Lunar nodal tide effects on variability of sea level, temperature, and salinity in the Faroe-Shetland Channel and the Barents Sea

https://doi.org/10.1016/j.dsr.2008.06.003Get rights and content

Abstract

The Faroe-Shetland Channel and the Kola Section hydrographic time-series cover a time period of more than 100 years and represent two of the longest oceanographic time-series in the world. Relationships between the temperature and salinity of Atlantic water from these two areas are examined in this paper, which also presents for the first time comparisons between them and annual mean sea levels in the region. The investigation was based on a wavelet spectrum analysis used to identify the dominant cycle periods and cycle phases in all time-series. The water-property time-series show mean variability correlated to a sub-harmonic cycle of the nodal tide of about 74 years, with an advective delay between the Faroe-Shetland Channel and the Barents Sea of about 2 years. In addition, correlations better than R=0.7 were found between dominant Atlantic water temperature cycles and the 18.6-year lunar nodal tide, and better than R=0.4 for the 18.6/2=9.3-year lunar nodal phase tide. The correlation between the lunar nodal tides and the ocean temperature variability suggests that deterministic lunar nodal tides are important regional climate indicators that should be included when future regional climate variability is considered. The present analysis suggests that Atlantic water temperature and salinity fluctuations in the Nordic Seas are influenced by forced tidal mixing modulated by harmonics of the nodal tide and influencing the water mass characteristics at some point “down stream” from the Faroe-Shetland Channel. The effects of the modulated oceanic mixing are subsequently distributed as complex coupled lunar nodal sub-harmonic spectra in the thermohaline circulation.

Introduction

Inflow of North Atlantic Water (NAW) passes from the North Atlantic through the Faroe-Shetland Channel and into the Norwegian Sea. The current continues north with a minor inflow to the Barents Sea. One part returns to the Greenland Sea, and one part has an inflow to the Arctic Ocean through the Fram Strait. From the Fram Strait, the current circulates in the Arctic Ocean and returns to the Greenland Sea return current (Fig. 1).

It has been well known for decades that this current has a major influence on the climate in northern Europe. To contribute to the study of regional climate variability, the temperature in the Faroe-Shetland Channel and the Kola section in the Barents Sea have been monitored for more than a 100 years; this represents two of the longest oceanographic time-series in the world. The relationship between these important time-series is examined in this study, which presents for the first time comparisons between the annual mean sea level, temperature and salinity of Atlantic Water in the Faroe-Shetland Channel and the Kola section of the Barents Sea.

The Faroe-Shetland Channel is the deep-water channel separating the Faroese plateau from the Scottish continental shelf. The northern entrance from the Norwegian Sea is approximately 1500–2000 m deep. Connections to the Atlantic lie to the southwest through the Faroe Bank channel, with maximum depths of 850 m, and across the Wyville-Thomson Ridge, with a sill depth of approximately 450 m. Two standard hydrographic sections across the Channel have been surveyed by the Fisheries Research Services Marine Laboratory since the start of the 20th century. These lines run approximately from Fair Isle to Munken (Faroe) and from Nolso (Faroe) to Muckle Flugga (Shetland). They were first surveyed by Dr. H.N. Dickson, contracted to the Fishery Board for Scotland, onboard the fishery protection vessel HMS Jackal. He performed the first hydrographic station of the Nolso-Flugga section on August 4, 1893. He went on to perform water bottle casts at four stations of the present day Nolso-Flugga section and at three stations of the Fair Isle-Munken section. Dickson resurveyed the same sections in 1896. Although fewer stations were performed during that survey, deeper casts were achieved, reaching >1000 m. Regular sampling at the full set of Nolso-Flugga and Fair Isle-Munken stations commenced in 1903 and has been performed approximately three times each year since that time, except for the war years and a 5-year period in the early 1980s.

A database of all available data collected along the two standard sections was constructed using data that existed in digital form from 1960 onwards and data from original manuscripts prior to that date. After the data was extracted, it was interpolated onto standard pressure levels. These procedures are more fully described in Turrell et al. (1999). In addition to data collected by the FRS Marine Laboratory, data from other sources—including Faroese, Norwegian, Swedish, Danish, English and Russian institutes—were obtained from ICES and entered into the database. Time-series derived from observations along the two sections have been employed in numerous previous publications (see Turrell et al., 1999).

