Seasonal and intraseasonal variability of teleconnection pattern modes

We examine the seasonal and intraseasonal variability of large-scale teleconnection pattern modes based both on barotropic instability theory and on empirical operators derived from observational data. For each month of the year we calculate the barotropic instability matrix for the zonally varying monthly averaged 300mb observed streamfunction fields. As well, we fit a linear first order stochastic model with temporal white noise forcing to the twice-daily observed data to form empirical matrix operators for each month. Time-dependent propagators covering the annual cycle are constructed from the monthly matrices by linear interpolation for both cases. The eigenvalues and some of the dominant eigenvectors of the barotropic and empirical propagators, termed finite-time normal modes (FTNMs) are calculated. The FTNMs exhibit intraseasonal variability in their structures, as well as longer period variations, and their amplification rates vary with time. In both cases, the relative amplification factors of FTNM number 1 have largest cumulative effects in boreal spring when the equatorward penetration of these disturbances is also the largest. The other 5 to 10 dominant FTNMs also have largest amplitudes during the first half of the year. We elucidate the roles of intra-modal and inter-modal interference effects in the evolution of FTNM1 and show both effects play significant roles particularly in boreal spring.

Our results suggest that a contributing cause of the boreal spring predictability and correlation barrier in observations and in coupled ocean-atmosphere models may be the fact that the amplitudes of the large scale atmospheric modes have peaks in boreal spring.