Green-up of deciduous forest communities of northeastern North America in response to climate variation and climate change

Abstract

Temporal shifts in phenology are important biotic indicators of climate change. Satellite-derived Land Surface Phenology (LSP) offers data for the study of vegetation phenology at landscape to global spatial scales. However, the mechanisms of plant phenological responses to temperature are rarely considered at broad spatial scales, despite the potential improvements to spatiotemporal predictions. Geographical gradients in community species composition may also affect LSP spatially and temporally. Using a modified survival analysis, we reveal how weather and climate relate to physiological chilling and heating requirements and affect deciduous forest green-up in New England, USA over 9 years (2001–2009). While warm daily temperatures lead to earlier green-up of deciduous forests, chilling temperatures had a larger influence on green-up. We also found that the effects of community composition across the landscape were as important as the effects of weather. Greater oak dominance led to later green-up, while sites with more birch tended to have earlier green-up dates. Projection into the future (2046–2065) with statistically downscaled, bias corrected climate model output suggested advanced green-up (8–48 days) driven by higher heating and chilling accumulations, but green-up in coastal areas may be delayed due to reduced chilling accumulation. This study provides an innovative statistical method combining plant physiological mechanisms, topographic spatial heterogeneity, and species composition to predict how LSP responds to climate and weather variation and makes future projections.