Need to Update the Tropical Wave/MJO Paradigm

Currently, most discrepancies between observations and the classical Matsuno-Gill theory of tropical waves are attributed to the neglect of coupling between the waves and moist convection, and given rise to the so-called Convectively Coupled Equatorial Wave (CCEW) paradigm. Although the waves and convection are undoubtedly related, the extent to which their observed association reflects true coupling versus forcing (or organization) of one by the other has never been adequately established. Our own theoretical and observational analysis shows that it is the neglect of spatial base state variations, rather than convective coupling, that is the primary limitation of the Matsuno-Gill theory. This suggests an updated paradigm in which models that include both effects, but are still linear, would continue to provide a useful framework for interpreting tropical variations. Such models would retain non-dispersive Generalized Kelvin waves and also zonally symmetric oscillations as robust eigenmodes of zonally symmetric base states with observed meridional and vertical structure.

Our observational analysis is based on Linear Inverse Modeling (LIM), which deduces the linear evolution operator for daily tropical anomalies using the time-lag covariances and cross-covariances of circulation and humidity fields in the ERA-Interim dataset. Remarkably, the MJO as well as other well-known tropical waves can be identified with specific individual eigenmodes of this linear operator. However, the eigenmodes (including the MJO eigenmode) are highly seasonally dependent, consistent with the strong modification of the wave dynamics by the seasonally varying base state. The eigenmodes also do not form an orthogonal set, as they do in the Matsuno-Gill or CCEW paradigm, and this has important implications for the predictable growth and decay of anomalies, especially those associated with the MJO. In particular, it implies that MJO evolution is strongly affected by its interaction with the traditionally neglected red noise background of tropical variability. We also show that although the circulation-humidity coupling is overall of secondary importance in tropical wave dynamics, it does significantly affect the evolution of some eigenmodes and the MJO.