University of Hawaii, International Pacific Research Center Honolulu, Hawaii, USA

maxy@soest.hawaii.edu

　

Climatological data on the oceanic and atmospheric variability are assimilated into a numerical model to study seasonal dynamics of the mode water production in the mid-latitudes of the North Pacific. The processed data sets include climatological fluxes of heat, salt and momentum at the ocean surface, Levitus hydrographic, TOPEX/POSEIDON altimetry and long-term surface drifter data. A variation data assimilation technique is used to retrieve variability in the open ocean region (18^o-42^oN, 142^oE-160^oW) bounded at 1000 m from below. By optimizing the open boundary values of oceanic fields in combination with initial conditions and atmospheric forcing, model solutions that are consistent with the climatological data sets within limits of observational errors were found.

　

Using this technique we estimated the mean geographical position of the Subtropical Mode Water (STMW) and Central Mode Water (CMW) formation sites, analyzed their structure, and obtained estimates of the mode water production rates. Our computations indicate that CMW formation is likely to occur 15^o-20^o west of the location diagnosed formerly without taking salinity data into account. The optimized seasonal cycle is characterized by the STMW and CMW production rates of 3.8±0.6 Sv and 3.1±0.5 Sv respectively. The KE annual mean transport in the upper 1000 m is diagnosed as 68±7 Sv with a maximum of 79±8 Sv in June-July and a minimum of 56±6 Sv in December-January. Analysis of the heat and salt budgets in the region has shown that atmospheric fluxes are counterbalanced by the horizontal divergence of the advective temperature and salinity transports. In the annual mean, horizontal diffusion plays a minor role in the budgets.

　

　

¹V.I. Il'ichev Pacific Oceanological Institute Vladivostok. Primorskii Krai, RUSSIA

trolia@poi.dvo.ru

　

²Faculty of Engineering, Kanazawa University Kanazawa, Ishikawa, JAPAN

　

The study is motivated by new findings on oceanography of the Japan Sea and research on climate change in the adjacent area. Long-term wintertime warming around the Japan Sea conditions an increase of vertical density stratification in the sea from mid to late 20^th century, while increased baroclinity affects sea circulation.

　

The hydrodynamic MHI model (Marine Hydrophysical Institute, Ukraine) is applied for numerical simulations. This is the primitive equation layered model suitable for simulating jet currents, fronts, and winter convection through entrainment/subduction and layer outcropping device. Seawater temperature, salinity, and buoyancy are allowed to vary horizontally in any layer, facilitating simulation of subduction and variation of water properties throughout the sea. The upper mixed layer model is incorporated, with buoyancy flux at sea surface approximated based upon external atmospheric characteristics. Two spin-up model runs are performed, starting from initial layer interface topography typical for 1950 (decreased baroclinity) and 1999 (increased baroclinity) based on data of oceanographic surveys in the Japan Sea.

　

Simulation results suggest that re-distribution of the Tsushima Current transport between western and eastern branches and regime of the northwest Japan Sea are highly affected by increased or decreased baroclinity. Intensification of the East Korean Warm Current in 1999 associated with increased baroclinity facilitates heat transport to the western and central sea area (through eddy streets device). Warm water is also transported to the area off Vladivostok from the east, with westward jets across the sea. The simulated circulation patterns are supported by observational evidence.

　

　

¹Computational Earth Sciences Division, Earth Simulator Center Yokohama, Kanagawa, JAPAN

sakuma@jamstec.go.jp

　

²Climate Variations Research Program, Institute for Global Change Research Yokohama, Kanagawa, JAPAN

　

Using a MOM3-based parallel OGCM code we optimized for the Earth Simulator, we have started a near global eddy-resolving (the horizontal resolution of 0.1 deg and 53 vertical layers) simulation, which is one of the Grand Challenge problems in earth sciences. The Earth Simulator is a vector parallel machine consisting of 5120 cpus with the total peak performance of 40 tera flops. Our code, though not so tightly optimized for the machine as our AGCM counterpart called T5-AFES, enables us to perform 10 model-year eddy-resolving simulation less than two days. As in the case of the North Atlantic sector simulation with the same horizontal resolution done by Smith and Maltrud, we noticed some realistic meso-scale features as well as pronounced western boundary currents emerged at the right places. Since our simulation produced huge amount of data, the analyses we got through with so far were fragmentary. Prior to the phenomena-oriented regional analyses, simulated basic heat and volume transports together with eddy activities in several key regions in the world oceans will be briefly discussed in addition to the attained computational performance of the code.