Dichothermal layer deepening in relation with halocline depth change associated with northward shrinkage of North Pacific western subarctic gyre in early 2000s
- 116 Downloads
In the western subarctic North Pacific, a wind-driven cyclonic circulation, called the western subarctic gyre (WSAG), exists. We examined year-to-year changes of the gyre and hydrographic structures, applying the altimetry-based gravest empirical mode (AGEM) method to hydrographic and altimetric sea surface height (SSH) data, and relation to the in situ variation of the temperature minimum layer, i.e., the dichothermal layer, depth at station K2 (47∘ N, 160∘ E). The AGEM-based geostrophic volume transport and the streamfunction of the WSAG in the top 1000-dbar layer show that the gyre changes substantially. From the late 1990s to the mid-2000s, the gyre shrunk northward. Due to the shrinkage, the halocline bottom, which is equivalent to the top of the main pycnocline, deepens at K2 outside the central part of the gyre. The downward displacement of the dichothermal layer at K2 was found to be significantly related to that of the underlying halocline due to the northward shrinkage of the WSAG.
KeywordsNorth Pacific western subarctic gyre Halocline Dichothermal layer K2 Hydrography Satellite altimetry AGEM
In the western subarctic North Pacific, a temperature minimum layer, called the dichothermal layer, exists in a layer between 100 and 200 m just above the main pycnocline (Dodimead et al. 1963; Favorite et al. 1976) and is occupied by the remnant of the mixed layer water from the preceding winter (e.g., Miura et al. 2002; Wakita et al. 2010, 2013). Wakita et al. (2010, 2013) reported a decadal variation of the dichothermal layer depth at stations KNOT (44∘ N, 155∘ E) and K2 (47∘ N, 160∘ E) in the western subarctic region from 1997 to 2011.
A regional cyclonic circulation, called the western subarctic gyre (WSAG), is embedded in the western part of the basin-scale subarctic cyclonic circulation (Dodimead et al. 1963; Favorite et al. 1976; Nagata et al. 1992). The WSAG involves the southwestward flowing western boundary current along the eastern coasts of the Kamchatka Peninsula and Kuril Islands (called the East Kamchatka Current or the upstream of the Oyashio), and the northeastward weak and wide return current in the North Pacific central region (Fig. 1a, b). The main pycnocline shoals toward the center of the WSAG (Miura et al. 2002) characterized by the minimum in sea surface dynamic height (Fig. 1a).
The variations of the subarctic circulations have been suggested to be caused by the change of wind stress curl field over the whole subarctic region (e.g., Sekine 1999; Qiu 2002; Kuroda et al. 2015). Recently, Kuroda et al. (2015) suggested a northward migration of the southern border of the Oyashio just southeast of Hokkaido, Japan, from the 1990s to the 2000s, which is presumably related to the long-term change of the basin-scale subarctic circulation caused by the change of the wind stress curl field. The variations of the intensity, extension, and shape of the circulations through the change of Ekman suction can yield the vertical displacement of the main pycnocline.
Because the strong density stratification in the main pycnocline prevents vertical mixing, the penetration of the developing winter mixed layer would be inhibited by the main pycnocline, as suggested by Miura et al. (2002, 2003). In other words, the depth of the main pycnocline would govern the potential maximal depth of the winter mixed layer, i.e., the dichothermal layer, in condition that the winter mixed layer develops sufficiently. If this suggestion by Miura et al. (2002, 2003) is valid in the real ocean, the variation of the main pycnocline depth due to the variation of the WSAG is identical to that of the dichothermal layer depth. However, because of a low spatial resolution and limited number of hydrographic observations in the western subarctic North Pacific by research vessels, the relationship between the variations of the WSAG and the dichothermal layer depth has never been verified yet by using time series data.
For the examination of changes of the horizontal and vertical features of the WSAG, data must be collected densely over a sufficiently broad area and its length must also be long. The Argo project was initiated in 2000, but until fairly recently, few Argo observations were performed in the western subarctic North Pacific region; therefore, the Argo data are not available for the examination of the long-term variation of the dichothermal layer shown by Wakita et al. (2010, 2013). Observations of the global sea surface height (SSH) have been performed with satellite altimeters, with resolution sufficiently fine to identify differences in the variations of circulations, as demonstrated by Nagano et al. (2013) in the Kuroshio region south of Japan and other regions. In particular, at high latitude regions, satellite tracks are closer to one another, and the horizontal resolution has been very high since the initial year of the observation by TOPEX/Poseidon (Ducet et al. 2000).
