The strength of flowering–temperature relationship and preseason length affect temperature sensitivity of first flowering date across space

Abstract

Temperature sensitivity (ST) of phenology, defined as the shift of phenophase with per unit change of preseason temperature, has been widely used to quantify phenological response to climate change. The spatial variation in ST of spring phenology has been well studied for several woody plants, but how to explain it became a challenge. Several hypotheses to explain the spatial variation in ST have been proposed, but studies examining all potential factors together were very limited. In this study, using first flowering date (FFD) data for five widespread woody plants at 47 stations in China and the other five species at 421 stations in central Europe during 1964–2014, we calculated ST by using the Sen's slope of FFD on mean preseason temperature for each species at each station. Subsequently, multiple regression analysis was applied to examining whether ST followed the geographical (latitudinal, longitudinal, or vertical) gradients. At last, four potential influencing factors of ST were tested by Spearman partial correlation analysis. We found that ST of FFD was higher at lower latitude for most species in China. However, in central Europe, the variation of ST did not follow the geographical gradient for most species. Only one species (Fraxinus excelsior) showed a latitudinal gradient, and one (Betula pendula) showed a longitudinal gradient. For most species in both regions, the strength of FFD–temperature relationship and the preseason length could account for the spatial variation of ST to a more considerable extent compared to preseason temperature variance and chilling conditions. Our results suggest that we need to consider the effects of multiple factors on phenological response to temperature when simulating future phenological change.