Sea Surface Height Variations of the Antarctic Circumpolar Wave

G. A. Jacobs and J. L. Mitchell
Naval Research Laboratory
941 Rue Ste.Anne
New Orleans, LA 70116, USA

email: jacobs@proteus.nrlssc.navy.mil

In the region of the Southern ocean, a single phenomenon dominates interannual sea surface height (SSH) variations. The eastward propagating Antarctic Circumpolar Wave (ACW) has a period of about 4 years and wavelength of 180 degrees in longitude which implies a wavespeed of about 10 cm/s. The ACW is apparent as a wave in SSH (or ocean pressure) which is forced by wind stresses. Variations in sea surface temperatures (SST) throughout the region also are associated with the wave. Recent research (1) indicates that the SST variations affect the ice edge extent along the entire Antarctic continent. The wave presents itself as a possible coupled ocean/atmospheric process and thus could be one of the fundamental modes of Earth's environment in the absence of interfering land masses.

Observations of global environment variations within the ocean and atmosphere have only become possible within the last decade. The systems that measure such variations range from large global networks of observation stations (2) to series of satellites which measure SST and SSH (3, 4). The most recent and most accurate satellite to measure SSH is the joint US/French TOPEX/POSEIDON (T/P) satellite (5). Since being placed in orbit in 1992, T/P has returned SSH measurements globally with an accuracy better than 5 cm RMS.

To extend the observational time series, we must unify data sets from altimeter satellites which cover separate time periods. In order to join additional altimeter data with T/P, all altimeter data must be referenced to a common mean sea surface. One possibility is to use a gravitational geoid as the reference surface. However, current independently derived gravitational geoids contain large errors on spatial scales smaller than the ground track separations of different satellites. Thus gravitational geoids do not allow accurate combination of separate altimeter data sets.

The mean SSH measured by T/P is accurate to better than 1.2 cm RMS. This accurate mean surface forms a reference to join other altimeter data sets over the time period from 1986 through 1995. The technique to join the altimeter data sets together requires using altimeter data at only the points where T/P and the other altimeter ground tracks cross. At these points the gravitational geoid is the same in both data sets. The drawback to this method is that only a small fraction (about 10 %) of the original data is used. However, this still provides a spatially dense network of data points across the Southern ocean for observing long time period and large spatial scale variations (Figure 1).

The variations in SSH, SST, and the atmospheric wind stress curl (WSC) are averaged over the latitude band from 45 degree S to 55 degree S. The data are then put into bins 5 degrees in longitude by 1 month in time. A filter is applied to remove the seasonal variability and to pass through only the ACW variations (Figure 2). Data points separated by more than two years' time are independent. The eastward propagation is apparent in all the fields in the Southern ocean and is by far the dominant signal on interannual time scales even before applying the filtering procedure.

Quasigeostrophic (QG) ocean dynamics provide a direct relation between the WSC which forces ocean circulation and pressure within the ocean which is measured through SSH. If an initial perturbation creates a large scale SSH anomaly, and the anomaly propagates freely without additional forcing then, according to the QG dynamics, a westward propagating Rossby wave should result (6). These waves have been observed throughout the Pacific ocean (7, 8), especially due to El Nino events at the American coasts (9). However, on the time scales considered here a freely propagating Rossby wave would cross a distance of only about 12 degrees longitude at 50 degree S in a time of 4 years. Thus Rossby waves are not expected to be an important contributor to ocean circulation variations on the time scales we are able to observe.

Since the ACW is eastward propagating, we immediately are able to determine that the ACW is not a freely propagating response to an initial perturbation, but the ACW is a forced wave. That is, the eastward propagating ACW requires some forcing external to sustain it. The WSC variations acting on the Southern ocean (Figure 2c) indicate an eastward propagation with the same wavenumber and frequency as the SSH variations of the ACW. An examination of the lag between the WSC and SSH (Figure 3) indicates that the SSH lags the WSC by 90 degrees phase, or 1 year's time. Since the wave has a 4 year period, the local maxima in the correlation function occur at -1 year and +3 years. If the Southern ocean is modeled by a channel with a mean flow of about 10 cm/s (the mean eastward drift over the region from 40 degree S to 60 degree S) then the quasigeostrophic SSH variations forced by plane waves in WSC indicate that SSH should indeed lag WSC by 90 degrees phase (one years' time) as observed.

