Search

As part of the verification phase of the TOPEX/POSEIDON mission, 10-day gridded fields of altimeter data derived from TOPEX geophysical data records are compared with 10-day gridded fields of dynamic height derived from more than 60 moorings of the Tropical Ocean and Global Atmosphere-Tropical Atmosphere Ocean (TOGA-TAO) array in the equatorial Pacific Ocean. Access to TAO data in real time permits the first 500 days of the TOPEX/POSEIDON mission to be placed in the context of complementary, in-situ measurements of surface winds, sea surface temperatures, and upper ocean thermal structure, as well as the time history of these variables prior to launch. Analysis of the space-time structure in the TOPEX and TAO surface topography data increases sea level variability primarily due to equatorial Kelvin wave activity generated by intense wind bursts west of the date line in association with the 1991-1993 El Nino. Cross correlations between the two data sets are generally greater than 0.7, with root mean square (RMS) differences less than 4 cm. However, for reasons not fully understood, correlations drop to less than 0.5 in certain regions off the equator in the eastern Pacific and RMS differences can be greater than 5 cm north of the equator in the central and eastern Pacific.

A linear, multimode model forced by observed winds is used to investigate anomolous heat transport and storage during the 1982-1983 El Nino. The study compliments the work of Wyrtki (1985) and of Zebiak and Cane (1987) and contains the ocean dynamics invoked by both these studies to explain heat content anomalies. Model hindcasts are compared with observational evidence derived from the island sea level network. The meridional distribution of heat storage and the components of heat transport are considered. It is found that the mechanisms contributing to heat transport out of bands of latitude symmetric about the equator are isolated and related to the wind anomalies and wave dynamics usually associated with ENSO events. It is noted that although the spatial and temporal distribution of oceanic heat anomalies necessary to initiate an ENSO event can be determined only by studying coupled ocean-atmosphere models, an examination of the oceanic component alone is useful in determining constraints imposed by ocean dynamics.

The applicability of satellite altimeter data for estimating zonal current variability at the equator is assessed using the meriodionally differenced form of the geostrophic balance. Estimates of geostrophic zonal flow anomalies in the equatorial Pacific have been deduced from 17-day collinear altimeter data during the first year of the Geosat Exact Repeat Mission. Altimeter-derived geostrophic estimates agree well with in situ zonal current variability. Comparison of flow-frequency near-surface zonal current observed from equatorial moorings at 165 deg E, 140 deg W, and 110 deg W yield correlations of 0.83, 0.85, and 0.51, respectively, with a mean rms difference of 23 cm/sec. The inclusion of up to 11 ascending and descending Geosat tracks within the 9-deg band for every 17-day repeat effectively reduced the temporal sampling interval to 1.5 days at 165 deg E and 140 deg W. The 6.8-km along track spacing of the altimeter measurements provides sufficient resolution for the effective filtering of small-scale meridional noise, both instrumental and oceanic.

A linear multiple vertical-mode model described by McPhaden et al. (1988) is used to examine potential errors due to data scatter around expendable bathythermograph (XBT) transects in the tropical Pacific. Two methods of sampling are compared. In the first, the model was sampled along approximately straight lines of grid points corresponding to the mean positions of XBT tracks in the eastern, central, and western Pacific; in the second, the model was sampled again at the dates and locations of actual XTB casts for 1979-1983. The model indicates that the data scattered zonally around XBT transects in general can lead to about 2 dyn cm error in dynamic height in composite sections of XBT data. Errors larger than 2 dyn cm occurred in regions where XBT sample spacing in the zonal direction was insufficient to resolve Rossby wave variations in the model.

This paper quantitatively evaluates simulations of the mean seasonal cycle in the tropical Pacific Ocean, using a multiple vertical-mode version of the Busalacchi and O'Brien (1980) linear model. The period studied (1979-1981) includes several coincident oceanic and surface wind data sets, as well as a fairly regular seasonal cycle leading into the 1982-1983 El Nino. Simulated mean seasonal cycles in dynamic height and sea level are compared with observed variations in expendable bathythermograph and island tide gauge data averaged over the same period. It was found that, although all simulations show characteristic features of the mean meridional ridge-trough structure in surface topography, simulated north and south equatorial ridges at 20 deg N and 20 deg S are much higher than those observed. Only weak equatorial ridges were generated near 4 N, and none of the simulations exhibited a significant equatorial trough. These discrepancies are attributed to limitations in model physics and to the wind forcing.