influence of air-sea interaction on the Madden-Julian Oscillation: the role of the seasonal mean state

• Published in: Climate dynamics 2007-06, Vol.28 (7-8), p.703-722
• Main Authors: Watterson, I. G, Syktus, J
• Format: Article
• Language: English
• Subjects: Air; Air-sea interaction; Annual variations; Anomalies; Belts; Climatology. Bioclimatology. Climate change; Earth, ocean, space; El Nino; El Nino phenomena; El Nino-Southern Oscillation event; Exact sciences and technology; External geophysics; General circulation models; Geophysics. Techniques, methods, instrumentation and models; Interannual variability; Interannual variations; Madden-Julian oscillation; Marine; Meteorology; Mixed layer; Modelling; Monsoon winds; Monsoons; Oceans; Principal components analysis; Propagation; Sea surface; Sea surface temperature; Sea surface temperature anomalies; Seasons; Southern Oscillation; Spectral analysis; Spectrum analysis; Surface temperature; Troposphere; Tropospheric winds; Wind
• ISSN: 0930-7575; 1432-0894
• DOI: 10.1007/s00382-006-0206-9

The CSIRO Mark 3 general circulation model at T63 resolution is used to explore the potential effect of air-sea interaction in enhancing the eastward propagation of the Madden-Julian Oscillation (MJO). Principal component analysis is used to define a seasonal lower-tropospheric wind signal. When the model is coupled with an interactive ocean, the monsoon wind anomalies in December-February (DJF) propagate from the Indian Ocean to the Pacific Ocean. In versions with a thinner mixed layer, the propagation speed approaches that seen in the observational ERA40 data set. However, in the non-interactive model with specified sea surface temperatures (SSTs) there is no propagation. Similar contrasts are seen in other seasons. The upper tropospheric long-wave signal determined through spectral analysis is also more realistic in the coupled model, although power around the 80 day period remains too large. Positive SST anomalies form to the east of low-level convergence, in part due to evaporative flux that is modified by the mean monsoon westerly belt in DJF. Interannual variations in this belt appear to have an effect on the propagation of the wind anomalies in the coupled model, while only the amplitude varies in the non-interactive model. This contrast is also seen in partitions of years by the state of ENSO. Propagation of the MJO signal is faster and extends farther into the Pacific in El Niño years in observations and the coupled model, although model biases, in particular a short westerly belt, appear to limit the effect. It is concluded that air-sea interaction is potentially very important to the MJO and its interannual variability, and that the westerly belt has an influence on its evolution.