4. Discussion and conclusions
Most studies of the responses of temperate zone plants to temperature increases have shown advancing spring phases (Chmielewski and Rötzer, 2001; Chmielewski et al., 2011, 2004; Fitter and Fitter, 2002; Grab and Craparo, 2011; Legave and Clauzel, 2006; Menzel et al., 2006; Parmesan and Yohe, 2003; Parmesan, 2007; Root et al., 2003; Wolfe et al., 2005). Only a small number of studies, on a few species and locations, has shown the opposite – delayed spring phases in response to warming (Cook et al., 2012; Elloumi et al., 2013; Kozlov and Berlina, 2002; Legave et al., 2013; Yu et al., 2010). Our results suggest that this apparent contrast can very well be explained by differential responses by plants to warming during the chilling and forcing phases. According to our research hypothesis, temperate zone plant phenology at cold-winter locations should be determined primarily by temperatures during the forcing period, whereas at sites with warmer winters, temperatures during the chill accumulation period should increase in importance, with warming leading to a delay in spring phases. Our analysis of apricot phenology along a wide temperature gradient supports this hypothesis. While forcing temperatures clearly emerged as the dominant driver of spring phenology in Jiamusi and Shihezi – the coldest locations in our study – spring phase timing at Guiyang was mainly determined by temperatures during the chilling phase. Apricot trees in Beijing and Xi’an showed an intermediate response pattern. Observed responses of apricots to temperature closely resembled the response curves shown in Fig. 1, lending supportto our hypothesis. Earlier results that compared responses of three different combinations of tree species and climate (chestnuts in Beijing, China, cherries in Klein-Altendorf, Germany, and walnuts in Davis, California, United States) showed similar patterns (Luedeling et al., 2013a). Also there, conditions during the chilling phase were much more important in the warm winters of northern California, comparedto the colder winters inChina and Germany.Near the extreme end of the climatically suitable range of temperate trees, Elloumi et al. (2013) already observed net delays in pistachio phenology in response to warming in Tunisia. Many other studies that have related spring phase timing primarily to spring warming were conducted in cold-winter climates, such as in the United Kingdom (Amano et al., 2010), Germany (Chmielewski et al., 2004),thenortheastern United States (Wolfe et al., 2005), Europe (Chmielewski and Rötzer, 2001; Menzel and Fabian, 1999) and Japan (Primack et al., 2009). Net advances in spring phases oftemperate zone plants were, on average, 6.0 days ◦C−1 in these studies, with a rate of even 16 days ◦C−1 reported for Japanese apricots in Japan (Doi, 2007). These are faster rates than what has been reported from South Africa, where winters are warmer and phenology of two apple varieties advanced by only 4.2 and 2.4 days ◦C−1 (Grab and Craparo, 2011).