A. Carillo á P. M. Ruti á A. Navarra Storm tracks and zonal mean ¯ow variability: a comparison between observed and simulated data Received: 30 March 1999 / Accepted: 8 July 1999 Abstract A statistical technique is used to analyze the relation between monthly mean zonal ¯ow and storm tracks activity in the observations and numerical simu- lations (ECHAM4 model). The singular value decom- position technique (SVD) has been used to correlate storm tracks and monthly mean wintertime anomaly ®elds. The analysis has been performed on data from January 1980 to December 1989 (NMC analyses) and on an ensemble of AGCM simulations with prescribed SST for the same period, separately in the Euro-Atlantic and Paci®c sectors. We found good correlation between storm tracks activity and zonal ¯ow in both regions. In both data and simulations the dominant SVD modes show that the storm tracks spatial displacement is in conjunction with jet shifts in the same direction. Our analysis suggests that the model is highly sensitive to the equatorial ocean forcing. Although the model produces an excessive response to El NinÄo and La NinÄa phases, it shows good capability of simulating the dynamical connection between storm tracks and jet. 1 Introduction The existence of zonally con®ned regions extending in the Atlantic and Paci®c basins, which constitute pre- ferred paths for storms, has been documented by several authors (Blackmon 1976; Lau and Wallace 1979; Wal- lace et al. 1988). These regions have been identi®ed with areas of large 2.5±6 day height ®eld variance, and are referred to as storm tracks (Blackmon et al. 1977). The principal storm tracks lie downstream and slightly poleward of the vertical and horizontal shear zones as- sociated with the quasi permanent jet streams over Ja- pan and the eastern United States seaboard, and they appear to be rather highly organized on monthly time scales. Modeling studies suggest that the structure and transport characteristics of transient eddies along the storm tracks may be interpreted in terms of the life cy- cles of eastward propagating baroclinic waves (Simmons and Hoskins 1978; Pierrehumbert 1984; Chang and Orlanski 1993; Lee and Mak 1995; Whitaker and Dole 1995). Using ECMWF data, Hoskins and Valdes (1990) showed that, in winter conditions, the regions of high baroclinicity over the Northern Hemisphere are well correlated to regions of high eddy activity. They also pointed out that the time mean ¯ow was consistent with the existence of the storm track. The storm tracks spatial displacements, in conjunc- tion with similar jet shifts, are linked to some of the best- known monthly averaged teleconnection patterns (Lau 1988; Metz 1989). Moreover the dipolar western Paci®c and western Atlantic patterns in the monthly mean ¯ow are seen to be accompanied by marked changes in the intensity of the storm tracks over the western oceans, whereas the Paci®c North American and eastern At- lantic teleconnection patterns are coincident with the north-south displacements of the storm track axes over the eastern oceans. Such a behavior had been previously observed by Blackmon et al. (1977), who noted a southern displacement of the storm tracks during 1965/ 66 and 1969/70 winters, occurring together with a simi- lar shift of the jets. Wallace and Gutzler (1981) related this oscillation to the Eastern Atlantic and Paci®c North American teleconnection (PNA) patterns. The former teleconnection pattern has been also found to result from a two dimensional wave train dispersion excited by initial preturbations placed in the tropics and midlati- tudes (Simmons et al. 1983). The transient eddies may play a very important role in determining the extra- tropical response to Paci®c forcing as El NinÄo (Held et al. 1989); anomalous upper tropospheric transients were found to provide a strong feedback on the Climate Dynamics (2000) 16:219±228 Ó Springer-Verlag 2000 A. Carillo (&) á P. M. Ruti Gruppo di Dinamica Atmosferica e Oceanica, Dipartimento Ambiente ± ENEA, Rome, Italy E-mail: adriana@mantegna.casaccia.enea.it A. Navarra IMGA-CNR Bologna, Italy