The ecological data revolution allows new avenues to examine phenological trends and provide insight on plant community response to climate change. This study examines the long-term phenological effects of climate change on Anemone acutiloba, Enemion biternatum, and Hydrophyllum virginianum, common forest perennials in the Midwestern and Eastern US that each express unique seasonal morphology. These plants flower in succession in spring and often co-occur, but whether phenological response would be congruent was unclear. Collection date and location were recorded for each species from two databases, iDigBio and the Consortium of Midwest Herbaria, and two physical herbaria, the Wartburg College Herbarium and the Ada Hayden Herbarium. Each sample was assigned a numerical code related to the reproductive stage (e.g. budding, flowering, or fruit set). Corresponding average temperature data for the three months prior to peak flowering (temp) of each species was downloaded from the National Oceanic Atmospheric Administration and linked to the phenological response based on the day of year and collection location for each species. Correlation matrices were developed for predictor variables (Day of Year, year, and temp), and multinomial logistic regressions were used to determine importance of predictor variables on phenological stages in response.
Results/Conclusions
Day of Year (DOY) of sample collection decreased over the years of the study (1985-2019), and temp increased over the same span in predictable congruence with climate warming. Further, the DOY decreased as temp increased for all models. Individual species demonstrated nuanced responses to logistic regression: Year was not an important predictor variable for Hydrophyllum, nor for earlier reproductive stages of Anemone and Enemion. For the earliest flowering Anemone, DOY and temp were important predictors; for Enemion, temp was not important for budding or full fruit stages; for later flowering Hydrophyllum, DOY and temp were important only for early reproductive stages. These nuances may be linked to organismal-level responses since variable vegetative phenology (wintergreen, ephemeral, or prolonged vegetative growth before flowering) among these species may buffer plants from fluctuations in temperature-triggered responses. Concurrently, nuance may be explained by landscape-scale responses such as variable effect size or direction of response at extremes of a species’ range.
These results indicate that large digital data sets can reveal patterns that cross regional boundaries, but in order to understand community-level response to climate change, we must adjust our scale of focus to zoom in on organismal-level responses and zoom out to the landscape scale for the full story.