In previous studies, time-series were constructed subjectively using manual methods from section plots and θS diagrams. Since the creation of the database, objective semi-automatic generation of time-series has been used in order to determine the characteristics of the different water masses observed along the sections. Two principal “types” of surface Atlantic water are always observed across the sections: a warmer and more saline type referred to as North Atlantic Water (NAW) and a fresher, cooler type referred to as Modified North Atlantic Water (MNAW). North Atlantic Water lies in the Slope Current, which flows northwards along the European shelf edge. Modified North Atlantic Water arrives at the standard sections after circulating around to the north of Faroe.

The Murman Fishery Research Expedition led by N. Knipovich occupied the Kola section for the first time in 1900 and was completed in 1906. During that period, the section was occupied one to six times per year. Subsequently, it was not occupied until 1921. In the 1920s and 1930s, the section was occupied from 1 (1924) to 10 (1937) times per year. The next gap in observations was during World War II. The section has been occupied more frequently since the middle of the 1950s, often with more than 12 surveys each year. In the 1970s and 1980s, the yearly number of occupations exceeded 15. A considerable decrease in occupation of the section occurred in the early 1990s. At that time and at present, PINRO took measurements along the section about 10 times per year. To date, the Kola section has been surveyed more than 950 times. The time-series of salinity began in 1951, owing to some uncertainties in the quality of the data collected during the first half of the 20th century. Quarterly and yearly temperature values from 1900 to 1920 and monthly temperature values from 1941 to 1981 were published by Bochkov (1982). Tereshchenko (1997) published monthly and yearly temperature and salinity values from 1951 to 1995.

The Kola section temperature time-series (Sts. 3–7, 0–200 m) can be used to assess seasonal and interannual variability of hydrographic conditions in the southern Barents Sea. There is a significant correlation (R=0.7–0.9) between this time-series and those for stations 1–3 and 8–10 of the Kola transect (Tereshchenko, 1997). There is also a significant correlation (R=0.7–0.9) between temperature time-series from the Kola section and other standard sections in the southern Barents Sea (North Cape-Bear Island section at the Barents Sea opening; Bear Island-East section (along 74°30′N) in the western Barents Sea; Cape Kanin-North section (along 43°15′E) in the eastern Barents Sea).

The Kola section time-series can serve as a reliable indicator of climate variability in the whole southern Barents Sea (coastal and Atlantic water domain). In addition, there is a significant correlation (R=−0.56 monthly data; R=−0.65 yearly data) between the time-series of temperature in the Kola section and ice coverage (percent of the total Barents Sea area). However, this is not true for the salinity time-series from stations 3 to 7, as correlation between temperature and salinity time-series from stations 3 to 7 is close to 0. This is probably due to the fact that there is a well-pronounced haline frontal zone (between 71°N and 72°N, Sts. 4–6) separating Atlantic water (warm temperature, high salinity) to the north and slightly warmer, considerably fresher coastal water to the south. Therefore, variability of salinity averaged for stations 3–7 may greatly depend on variations of the haline front position, but this is yet to be studied.

Stations 3–7 are usually associated with the Murman Current, which is the continuation of the North Cape Current entering the Barents Sea through its western boundary. The Murman Current is related to the haline/density frontal zone. According to calculations by a numerical model (Trofimov, 2000), the core of the current in the section is located at about 72°00′N.

The Kola section temperature and salinity time-series are often used in fisheries research and regional climate variability studies. Some examples can be found in the papers by Izhevskii, 1961, Izhevskii, 1964, Loeng (1989), Loeng et al. (1995), Ottersen et al., 1994, Ottersen et al., 2000, Yndestad, 1999, Yndestad, 2003, Yndestad, 2006.

The North Atlantic Water temperature and the Kola section water temperature represent long-term regional climate indicators. Better understanding of the causes of fluctuations of these indicators may lead to a better understanding of climate variability and forecast climate change. In this analysis, the variability in the climate indicators was characterised by the time-series spectrum properties. This spectrum may have information about the source of the deterministic properties that cause water mass characteristic fluctuations, including temperature. If these time-series have temporary deterministic properties, there is a possibility of forecasting dynamic change in the North Atlantic Water temperature, the Barents Sea water temperature and regional climate.

The present study compares the time-series of North Atlantic Water properties from the Faroe-Shetland Channel and the Kola section time-series. The cross-correlation between time-series of NAW temperature and Kola section temperature was estimated to be R=0.05. This small correlation showed that the time-series fluctuations had different amplitude and phase.

