Diagnoses of simulated water-mass subduction/formation/transformation in the Japan/East Sea (JES)

Abstract

Abstract The impacts of surface atmospheric forcing of different time–space scales on the simulation of water-mass formation and spreading of formed water are investigated by quantifying water-mass subduction/formation/transformation for the Japan/East Sea (JES). The Princeton Ocean Model (POM) was implemented for the JES (JES-POM) to simulate interannual, seasonal, and mesoscale variations in velocity and mass fields. Three sets of atmospheric flux data were used; (1) 6-h fluxes calculated from 6-h atmospheric variables (syn), (2) monthly means of 6-h fluxes (empm), and (3) fluxes calculated from monthly averaged atmospheric variables (mont). The mass exchange between the mixed layer and interior in the JES was diagnosed in terms of annual subduction rate (Sann). Three areas of local maximum Sann (>500 m/yr) occurred: area V (41–43°N west of 137°E), K (36–39°N west of 132°E), and KB (near Korea Bay). Area V corresponds to the “flux center” (i.e., maximum heat and momentum fluxes) described by Kawamura and Wu [1998. Formation mechanism of Japan Sea Proper Water in the flux center off Vladivostok. J. Geophys. Res. 103 (C10), 21611–21622] the subduction region suggested by Senjyu and Sudo [1994. The upper portion of the Japan Sea proper water: its source and circulation as deduced from isopycnal analysis. J. Oceanogr. 50, 663–690; 1996. Interannual variation of the upper portion of the Japan proper water and its probable cause. J. Oceanogr. 52, 72–42] and Yoshikawa et al. [1999. Formation and circulation processes of intermediate water in the Japan Sea. J. Phys. Oceanogr. 29 (8), 1701–1722], and the wintertime convection location identified by Seung and Yoon [1995. Some features of winter convection in the Japan Sea. J. Oceanogr. 51, 61–73]. With monthly forcing (mont), there is no localized maximum value off Vladivostok, while with forcing influenced by synoptic events (monthly; empm and 6-h; syn), either one or two localized areas with a high subduction rate occur with year-to-year variations. The presence of simulated dense surface water ( > σ θ = 26.8 ) in the northern JES is due to the effects of synoptic events during winter months when large heat loss and strong wind stress associated with cold-air outbreaks and extratropical cyclones occur. The convection magnitudes as a result of air–sea interaction are 3.0, 4.0, and 6.0 Sv for mont, empm, and syn, respectively. A net flux of mass from the interior to the mixed layer (entrainment) occurs in the density range between σ θ = 24.0 and σ θ = 26.2 for both empm (3.5 Sv) and syn (3.0 Sv), while it occurs in the density range between σ θ = 24.0 and σ θ = 25.8 for mont (2.8 Sv). The diffusive fluxes across the winter mixed layer base are about 0.8, 1.0, and 1.4 Sv for mont, empm, and syn, respectively. An undiagnosed eddy-induced turbulent mixing (the residual of the balance after removing discretization error) for syn is almost twice that for the other two cases (mont and empm). These results indicate that without synoptic atmospheric forcing, the diagnostics using the numerical circulation model may significantly underestimate buoyancy loss at the surface, and, hence, water-mass formation, as well as mixing and spreading of the formed water mass.