Search for a command to run...
Climate change is known to affect regional weather patterns and phenology; however, we lack understanding of how climate drives phenological change across local spatial gradients. This spatial variation is critical for determining whether subpopulations and metacommunities are changing in unison or diverging in phenology. Divergent responses could reduce synchrony both within species (disrupting gene flow among subpopulations) and among species (disrupting interspecific interactions in communities). We also lack understanding of phenological change in environments where life history events are frequently aseasonal, such as the tropical, arid, and semi-arid ecosystems that cover vast areas. Using a 33-year-long dataset spanning a 1,267-m semi-arid elevational gradient in the southwestern United States, we test whether flowering phenology diverged among subpopulations within species and among five communities comprising 590 species. Applying circular statistics to test for changes in year-round flowering, we show flowering has become earlier for all communities except at the highest elevations. However, flowering times shifted at different rates across elevations likely because of elevation-specific changes in temperature and precipitation, indicating diverging phenologies of neighboring communities. Subpopulations of individual species also diverged at mid-elevation but converged in phenology at high elevation. These changes in flowering phenology among communities and subpopulations are undetectable when data are pooled across the gradient. Furthermore, we show that nonlinear changes in flowering times over the 33-year record are obscured by traditional calculations of long-term trends. These findings reveal greater spatiotemporal complexity in phenological responses than previously recognized and indicate climate is driving phenological reshuffling across local spatial gradients.