In this study, the time-series was analyzed using a wavelet transform. This method made it possible to identify the amplitude and phase of dominant periodic cycles in each of the time-series and between time-series. The wavelet analysis showed that single dominant fluctuations were correlated to the lunar nodal tidal cycles of 18.6 and 9.3 years. The lunar nodal tide cycles have an astronomical property that made it possible to introduce a deterministic periodic reference in the cycle analysis, which added new information about the time-series properties.

Section snippets

Faroe-Shetland Channel time-series

It has been shown that the best definition of the properties of North Atlantic Water, found within the core of the Slope Current on the Scottish side of the Faroe-Shetland Channel, is the temperature and salinity at the standard depth, which exhibits the maximum salinity within an individual survey of the first two stations in both standard sections on the Scottish side of the Channel. As the standard section surveys in the past were performed at quite different times of year in order to remove

Aberdeen annual mean sea level

Fig. 2 shows the annual mean sea level x(t) at Aberdeen, the dominant 18-year wavelet cycle W18(t) and the astronomic 18.6-year lunar nodal amplitude tide uT(t). The long-term fluctuation of the annual sea level was analyzed from the wavelet spectrum W1:N/2(t) of the Aberdeen sea level time-series (Eq. (3)). This transform represented a moving linear filter that separated periodic cycles in the time-series. The fluctuation had maxima at about t={1950, 1990} and minima at about t={1940, 1975}.

Discussion

The analysis was based on high-quality time-series, and wavelet analysis was used to identify single fluctuation periods and phase in time variant stochastic time-series. The wavelet method has limitations at the beginning and end of a time-series. A potential source of phase error exists from about 1900 to 1925, when the NAW time-series and the Kola time-series had some missing elements. Between these years, time-series gaps were filled using either cubic interpolation for the Faroe-Shetland

Conclusions

Wavelet analysis of some of the longest oceanographic time-series in the world indicated that the 18.6-year lunar nodal amplitude tide and the 9.3-year phase tide may influence the variability of temperature, salinity and sea level in the Faroe-Shetland Channel and the Barents Sea. The time-series had a mean variability correlated to the sub-harmonic cycle of about 74 years and an advective delay to the Barents Sea of about 2 years. The correlation between the lunar nodal tides and the ocean

References (42)

  • I. Daubechies

    Ten lectures of wavelet

    SIAM Journal on Mathematical Analysis

    (1992)
  • Daly, F., Hand, D.J., Jones, M.C., Lunn, A.D., MaConway, 1995. Elements of Statistics. Addison-Wesley Publishing...
  • G.D. Egbert et al.

    Significant dissipation of tidal energy in the deep ocean from ocean inferred from satellite alitimeter data

    Nature

    (2000)
  • Ekman, M., 2003. The World's Longest Sea Level Series and a Winter Oscillation Index for Norteren Europe 1774–2000....
  • Helland-Hansen, B., Nansen, F., 1909. The Norwegian Sea: its physical oceanography based on Norwegian researches...
  • G.K. Izhevskii

    Oceanological Principles as Related to the Fishery Productivity of the Seas

    (1961)
  • Izhevskii, G.K., 1964. Forecasting of Oceanological Conditions and the Reproduction of Commercial Fish. Moscow:...
  • C.D. Keeling et al.

    Possible forcing global temperature by oceanic tides

    Proceedings of the National Academy of Sciences of the United States of America

    (1997)
  • J.W. Loder et al.

    The 18.6-year cycle of the sea surface temperature in the shallow seas due to tidal mixing

    Journal of Geophysical Research

    (1978)
  • H. Loeng

    The influence of temperature on some fish population parameters in the Barents Sea

    Journal of Northwest Atlantic Fishery Science

    (1989)
  • H. Loeng et al.

    Larval fish growth in the Barents Sea

  • Cited by (34)

    • Influence of the 18.6-year lunar nodal cycle on the tidal resource of the Alderney Race, France

      2020, Applied Ocean Research
      Citation Excerpt :

      This change of inclination modulates the tides, which has implication on coastal processes. As examples, Levoy et al. [14] showed that the 18.6-year lunar nodal cycle influences the morphodynamics of the Mont-Saint-Michel Bay and Yndestad et al. [27] investigated the effect of the lunar nodal tide on variability of sea level, temperature and salinity in the Faroe-Shetland Channel and Barents Sea. However, the extent to which the 18.6-year lunar nodal cycle modifies the tidal resource has yet received very little (or no) attention in the scientific literature.

