Steven WALKER

Scripps Institution of Oceanography


We are currently using anthropogenic chlorofluorocarbons (CFCs) as diagnostic tracers in large scale ocean models. This is accomplished by utilizing an historic atmospheric CFC source function to prescribe model surface water CFC concentrations. The model is then run in a diagnostic fashion so that the simulated CFC profiles are compared with available cruise track observations.

Simulations have been performed using the 22 layer Hamburg Large Scale Geostrophic (LSG) model which is a prognostic primitive equation model. However, in the South Atlantic and the Pacific, the LSG model forms deep water in the open ocean rather than along the continental shelves, and mixing occurs far more rapidly and over larger areas than in the real ocean. This is partly because of the large grid spacing and because of the type of convection parameterization used in the LSG. CFCs are therefore injected into the deep Southern Ocean incorrectly, making it difficult to diagnose the subsequent flow pattern correctly.

We have recently begun collaborative efforts with Reiner Schlitzer's research group at the Alfred Wegner Institute in Bremerhaven. He has developed an ocean model that combines a traditional box model with an Adjoint model. During the solution procedure, comparison is made between the model solution and existing hydrographic and CFC data. The adjoint section of the model then determines optimal changes in certain poorly constrained parameters (flow field, mixing coefficients and fluxes) which will minimize the overall error in the model solution and keep all parameters close to a priori estimates. We therefore arrive at a final model state which matches both the observations and the physics very closely. We are beginning our work using this model to perform relatively high resolution simulations for the Atlantic, but we plan eventually to expand this work to include Pacific and Indian simulations in a global Adjoint model to simulate world ocean circulation and hydrography.

We believe that, although such a purely diagnostic model cannot be used to simulate the effects of climate change, the Adjoint model incorporating existing observations is superior to traditional prognostic ocean models in determining today's flow field.


To Abstract List

To Pacific Workshop

To WOCE Home Page

uswoce@astra.tamu.edu