Record Details
Numerical and assimilative studies of the equatorial Pacific cold tongue
ScholarsArchive at Oregon State University
| Field | Value |
|---|---|
| Title | Numerical and assimilative studies of the equatorial Pacific cold tongue |
| Names |
Perez, Renellys Christine
(creator) Miller, Robert N. (advisor) |
| Date Issued | 2005-10-31 (iso8601) |
| Note | Graduation date: 2006 |
| Abstract | Numerical model and assimilation experiments were conducted in the tropical Pacific Ocean to obtain a better understanding of the processes that control the cold tongue surface mixed layer temperature balance during August 1999 to July 2004. The numerical model was first applied to test two hypotheses (asymmetric background currents and asymmetric wind forcing) for the observed asymmetry of annual equatorial Rossby waves. The model with asymmetric background currents perturbed with symmetric annually-varying winds consistently produced asymmetric Rossby waves, and simulations with symmetric background currents perturbed by asymmetric annually-varying winds failed to produce the observed Rossby wave structure unless the perturbation winds were strong enough for nonlinear interactions to become important. The observed latitudinal asymmetry of the westward phase speed was found to be critically dependent on the inclusion of realistic coastline boundaries. To measure the cold tongue sensitivity to errors in wind forcing, the next study compared the seasonal cycle response of the model driven by different wind stress products. The FSU wind stress produced the least realistic cold tongue, and both the ECMWF and QuikSCAT wind stress driven model runs exhibited cold tongue annual cycles, tropical instability waves, and annual equatorial Rossby waves that compared well with observations. The highest realism, however, was obtained with QuikSCAT wind forcing. In the final modeling study, mean dynamic height biases resulting from climatological drift away from the Levitus initialization were discovered in the waveguide. The assimilation experiments combined the model driven by 5-day QuikSCAT winds with 5-day Tropical Atmosphere Ocean dynamic height anomalies via a reduced state space Kalman filter. Assimilation improved the interannual and intraseasonal variability of sea surface height, reduced the cold tongue bias in the waveguide, increased the core strength of the Equatorial Undercurrent, and produced more realistic albeit weak tropical instability waves. An autoregressive model added to the innovation sequence further optimized the assimilation scheme, but did not correct the pre-existing cold tongue thermal biases. Despite the decrease in positive (warming) high-frequency horizontal advection associated with TIWs, the assimilation run with the autoregressive model did not alter the mean balance significantly as there was a compensatory decrease in magnitude of the cooling by the low-frequency horizontal advection. Based on comparisons with observations, the annual cycle of the model tendency was too weak in the eastern Pacific giving rise to sea surface temperatures that were too cold in the spring and summer months and during the 2002-2003 El Niño event. Errors in the simulated net surface heat flux, vertical entrainment, and diffusion were identified as sources for the unrealistically low annual amplitudes of sea surface temperature and tendency in the model cold tongue. |
| Genre | Thesis/Dissertation |
| Topic | Ocean currents -- Pacific Ocean -- Mathematical models |
| Identifier | http://hdl.handle.net/1957/22945 |
