Properties of winter mixed layer variability on the shelf-slope region facing the Kuroshio—study of Tosa Bay, southern Japan

Abstract

To examine the properties of winter mixed layer (ML) variability in the shelf-slope waters facing the Kuroshio, we analyzed historical temperature records and the simulated results of a triply nested high-resolution numerical model. As a candidate of the shelf-slope waters, we focused on Tosa Bay, off the southern Japan. A time series of observed monthly mean ML temperatures and depths in the bay exhibits a remarkable seasonal variation. The period when the ML develops can be divided into two regimes: from September to November, when the sea surface cooling is gradually enhanced, the ML temperature and depth decreases and increases, respectively; from January to March, the ML temperature and depth are kept nearly constant, while the sea surface cooling in January reaches its annual maximum. In the latter regime, variance for the monthly mean ML depth is the largest of the year. To further study the ML properties in the latter regime corresponding to winter, we examined simulated results. It was found that the largest variance for ML depth is attributed to a dominant intramonthly variation. This is related to a submesoscale variation with typical spatial scales of 10–20 km, induced by the Kuroshio and its frontal disturbances. Simulated monthly mean heat balance within the ML showed that heat advection balances with heat flux at the sea surface and entrainment through the ML bottom. Moreover, the monthly mean heat advection is determined mainly by the intramonthly eddy heat advection, suggesting that the high-frequency intramonthly variation related to submesoscale variations contributes significantly to the low-frequency monthly variations of the ML in winter.

Appendix

In this section, heat balance within the ML is briefly described in terms of spatial mean. Figure 18 shows the 12-hourly mean of major four terms in Eq. (1), which is averaged over Tosa Bay (133°30′–134° E, 32°50′–33°30′ N). The tendency and advection terms seem to mainly balance with each other as well as Fig. 14, whereas magnitudes of the two terms are smaller than over the bay (Fig. 18) than at station A (Fig. 14). The smaller magnitude signifies that effects of submesoscale variations on the heat balance are reduced or compensated by the spatial-mean procedure, implying the importance of submesoscale ML variability at a local position in Tosa Bay. Moreover, it was confirmed (not shown) that the monthly mean of heat balance averaged over Tosa Bay is almost the same as that at station A (Fig. 13), indicating that the monthly mean heat balance at station A, from which effects of submesoscale variations are excluded, can be representative of that over the bay.

Fig. 18

Same as Fig. 14, except for spatial mean over Tosa Bay