Abstract:

The main objectives in this dissertation are to examine the impact of the seasonal cycle on El Ni?o-Southern Oscillation (ENSO), and to provide a quantitative validation of the mechanisms for the impact proposed by previous studies. Hypothesis-validation studies are performed in the control environment provided by ENSO events simulated with a coupled general circulation model (CGCM). Using this approach the simulated seasonal cycle and interannual variability can be separated and artificially modified in view of the aspect selected for examination. The CGCM used in this study is the tropical Pacific version of the UCLA CGCM.

An analysis of the energy balance in the tropical Pacific Ocean in simulations by the CGCM and by its OGCM component with forcing based on observational data address the irregularities of ENSO. The results suggest that ENSO irregular features manifested in the evolution of energy components on interannual time scales are due to the impacts of seasonal cycle-ENSO interactions.

The impact on ENSO of the seasonal cycle in the upper ocean current system and thermocline structure is tested with uncoupled OGCM by shifting the phase relation between the seasonal and interannual components of the surface forcing. It is found that such a shift can significantly affect the evolution of heat content anomalies in the equatorial eastern basin. The impact is primarily due to changes in the anomalous zonal advection of temperature.

The following hypotheses about the impact on ENSO of the atmospheric seasonal cycle are validated in a coupled setup with the CGCM: (1)?the seasonal warming of the cold tongue during January-April favors the initial growth of warm events; (2)?the northward migration of the convergence zone in the western part of the basin during April-May plays a significant role in the initial growth of warm events; and (3)?the southward migration of the convergence zone in the western part of the basin during December-January favors the demise of ongoing warm event. The methodology used is based on CGCM simulations in the "anomaly coupling" mode. This allows for comparisons of runs with fixed and time-varying seasonal conditions in the model's atmospheric component only. The results confirmed the validity of hypotheses (1) and (3), but do not provide evidence to validate (2). Consideration of "random" atmospheric noise in these experiments does not affect these conclusions, albeit we find evidence that "random" atmospheric forcing can greatly alter the evolution of the anomalies in some cases.