Temporal asynchrony among interacting species is an important mechanism by which biodiversity can stabilize ecosystem function. But what drives synchrony in natural communities? Results from previous analyses are mixed: in some cases, environmental fluctuations seem to drive synchrony, while in others interspecific competition plays a stronger role. In all cases, however, inference on the drivers of community synchrony was indirect and influenced by biased null expectations. We take a direct approach by using stochastic, multi-species population models fit to demographic data from five semiarid plant communities to determine the importance of the three forces that drive community synchrony: environmental fluctuations, interspecific competition, and demographic stochasticity. Our main focus is on synchrony of species’ per capita growth rates, rather than abundance, because per capita growth rates represent the immediate response of species to competition and the environment. We used simulations with environmental fluctuations, interspecific competition, and demographic stochasticity turned on/off to determine the drivers of synchrony in our focal grasslands. We quantified species’ environmental responses as temporal random effects on intercepts in vital rate regressions for survival, growth, and recruitment and as the synchrony of species when simulated in monoculture.
Species' environmental responses were positively correlated in each community, ranging from Pearson’s ρ = 0.23 to ρ = 0.38, and synchrony of species in monoculture was strongly associated with synchrony of species in polyculture for growth rates (rank correlation = 1) and percent cover (rank correlation = 0.8). As a result, community-wide synchrony of per capita growth rates (φr), which ranges from 0 (perfect asynchrony) to 1 (perfect synchrony), was higher when simulated with environmental variation present (average φr = 0.62) than absent (average φr = 0.43). Removing interspecific competition, which was weak in all communities, did not affect synchrony, and demographic stochasticity only played a role in small populations. Our results suggest that all five communities represent the theoretical limiting case where demographic stochasticity and competition are weak, and community-wide synchrony is approximately equal to the synchrony of species' environmental responses. However, our results contrast with a recent analysis showing that synchrony was driven by interspecific competition rather than environmental fluctuations in recently assembled experimental communities. The difference between our studies suggests that strong transient competition early in community assembly can counteract the synchronizing effect of environmental fluctuations that dominates in older communities.