At present, altimetric observations of SSH have been conducted for more than a decade, and these data are available for investigating the interannual or longer timescale variations. In addition, hydrographic data were collected at and around the K2 and KNOT stations. Altimetric SSH variation principally consists of the baroclinic component associated with the variation of the main pycnocline depth (e.g., Kakinoki et al. 2008; Nagano et al. 2013). Other contributions such as the barotropic component and the near-surface thermal expansion are generally much smaller for year-to-year variations. By adopting the altimeter-derived gravest empirical mode (AGEM) method, temporal variations of the main pycnocline depth and the geostrophic volume transport can be estimated from altimetric SSH and hydrographic data, as Sun and Watts (2001), and others have applied in high-latitude areas where salinity is the dominant factor controlling the potential density and its variation is perceptible by altimetric SSH. We expect that the combined use of altimetric SSH and hydrographic data in the western subarctic North Pacific will make it possible to elucidate the variation of the vertical feature of the WSAG in addition to that of the horizontal feature.
In this study, we examine the variation of the WSAG on the basis of SSH, AGEM-based hydrographic structures, geostrophic volume transport and streamfunction of the gyre, and the relationship to the depth of dichothermal layer at station K2. After describing altimetric SSH and hydrographic data and the AGEM method in Section 2, we show the variation of the WSAG by the empirical orthogonal function (EOF) of SSH between 1992 and 2010 in Section 3. In Section 4, we estimate the temporal variations of the geostrophic volume transport and streamfunction of the WSAG by combining the SSH and conductivity-temperature-depth (CTD) data using the AGEM method, and the depth of the halocline bottom, which is equivalent to the top of the main pycnocline, at K2, and identify the relationship between the variation of the halocline bottom depth associated with the gyre change and that of the in situ dichothermal layer depth. A summary and conclusion of our results are provided in Section 5.
2 Data and method
We collected weekly SSH anomaly data during 1992–2010 from the Archiving, Validation and Interpretation of Satellite Oceanographic (AVISO) delayed-time updated mapped data. Because the SSH anomaly is generally less reliable near the coast, the data in regions shallower than 1000 m were eliminated. We added the SSH anomaly from the mean values during 1993–2007 to the climatological sea surface dynamic height relative to 1000 dbar based on the World Ocean Atlas 2005 (Fig. 1a) (Locarnini et al. 2006; Antonov et al. 2006) and obtained absolute SSH (HALT).
Cruise names, observation periods, and numbers of data points of the R/V Mirai observation data used in the present paper
Number of points
Oct 30 – Dec 15, 1998
May 7 – May 30, 1999
Jan 5 – Feb 6, 2000
May 9 – Jun 9, 2000
Jun 4 – Jul 18, 2001
Aug 28 – Sep 14, 2001
Oct 11 – Nov 6, 2002
Feb 20 – Mar 30, 2003
Mar 26 – Apr 16, 2004
Oct 14 – Nov 9, 2004
Feb 28 – Mar 24, 2005
May 26 – Jun 18, 2006
Feb 16 – Mar 26, 2007
Oct 11 – Nov 7, 2008
Jan 19 – Feb 24, 2010
Oct 23 – Nov 3, 2010
CTD sensors were calibrated before and after the research cruises. Accuracies for the temperature and conductivity are better than the nominal accuracies of 0.001∘C and 0.0003 S m−1, respectively. Water was sampled by Niskin bottles mounted on a frame bearing the CTD profiler and water sampler. CTD salinity values were corrected using salinity data from the bottles. The accuracy of the corrected salinity is better than 0.001 psu. The data are publicly available on the websites of hydrographic data during JAMSTEC Cruises at stations KNOT and K2 (http://www.godac.jamstec.go.jp/k2/index.html) and the JAMSTEC Environmental Biogeochemical Cycle Research Program (http://ebcrpa.jamstec.go.jp/k2s1/en). Detailed information about the cruises and the CTD data is provided on the websites.
Sea surface dynamic height based on the CTD data (HCTD) for comparison with HALT was computed in meter units. In this estimation, vertical integration was performed from the sea surface to the reference pressure level of 1000 dbar. The error in HCTD from CTD temperature (0.001∘C) and salinity (0.001) was estimated to be at most 0.1 cm, negligibly small compared to HALT.
