ATOC STUDY
A similar study was also carried out in the North Pacific using data collected by the ATOC project (The ATOC Consortium, 1998). Five acoustic sections were used in the analysis, in addition to altimetric, meteorological, and hydrographic data. The various data sets were compared along the five ATOC sections. The major feature in all but one time series is the seasonal heating and cooling. But there are also differences due to errors in the obervations and to processes other than heating and cooling which are invisible (inaudible?) to the acoustic data.
To summarize the differences, the altimetric sea surface height differs from that inferred by either XBTs or ATOCby about 2 to 3 cm rms. Furthermore, the altimetric annual cycle amplitude is about 1 to 2 cm greater than that obtained from heat content estimates alone. The differences emphasize the complementarity of the data sets and can be explained by a number of physical processes.
First, the altimeter sees effects of salinity which are not seen by XBTs or acoustic data. A recent study (Gilson et al. 1999) based on repeat XCTD sections from Taiwan to San Francisco, concurrent with the TOPEX/POSEIDON period, reported a drift in the salinity field relative to Levitus of about 2 cm surface equivalent, and a cruise to cruise standard deviation of about 1 to 2 cm surface equivalent.
Second the altimeter sees barotropic Rossby waves and a lower frequency barotropic signal due to gyre wobbling. From a numerical simulation we estimate that in the ATOC region, the annual harmonic of the barotropic signal is about 1 cm and has the right phase to explain about half of the observed TOPEX/POSEIDON to ATOC difference at the annual cycle.
Third the hydrological cycle also has a distinct annual cycle with the right phase for explaining in part the observed difference between the altimetric and the tomographic data. A recent study by Minster et al. (1999) reported that there is an annual cycle in the total mass of the water contained in the oceans which corresponds to a signal with an amplitude of about 7 mm at the surface. The largest contribution is from the Northern Hemisphere land-mass, which is driest at the end of the boreal summer, causing a maximum in the mass of the Oceans towards the end of September, and hence further contributing to the TOPEX/POSElDON to ATOC difference at the annual cycle.
In a way similar to what was done in the Mediterranean, the various data were combined together and with a numerical circulation model to produce best estimates of heat content and of the circulation. We found that the model alone was surprisingly good at estimating spatial characteristics of large scale changes, but that, in general, the model underestimated the magnitude of these changes. Inter-annual changes of up to 0.2 deg C vertically averaged over 4000 m were odserved, and were of the same order of magnitude as the inferred annual cycle of heating and cooling.
