Abstract
The Madden-Julian Oscillation (MJO) is the largest source of intraseasonal variability in the tropics. The MJO affects weather and climate phenomena both in the tropics and extratropics and is a key to accurately forecasting in the subseasonal time scale. There are several theories for the maintenance and propagation of the MJO, but little consensus on why it initiates. Here we explore several aspects of the MJO through the lens of two MJO events that occurred during the Dynamics of the MJO field campaign. We examine the role of the intertropical convergence zone (ITCZ) in the initiation of the MJO. We find that almost half of MJO events initiate after an active ITCZ was present over the Indian Ocean. Convection associated with a MJO can more quickly grow wider and taller when an ITCZ precedes initiation than when an ITCZ is not present. The ITCZ can provide a more moist environment favorable to convective growth prior to onset of MJO initiation. The ITCZ, however, is neither a necessary nor sufficient condition to the onset of MJO initiation. We investigate the importance of atmosphere-ocean coupling during the first two DYNAMO MJO events using a regional model. Despite the model setup not being optimally tuned and experiencing an ocean cooling bias during the simulations, we find evidence to suggest the first MJO event was less reliant on coupling than the second event. The October MJO event propagates further in the uncoupled model run than in the coupled. Both the uncoupled and coupled run show an increase in rain over the Indian Ocean during the time when the November event should initiate, but neither propagate the rain. This suggests that the boundary conditions provide some signal for the model simulations to rain at the appropriate time. Correlation relationships suggest that the coupled model experiment often has relationships close to those found in the observations and reanalysis, suggesting if the ocean cooling bias could be corrected the regional model has the potential to successfully simulate the MJO. Finally, we compare an ensemble of reforecasts of GCM form the Subseasonal Experiment (SubX) during the two DYNAMO MJO events to our regional model simulations. We find that the ensemble has a difficult time propagating the October MJO, but more successfully initiates and propagates the November event. We discuss the differences between the “uncoupled” October MJO event and the “coupled” November MJO event.