Salinity increases with depth below the minimal- 𝜃 layer, i.e., in the halocline, where no significant seasonal variation is present, compensating for the temperature inversion to maintain the density stratification (Fig. 3b). The main pycnocline is present just below the halocline, and their boundary, i.e., the halocline bottom or the main pycnocline top, corresponds to the depth of salinity value of 33.8 (white contour in Fig. 3a), which is equivalent to that identified by Ohtani (1973). These characteristics of the AGEM fields are consistent with the climatological structures along 49∘ N in Fig. 6 prepared by Miura et al. (2002). Furthermore, the 33.8-isohaline depth is clearly dependent of HCTD. Thus, in this paper, the 33.8 isohaline is defined as depth of the pycnocline top, and the year-to-year variation will be discussed in Section 4.
For HCTD higher than approximately 1.1 m (HALT>1.26 m, as a result of a linear regression described below), no temperature inversion is present (Fig. 3a), suggesting strong influence by water originated in the subtropical region. The contours of salinity greater than approximately 33.8 and the contours of the potential density greater than approximately 26.8σ𝜃, i.e., the main pycnocline, deepen as the HCTD increases (Fig. 3b). As in Fig. 7, high HALT (>1.26 m) is almost always outside the WSAG. Thus, the AGEM analysis based on the CTD data makes the identification of the year-to-year variation of the halocline bottom depth in the western subarctic region possible.
Note that HCTD distributes narrowly in comparison with HALT, resulting in a positively discrepant slope in the linear regression, i.e., a, from unity. HALT includes the variation in the layer deeper than 1000 dbar, whereas HCTD does not. The small variation in HCTD is attributable mainly to the absence of any calculation for the variation in the deep layer. Nevertheless, the high correlation between HALT and HCTD suggests that the variation in the deep layer is in phase with the layer above it. To evaluate the variation in the WSAG in the top 1000 dbar, converting HALT to potential temperature and salinity in the top 1000 dbar via (1) can provide useful information.
3 Altimetric SSH change
The zonal extension of the low- HALT region began to shrink around the year 2000 (Fig. 7d, e). Subsequently, a local cyclonic circulation was generated east of the Kamchatka Peninsula (Fig. 7f). In the mid-to-late 2000s, the local circulation intensified and is clearly recognizable east of the peninsula from the HALT values lower than 0.9 m (Fig. 7g, h). Thus, the circulation current south of the peninsula was weakened and was associated with the northward shrinkage of the low- HALT region. The variation of HALT is considered to be generated by the change of the WSAG. Meanwhile, the outer limb of the low- HALT region (HALT>1.2 m) migrates northward only in the region off the southeastern coast of Hokkaido, as is consistent with the northward shift of the southern border of the Oyashio reported by Kuroda et al. (2015).
4 Circulation and halocline bottom depth changes
However, for the diminished WSAG, as observed in 2004 (Fig. 10b), the eastward current at the M-line possibly originates in the southward current east of the Kamchatka Peninsula. In addition, the current at the Z-line is directed northward and proceeds to the southern part of the eastward current at the M-line. By removing the volume transport from the Z-line, the volume transport of the local cyclonic circulation at the M-line reaches approximately 1.3 Sv around 2004, equivalent to that of the extended WSAG around 1996. Thus, the volume transport of the WSAG is suggested not to be reduced substantially, despite the gyre being shrunk to the northern local circulation east of the Kamchatka Peninsula. The maximal volume transport of the gyre is on the same order of magnitude of the Oyashio transport southeast of Hokkaido (1–4 Sv) with the reference level of 1000 dbar by Hata (1965), but tends to be smaller than the absolute Oyashio transport in the top 1000 dbar layer (0.5–12.8 Sv) (Uehara et al. 2004) because of the moderately baroclinic current structure of the gyre.
It should be noted that from 2007 to 2009, the dichothermal layer is unexpectedly shallow despite of the very deep halocline bottom. Presumably, sea surface heat, freshwater, and momentum forcing in the Being Sea and the northern part of the western subarctic region might not be sufficiently strong for the winter mixed layer to reach the potential maximal depth of the dichothermal layer set by the main pycnocline depth. Interestingly, the variations of the dichothermal layer depth and the halocline bottom depth seem to be discrepant after the La Niña event of 2007/2008. In La Niña, sea-level winds in the North Pacific subarctic region tend to be reduced by the attenuation of the Aleutian Low (e.g., Horel and Wallace 1981). Therefore, the La Niña-related weak winds might reduce the mechanical and convective mixing in the sea surface layer along paths of water column and suppress the winter mixed layer development at K2.
