Ecosystem functioning and carbon or water cycling are affected strongly by climate. In turn, climate can experience change if vegetation changes, such that both negative feedbacks (climate change buffered by vegetation) and positive feedbacks (climate change enhances vegetation change and vice versa) can occur. Mechanisms acting on long timescales are fairly well understood and are being included into global climate models. At shorter time scales, such as seasonal (phenological) change or interannual variability, these feedbacks are not yet well understood. Examples include the effects of changing timing of leaf emergence, or extreme dry seasons in the Amazon, changing regional moisture transport and enhancing effects on vegetation. Seasonality of vegetation and interannual variability are also included in global climate models, but only in a simple, empirical way, mainly for temperature-induced phenology. Drought-induced phenology (e.g. Amazon dry season dependence) and interannual variability are generally not captured adequately in those models. In this project we will primarily investigate the magnitude of feedbacks between changed seasonal and interannual variability on one hand and climate on the other. In addition, we will contribute novel model components to represent especially drought-induced seasonal and interannual variability, starting from the hypothesis that productivity in one particular moment of time affects the carbon and water uptake capacities in subsequent periods, and that leaves are shedded if their productivity becoms negative, leading to variability in carbon productivity and water use over time. Preceding the model development we will test the sensitivity of existing coupled climate models to the representation of phenology. Towards the end of the project we will have the newly developed components coupled to the EC Earth system and evaluate them for three key regions: the Amazon, Europe and Siberia, where good and substantial data sets are available to validate. Finally we will use future climate projections in the coupled and decoupled model versions to assess changes in the potential strength and sign of feedbacks.