Nonlinear responses of oceanic temperature to wind stress anomalies in tropical Pacific and Indian Oceans: A study based on numerical experiments with an OGCM

Abstract

As a highly nonlinear dynamic system, oceanic general circulation models (OGCMs) usually exhibit nonlinear responses to prescribed wind stress forcing. To explore mechanisms for these nonlinear responses, we designed and conducted three idealized numerical experiments with an OGCM with modified wind stress forcing. In the experiments, the climatological mean wind stress was identical, and the only differences in external forcing were wind stress anomalies. The wind anomalies were set to zero in a control run, and the observed wind stress anomalies with and without reversed signs were superimposed on the mean climatology in two sensitivity experiments. Forced by the prescribed wind stress anomalies in sensitivity runs, the OGCM well reproduced the El Niño-Southern Oscillation (ENSO) and the Pacific and Indian Ocean Dipole (IOD) in the Indian Ocean, as well as the asymmetry between positive and negative phases of these modes. Relative to the control run, the two sensitivity runs exhibited almost identical changes in the mean climate state, although the wind stress anomalies were reversed in these two experiments. Thus, it was concluded that the asymmetry of wind stress anomalies contributes only slightly to the mean state changes and ocean internal dynamics was the main contributor. Further heat budget analysis suggested that nonlinear temperature advection terms, including both mean advection and perturbed advection, favor the ENSO/IOD rectified effect on the mean state. Citation: Hua Lijuan and Yu Yongqiang, 2015: Nonlinear responses of oceanic temperature to wind stress anomalies in the tropical Pacific and Indian Oceans: A study based on numerical experiments with an OGCM.

The SST–Wind Causal Relationship during the Development of the IOD in Observations and Model Simulations

“…Current state-of-the-art climate models can adequately simulate certain features and behaviors of the IOD, such as its phase locking (the IOD peak always occurs in autumn), asymmetry (i.e., the asymmetry of the SSTA amplitudes between the positive and negative phases of the IOD), and period (the interval between IOD occurrences) (e.g., Yao et al, 2016 [20]; Hua and Yu, 2015 [21]; Wang et al, 2014 [22]). However, large errors still exist between IOD observations and the simulations using current numerical models; these errors manifest mainly via a warming bias throughout the IO in model simulations [23,24].…”