1 INTRODUCTION

The Okhotsk Sea (OS) is one of the largest marginal seas connected with the subarctic gyre of the Northwestern Pacific and separated from the ocean by the Kuril Islands chain (Fig. 1). It is connected with the Japan Sea through the Soya/La-Perouse and Nevelskoy straits. The Soya Warm Current carries warm and saline Japan Sea Water through the Soya/La-Perouse Strait, flows along the coast of Hokkaido and gets into the ocean through the southernmost Kuril straits. The exchange between the seas through the very narrow Nevelskoy Strait is negligible.

Fig. 1.
figure 1

Bathymetry map of the Okhotsk Sea with a scheme of the main currents and geographic features. The Nevelskoy Strait (NS) connects the Japan and Okhotsk seas.

Full size image

The surface circulation in the OS is generally cyclonic with the West Kamchatka and East Sakhalin currents to be the eastern and western parts of this circulation [1]. The Kuril Basin and the southern OS are the areas populated with mesoscale (mainly anticyclonic) eddy. The Northwestern Pacific Water (NWPW) enters into the Sea through the northern Kuril Straits, mainly through the Fourth and Kruzenshtern straits [2, 3], and propagates along two main paths [4]. It is transported by the West Kamchatka Current, reaching the northern coast, and then this water is advected by the North Okhotsk Current to the west and southwest. The second pathway is provided by the Middle Current which follows the 500 m isobath and delivers the NWPW to the northeastern shelf of Sakhalin where it immediately flows into the East Sakhalin Current. It is considered that the transport along both the pathways is carried out in the surface layer [4].

The NWPW, transported to the northwestern OS shelf along the first path, is transformed into the Dense Shelf Water due to ice formation, intensive tidal mixing and invasion of continental cold air masses during the cold period of a year [5]. The Dense Shelf Water subducts into the intermediate layer (below 200 m) and is advected to the Kuril Basin by the East Sakhalin Current [6]. Then it mixes in the Kuril Basin with the forerunner of the Soya Warm Current Water and Western Subarctic Water [7] being accompanied by mixing due to strong tides [8]. As the result, the Okhotsk Sea Mode Water enters into the ocean through the southern Kuril Straits and is modified into the North Okhotsk Intermediate Water [9–11].

The conventional schemes of the OS circulation [1, 4] seem to be incomplete for the following reasons. (1) The North Okhotsk Current has a more complex structure, consisting of two branches, the first one above the continental slope and the second one along the northern shelf. The West Kamchatka Current transports most of the NWPW to the slope branch of the North Okhotsk Current which is in agreement with the oceanographic reanalysis by the HYCOM, ECCO2 and JCOPE2. (2) The NWPW flows into the OS not only in the surface layer, but also in the subsurface and intermediate layers. In different layers, its transport pathways to the northern OS shelf might be different.

In this paper, we use the eddy-permitting circulation model RIAMOM to simulate general circulation in the OS and to estimate and analyse the values of the simulated monthly-averaged volume transport of the main currents in different seasons.

2 MODEL

The RIAMOM has been developed at the Research Institute for Applied Mechanics of the Kyushu University by [12] as a 3D z-coordinate primitive equation OGCM with a free surface boundary condition for the sea surface height. The model is based on nonlinear primitive equations assuming the hydrostatic and Boussinesq approximations. It was described in detail by [13] and recently by [14]. So only a brief description is given here.

The model domain is within the boundary from 129° E to 165° E in longitude and from 40° N to 65° N in latitude. Horizontal resolution is 1/18° (within the range of 3–6 km) with 70 levels resolving vertical structure in ocean properties. The vertical grid intervals are from 10 to 125 m from the surface to 2500 m and are 250 m below 2000 m depth. This resolution is enough for eddy-permitting in accordance with recent estimates of the Rossby radius of deformation in the OS [15].

