Dynamics of future seasonal temperature trends and extremes in Europe: a multi-model analysis from CMIP3

European surface temperatures have increased during the past decades. According to climate projections, this warming is expected to continue in future years under enhanced radiative forcings. In addition to the mean increase, changes in temperature variability are likely to occur, with more frequent extreme seasons such as those recently observed in the past decade (e.g. summer 2003, autumn 2006). Yet, most of the processes driving such long-term tendencies remain unidentified and unexplored. A particularly important issue is how changes in the atmospheric dynamics over the North-Atlantic and Europe (NAE) contribute to trends in both mean and extreme temperatures. The high occurrence of the positive phase of the North Atlantic Oscillation (NAO) in the 1980s-1990s suggested that the European warming could result from a re-organization in the main structures of the NAE dynamics. However, the 2000s have rather revealed an inconsistency between atmospheric circulation conditions and European temperatures. Here we investigate the relationship between sea-level pressure (SLP) and 2-m temperature (T2m) using a flow-analogues method applied over both observations and future projections. We use 13 models of the Third phase of the Coupled Model Intercomparison Project (CMIP3) over 1961-2000, 2046-2065 and 2081-2100, from which we extract seasonal subensembles with respect to their representation of observed SLP-T2m seasonal relationships. First we show that the distribution of SLP does not undergo major changes in future climate according to these seasonal model-ensembles, albeit a general decline of the variability is observed for all seasons. Then, using the flow-analogues, we conclude that the projected European warming appears disconnected from changes in the NAE dynamics. Only in winter a slight shift towards positive NAO conditions could partially contribute to the future temperature increase. Finally, a focus over unusually warm/cold seasons reveals that future temperature extremes should likely to be associated with similar atmospheric circulations as observed during recent ones.