Sensitivity of eddy-induced heat transport to diabatic forcing

Research output: Contribution to journalArticleAcademicpeer-review


Compensation of the poleward eddy heat transport by the heat transport of an eddy-induced mean meridional overturning cell is a common feature in many eddy-resolving ocean models. It has been argued that this is the result of the weak thermal driving of the ocean. As the actual air/sea coupling is scale dependent, it might be questioned whether the approximation of weak thermal driving is relevant for the oceanic eddy field. In this paper the role of diabatic forcing in modifying eddy-mean flow interaction is investigated. Emphasis has been placed on the sensitivity of the eddy-induced change in heat transport to the sea surface thermal boundary condition. Experiments have been performed with a multilayer isopycnic primitive equation model of an idealized North Atlantic subtropical and subpolar gyre. For different values of the air/sea coupling, solutions with and without transient eddies have been compared. The air/sea coupling mostly affects the upper ocean thermal and velocity fields. A decrease of the coupling timescale pushes the separation point of the midlatitude jet further northward and induces a tight recirculation southwest of the separated jet. These effects are enhanced by the eddies. In the present model there is compensation of the eddy heat transport for sea surface temperature (SST) relaxation times longer than 150 days; a breakdown of the compensation occurs for SST relaxation times shorter than 50 days (the average upper layer depth is 200 m). In between is a transition regime. For strong thermal driving the eddy-induced change in total heat transport is of the same order as the eddy heat transport.
Original languageEnglish
Pages (from-to)18481
JournalJournal of Geophysical Research
Issue numberC9
Publication statusPublished - 1994


Dive into the research topics of 'Sensitivity of eddy-induced heat transport to diabatic forcing'. Together they form a unique fingerprint.

Cite this