Effects Of The Kuroshio Extension Bimodal States On Temperature And Salinity And The Predictability Studies Of Its State Transition

The Kuroshio Extension(KE)exhibits the typical bimodal states –– a stable state and an unstable state –– on the decadal time scales.The anomalous temperature and salinity variations in the different Kuroshio Extension(KE)states have significant effects on the atmospheric circulation,mode water and survival of fish.Therefore,in this dissertation,observation and reanalysis data are used to explore the effects of the KE bimodal states(the KE decadal variations)on the seasonal changes of anomalous temperature and salinity in the surrounding region of the KE.Our analysis reveals that decadal variations of the KE play an important modulation role in the seasonal changes of the mixed-layer salinity anomaly(MLSA)and mixed-layer temperature anomaly(MLTA)in the Kuroshio-Oyashio Confluence Region(KOCR).Both variables have similar seasonal variation trend under the KE bimodal states.In the stable state of the KE,they increase first and then decrease,while in the unstable state,they change in the opposite trend.However,the factors causing their respective seasonal variations are different.For the MLSA,the contribution of ocean advection is primary and that of the atmospheric forcing(freshwater flux)is secondary.For the MLTA,the ocean dynamic processes,including ocean advection and entrainment process,have a significant impact on its seasonal variations,while the atmospheric forcing(heat flux)has a negative feedback,but not obvious.Due to the important role of the KE bimodal states in the North Pacific climate system,we adopt an eddy-resolving Regional Ocean Modeling System(ROMS)and the conditional nonlinear optimal perturbation(CNOP)method to investigate the predictability of the KE transition process and conduct its target observation.In this dissertation,the KE state transition is successfully simulated using ROMS model,and the simulation is consistent with the observation in time.Based on the results of the simulation,a nonlinear optimization system is established to obtain the fastest-growing initial errors(CNOP-type initial errors)in the KE transition from the stable to the unstable state.Our results indicate that the large values of this kind of initial errors are mainly located in the upstream of the KE over the upper 500 m.Besides,from the viewpoint of the relative vorticity and eddy-energy,the dynamic mechanism of the error growth is explored in terms of the error maintenance and development during the KE transition period.The results indicate that the errors are transported eastward from the south of Japan to the KE first quasi-stationary meander by linear advection.And the nonlinear advection slows down the moving speed of these errors,which lead to the long-term existence of them in the KE first quasi-stationary meander.In this situation,errors have enough time to absorb energy from the background field through barotropic energy conversion continuously.Ultimately,errors grow rapidly and the KE strength is overestimated.Furthermore,based on the spatial structure of the fastest-growing initial errors,the sensitive areas of target observation of the KE state transition are identified.The sensitivity experiments in terms of the spatial locations and patterns of the initial errors are conducted.And the validity of the sensitive areas and the importance of the structures of the CNOP-type initial errors in this area are verified in the prediction of the KE state transition.Moreover,the target observation improvement is evaluated theoretically with the help of Observing System Simulation Experiment(OSSE).The results show that the target observation is relatively suitable when it is implemented in the sensitive area,which can effectively reduce the prediction uncertainty and improve the predicted results of the KE state transition to a larger extent.