Extension of the period of useful forecasts by dynamical- numerical models of the atmospheric circulation is the goal of a rapidly growing body of ongoing and proposed research in large scale dynamic meteorology. Recent success in improving the accuracy and utility of short (0-5 day)-, and medium (5-10 day)-range numer- ical forecasts has stimulated intensified interest in the problems associated with developing models that can exhibit forecast skill in the extended (10 day to seasonal)-range and beyond. One line of research concerns the set of observed atmospheric flow phenomena with characteristic periods of variation in the extended range, and generally of hemispheric/global spatial scale (in particular, mid- latitude "blocking", teleconnection patterns, and zonal index cycles of the jet-stream and its associated storm-track regions). The PI's propose to study the processes leading to the onset and evolution of similar features in a suite of simple theoretical models which have exhibited similar behavior in their prior research. Understanding these processes in such simple models will help to guide the interpretation of forecast errors in the solu- tions of the more complex quasi-operational models now being used in experimental dynamical extended-range forecasts, and the step- wise elimination of the sources of these errors.
