Best visible in variables that relate to global integrals of motion (interhemispheric exchange of atmospheric mass, relative angular momentum of the atmosphere, enstrophy), the GCM's dynamic 'skeleton' of the season turns out to consist of a complete, yet inverse, route to (resp. out of) planetary-scale "monsoon chaos" - to which the system is kicked at monsoon onset by a hard jump when passing a subcritical Hopf bifurcation in boreal spring (recall the "spring predictability barrier"). The seasonal cycle inflates into a torus segment there, i.e. it develops a second, independent ('poloidal') oscillation, which is not astronomically driven but rather a self-maintained natural mode of variability that is organized in a complex, planetary-scale feedback cycle - the well-known 30-60 day monsoon activity cycle. The very existence of this regime of the 'interactive' monsoon, with its inherent tropical-extratropical interactions, may be traced back in the GCM to elementary conditions in the status of two of the global variables used to represent it: atmospheric mass balance between the hemispheres and sub-, co- or superrotational status of the atmosphere with respect to the surface (of a model Earth without dynamic ocean and other contributors to the real-world angular momentum).
As the season advances, the backwards running period multiplication scenario reaches a period-one oscillation via changing synchronizations of planetary waves (at lowest-possible rational frequency relationships, indicating a high degree of internal synchronization) and ends up in a wrinkling torus segment before it ceases to exist at the date of - structural - monsoon retreat. The wrinkle itself originates in the remainder of an 'enstrophy stage' at the transition from period-two to period-one cycles and is the host then of a slow, degenerate 'monsoon' oscillation after retreat, an irregular wander between summer and winter fixpoints which shows up marked signs of the Southern Oscillation.
Apart from these and other attractor studies addressing the topological aspects of present-day monsoon dynamics in boreal summer, short- and longer-term seasonal runs under varied external conditions (e.g. changing insolation at top of the GCM's atmosphere) show the impact of the intraseasonal dynamic object uncovered. Three monsoon retreat categories may roughly be distinguished (termed as of "La Nina prone", "Indian summer" and "El Nino trigger" type) which stamp their signatures on the GCM's interannual to centennial climate dynamics and may generate different variability regimes under changing conditions. Paleoclimatic aspects are also related to the existence or non-existence of the 'interactive' monsoon itself: in pre-monsoon attractor soundings a climate regime has been found that calls into mind the "Green Sahara" regime, for example.
A topologically founded "monsoon hypothesis" has been developed, and updated over the decades since its first detection in the model. That the GCM's high sensitivity to the smoke clouds rising in 1991 in the neighborhood of the strongest equator-crossing circulation, the Somali Jet (which feeds the South Asian monsoon system in boreal summer), has been taken seriously at that time, is also due to the impact that the fundamental work published by Peter Webster in the second half of the 1980s had - work that has been very helpful in understanding what was going on there in the model, and for which great respect is due to him.