The model bottom topography is based on the GEBCO One Minute Grid. The data for initial conditions and boundary conditions along the southern and eastern open boundaries are taken from simulation results obtained using the Pacific Ocean 1/6° RIAMOM model by [16].

The net heat flux at the surface \({{Q}_{{NET}}}\) is parametrized using the formula by [17]:

$$\begin{gathered} {{Q}_{{NET}}} = ({{Q}_{{SOL}}} + {{Q}_{{IR}}} + {{Q}_{{LAT}}} + {{Q}_{{SENS}}}) \\ \times \,\,\,\left( {1 - {{C}_{{ICE}}}} \right) + \,\,\frac{1}{{{{t}_{T}}}}\rho {{C}_{P}}\Delta {{z}_{1}}({{T}_{{SAT}}} - T(t)). \\ \end{gathered} $$
(1)

The surface net solar radiation QSOL, surface net thermal radiation QIR, surface latent heat flux QLAT, surface sensible heat flux QSENS are taken from the ECMWF ERA-40 reanalysis [39]. CP = 4.18 × 107 \(\frac{{{\text{c}}{{{\text{m}}}^{2}}}}{{({{{\text{s}}}^{2}}\,{\text{K}})}}\) is the specific heat of sea water. TSAT is a satellite sea surface temperature (SST) in Celsius degrees, CICE is a sea ice concentration. TSAT and CICE are taken from the NOAA Optimum Interpolation of the 1/4° daily SST Analysis [18]. T(t) is the model surface temperature in Celsius degrees, Δz1 is the thickness of the first layer of the model, tT = 10 and tS = 30 are restoring time scale in days.

The salinity flux F is:

$$\begin{gathered} F = S(t)({\text{E}} - {\text{P}} - {\text{R}})(1 - {{C}_{{ICE}}}) \\ + \,\,\frac{1}{{{{t}_{S}}}}\Delta {{z}_{1}}({{S}_{{CLIM}}} - S(t)), \\ \end{gathered} $$
(2)

where the evaporation rate E and precipitation rate P are taken from the ECMWF ERA-40 reanalysis.

The climatological monthly Amur River runoff data R are taken from the Global River Discharge, 1807-1991, V.1.1 [19]. SCLIM is the climatic sea surface salinity (SSS) taken from the World Ocean Atlas 2001 [20]. S(t) is the model surface salinity.

We assume that water temperature cannot be below –2.0°C, and no heat flux through the ocean surface is set when the ice concentration equals to 1. The 2D fields of the net solar radiation, net long wave radiation, latent heat flux, sensible heat flux, evaporation, precipitation rate and wind are taken from the ECMWF ERA-40 reanalysis.

The 2D wind stress fields \({{\tau }_{{x,y}}}\) at the water surface were calculated using formula

$${{\tau }_{{x,y}}} = {{C}_{D}}{{w}_{{x,y}}}W(1 - {{C}_{{ICE}}}),$$
(3)

where

$$W = \sqrt {w_{x}^{2} + w_{y}^{2}} ,$$

CD is the wind-drag coefficient [21]:

$${{C}_{D}} = \left\{ \begin{gathered} 2.18\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,W \leqslant 3 \hfill \\ \left( {0.29 + \frac{{3.1}}{w} + \frac{{7.7}}{{{{w}^{2}}}}} \right)\,\,\,\,\,\,\,\,3 \leqslant W \leqslant 6 \hfill \\ (0.6 + 0.07W)\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,W \geqslant 6 \hfill \\ \end{gathered} \right.,$$
(4)

\({{w}_{x}},{{w}_{y}}\) are components of the wind speed in meters per second.

We use a restoring of surface temperature and salinity observation data in numerical experiments with the circulation model, taking into account the influence of ice with various concentrations on the momentum, heat and salt fluxes at the sea surface. The RIAMOM was run using the standard relaxation method by [17] for restoring of surface salinity and temperature from the observation data. Surface salinity and temperature were taken from the World Ocean Atlas 2001 [20] and NOAA Optimum Interpolation data [18], respectively.