5 Summary and conclusion
Using hydrographic data collected by the R/V Mirai and altimetric SSH data, we examined the year-to-year changes of SSH, AGEM-based geostrophic volume transport, and streamfunction in the top 1000 dbar in the western subarctic North Pacific, and the halocline bottom depth (equivalently, the main pycnocline depth) at station K2. The low SSH region was found to have shrunk northward from the late 1990s to the mid-2000s, being associated with the locally and anomalously low SSH region to the east of the Kamchatka Peninsula. Associated with the SSH variation, the WSAG was greatly reduced in its extension, but not in intensity (volume transport). Meanwhile, the southern border of the basin-scale subarctic circulation surrounding the WSAG, i.e., the Oyashio, migrates northward only in the region off the southeastern coast of Hokkaido, as reported by Kuroda et al. (2015).
The K2 station was found to stay in the region of large SSH variability associated with the substantial northward shrinkage of the WSAG. The downward displacement of the halocline bottom at K2 with the northward shrinkage of the WSAG was found to be significantly related to the downward displacement of the temperature minimum layer, i.e., dichothermal layer, around K2 based on the in situ shipboard CTD and Argo data processed by the method that Wakita et al. (2010, 2013) adopted. Thus, it is concluded that the dichothermal layer depth at K2 is significantly linked to the vertical displacement of the main pycnocline associated with the variation of the gyre if the vertical mixing by the sea surface heat, freshwater, and momentum forcing are sufficiently strong.
The authors thank the members of the board, the R/V Mirai (cruise number: MR98-K01, MR99-K02, MR00-K01, MR00-K03, MR01-K03, MR01-K04 leg2, MR02-K05 leg2, MR03-K01, MR04-02, MR04-06, MR05-01, MR06-03 leg1, MR07-01, MR08-05, MR10-01, and MR10-06) for collecting the CTD data. The authors also thank the JAMSTEC Argo team for providing Advanced automatic QC Argo Data, and Ssalto/Duacs and AVISO for providing the altimetric SSH data with support from Cnes. The thanks are extended to Prof. Y. Michida (Atmosphere and Ocean Research Institute, The University of Tokyo), Dr. R. J. Greatbatch (Associate Editor of Ocean Dynamics), and anonymous reviewers for their helpful comments. This work was partly supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Grant-in-Aid for Scientific Research on Innovative Areas (22106007, 25106709, 15H02835).
- Antonov JI, Locarnini RA, Boyer TP, Mishonov AV, Garcia HE (2006) World Ocean Atlas 2005, Volume 2: Salinity, U.S. Government Printing OfficeGoogle Scholar
- Dodimead AJ, Favorite F, Hirano T (1963) Winter oceanographic conditions in the central Subarctic Pacific, . Bull Int North Pacific Comm 13:1–195Google Scholar
- Favorite F, Dodimead AJ, Nasu K (1976) Oceanography of the subarctic Pacific region,. Bull Int North Pacific Comm 33:1–187Google Scholar
- Hata K (1965) Seasonal variation of the volume transport in the Oyashio area. J Oceanogr Soc Japan 21(5):193–201Google Scholar
- Kuroda H, Wagawa T, Shimizu Y, Ito S, Kakehi S, Okunishi T, Ohno S, Kusaka A (2015) Interdecadal decrease of the Oyashio transport on the continental slope off the southeastern coast of Hokkaido. Japan J Geophys Res 120. doi:10.1002/2014JC010402
- Locarnini RA, Mishonov AV, Antonov JI, Boyer TP, Garcia HE (2006) World Ocean Atlas 2005, Volume 1: Temperature, U.S. Government Printing OfficeGoogle Scholar
- Ohtani K (1973) Oceanographic structure in the Bering Sea. Mem Fac Fish, Hokkaido Univ 21:61–106Google Scholar
- Rintoul SR, Sokolov S, Church J (2002) A 6 year record of baroclinic transport variability of the antarctic circumpolar current at 140∘E derived from expendable bathythermograph and altimeter measurements . J Geophys Res 107 (C10). doi:10.1029/2001JC000787
- Sokolov S, King BA, Rintoul SR, Rojas RL (2004) Upper ocean temperature and the baroclinic transport stream function relationship in Drake Passage. J Geophys Res 109(C5). doi:10.1029/2003JC002010
- Thomson RE, Emery WJ (2014) Data Analysis Methods in Physical Oceanography, third ed. ElsevierGoogle Scholar