Abstract
While commonplace in other parts of the world, long‐term and ongoing observations of the phenology of native tree species are rare in North America. We use 14 years of field survey data from the Hubbard Brook Experimental Forest to fit simple models of canopy phenology for three northern hardwood species, sugar maple ( Acer saccharum ), American beech ( Fagus grandifolia ), and yellow birch ( Betula alleghaniensis ). These models are then run with historical meteorological data to investigate potential climate change effects on phenology. Development and senescence are quantified using an index that ranges from 0 (dormant, no leaves) to 4 (full, green canopy). Sugar maple is the first species to leaf out in the spring, whereas American beech is the last species to drop its leaves in the fall. Across an elevational range from 250 to 825 m ASL, the onset of spring is delayed by 2.7±0.4 days for every 100 m increase in elevation, which is in reasonable agreement with Hopkin's law. More than 90% of the variation in spring canopy development, and just slightly less than 90% of the variation in autumn canopy senescence, is accounted for by a logistic model based on accumulated degree‐days. However, degree‐day based models fit to Hubbard Brook data appear to overestimate the rate at which spring development occurs at the more southerly Harvard Forest. Autumn senescence at the Harvard Forest can be predicted with reasonable accuracy in sugar maple but not American beech. Retrospective modeling using five decades (1957–2004) of Hubbard Brook daily mean temperature data suggests significant trends ( P ≤0.05) towards an earlier spring (e.g. sugar maple, rate of change=0.18 days earlier/yr), consistent with other studies documenting measurable climate change effects on the onset of spring in both North America and Europe. Our results also suggest that green canopy duration has increased by about 10 days (e.g. sugar maple, rate of change=0.21 days longer/yr) over the period of study.