MJO in the BMRC Atmosphere and Coupled Models
Bureau of Meteorology Research Centre
(*PhD student, Mathematics Dept, Monash University)
Introduction
The Madden-Julian Oscillation (MJO; Madden and Julian, 1972) is the strongest signal of intra-seasonal variability of the tropical atmosphere. The ENSO phenomenon is known to be the single most prominent signal in the inter-annual variability of the earth’s climate (Lau and Chan, 1986). Until about 20 years ago, these two phenomena had traditionally been considered as separate events due to the vastly different time scales associated with the recurrence and duration of each. However more recently scientists have hypothesised possible connections between MJO and ENSO. There are currently many competing theories of the link between the two phenomena (Zhang et al., 2001).
The MJO plays a particularly important role for climate variations in Australia. Significant impacts on the onset and breaks of the summer monsoon and tropical cyclone genesis in the Australian sector are well documented. Ongoing research indicates potentially important effects of the MJO for onset and evolution of ENSO, for interannual variations of the monsoon, and for non-summertime extra-tropical rainfall episodes. Ample scope exists to improve short to medium range predictions in the Australian sector by proper monitoring and prediction of the large-scale circulation and convection anomalies associated with the MJO. Incorporation of MJO-activity into coupled forecasts will increase spread and may improve estimates of forecast uncertainty.
In this study the ability of the Bureau of Meteorology's atmosphere model (BAM) to similate the MJO was explored using long AMIP style integrations of the atmosphre model with prescribed SSTs.
MJO in BAM
Two versions of BAM were investigate. Firstly, the standard version of BAM3 (Colman et al 2002), which used a moisture convergence convection closure. A second version was the same as the first except that it used a CAPE convection closure with a convection trigger delay.
(a) NCEP Re-analysis (b) BAM: Moisture convergence closure (b) BAM: CAPE convergence closure
Figure 1. Relative spectra define as ratio of spectral power to background power for surface zonal wind averaged 10S to 10N Figure 1a shows the relative power spectrum for the 10m wind from the NCEP reanalysis. The relative spectrum is calculated by the dividing the total power spectrum by a background spectrum, as done by Wheeler and Kiladis (1999). In our case the background spectrum is calculated by applying a box-scar smoother to the raw-spectrum 20 times. This enables the background "red-noise" to be removed from the spectral plot.
The NCEP spectrum shows a clear spectral peak between about 30 and 90 days for eastward propagating wave number 1, with relative spectral values reaching well over 1.5. This peak is due to the MJO. There is also a peak in westward propagation at low wave numbers, characteristic of atmospheric Rossby waves. The relative power associated with these Rossby wave is generally less than 1.2, much less than that associated with the MJO.
Figure 1b shows the 10m wind zonal spectrum from a 20 year integration of BAM with moisture convergence closure. There is no clear peak at MJO time and space scales. In fact, eastward propagation at MJO scales is weaker than westward propagation.
Figure 1c shows the 10m wind zonal spectrum from a 20 year integration of BAM with the CAPE convection closure. There is a clear peaks at MJO scales, one at around 30 days with relative power reaching over 1.5 and one at around 60 days with relative power also reaching 1.5. This model shows significantly greater eastward than westward propagation at MJO scales.
The variability in the CAPE integration has been examined in more detail. It appears to have some of the characteristics of the MJO. It also captures the observed seasonality (not shown). For more details see the documents below.
Future Plans
The intra-seasonal variability in BAM with CAPE convection closure is in general at least comparable with the best international models. It has variability characteristic of the MJO and captures the correct seasonality. Futher work will continue by the BMRC model development group to improve the simulation of the MJO. However, this version of BAM is good enough to use it in both atmosphere only and coupled modes to investigate the role of MJO in climate. We will be using this version of BAM to investigate:
- The role of coupling on the MJO using slab ocean layers, ocean mixed layers and full ocean models.
- The role of the MJO on the predictability of ENSO using the coupled model used for POAMA.
- Techniques to sample the uncertainty in the POAMA forecasts due to intra-seasonal variability.
Documents/Papers
Two BMRC internal reports are presently being prepared as well as a paper to be submitted to the Australian Meterorological Magazine. This will be made available here as soon as they are ready.
References
List of references used in this site
For further information: email the POAMA group
© Copyright Commonwealth of Australia 2002, Bureau of Meteorology.
Users of these Web pages are deemed to have read and accepted the conditions
described in the copyright declaration.
