Modeling leaf senescence of temperate deciduous tree species

The timing of phenological events, such as leaf emergence and senescence, is changing in temperate forests and has important implications for a wide range of ecosystem properties and processes. While a large body of research has focused on how climate change is altering the timing of leaf emergence, the complexity of ecophysiological processes and lack of information related to leaf senescence has limited progress in understanding how end-of-growing season processes in forested ecosystems will change in the coming decades. In this study we examine how variability in bioclimatic controls affect the timing of leaf senescence. To do this, we used a hierarchical Bayesian model (HBM) in association with two data sets: (1) the 30+ year record of field observations for 12 temperate deciduous tree species at Harvard Forest (dominated by Acer species, Quercus species, Betula species); and (2) a 5-year record of land surface phenology derived from remote sensing at 30 m spatial resolution covering New England. By estimating the HBM using the Harvard Forest field observations, we show that air temperature and photoperiod are the dominant factors that control the timing of leaf senescence, but that photoperiod exerts stronger control on senescence than air temperature across all 12 species. Species exhibiting the strongest dependence on photoperiod, especially subspecies of Acer, showed low inter-annual variation and no long-term trends in the timing of senescence. Subspecies of Quercus, which exhibited greater dependence on air temperature, showed statistically significant trends in the timing of senescence in response to systematic warming during fall that occurred over the 30-year record. At regional scale, high spatial resolution land surface phenology data from remote sensing show corroborating patterns over the New England region. Specifically, sub-regions of New England where Acer subspecies are more abundant had lower interannual variability in the timing of leaf senescence relative to areas dominated by Quercus species. Results from this study demonstrate that quantitative understanding of climate controls on the timing of leaf senescence is needed at the species-level to accurately forecast how the timing of senescence in temperate forests is likely to change in the future.