Sensitivity of a general circulation inverse model to sub-grid scale parametrization coefficients

Abstract

Ocean modellers seek to understand the circulation of the oceans, or portions thereof, by developing models of the ocean they can solve. This tractability constraint forces ocean modellers to make choices. Naturally, they hope to make intelligent choices, but whenever a new model is being developed or an existing one extended, the issue of tractability lurks. The large-scale, basin-wide, circulation of the oceans can be divided into two components, classified by their driving force. The wind-driven circulation, whose flow occurs mainly above the thermocline, was first explained qualitatively by Stommel (1948) with a simple, elegant analytical model. The other component of the oceans' circulation, the density-driven, or thermohaline circulation, flows below the thermocline. Again, the first simple analytical model for the deep thermohaline flow was proposed by Stommel (1958) and developed by Stommel and Aarons (1959) whose basic ideas underlie even the most recent conceptual models of the large-scale circulation. The details of the thermohaline circulation and its interaction with the wind-driven circulation in a realistic ocean basin is a problem which is not tractable analytically. This has driven ocean modellers interested in this aspect of the oceans' circulation to numerical models: ocean circulation models.