Mapping range-wide budburst gradients in 22 temperate tree species across the eastern U.S. using quantile regression

Phenology plays a pivotal role in ecosystem functioning, and shifts in the timing of phenological events are a key indicator of climate change impacts on the biosphere. However, the sparseness of species-specific observations limits our ability to track phenological changes and their ecological consequences comprehensively. This study addresses this gap by developing a gradient-based spatial prediction approach to map range-wide budburst patterns of temperate tree species in the eastern U.S. Using sparse yet extensive observations from the USA National Phenology Network, I mapped range-wide phenological gradients for 22 species using quantile regression models driven by mean spring temperatures (1991–2020). Despite the simplicity of the modeling approach, the predicted gradients were highly consistent within each species between training and testing datasets, achieving prediction accuracy within 2 to 7 days. The predicted budburst gradients confirm established phenology-climate relationships across space, while revealing species-level variations. Gradients of phenological variability (inter-decile range) revealed decreasing plasticity in budburst timing toward colder climates, potentially due to local adaptation to narrower windows of frost-safe temperatures. This approach also supported quantifying phenological shifts with geographic factors, highlighting species-specific effects of latitude and elevation. Moreover, I demonstrated how interspecific comparisons of phenological surfaces in overlapping ranges could be made to help identify climate-driven mismatches, which may inform studies of species interactions. By integrating range-wide, intra- and interspecific variation across a climatic gradient, this study advances the understanding of phenological patterns in temperate tree species and provides an approach to tracking species-specific phenological changes in a geographically complete manner.