The wind-forced response of the Southern Ocean

Typescript (photocopy). In an analysis of satellite-tracked drifting surface buoys released in the Southern Ocean, buoy velocities are averaged for 90 days. The averaged velocity vectors show circulation that generally agrees with views of the large-scale circulation. The present analysis indicates...

Full description

Bibliographic Details
Main Author: Johnson, Mark Andrew
Other Authors: Nowlin, Worth D., Brooks, David A., McGuirk, James P., Reid, Robert O., Stecher, Michael J.
Format: Thesis
Language:English
Published: Texas A&M University. Libraries 1987
Subjects:
Online Access:https://hdl.handle.net/1969.1/DISSERTATIONS-26995
Description
Summary:Typescript (photocopy). In an analysis of satellite-tracked drifting surface buoys released in the Southern Ocean, buoy velocities are averaged for 90 days. The averaged velocity vectors show circulation that generally agrees with views of the large-scale circulation. The present analysis indicates that the spatial structure of an "eddy", defined as the average perturbation about the 90-day mean, has a meridional to zonal wavenumber ratio of 1.5. Regionally-averaged Lagrangian spectra of buoy velocities have largest amplitudes in the Indian Ocean and smallest amplitudes in the Pacific Ocean. For the southern hemisphere oceans combined, the period of the most energetic response is less than 30 days. To simulate buoy trajectories, a non eddy-resolving numerical model with a recirculating domain is constructed. Trajectories are numerically computed from the finite-difference, quasi-geostrophic, one-layer model forced by the annual average of the wind stress curl, scaled assuming a Sverdrup balance. Model streamlines are in reasonable agreement with a South Atlantic subtropical gyre, a Weddell Gyre, and a circumpolar current. However, compared to velocities of the surface drifting buoys, the model velocities are slower. Power spectra of accelerations, computed from trajectories simulated during the model spin-up, show statistically significant peaks at 25 days. This 25-day period is in the range of expected periods of wind-induced Rossby waves. The model is also forced with a wind stress field having negative wind stress curl at only one latitude and zero curl elsewhere. Numerical simulations have volume transport variations of 11% when the latitude of forcing changes by 100 km. A time series of the latitude of the maximum (negative) curl from the twice-daily winds is visually correlated with model volume transport and with transport from measured pressure differences across Drake Passage, when variations in transport lag 25 to 30 days behind variations in the latitude of the wind stress curl. This study shows ...