Ocean acoustic tomography: Mesoscale velocity

Abstract

The travel times of acoustic pulses transmitted in opposite directions over a 300‐km distance in mid‐ocean have been used to measure the fields of sound speed and absolute water velocity with mesoscale resolution. The experiment, which took place west of Bermuda during August and September 1983, consisted of transceivers on two moorings that sent and received signals every 2 hours for 21 days. Ray paths in opposite directions were found to be very nearly reciprocal; effects due to internal waves and mean currents were small. Sum travel times were inverted to obtain sound speed at 2‐day intervals, and differential travel times were inverted to obtain absolute current velocity u . The principal results are (1) range‐averaged sound speed perturbation δ C ( z ) decreased over the 21 days by 5 ±0.2 m s −1 at 1000m depth, corresponding to a temperature decrease of approximately 1°C; similar (but lesser) changes are measured at other depths. The changes are associated with the first baroclinic mode. (2) The field δ C ( x , z ) determined from expendable bathythermograph (XBT) measurements at the start of the experiment agrees with the field determined using the acoustic data within estimated errors of ± 1.6 m s −1 . It is thus possible to obtain range‐dependent δ C information from acoustic measurements made in a single vertical slice. (3) During the 21 days the range‐averaged barotropic velocity mode increased by a factor of 3, while the baroclinic mode decreased by a factor of 2. The estimated error in range‐averaged u ( z ) is ±2 cm s −1 above the main thermocline and ±1 cm s −1 below, while for range‐dependent u ( x , z ) the errors are 5 and 3 cm s −1 , respectively. No significant range dependence in u ( x , z ) was found. The baroclinic velocity agrees with geostrophic velocities inferred from the XBT measurements at deployment and aircraft‐deployed XBT measurements 11 days after the end of the reciprocal transmissions.