many receivers in an ocean basin), each acoustic path can be occupied briefly by sound as frequently as desired. Measurements of temperature along several paths in the North Pacific have been acquired every four hours, for example, but daily or weekly sampling may prove to be adequate.
Observations
The Acoustic Thermometry of Ocean Climate (ATOC) project commenced transmissions in December 1995 from a source at sound channel depth on Pioneer Seamount, 100km offshore from San Francisco. Sound signals were received on US Navy Sound Surveillance System (SOSUS) and other hydrophone arrays. The SOSUS arrays are bottom mounted, while two others were 40-element vertical line arrays. A total of 772 transmissions were made in 43 groups (typically 4 hours apart, for four days, two to four times a month) between December 1995 and March 1997. A clear correspondence between observed and predicted ray arrivals shows that at ranges of 3000 to 5000km, ray arrivals are resolvable, identifiable and stable. Ray arrivals were tracked and used to infer range-averaged profiles of sound speed and temperature along each section. The effects of internal waves and mesoscale eddies are smoothed out by the large scale, long range integration.
The travel times for all but one section decreased in summer and increased in winter, consistent with anticipated seasonal heating and cooling of the surface layer. The exception shows that subsurface warming in late winter near receiver v1 (Hawaii) more than compensated for surface cooling near the source.
The TOPEX/POSEIDON altimeter produces precise measurements of sea surface topography globally at a 10-day repeat period, which can be used to infer circulation and ocean structure. The structure is a complex sum of different physical processes, some barotropic and some baroclinic. By combining data from other sources, acoustic, XBT and numerical generai circulation models, it is possible to separate the various components contributing to the observed sea level.
Direct temperature measurements of the upper 800m of the ocean along the section from source to v1 were taken using XBT's as a form of "sea-truth". They make an informative comparison with the acoustically and altimetrically derived temperature.
Data Comparison
Estimates of temperature or sea level can be made from all three types of measurements (acoustic, alimetric and XBT). Here we choose to compare derived sea level, because of the acknowledged complexity of estimating temperature from satellite altimetry. If we compare all three derived sea levels with the climatologicai average (say from Levitus' well-known atlas), we see that they all track each other remarkably closely over the 15 months of ATOC transmissions.