We have also used the simple method for assimilation of sea ice concentration from the satellite observation data (NOAA/National Climatic Data Center) to calculate the heat, salt and momentum fluxes at the water surface under ice with different values of its concentration. These fluxes were considered to be proportional to concentration of sea ice. An increase in sea water salinity due to a brine rejection during the growth of sea ice from December to February and a decrease in surface water salinity during melting of ice in March–April were taken into account by a restoring of observed surface salinity from the World Ocean Atlas 2001. Using the data on concentration of sea ice, taken from the NOAA/National Climatic Data Center, the wind stress, the heat and fresh water fluxes were recalculated in accordance with Eqs. (1)–(3).

Daily mean data (ice concentration (ICE), SST, wind velocity (at 10 m above the surface), heat flux and evaporation-precipitation (EP) flux) for the period from January 1, 1991 to December 31, 2000 and climate monthly mean data of SSS are used as forcing. Monthly mean data (velocity, sea temperature, salinity sea level) for the period from January 1, 1991 to December 31, 2000 are used for boundary conditions on the open boundaries. Daily data were processed by a moving five-day average. The RIAMOM outputs for each day from 1991 to 2000 were analysed in this study. The simulated current velocity, temperature and salinity fields were compared with observation data for this period, when there were enough oceanographic observations to validate the model.

3 GENERAL CIRCULATION IN THE OKHOTSK SEA

An overview of the surface circulation in the OS is represented by the currents in Fig. 1 which exist throughout a year. The NWPW enters to the Sea mainly through the northern Kuril Straits (to the north of the Nadezhda Strait, including this strait) and forms the West Kamchatka and Middle currents. The West Kamchatka Current follows the continental slope splitting into two branches in the region of the TINRO Basin. The first one, the so-called Northern Branch, flows into the Shelikhov Bay. The second one forms an over-slope branch of the North Okhotsk Current which follows the northern continental slope between the TINRO Basin and Kashevarov Bank (Fig. 1). The Middle Current follows the 500 m isobath in the north-western direction. The Middle Current and the over-slope branch of the North Okhotsk Current converge in the area of the Kashevarov Bank and are transformed into an over-slope branch of the East Sakhalin Current. On the northern shelf, there is a coastal branch of the North Okhotsk Current. On the western shelf, there is a coastal branch of the East Sakhalin Current. The circulation in the south-western part of the Sea is dominated by the Soya Warm Current. The circulation is more complex in the other OS areas, including the Kashevarov Bank, TINRO Basin, Deryugin Basin, Shelikhov Bay and the Kuril Basin with a plenty of eddies of different size and polarity, especially in the Kuril Basin.

There is a strong seasonal variability in the OS circulation with the four different types, each one for its own season, whereas the intra-seasonal variability is weak. We summarize in Table 1 the values of the simulated monthly-averaged volume transport of the main OS currents: the coastal branch of the East Sakhalin Current (ESC1), the over-slope branch of the East Sakhalin Current (ESC2), the coastal branch of the North Okhotsk Current (NOC1), the over-slope branch of the North Okhotsk current (NOC2), the Middle Current (MC) and the West Kamchatka Current (WKC).

Table 1.   The simulated monthly-averaged volume transport of the main currents in the Okhotsk Sea in Sverdrup: the coastal branch of the East Sakhalin Current (ESC1), over-slope branch of the East Sakhalin Current (ESC2), coastal branch of the North Okhotsk Current (NOC1), over-slope branch of the North Okhotsk current (NOC2), Middle Current (MC) and West Kamchatka Current (WKC)
Full size table

