The Many Phases of Phenology : Geographic and Inter-Specific Differences in Phenological Between-Year Variation

Abstract: As climate-driven changes in phenology are becoming more apparent, the need to quantify these changes is increasing. An important challenge in detecting phenological changes is that between-year variation in phenology is large. Between-year variation determines the statistical power of comparisons between contemporary and historical observations. For 44 plant species with different distributions across Sweden, geographicand inter-specific differences in between-year variation in different phenophases (budburst, flowering, ripe fruits, and leaf fall) was studied. I also modeled and evaluated theresponse of bud burst, and flowering, to temperature using three different temperature sum models over a latitudinal gradient. The data used was a sub-sample from a dataset collected by a Swedish nation-wide phenology network between 1873-1917. In agreement with previous studies, I show that early spring phases have a higher variability than phases occurring later in the season. However, the relation between onset and variation was not monotonically decreasing. In the geographical analyses, a unimodal relation between between-year variation and latitude was found, that is, the between-year variation decreased along the latitudinal gradient for early- and late season events, while it increased over latitude for summer events. These patterns are, to a great extent, reflections of patterns in air temperatures which is discussed using meteorological data from adjacent climate stations. Models were evaluated using Akaike's Information Criterion, and in 60% of all fits, the Spring warming CF2 model (SWCF2; the model with the least number of parameters) was selected as the best model to describe the data. For Sorbus aucuparia bud burst, in the two parameter model SWCF2, both parameters (threshold temperature andtemperature sum) correlated with latitude. However, future analysis using more locations and a wider span of species will be needed to understand the generality in these findings. In conclusion, future efforts to detect and quantify phenological changes need to consider differences in between-year phenological variability along geographical gradients and among species with different phenology.