Recent changes in the onset of spring across northern hemisphere temperate land areas

Understanding atmosphere-biosphere interactions is a crucial part of efforts to improve global change simulation models, monitor variations in the growing season, and calculate the carbon budget. Satellite-derived information has a role to play in the development of global biospheric databases, but no comprehensive surface phenology network exists to calibrate these data. One type of phenological measure, the first appearance of spring foliage (commonly called the "green wave," "start of season," or "onset of spring"), is particularly important because it is crucial for accurate assessment of many processes, and is among the most sensitive plant-response measures of climate change.

Since satellite data are available for only several decades, and may not provide the details needed for many studies, and a global phenology network is not yet functional, alternatives must be employed to measure changes in the onset of spring at the global spatial-scale and century timescale. The Spring Indices (SI) phenology models have been developed to simulate the spring phenology of representative understory shrubs, using only daily maximum-minimum temperature data as input. They have been rigorously tested in a variety of regions and continents. While not capable of reproducing all the detailed information that would be obtained from multi-species phenology data, they can process weather data into a form where it can be applied as a baseline assessment of some aspects of a location’s phenological response over time.

In this study, a 1417-station daily maximum-minimum temperature database is used to assess the onset of spring across Northern Hemisphere temperate land areas over approximately the last 50 years. Linear trends for individual stations are evaluated during the 1961-2000 period, and hemisphere-wide area-weighted average yearly values and trends are computed for 1955-2002. SI first leaf dates, constituting a measure of change in early spring, are getting earlier in virtually all parts of the Northern Hemisphere, at an average rate of 1.2 days per decade. SI first bloom dates, representing change in later spring, show more diversity across the hemisphere, with small portions of central North America even showing dates getting later. Overall, however, most areas are still getting earlier, at an overall average rate of 1.0 days per decade. The length of the -2.2°C freeze period (driven primarily by earlier last freeze dates in spring that are getting earlier on average at a rate of 1.5 days per decade) is decreasing, with the most dramatic change occurring in east Asia. The time period between last freeze dates and the onset of plant growth is lengthening in east Asia and shortening in central Europe, which may indicate a lessening and aggravation of the potential for plant frost damage, respectively, in these two regions. Average monthly and seasonal temperature values are related to these springtime changes, but do not adequately account for all of the observed patterns.