The variability of Atlantic-European blocking as derived from long SLP time series

Blocking as an extreme event in European weather is addressed. A daily blocking index is applied to analyses of daily geopotential height at 500 hPa from 1949 to 1995. In order to allow for a reconstruction of these blocking time series back to 1880, daily SLP analyses are used to extract similar indices, which are optimally correlated to those based on 500 hPa. Daily indices are also determined for the geopotential height fields of 5 runs with the ECHAM3/T21 GCM, driven by monthly means of observed SST and sea ice boundaries for the years 1950-1994. These simulations differ only in their initial conditions, making it feasible to separate external variance (induced by SST and sea ice boundaries) from the atmospheric, model generated, internal variance. The index series reveal high variability both on time-scales of years to decades and interannual. Time series of spatial means show poor correlations not only between observations and simulations but also among the single runs of the model. Together with an analysis-of-variance (ANOVA) this leads to the conclusion, that Atlantic-European blocking is barely influenced by SST and sea ice distribution. The blocking index time series from observations is weak, yet significantly, negatively correlated to the NAO index time series favouring blocking situations during weak NAO periods.