The vertical section along 51° N is taken for the analysis of the ESC1 (within 143° E–145° E) and ESC2 (within 145° E–147° E). The width of the ESC1 is in the range 60–90 km with a minimum in summer and a maximum in winter. The width of the ESC2 varies from 100 km in the spring-summer period to 140 km in the late autumn-winter period. The vertical section along 53° N is taken for the analysis of the MC and WKC. The main stream of the MC flows approximately along 151° E–153° E. Its minimum width is 140 km in August-October and the maximal one is about 240 km in April. The main stream of the WKC is within 153° E–155° E. The minimal width of the WKC is 100 km in June-September and a maximal one is 250 km in October and February. The vertical section along 148.5° E is taken for the analysis of the over-slope branch of the NOC2. The main stream of the NOC2 flows along 55° N–57° N and its width is about 220 km. The vertical section along 141° E is taken for the analysis of the NOC1 (within 57° N–59° N) with the width of 10 km. Figure 2 shows the simulated velocity fields at the surface (the 1st layer) averaged during 1991–2000 for January, April, July and October.

Fig. 2.
figure 2

The simulated velocity field at the surface (1st layer) averaged for January, April, July, and October.

Full size image

The strongest currents are observed in winter and spring weakening in summer and autumn. In winter, all currents, excepting for the coastal branch of the North Okhotsk Current, demonstrate a maximal volume transport. A comparatively small transport of the coastal branch of the North Okhotsk Current is associated with an increased concentration of ice in the area in the cold period. The large transport of the coastal branch of the East Sakhalin Current is related to the fact that the ice sheet is set a little later and has not have enough time to impact strongly on the current. In spring, transport of the Middle Current, over-slope branch of the North Okhotsk Current and over-slope branch of the East Sakhalin Current are maximum, while transports of the West Kamchatka Current, coastal branches of the East Sakhalin and North Okhotsk currents do not vary significantly. In summer, practically all currents weaken, except for the over-slope branches of the North Okhotsk Current and East Sakhalin Current. The autumn season is characterized by an intensification of the coastal branches of the North Okhotsk and East Sakhalin currents. The autumn volume transport of the other over-slope currents is minimum.

4 MODEL VALIDATION

In this section we compare the corresponding model results with multi-year estimations of (1) volume transport between the Okhotsk and Japan seas through the Soya/La-Perouse Strait, (2) with the measured data on surface velocity on the eastern Sakhalin shelf and (3) with the measured data on subsurface temperatures in the key regions of the Sea.

4.1 Volume Transport through the Soya/La Perouse Strait

Historical review of the instrumental measurements from the beginning of the 20th century gave the volume transport from the Japan Sea to the OS in August to be 1.1 Sv [22]. It was confirmed by the measurements by [23, 24] and [25, 26]. The later measurements by [27] showed that the average annual volume transport was about 0.5 Sv with a flow from the Japan Sea to the OS, although the opposite situation was possible in winter.

The seasonal variation of the volume transport through the Soya Strait was estimated using the difference in sea level observed near Crillon (Sakhalin, Russia), and Wakkanai (Hokkaido, Japan) in the period 1975–1988 [28]. The volume transport was found to vary from –0.01 to 1.18 Sv with the annual mean of 0.61 Sv. The seasonal variation of the monthly mean volume transport was found to have a unimodal distribution with a maximum in August and a minimum in December–February.

The volume transport through the Soya Strait was previously thought to be about 1 Sv in summer, but almost nil in winter [29]. However, it has been revealed recently from the five across-current ADCP stations between 1999 and 2002 off the Hokkaido coast in the OS [3] that the volume transport of the Soya Warm Current varied in the range of 0.5–1.5 Sv. The volume transport, estimated with other measurements, was about 1.2–1.3 Sv in August, 1998 and 1.5 Sv in July, 2000 [30]. The ADCP data and high-frequency radars showed that the volume transport of the Soya Warm Current had a minimum in winter and a maximum in autumn with the annual averages in the range of 0.94–1.04 Sv [31].

