The timing of autumn senescence is affected by the timing of spring phenology

Implications for predictive models

Trevor F. Keenan*, Andrew D. Richardson

*Corresponding author for this work

Research output: Contribution to journalArticle

115 Citations (Scopus)


Autumn senescence regulates multiple aspects of ecosystem function, along with associated feedbacks to the climate system. Despite its importance, current understanding of the drivers of senescence is limited, leading to a large spread in predictions of how the timing of senescence, and thus the length of the growing season, will change under future climate conditions. The most commonly held paradigm is that temperature and photoperiod are the primary controls, which suggests a future extension of the autumnal growing season as global temperatures rise. Here, using two decades of ground- and satellite-based observations of temperate deciduous forest phenology, we show that the timing of autumn senescence is correlated with the timing of spring budburst across the entire eastern United States. On a year-to-year basis, an earlier/later spring was associated with an earlier/later autumn senescence, both for individual species and at a regional scale. We use the observed relationship to develop a novel model of autumn phenology. In contrast to current phenology models, this model predicts that the potential response of autumn phenology to future climate change is strongly limited by the impact of climate change on spring phenology. Current models of autumn phenology therefore may overpredict future increases in the length of the growing season, with subsequent impacts for modeling future CO2 uptake and evapotranspiration.

Original languageEnglish
Pages (from-to)2634-2641
Number of pages8
JournalGlobal Change Biology
Issue number7
Publication statusPublished - 1 Jul 2015


  • Budburst
  • Climate change
  • Leaf senescence
  • Senescence model
  • Spring flushing
  • Tree phenology

Fingerprint Dive into the research topics of 'The timing of autumn senescence is affected by the timing of spring phenology: Implications for predictive models'. Together they form a unique fingerprint.

Cite this