Predictability of the tropical Atlantic Ocean
In: Journal of Marine Systems. Elsevier: Tokyo; Oxford; New York; Amsterdam. ISSN 0924-7963; e-ISSN 1879-1573, more
Also appears in:Nihoul, J.C.J. (Ed.) (1990). Coupled ocean-atmosphere modeling: proceedings of the 21th International Liège Colloquium on Ocean Hydrodynamics, Liège, May 8-12, 1989. Journal of Marine Systems, 1(1-3). Elsevier Scientific: Amsterdam. 313 pp., more
|  |
| Abstract |
We have used a primitive equation multi-level model of the tropical Atlantic Ocean to calculate the time evolution of sea surface temperature and ocean circulation due to a prescribed surface stress forcing for a given year. We have then repeated the model integration starting from the same initial oceanic state as the first integration but with quite different surface stress forcing given by atmospheric observations from another year. We have examined the rate at which the differences between the two model-simulated ocean states grow with time and find that it takes only about three months for the differences to grow from zero to their saturation value. We have also examined the time growth of differences between two ocean model simulations for which the atmospheric forcing of surface stress was identical but the initial ocean states were quite different. In this case also, we find that it takes only about three months for the initial large differences to decay to their minimum value.The results of these experiments lead to the following conclusions: (1) In the absence of accurate surface forcing, an accurate estimate of the ocean circulation at some initial time will lead to a reasonable estimate of circulation for only a few months. (2) Given accurate surface forcing, errors in the thermal and velocity structure of the ocean decrease rapidly within a few months. However some residual error will remain. |
|
