Applications, Sensitivities, and Development of an RMM-like Index for the MJO Skeleton Model

It is well known that global climate models (GCMs) continue to struggle with simulating many of the observed features of the Madden-Julian Oscillation (MJO). The MJO “skeleton model”, originally developed by Majda and Stechmann (2009), is a low-order dynamic model that is capable of reproducing several salient features of the MJO despite its many simplifications relative to higher-order GCMs. Before the skeleton model can be used for standalone predictions, however, it must be thoroughly evaluated with hindcast experiments, reanalyses, satellite products, and other observations.

This study presents a necessary first step for a meaningful comparison of simulated MJO activity to observations, namely, the development of an MJO index for the skeleton model so that it can be successfully evaluated in a manner consistent with the evaluations of GCMs used for weather and climate simulation/prediction. As such, it would be advantageous if the MJO index for the skeleton model utilized the same principles as the commonly used Real-time Multivariate MJO (RMM) index based on a combined EOF analysis of near-equatorial OLR and upper- and lower-tropospheric zonal winds. For this study, single variable and multivariate empirical orthogonal functions and principal components are computed and examined for the stochastic version of the skeleton model using zonal wind at a single level and synoptic-scale wave activity envelop amplitude as proxies for MJO circulation and convection anomalies, respectively. MJO activity in the skeleton model is also evaluated using wavenumber-frequency filtering for both zonal winds and the envelope of convective heating with comparisons made to those results using the RMM-like index. Finally, early applications of the model index are presented that include documenting the frequencies and locations of MJO event initiation and termination in the skeleton model compared to a climatology using the observed RMM index.