Plant phenological anomalies in Germany and their relation to air temperature and NAO

This paper analyses long-term (1951-2000) phenological observations of 20 plant seasonal phases recorded within the phenological network of the German Weather Service in relation to climate data and NAO. Phenological inter-annual variability and temporal trends were determined by using mean anomaly curves for Germany. For all phases, the mean trends derived by this method are similar to German averages of linear trends of single station records. Trend analysis using anomaly curves appears to be effective in relating seasonal phenological trends to climate or satellite data: Spring and summer phenological anomalies, such as leaf unfolding and flowering of different species, strongly correlate with temperature of the preceding months (R² between 0.65 and 0.85, best one-variable model) and their onsets have advanced by 2.5 to 6.7 days per °C warmer spring. Fruit ripening of Sambucus nigra and Aesculus hippocastanum, key phenophases of early and mid autumn, correlate well with summer temperature (R² 0.74 and 0.84) and also advance by 6.5 and 3.8 days per °C (April-June). But the response of autumn colouring to warmer climate is more complex because two opposing factors influence autumn colouring dates. Higher spring and early summer temperatures advance leaf colouring, whereas warmer autumn temperatures delay leaf colouring. The percentage of variance explained by temperature (R² 0.22 to 0.51, best one-variable model) is less than for spring and summer phases. The length of the growing season is mainly increased by warmer springs (R² 0.48 to 0.64, best one-variable model) and lengthened by 2.4 to 3.5 days/°C (February-April). The North Atlantic Oscillation Index (NAO) of January-March correlates with spring phenological anomalies (R² 0.37 to 0.56, best one-variable model), summer to mid autumn phases respond to NAO of February-March (R² 0.23 to 0.36) (both negative correlations). Leaf colouring is delayed by higher NAO of (August) September (R² 0.10 to 0.18). NAO of January-February explains 0.41 to 0.44% of the variance of the length of the growing season.