The direct mechanisms by which the SSH and wind stresses produce the observed SST variations have not been fully explored. The unique spatial and temporal character of the ACW does appear strongly in the SST, thus the data imply that the SST is connected to the ACW. In general, the heat content of the upper ocean has been significantly correlated to SSH measured by tide gauges (10) and altimeter data (11). Because the ACW signal appears in atmospheric variables which drive ocean circulation (the WSC) and in oceanic variables which drive atmospheric circulation (the SST), the process presents itself as a possible coupled ocean/atmospheric process. The Southern ocean is the one region of the world where such a process could freely develop. In other areas, land masses would act to interfere with the surface temperatures which drive the atmosphere. Thus the ACW possibly represents a fundamental mode of variation in Earth's environmental system. A thorough study of this entire process requires more model and observational work than presented here. Here we have shown the connection from the atmosphere to the ocean and possible effects on SST due to the newly discovered ACW in the Southern ocean.

Figure Captions

Figure 1. The crossover points of the TOPEX/POSEIDON ground tracks and the Geosat-ERM ground tracks. The geoid height measured by each satellite at these crossover points is the same. Thus long time period SSH variations are observable at these points.

Figure 2. (a) SSH (greyscale range from -2 to 2 cm), (b) SST (greyscale range from -.5 to .5 degree C), and (c) WSC (greyscale range from -5 10^-6 to 5 10^-6 N/m^3) variations across the Southern ocean. The data from 45 degree S to 55 degree S have been binned into 5 degree longitude by 1 month bins and filtered to pass the variability due to the ACW. Data separated by more than two years are independent. The solid lines indicate a speed of 10 cm/s with a wavenumber of 2 cycles about the globe. The eastward propagation is the dominant variability on interannual time scales.

Figure 3. The lagged cross-correlation of SSH and WSC (WSC lagged relative to SSH). Local maxima in the correlation occur at times of -1 year and +3 years as the ACW has a period of 4 years. Thus the SSH lags the WSC by 90 degrees in phase, or one years' time. This one year time lag is predicted by the quasigeostrophic SSH variations produced by a plane wave in WSC.

Bibliography:

(1) Peterson, R. G & White, W. B. Nature (1996).

(2) Hayes, S. P., Mangum, L. J., Picaut, J., Sumi, A., & Takeuchi, K. Bull. Am. Meteorlol. Soc., 72, 339-347 (1991).

(3) Reynolds, R. W. & Marsico, D. C. J. Clim., 6, 768-774 (1993).

(4) Fu, L. L., and R. E. Cheney, Rev. of Geophysics 213 -- 223, 1995. (5) Fu, L. L., Christensen, E. J., Yamarone, C. A, Lefebvre, M., Menard, Y., Dorrer, M., & Escudier, P. J. Geophys. Res., 99(C12), 24,369-24,381 (1994).

(6) Hurlburt, H. E., Kindle, J. C., & O'Brien, J. J. J. Phys. Oceanogr., 6, pp 621-631 (1976).

(7) White, W. B., & Saur, J. F. T. J. of Phys. Oceanogr., 11, 1452-1462 (1981).

(8) Meyers, G., J. of Phys. Oceanogr., 9, 633-674 (1979).

(9) Jacobs, G. A., Hurlburt, H. E., Kindle, J. C., Metzger, E. J., Mitchell, J. L., Teague, W. J., & Wallcraft, A. J. Nature, 370, 360--363 (1994).

(10) Rebert, J. P., J. R. Donguy, & G. Eldin, J. of Geophys. Research, 90, 11,719-11,725 (1985).

(11) Carnes, M. R., W. J. Teague, & J. L. Mitchell, J. of Atmosph. and Oceanic Tech., 11, 551--566 (1994).

Note added November 1996: this work has since been published in:

Jacobs, G. A., and J. L. Mitchell, Ocean circulation variations associated with the Antarctic Circumpolar Wave. Geophys. Res. Let., Vol. 23, No. 21, pp. 2947-2950, 1996.