The Predictability Studies Of The Kuroshio Large Meander Path South Of Japan And Its Adaptive Observation

Based on the high-resolution Regional Ocean Modeling System(ROMS)and Conditional Nonlinear Optimal Perturbation(CNOP)method,this paper conducted the predictability studies on the Kuroshio large meander south of Japan and its adaptive observation,the main content and conclusions are as follows:Firstly,we simulated the Kuroshio path variations south of Japan by one-way nesting,and the results show that ROMS not only can capture the bimodal characteristics of Kuroshio south of Japan but also successfully reproduces the Kuroshio Extension system,with a reasonable recirculation gyre located in the Shikoku Basin and two quasi-stationary meanders east of 140°E.Furthermore,by comparing with the satellite observation data,it is found that the formation process of the simulated Kuroshio large meander path is similar to the 2004/05 large meander,including the trigger meander southeast of Kyushu Island and its propagation along Japan coast.Based on the above analysis,it is concluded that ROMS has a good simulation ability for the Kuroshio path variation,especially,especially the formation process of the large meander.Secondly,in order to explore what initial perturbation is most likely to trigger the transition of the Kuroshio from the non-large meander path(straight path)to the large meander,i.e.,the optimal triggering perturbation of the Kuroshio large meander,the non-large meander path is used as the reference field,two cases are selected,which denoted as Case 1 and Case 2 respectively.By means of the nonlinear optimization system based on ROMS,we calculate the optimal triggering perturbation of the large meander path,its large amplitudes are mainly located in the upper 2500 m in the southeastern area of Kyushu(29°–32°N,131°–134°E),indicating the perturbations in this area is vital to the formation of the large meander path.Further,by exploring the nonlinear evolution of the optimal triggering perturbation,and by means of the depthintegrated vorticity equation,the role of the nonlinear physical processes(the interaction between the perturbation and the perturbation)is revealed.The results have shown that the linear advection(the interaction between the perturbation and the reference field)tends to move the cyclonic eddy in the center of the meander path eastward,while the nonlinear processes associated with the interaction between the perturbation and the perturbation tends to move the cyclonic eddy westward.It is the nonlinear processes that largely offset the effect of the linear advection,thereby slowing the eastward movement of the cyclonic eddy,allowing it to be fully developed,and ultimately resulting in the Kuroshio large meander path.Then,in order to explore what initial error can lead to the largest forecast error in the prediction of the Kuroshio large meander path,two large meander formation processes caused by the nonlinear optimal triggering perturbation are treated as the new reference fields(re-represented as Case 1* and Case 2*),we calculated the optimal initial errors in the prediction of the large meander path,and explored their growth mechanism.For each case,two types of initial error(denoted as CNOP1 and CNOP2),which have almost opposite spatial structures,are obtained,and their large amplitudes are mainly located in the upper 2500 m southeast of Kyushu i.e.,the upstream region of the large meander.By analyzing the patterns and nonlinear evolution of two CNOPtype errors,it is found that CNOP1 tends to strengthen the large meander path through southwestward extension,while CNOP2 tends to weaken the large meander path through northeastward contraction.Through the perturbation-energetics analysis,it is found that both the barotropic instabilities and baroclinic instabilities both contribute to the transfer of the energy from the reference field to the error field,resulting in rapid development of the error.Therefore,both barotropic and baroclinic processes play important roles in the growth of CNOP-type optimal initial errors.In addition,the paper also verifies the similarity of the optimal initial error and the nonlinear optimal trigger perturbation in spatial distribution.Finally,based on the similarity of the optimal triggering perturbation and the optimal initial error,the adaptive observation sensitive areas for Case 1* and Case 2* are identified.By the sensitivity experiments,we verify the validity of the sensitive areas and investigate the effects of the spatial patterns and the locations of initial errors on the large meander path prediction.Further,by means of the Observing System Simulation Experiment(OSSE),the improvement of the adaptive observation is theoretically evaluated.The results show that the adaptive observation in the sensitive area can effectively reduce the forecast error and improve the forecast result to a large extent.Specially,the improvement for Case 1* is 27% or so and for Case 2* is approximately 20%.The results derived in this study further deepen the understanding on the formation mechanism of the large meander path,and provide effective theoretical guidance for the prediction and observation of the Kuroshio large meander path.