatments were idiosyncratic. For instance, some species showed an increase in growth rates with experimental warming (compared to ambient temperatures) while also showing a depressed growth rate under combined warming and dominant species removal. Many species, however, showed no effect of warming or a slight depression of growth rate in response to warming. Both direct and higher-order species interactions varied in magnitude and direction among species and treatments. Some species which showed altered growth rates under warmed conditions also had more variable species interactions, sometimes switching from neutral to strongly competitive or facilitative. For example, one species that increased its growth rate under warming conditions also had stronger direct interactions with neighboring species; such stronger direct interactions may ultimately reduce its fitness. Despite the many species-specific effects of warming observed here, our study highlights the importance of including species interactions in predictions of communities under climate change. Moreover, by quantifying direct and higher-order species interactions, we can better estimate the total effects of increased temperature and varying competition on alpine plant communities.