TitanWRF general circulation model simulations performed without sub-grid-scale horizontal diffusion of momentum produce roughly the observed amount of superrotation in Titan’s stratosphere. We compare these results to Cassini–Huygens measurements of Titan’s winds and temperatures, and predict temperature and winds at future seasons. Weuse angular momentum and transformed Eulerian mean diagnostics to show that equatorial superrotation is generated during episodic angular momentum ‘transfer events’ during model spin-up, and maintained by similar (yet shorter) events once the model has reached steady state. We then use wave and barotropic instability analysis to suggest that these transfer events are produced by barotropic waves, generated at low latitudes then propagating poleward through a critical layer, thus accelerating low latitudes while decelerating the mid-to-high latitude jet in the late fall through early spring hemisphere. Finally, we identify the dominant waves responsible for the transfers of angular momentum close to northern winter solstice during spin-up and at steady state. Problems with our simulations include peak latitudinal temperature gradients and zonal winds occurring ?60 km lower than observed by Cassini CIRS, and no reduction in zonal wind speed around 80 km, as was observed by Huygens. While the latter may have been due to transient effects (e.g. gravity waves), the former suggests that our low (?420 km)model top is adversely affecting the circulation near the jet peak, and/or thatwerequire active haze transport in order to correctly model heating rates and thus the circulation. Future work will include running the model with a higher top, and including advection of a haze particle size distribution.