4.2 Surface Velocity on the Eastern Sakhalin Shelf

Numerous instrumental measurements of currents have been carried out on the eastern Sakhalin shelf [27, 32–35]. We choose 8 long-term stations, where the measurements have been made over several months (Fig. 3 and Table 2). These measurements were made in the surface 10 m layer from an oil platform by the Regional Center for Oceanographic Data at the Far Eastern Regional Hydrometeorological Research Institute (Vladivostok, Russia). The tidal component was removed from the time series. After that, these data were compared with the model results for the same period of time and at the same points (Fig. 4), and the corresponding statistics were calculated. Statistical analysis, using the p-value method, shows that the correlation between the model results and measurements is significant, except for the flow direction at the station no. 4.

Fig. 3.
figure 3

Locations of 8 hydrographic stations with measurement of velocity in the surface layer on the eastern Sakhalin shelf. Bathymetric contours are drawn.

Full size image
Table 2.   Locations and dates of hydrographic stations with measurement of velocity in the surface layer on the eastern Sakhalin shelf
Full size table
Fig. 4.
figure 4

Comparison of measured and simulated velocities in the surface layer at 8 CTD stations on the eastern Sakhalin shelf. Here \({{r}_{\varphi }}\) and \({{r}_{U}}\) are the Pearson correlation coefficients between measured and model values of the flow velocity direction and between the measured and simulated values of the velocity modulus, respectively, with the green (red) colour \({{r}_{\varphi }}\) and \({{r}_{U}}\) corresponding to statistically significant (weak) correlation.

Full size image

4.3 Subsurface Temperature

The observed CTD temperature data, for the period from 1991 to 2000, were used for model validation. They were taken from the Regional Center for Oceanographic Data of the Far Eastern Regional Hydrometeorological Research Institute, Vladivostok, Russia (RODS FERHRI). The OS is conditionally divided into four geographic regions: the western, eastern, northern and southern ones. The two branches of the East Sakhalin Current are present in the western region. In the eastern region, there are the West Kamchatka and Middle currents. The two branches of the North Okhotsk Current are present in the northern region. The southern region is characterized by mesoscale eddies in the Kuril Basin and the presence of the Soya Warm Current.

The CTD data of oceanographic surveys were selected for each month from 1991 to 2000. Those months were selected for which the number of CTD stations allowed revealing the spatial temperature distribution. Among the months when the grid of CTD stations was relatively dense, there were 21 for the southern part of the OS, 19 for the northern, 12 for the western and 26 for the eastern ones. For example, in the northern part those months are: May and June 1991, May, June and July 1992, November 1993, September 1994, May and December 1995, May, July, August and September 1997, May and September 1998, May and September 1999, May and September 2000.

The data were vertically interpolated from the surface to bottom onto standard depth levels used in RIAMOM. Then the data were horizontally interpolated on a uniform horizontal grid. Finally, the measured fields were compared with the monthly mean model fields for the corresponding month of the corresponding year.

The distribution of model temperature in the eastern part of the OS correlates with the results of the CTD survey at the depth of 30 m and shows a flow of relatively warm water along the western Kamchatka shelf (Fig. 5). This stream corresponds to the West Kamchatka Current. The temperature distribution in the northern part of the sea shows a flow of comparatively warm water along the continental slope and a flow of cold water along the central shelf break (Fig. 6). The temperature distribution in the southern part of the Sea shows a complex eddy dynamic (Fig. 7). The warm water in the southwestern corner of figure is transported by the Soya Warm Current, and the flow of cold water from the north corresponds to the East Sakhalin Current. The World Ocean Database 2018 and the database of the Pacific branch of the Federal State Budget Scientific Institution “Russian Federal Research Institute of Fisheries and oceanography” (Vladivostok, Russia) were additionally used for a longer period of time, from 1980 to 2018. These oceanographic databases support the above-mentioned conclusion.

Fig. 5.
figure 5

(a) Measured and (b) simulated temperature at 30 m in April 1992 on the western Kamchatka shelf. The points are locations of the CTD stations.

Full size image
Fig. 6.
figure 6

(a) Measured and (b) simulated temperature at 30 m in May 1997 on the northern OS shelf. The points are locations of the CTD stations.

