•The life span of atmospheric anomalies depends on whether they are coupled, and on the type of interaction they have, with the underlying ocean anomalies.
•
•The dynamical rule applied to the reanalysis data gives results consistent with previous studies and with the lag-correlation between 850mb RV and SST: Ocean drives the atmosphere in the tropics and the reverse occurs in the extratropics.
•
•The AMIP run from the NCEP shows fewer long lasting anomalies than observed. This error is associated with ignoring the feedback effect of the atmosphere on the ocean.
Summary
•The dynamical rule proposed by Mo and Kalnay (1991) to determine the forcing direction in locally coupled anomalies gives results consistent with previous studies and with the lag-correlation between 850mb RV and SST. Statistics of the annual number of locally coupled anomalies indicate that “ocean-driven” anomalies prevail in the tropics, while “atmosphere-driven” anomalies prevail in the extratropics. Similar distribution of the forcing direction is obtained with five-day average data from ERA14 and with monthly data. It remains to be shown how the distribution of ocean-driven anomalies depend on the season and with the ENSO phases.

•It remains to be shown how the distribution of ocean-driven anomalies depend on the season and with the ENSO phases.