CMIP Subproject: Southern Mid-To-High Latitude Variability on Interseasonal and Interannual Time Scales

Dr. Wenju Cai
CSIRO Division of Atmospheric Research
Aspendale Vic 3195 Australia
wjc@dar.csiro.au

Background

Recently, an interactive phenomenon in the southern mid-to-high latitudes has been observed with eastward propagating signals in both oceanic and atmospheric variables. This has been termed the Antarctic Circumpolar Wave (ACW) by White and Peterson (1996) (hereafter WP). There are several well defined features. First, local signals in fields of sea surface temperature (SST), pressure at sea level (PSL), meridional wind stress, and sea ice extent are generated every 4-5 years. Second, these signals are phase-locked and move eastward, encircling the globe in about 8-10 years. Third, at any instant, there is a predominant wavenumber 2 spatial pattern. WP hypothesized that there exists a teleconnection between the ACW and El Nino Southern Oscillation (ENSO) events. In a separate study, Jacobs and Mitchell (1996) confirmed the wavenumber 2 pattern in an analysis using altimeter data.

Subsequently, ACW-like phenomena have been found in coupled ocean-atmosphere models (Christoph et al. 1997, hereafter CBR; and Motoi et al. 1997; Cai et al. 1998 hereafter CBG). In the CBR and CBG studies, the eastward movement appears in the SST anomalies but not in other aforementioned fields. In those fields the signal is dominated by standing oscillations and there is no phase- locking. Although local signals are generated every 4-5 years as in the observed, the SST signal takes 12-16 years to complete a circuit around the pole, and there is a predominant wavenumber 3 pattern. The results from the Motoi et al. (1997) coupled model show propagating signals in the SST together with standing oscillations in other fields, and in their model there is a preferred wavenumber 2 pattern. But SST anomalies would take 20-30 years to complete a circuit. They argued that their modelled ACW is closely related to ENSO activities.

Earlier, the feature of a dominant quasi-stationary wavenumber 3 pattern in southern mid-to-high latitude variability has been reported in many observational studies on day-to-day, interseasonal, and interannual time scales. In particular, Mo and White (1985) (hereafter MW) demonstrated that there are preferred locations (centers) for the strongest anomalies for the southern winter season. They correlated time series of PSL at a base point with values at other grid points and showed a wavenumber 3 pattern. When the base point was moved away by 10 degrees in either direction, the pattern disappeared. A similar feature was produced by the CSIRO coupled model (CBG).

Objectives

This subproject intends to examine relevant issues by using the monthly mean SST and SLP global fields over a time length of 30 years that some of the modeling groups have submitted for CMIP. In particular, focus will be placed on the wavenumber-pattern of anomalies, the presence or the lack of the phase-locking of the oceanic and atmospheric anomalies, and the possible teleconnection with ENSO in these models (although one understands that ENSO is not properly simulated in these models). As an entry point, the dominant pattern of variability on interseasonal time scale can also be examined and bench-marked against the results of the MW study. It is hoped that this study will give a cursory feeling for the ACW in the models. This may point to future more detailed analyses with focus on reconcilling the differences between the modelled and observed characteristics, and eventually on its climatic impact and predictability.

Methodology

  1. To examine the phase-locking of oceanic and atmospheric anomalies, the project will construct Hovmoeller diagrams of anomalies of SST and PSL and examine their interrelationship.
  2. To examine the dominance of a specific wavenumber-pattern, correlation analysis following WP, or/and by applications of empirical orthogonal function analysis will be performed. This will provide time series of the pattern index (temporal cofficient in the case of EOF analysis).
  3. To examine any possible connection with the variability in the tropics, lag correlation between the pattern index and time series of Nino3 SST anomaly (or southern oscillation index) will be performed.

References

Cai, W.J., Peter G. Baines and H.B. Gordon, 1998: Southern mid-to-high latitude variability, a wavenumber 3 pattern, and the Antarctic Circumpolar Wave in a coupled model. J. Climate, submitted.

Christoph, M., T.P. Barnett, and E. Roeckner, 1997: The Antarctic Circumpolar Wave in a coupled ocean-atmosphere GCM, MPI-Report 235, 28pp, Max-Planck-Institut fur Meteorologie, Hamburg.

Jacobs, G. A., and J. L. Mitchell, 1996: Ocean circulation variations associated with the Antarctic Circumpolar Wave. Geophys. Res. Lett., 23, 2947-2950.

Mo, K. C., and G. H. White, 1985: Teleconnections in the Southern Hemisphere. Mon. Wea. Rev., 113, 22-37.

Motoi, T., A. Kito, and H. Koide, 1997: Antarctic Circumpolar Wave in a coupled ocean-atmosphere model. Annals Glaciology, in press.

White, B. W., and R. G. Peterson, 1996: An Antarctic circumpolar wave in surface pressure, wind, temperature and sea-ice extent. Nature, 380, 699-702.