Numerical simulation of tidally induced eddies in the Bungo Channel: A possible role for sporadic Kuroshio-water intrusion (kyucho)

Abstract

It is well known that the sudden intrusion of Kuroshio warm water into the Bungo Channel (kyucho) is regulated by spring–neap tidal forcing. In order to clarify the physical background behind this regulation, numerical experiments are carried out using a high-resolution non-hydrostatic three-dimensional model. We first reproduce the strong mixing region off the east coast of the Bungo Channel resulting from tidal flow interaction with complicated land configurations during spring tides; behind islands and headlands, small-scale eddies satisfying an approximate cyclostrophic balance are generated. As a result, averaged over the whole model domain, the tidal-mean energy dissipation rate reaches ≈1.6 × 10⁻⁶ W kg⁻¹. The model predicted energy dissipation rates at the location and times of direct microstructure measurements in the Bungo Channel are comparable to the observed values. We next examine whether or not strong tidal mixing thus reproduced can inhibit the northward intrusion of Kuroshio warm water in the Bungo Channel. It is shown that the Kuroshio warm water can (or cannot) pass through the tidal mixing regions off the east coast of the Bungo Channel during periods of weakened (or enhanced) tidal mixing at neap (or spring) tides. This indicates that taking into account the realistic spring–neap modulation of tidal mixing intensity is indispensable to further increase the ability of the existing forecast system for kyucho in the Bungo Channel.

Appendix: Field measurement

In order to obtain real data of the ocean which provide useful information on the present numerical simulation, we carried out the first microstructure measurement in the Bungo Channel during spring tides (September 13, 2011) on board the R/V “Isana” of the Center for Marine Environmental Studies (CMES) of Ehime University. During the maximum ebb [between 1200 and 1300 (JST)], microscale velocity shear as well as temperature and pressure were measured by deploying a loosely tethered vertical microstructure profiler (“VMP-5500” manufactured by Rockland Scientific International Inc.) four times successively over 1 h at a location 20 km offshore of Uwajima (33.21°N, 132.26°E) (Fig. 1a).

The obtained microstructure velocity data was divided into consecutive segments of 8,192 data points, corresponding to a bin height of approximately 10 m. For each segment, the power spectrum was calculated using Welch’s averaged periodogram method with an overlapping fast Fourier transform (FFT) window length of 512 data points. The approximate value of dissipation rate ε was calculated by integrating the resulting vertical wavenumber power spectrum $\varphi \left( k \right)$ such that

$$ \varepsilon = \frac{15}{2}v\int\limits_{{k_{1} }}^{{k_{2} }} {\varphi \left( k \right){\text{d}}k} $$

(A1)

where k is the vertical wavenumber and is the kinematic viscosity. The lower integration limit k ₁ was set to 1 cpm, whereas upper limit k ₂ was set to the highest vertical wavenumber free from instrument’s vibration noise. For more details about the microstructure data analysis, readers are referred to Nagasawa et al. (2007). The vertical profiles of temperature and dissipation rates so obtained are shown in Figs. 3 and 4.