AMIP simulations from five GCMs (GFDL, ECMWF, NMC, MPI, MGO) were used in analysis of the surface air temperature, sea-level pressure and geopotential heights at 500 hPa over the Nothern Hemisphere. Four regions with specific large-scale patterns were selected for analysis. These are the Icelandic low and the Azore high which are mainly located over the ocean, the Siberian high and the patterns of the large-scale flow extending over the Eastern Europe and the Western Siberia.
The current atmospheric GCMs produce the annual cycle of the mean sea-level pressure with some bias which is particularly pronounced over the continents. The models also simulate the annual cycle of variability in the surface air temperature and the geopotential heights at 500 hPa.
The 1981 drought summer, the 1984/1985 cold winter and the 1982/1983
warm winter observed over the eastern Europe were selected for analysis.
The correlation between simulation and observation of the monthly mean
air temperature and sea-level pressure does not reveal a meaningful signal
in computed anomalous patterns over the considered region. It implies that
either GCMs are not able to simulate such anomalies in the middle latitudes,
or such anomalies are not predictable.