Full size image
Fig. 7.
figure 7

(a) Measured and (b) simulated temperature at 30 m in September 1997 on the northern Hokkaido shelf and in the Kuril Basin. The points are locations of the CTD stations.

Full size image

Some difference between the CTD data and the simulation results is explained as follows (for example, at the Kamchatka shelf in Fig. 5). The simulation results are monthly averaged data for April 1992. The CTD data were obtained during April 1992. The research vessel passed CTD stations from the south to north. The survey began in early April and ended in late April. Therefore, the temperature values at the southern CTD stations were lower than the monthly mean model results, and the temperature values at the northern CTD stations were higher than the monthly mean model results. That is why it seems that the observed Western Kamchatka Current propagates in the OS much norther than it really does.

5 DISCUSSION

The results of measurements and diagnostic calculations based on CTD-data do not allow a detailed analysis of the seasonal variability of the OS circulation [1]. The disadvantages of the Earth remote sensing from space also do not allow us to do that. The exceptions are the Soya Warm [30, 31] and East Sakhalin currents [33], the seasonal and interannual variability of which are relatively well studied by measurements. Therefore, it is necessary to use numerical models to analyze the OS seasonal variability. Available publications, where the results of numerical experiments have been presented, were focused mainly on the formation of dense shelf water [15, 36–40].

Therefore, for a more complete understanding of the features of water circulation, a comparison of different oceanographic reanalyses between themselves and with the results of RIAMOM has been carried out. The next reanalyses have been used:

(1) ECCO2/MITgcm with the horizontal resolution of 1/4 degree for the period 1993–2012 [41].

(2) FORA-WNP30 with the horizontal resolution of 1/6–1/10 degrees for the period 2001–2015 [42].

(3) JCOPE2 with the horizontal resolution of 1/12 degrees for the period 2001–2018 [43].

(4) GLORYS2V4 with the horizontal resolution of 1/4 degrees for the period 1993–2015 [44].

All the above data were monthly averaged over the indicated periods of time. Then, the volume transport of each of the main OS currents was calculated. The boundaries of the vertical sections, along which the transport was considered, were the same as for the RIAMOM (see Sec. 3). Separately, the transport of water over the northern shelf and slope was calculated through the section along 149E within 55N-60N (NT—Northern Transport).

All the reanalyses show the same set of main currents (ESC1, ESC2, MC, WKC, NOC1, NOC2) as the RIAMOM, with the exception of FORA-WNP30, in which there the Middle Current is absent. Strengthening of all the currents occurs between winter and spring. Weakening of all the currents occurs during the summer-autumn period. The single exception is JCOPE2, where all the currents intensify in fall.

The ESC1 reaches its maximum in autumn and winter (Fig. 8a). In spring, its intensity decreases, and in summer it reaches a minimum. The most intense ESC1 is obtained with JCOPE2 (with the annual average value of 1.6 Sv), the least intense with ECCO2/MITgcm (with the annual average value of 0.7 Sv).

Fig. 8.
figure 8

Seasonal variability of volume transport of the main current in the Okhotsk Sea: ESC1 (a), ESC2 (b), NOC1 (c), NOC2 (d), NT (e), MC (f), WKC (g).

Full size image

The ESC2 reaches a maximum in winter and spring and a minimum in autumn (Fig. 8b). The exception is JCOPE2 data where minimum is reached in March and maximum in September–October. The most intense ESC2 is obtained with JCOPE2 (with the annual average value of 2.7 Sv), the least intense – with FORA-WNP30 (with the average annual value of 0.8 Sv).

The NOC1 reaches a maximum between October and January and a minimum in July (Fig. 8c). The most intense NOC1 is obtained according to FORA-WNP30 (with the annual average value of 0.7 Sv), whereas the other sources show proportional values of average annual values (0.5 Sv).

