Species’ phenological responses to climate change have large implications for future species distributions, trophic interactions, and ecosystem processes. Analyses of historical databases have shown that these responses are often species-specific and spatially variable. This variability makes predicting future responses more challenging. At the root of this challenge is the fundamental problem in ecology of how locally variable processes scale up to yield regional patterns. In this study, we show how hierarchical, process-based models of species phenological responses to climate may help address this challenge. Using long-term datasets (1953-2005) of daily climate and plant and animal phenological events from 150 sites across Japan and South Korea, we model plant species responses to climate based on underlying relationships with chilling and forcing temperatures.
Results/Conclusions
In agreement with previous studies, we found that spring phenology is advancing more rapidly for some taxa compared to others, but exhibits significant spatial variability. We use the process-based models to show how this spatial variability results in part from different chilling and forcing requirements for the plants in different parts of their ranges. Hierarchical models are used to translate these different local processes into realistic predictions of species dynamics across the region and at unmeasured locations. Finally, we compare these results to more standard analyses correlating events with mean monthly temperatures to contrast the predictive value of the more detailed models. Our results emphasize the utility of hierarchical, process-based models for understanding and forecasting regional changes in phenology for multiple taxa.
