Abstract
Vegetation phenology-the seasonal timing and duration of vegetative phases-is controlled by spatiotemporally variable contributions of climatic and environmental factors plus additional potential influence from human management. We used land surface phenology derived from the Advanced Very High Resolution Radiometer and climate data to examine variability in vegetation productivity and phenological dates from 1989 to 2014 in the U.S. Northwestern Plains, a region with notable spatial heterogeneity in climate, vegetation, and land use. We first analyzed interannual trends in six phenological measures as a baseline. We then demonstrated how including annual-resolution predictors can provide more nuanced insights into measures of phenology between plant communities and across the ecoregion. Across the study area, higher annual precipitation increased both peak and season-long productivity. In contrast, higher mean annual temperatures tended to increase peak productivity but for the majority of the study area decreased season-long productivity. Annual precipitation and temperature had strong explanatory power for productivity-related phenology measures but predicted date-based measures poorly. We found that relationships between climate and phenology varied across the region and among plant communities and that factors such as recovery from disturbance and anthropogenic management also contributed in certain regions. In sum, phenological measures did not respond ubiquitously nor covary in their responses. Nonclimatic dynamics can decouple phenology from climate; therefore, analyses including only interannual trends should not assume climate alone drives patterns. For example, models of areas exhibiting greening or browning should account for climate, anthropogenic influence, and natural disturbances. Investigating multiple aspects of phenology to describe growing-season dynamics provides a richer understanding of spatiotemporal patterns that can be used for predicting ecosystem responses to future climates and land-use change. Such understanding allows for clearer interpretation of results for conservation, wildlife, and land management.