Alumni Project

Decadal Variability in the Coupled Ocean-Atmosphere System

Paola Cessi, Scripps Institution of Oceanography - UCSD

Summary

It is shown that on decadal time-scales the ocean has a global response to surface forcing that includes changes in sea-level in basins very distant from the region of direct forcing. Moreover, the spatial pattern of remote response depends very little on the pattern of the forcing, and is thus quite predictable. In a separate research project, the role of time-dependent waves and vortices on the scale of 10 to 100 km in shaping the vertical distribution of oceanic density is examined. We show that these flows, usually not included in global climate simulations, play a fundamental role in the global heat and energy budget of the ocean.

There are two main themes in this DOE project: one concerns the decadal global modes of variability of the upper oceanic circulation and their role in the long period fluctuations of the coupled ocean/atmosphere system. The other theme is the role of mesoscale eddies, i.e. oceanic flows on scales of 10 to 100 km, in the large-scale heat and energy budget of the ocean.

Global modes of variability

In this part of the project, we are focusing on decadal, large-scale oceanic basin modes, which arise from the planetary geostrophic balance and the constraint of global mass conservation. The mechanism through which these modes are generated is the following: In a closed basin, a localized displacement of density surfaces produces a global response by exciting rapidly moving gravity waves which readjust the local mass imbalance. When the pressure signal thus generated reaches the eastern boundary it sheds Rossby waves, which are much slower and cross ocean basins on decadal scales in mid-latitudes. As the Rossby wave front slowly progresses westward, the boundary pressure continuously adjusts in order to maintain mass balance, and thereby replenishes the source of Rossby waves. The net result is a sequence of oscillating patterns which decay in time extremely slowly. The least damped mode has a period which coincides with the transit time of the slowest baroclinic Rossby wave, and the other modes have frequencies that are integer multiples of the gravest mode. Atmospheric forcing near these frequencies can resonantly excite these modes.

The large scale structure and low frequency of these modes implies global oceanic teleconnections with signatures in all the basins of the world ocean. In particular a global response can be excited by an appropriate time- dependent wind-forcing or heating-cooling pattern, even if confined to a single hemisphere of a single basin. The remote signal is characterized by a heaving of the density surfaces, accompanied by fluctuations of the sea-level with a characteristic spatial pattern that has the largest variations in the Western corner of the high latitude region in each basin.

The existence of remote response to localized forcing has several implications for sea-level changes: Firstly, the sea-level change observed at one location might be generated by forcing which is very distant in space and that occurred several years earlier: secondly, away from the direct region of forcing, the response is largest in high latitudes and smallest in the tropics.

The eddy-driven stratification

The role of baroclinic eddies in transferring thermal gradients from the surface of the ocean to the abyss, and thus determining the stratification, is examined. Our hypothesis is that the density difference imposed at the surface by differential insolation is a large source of available potential energy which can be released by mesoscale eddies with horizontal scales of the order of 10 km. Eddy-fluxes of heat deepen the region of horizontal gradients (i.e. the thermocline). This conjecture is in contrast with the current thinking that the deep stratification is determined by a balance of turbulence at smaller scales and the planetary-scale ocean circulation.

Eddy-processes are analyzed in the context of a rapidly rotating primitive-equation model driven by specified surface temperature, in the presence of isotropic diffusion and viscosity. The model solutions are obtained by long numerical simulations (each simulation spans several model-decades) on a fine three-dimensional grid (typically 128 x 128 x 100) points. The simulations are largely performed on a multi-processor computer at Oak Ridge National Laboratory. Currently, the performance of the model on a multi-CPU, multi-node machine is limited by the use of an iterative algorithm which is not efficiently parallelized. In the next year, we plan to replace this algorithm with one that takes advantage of the multi-processor architecture.

The goal of these simulations is to establish scaling-laws for the depth of the thermocline and for the mean poleward heat transport as a function of the external parameters. In turn the scaling laws can be used to parametrize the eddy-fluxes in non eddy-resolving climate models.

Publications to date:

Ferrari, R. and P. Cessi, 2003. Seasonal synchronization in a chaotic ocean-atmosphere model. In press. J. Climate.

Spydell, M. and P. Cessi, 2003. Baroclinic modes in a two-layer basin. J. Phys. Oceanogr., Vol. 33, 610-622.

Cessi, P. and P. Otheguy, 2003. Oceanic teleconnections: Remote response to decadal wind forcing. (Accepted for publication in J. Phys. Oceanogr.)

Cessi, P. and M. Fantini. The eddy-driven thermocline. In preparation.

For further information contact:

Prof. Paola Cessi
Scripps Institution of Oceanography
UCSD-0213
La Jolla, CA 92093-0213
pcessi@ucsd.edu

back to project page

 

Home  |  ASCR  |  Contact Us  |  DOE disclaimer

URL http://www.scidac.gov/BER/BER_decvar/reports/decvar2003Annual.html
Updated: Friday, 18-Aug-2006 16:48:58 EDT