[1] Long-term spring and autumn phenological observations from Switzerland and Burgundy (eastern France) as well as long-term Swiss monthly and seasonal temperature measurements offer a unique possibility to evaluate plant phenological variability and temperature impacts over the last 250 years. We compare Pearson correlation coefficients and linear moving window trends of two different lengths with a Bayesian correlation and model comparison approach. The latter is applied to calculate model probabilities, change-point probabilities, functional descriptions, and rates of change of three selected models with increasing complexity and temperature weights of single months. Both approaches, the moving window trends as well as the Bayesian analysis, detect major changes in long-term phenological and temperature time series at the end of the 20th century. Especially for summer temperatures since the 1980s, Bayesian model-averaged trends reveal a warming rate that increased from an almost zero rate of change to an unprecedented rate of change of 0.08°C/a in 2006. After 1900, temperature series of all seasons show positive model-averaged trends. In response to this temperature increase, the onset of phenology advanced significantly. We assess the linear dependence of phenological variability by a linear Pearson correlation approach. In addition we apply the Bayesian correlation to account for nonlinearities within the time series. Grape harvest dates show the highest Bayesian correlations with June temperatures of the current year. Spring phenological phases are influenced by May temperatures of the current year and summer temperatures of the preceding growing season. For future work we suggest testing increasingly complex time series models such as multiple change-point models.Citation: Schleip, C., T. Rutishauser, J. Luterbacher, and A. Menzel (2008), Time series modeling and central European temperature impact assessment of phenological records over the last 250 years,

A recent lengthening of the growing season in mid and higher latitudes of the northern hemisphere is reported as a clear indicator for climate change impacts. Using data from Germany (1951–2003) and Slovenia (1961–2004), we study whether changes in the start, end, and length of the growing season differ among four deciduous broad-leaved tree species and countries, how the changes are related to temperature changes, and what might be the confounding effects of an insect attack. The functional behaviour of the phenological and climatological time series and their trends are not analysed by linear regression, but by a new Bayesian approach taking into account different models for the functional description (one change-point, linear, constant models). We find advanced leaf unfolding in both countries with the same species order (oak > horse chestnut, beech, and birch). However, this advance is non linear over time and more apparent in Germany with clear change-points in the late 1970s, followed by marked advances (on average 3.67 days decade−1 in the 2000s). In Slovenia, we find a more gradual advance of onset dates (on average 0.8 days decade−1 in the 2000s). Leaf colouring of birch, beech, and oak has been slightly delayed in the last 3 decades, especially in Germany, however with no clear functional behaviour. Abrupt changes in leaf colouring dates of horse chestnut with recent advancing onset dates can be linked across countries to damage by a newly emerging pest, the horse chestnut leaf-miner (Cameraria ohridella). The lengthening of the growing season, more distinct in Germany than in Slovenia (on average 4.2 and 1.0 days decade−1 in the 2000s, respectively), exhibits the same species order in both countries (oak > birch > beech). Damage by horse chestnut leaf-miner leads to reduced lengthening (Germany) and drastic shortening (Slovenia) of the horse chestnut growing season (-12 days decade−1 in the 2000s). Advanced spring leaf unfolding and lengthening of the growing season of oak, beech and birch are highly significantly related to increasing March temperatures in both countries. Only beech and oak leaf unfolding in Germany, which is generally observed later in the year than that of the other two species, is more closely correlated with April temperatures, which comparably exhibit marked change-points at the end of the 1970s.

Running title (less than 45 characters):Bayesian analysis of seasonal change points 20 6. The Bayesian model approach allows not only the calculation of phenological change points during the year but also estimates the probability of changes occurring on a particular day. This method leads to a higher accuracy in estimating phenological events in the growing season, especially when handling low quality webcam data. 10

No abstract

This study presents an approach based on the Bayesian paradigm to identify and compare observed changes in the timing of phenological events in plants. Previous studies have been based mostly on linear trend analyses. Our comprehensive phenological dataset consists of long-term observational records (> 30 yr) within the 1951-1999 period across central Europe, from which we selected 2600 quality-checked records of 90 phenophases (mostly in spring and summer). We estimated the model probabilities and rates of change (trends) of 3 competing models: (1) constant (mean onset date), (2) linear (constant trend over time) and (3) change point (time-varying change). The change point model involves the selection of 2 linear segments which match at a particular time. The matching point is estimated by an examination of all possible breaks weighted by their respective change point probability. Generally we found more pronounced changes in maritime Western and Central Europe. The functional behaviour of all 2600 time series was best represented by the change point model (62%), followed by the linear model (24%); the constant model was the least preferred alternative. Therefore, non-linear phenological changes were by far the most commonly observed feature, especially in Western Europe. Regression analyses of change point model probabilities against geographic coordinates and altitude resulted in some significant negative regression coefficients with longitude; in contrast, the constant model probabilities increased with longitude. Even when differences between locations across Europe existed, an overall trend towards earlier flowering was determined at most locations. Multiple regressions confirmed that mean advancing trends in the 1990s were stronger in the northwestern part of the study area.