Spring onset variations and trends in the continental United States: past and regional assessment using temperature-based indices

Abstract

Phenological data are simple yet sensitive indicators of climate change impacts on ecosystems, but observations have not been made routinely or extensively enough to evaluate spatial and temporal patterns across most continents, including North America. As an alternative, many studies use weather-based algorithms to simulate specific phenological responses. Spring Indices (SI) are a set of complex phenological models that have been successfully applied to evaluate variations and trends in the onset of spring across the Northern Hemisphere's temperate regions. To date, SI models have been limited by only producing output in locations where both the plants' chilling and warmth requirements are met. Here, we develop an extended form of the SI (abbreviated SI-x) that expands their application into the subtropics by ignoring chilling requirements while still retaining the utility and accuracy of the original SI (now abbreviated SI-o).The validity of the new indices is tested, and regional SI anomalies are explored across the data-rich continental United States. SI-x variations from 1900 to 2010 show an abrupt and sustained delay in spring onset of about 4-8 d (around 1958) in parts of the Southeast and southern Great Plains, and a comparable advance of 4-8 d (around 1984) in parts of the northern Great Plains and the West. Atmospheric circulation anomalies, linked to large-scale modes of variability, exert modest but significant roles in the timing of spring onset across the United States on interannual and longer timescales. The SI-x are promising metrics for tracking spring onset variations and trends in mid-latitudes, relating them to relevant ecological, hydrological, and socioeconomic phenomena, and exploring connections between atmospheric drivers and seasonal timing.

The False Spring of 2012, Earliest in North American Record

“…Despite uncertainties associated with seasonal forecasts [National Research Council, 2010] Schwartz et al [2006] and Schwartz et al [2013] [cf. de Beurs and Henebry, 2005 several weeks, perhaps even a season, in advance (http:// www .esrl .noaa .gov/ psd/ csi/ events/ 2012/ marchheatwave/ anticipation .html).…”

“…Although phenological forecasting on comparable time scales is still in its infancy, several recent applications of phenological metrics, called phenometrics, are beginning to lay the groundwork for relating spring variations in the atmosphere to large-scale ecological effects [e.g., Cayan et al, 2001;Russell and Wallace, 2004;de Beurs and Henebry, 2008;Ault et al, 2011;McCabe et al, 2012;Schwartz et al, 2013]. These studies suggest that hemispheric patterns of variability in the atmosphere and ocean-which are sometimes predictable-play a fundamental role in governing the timing of spring throughout North America and elsewhere.…”

“…Most important, however, these models capture the timing of a major transition into spring growth shared by many plant species whose leaf-out and flowering vary primarily with temperature. Such "spring indices" (SIs) have been used to characterize spring onset through time and across a wide range of geographies [Schwartz et al, 2006;Schwartz et al, 2013]. Essentially, these phenometrics allow scientists to model what the Caprio network of observations would look like if it included every site where daily meteorological data are available from the last century.…”

“…Continental-scale variations in phenometrics occur on interannual to decadal time scales as responses to organized structures of variability in the ocean and atmosphere [Cayan et al, 2001;Russell and Wallace, 2004;de Beurs and Henebry, 2008;Ault et al, 2011;McCabe et al, 2012;Schwartz et al, 2013]. For example, twentieth century trends toward earlier spring in western North America (1.5 to 3 days per decade [Schwartz et al, 2006;Ault et al, 2011]) outpaced global averages (~1.2 days per decade [Schwartz et al, 2006]), while trends in the southeastern United States were toward later spring onset dates [Schwartz et al, 2013].…”

“…For example, twentieth century trends toward earlier spring in western North America (1.5 to 3 days per decade [Schwartz et al, 2006;Ault et al, 2011]) outpaced global averages (~1.2 days per decade [Schwartz et al, 2006]), while trends in the southeastern United States were toward later spring onset dates [Schwartz et al, 2013].…”