    • Lunar fingerprints in the modulated incoming solar radiation: In situ insolation and latitudinal insolation gradients as two important interpretative metrics for paleoclimatic data records and theoretical climate modeling

      2018, New Astronomy
      Citation Excerpt :

      We further restrict ourselves to the lunar nodal forcing scale of 18.6 years for the deterministic modulation of the incoming solar radiation while leaving other more complex modulation timescales that were recognized and discussed for example in Cionco and Soon (2015) and Yndestad and Solheim (2017) for future explorations. A second purpose of this paper is to propose and explore a new interpretation of the lunar signals and fingerprints often found wanting in various instrumental records covering ocean temperature, rainfall, to sea level (Loder and Garrett 1978; Yasuda et al., 2006; McKinnell et al. 2007; Gratiot et al., 2008; Yndestad et al., 2008; Agosta 2013; Malherbe et al., 2014; Osafune et al., 2014; Spada et al., 2014; Hansen et al., 2015 and references therein) as well as in paleoclimatic records (Cook et al., 1997; Black et al., 2009; Yasuda 2009; Davis and Brewer 2011 and references therein). Both Munks and Bills (2007) and Ray (2007) explored a mechanistic understanding and interpretation of the lunar signals, particularly the lunar nodal modulation of 18.6 years and concluded that the typically assumed connection arising from enhanced tidal vertical mixing or other tidal impacts may be not only inadequate but likely improbable.

    • Periodicities in mid- to late-Holocene peatland hydrology identified from Swedish and Lithuanian tree-ring data

      2016, Quaternary Science Reviews
      Citation Excerpt :

      Although convincing evidence for the influence of a 18.6-year lunar nodal tide on climate has not been presented so far (Ray, 2006), periodicities of about 18.6 years have been detected in rainfall variations in India (Mitra et al., 1991), dryness-wetness indices from China (Currie, 1995), drought records in North America (Woodhouse et al., 1998), SSTs along the North American west coast (McKinnell and Crawford, 2007), as well as tree-ring data along the coast of the Barents Sea (Raspopov et al., 2004) and Scandinavia (Linderholm, 2001). Moreover, Yndestad et al. (2008) found correlations (r = 0.7) between dominant Atlantic water temperature periodicities and the 18.6-year lunar nodal tide. This correlation, in combination with previously discussed observations of 18–19-year periodicities suggests that deterministic lunar nodal tides may indeed influence regional climate variability.

    • Abrupt changes, multidecadal variability and long-term trends in sea surface temperature and sea level datasets within the southeastern Bay of Biscay

      2013, Journal of Marine Systems
      Citation Excerpt :

      The spectrum shows a prominent peak around 2.6 years, suggesting that the QBO appears to be a real characteristic of the SLP and SL data in the Bay of Biscay. The remainder of the significant peaks in the Santander SL spectrum are attributed to the 9.3-year oscillation (associated with the lunar nodal phase tide) and the 18.61-year lunar nodal tide (Araújo, 2005; Yndestad et al., 2008). The series for St. Jean de Luz has a length of 34 years; hence the nodal constituent with a period of 18.6 years cannot be evaluated accurately.

    • The long sunspot cycle 23 predicts a significant temperature decrease in cycle 24

      2012, Journal of Atmospheric and Solar-Terrestrial Physics
      Citation Excerpt :

      The temperatures in the North Atlantic and on adjacent land areas are controlled by the Atlantic Multidecadal Oscillation (AMO), which has in the instrumental period from 1856 exhibited a 65–80 year cycle (0.4 °C range), with warm phases at roughly 1860–1880 and 1930–1960 and cool phases during 1905–1925 and 1970–1990 (Kerr, 2000; Gray et al., 2004). This period is proposed related to a 74 year period, which is a sub harmonic of the 18.6 years lunar-nodal-tide (Yndestad et al., 2008). The AMO may affect the temperatures at the locations we have analyzed, but a preliminary investigation by the authors have found considerable variations in dominant periods from place to place, which we interpret as due to local variations related to geography, dominant weather patterns, and distance from the ocean currents.

    View all citing articles on Scopus
    1

    Tel.: +44 1224 876544; fax: +44 1224 295511.

    2

    Tel.: +7 8152 47 32 80; fax: +7 8152 47 33 31.

    View full text