That is, strong eddies mix the potential vorticity effectively, as well as provide strong eddy forcing, Ieading to the intensification of the recirculation gyre. Then, the intensified recirculation interacts with the Kuroshio current to enhance the meander, which further induce even stronger eddies and fluctuations.
Figure 7 indicates the latitude of the Kuroshio path, defined as the maximum velocity off the south coast of Japan, when the inflow transport is increrased very slowly from 25 Sv to 45 Sv with the rate of 5Sv increase per year. The boundary forcing is considered to be quasi-steady because of this slow changes in the inflow condition. In the case of increasing transport, the Kuroshio starts to take the meander path when the transport passes about 40Sv. In the case of decreasing transport, the Kuroshio meanders initially, it then suddenlytakes the non-meander path from about 30 Sv. Therefore, there are multiple states in the Kuroshio path in the inflow transport range between 30 Sv and 40 Sv in this experiment. We can also see that there is stronger variability in the axis position when the Kuroshio took the large meander path. This implies that this multiple states arise from the non-linear nature of the recirculation whose potential vorticity is markedly homogenized as discussed above.
Fig 7: Relation between the inflow transport and the position of the south-ernmost Kuroshio axis, indicating the amplitude of the meander.
4. Conclusion
We have shown that the high resolution regional model successfully simulate the multiple paths of the Kuroshio. In this paper, it has been particularly highlighted that the non-linear nature of the Kuroshio path in this model arises from the potential vorticity mixing in the recirculation gyre off Shikoku; potential vorticity there is markedly homogenizd when the Kuroshio takes the large meander path.
The potential vorticity mixing is accompanied with strong 50 - 100 day variations in transport. Such variations have been indeed observed along the south coast of Japan.