NOC2 reaches a maximum in spring and a minimum in autumn (Fig. 8d). The exception is JCOPE2 data. According to them, minimum is reached in March and maximum rin September-October. The most intense NOC2 is obtained according to JCOPE2 (with the average annual value of 1.7 Sv) and the least intensive one according to GLORYS2V4 (with the average annual value of 0.2 Sv).

MC reaches a maximum in spring and a minimum during the summer–autumn period (Fig. 8f). In the FORA-WNP30 project, the MC is absent, while in JCOPE2 its seasonal variation is weakly expressed. The most intense MC is obtained according to ECCO2/MITgcm (annual average value of 1.25 Sv), and the least intense one according to GLORYS2V4 data (average annual value of 0.7 Sv).

WKC reaches a maximum between December and March and a minimum between July and September (Fig. 8g). In JCOPE2 data, the maximum is reached in fall. The most intense WKC is obtained according to JCOPE2 (average annual value of 1.35 Sv) and the least intense one according to GLORYS2V4 (annual average of 0.14 Sv).

The RIAMOM results have been found to be closest to ECCO2/MITgcm and, to a lesser extent, to GLORYS2V4. The main difference between the RIAMOM and FORA-WNP30 is the absence of the MC in the latter. The difference between the RIAMOM and JCOPE2 is in the seasonal variability.

The low transport of the NOC1 is due to the overestimated influence of the ice cover (Formulae 3 and 4), while the RIAMOM-based total flow over the northern shelf and slope (NT) is of the same order as in other models (Fig. 8e). Numerical experiments with the RIAMOM [6] show that the influence of tides increases volume transport of the NOC1 three times.

The North Okhotsk Current has a complex structure and consists of two branches, the coastal and over-slope ones. The coastal branch (NOC1) is more intense during the fall-winter period. The over-slope branch (NOC2) is more intense during the winter-spring period. The second fact is confirmed by all of the above oceanographic reanalyses (with the exception of JCOPE2). In addition, the presence of a current that exists over the continental slope of the northern part of the OS is indicated by the temperature distribution according to CTD data (WOD2018, RODS FERHRI, VNIRO DATASET). In particular, for the period from 1991 to 2000, according to RODS FERHRI, 11 cases from 19 show the obvious presence of a tongue of warmer water over the northern continental slope. Numerical experiments with the RIAMOM [6] show that the influence of heat and salt fluxes and the absence of relaxation do not affect the volume transport. The absence of wind reduces the volume transport of this current by half. The influence of tides reduces the volume transport of this current by one and a half times. Thus, there exists the over-slope branch of the North Okhotsk Current even if there is no external forcing.

6 CONCLUSION

The eddy-permitting model with the horizontal resolution of 3–6 km based on RIAMOM was applied to study the circulation in the OS during the period from 1991 to 2000. The model results have been found to be in a reasonable agreement with available instrumental measurements. The North Okhotsk Current was shown to have a more complex structure than it was presented on traditional circulation schemes, such as the Chernyavsky’s scheme [1]. We have found that the North Okhotsk Current consists of the branch over the continental slope between the TINRO Basin and Kashevarov Bank and the branch along the northern shelf. The Sea has four types of circulation, each one for its own season. A strengthening of all currents occurs in winter. A strengthening of the over-slope branches of the East Sakhalin and North Okhotsk Current and West Kamchatka Current occurs in spring. A weakening of all currents occurs in summer. A strengthening of the coastal branches of the East Sakhalin and North Okhotsk currents occurs in autumn. A cyclonic circulation is observed in the Shelikhov Bay from June to October and an anticyclonic one from November to May. We have found an anticyclonic circulation in the Kuril Basin from September to March and a complex eddy dynamics there from April to August with mesoscale eddies of various size and polarity. The RIAMOM-based volume transports of each of the main OS currents were compared to the corresponding transports obtained with the ECCO2/MITgcm, FORA-WNP30, JCOPE2 and GLORYS2V4 oceanographic reanalyses.