Our study presents new North Atlantic paleoceanographic results over the last deglaciation, which was a time of abrupt climatic changes (general warming superimposed by sudden short glacial reversals). Atlantic thermohaline circulation or North Atlantic meridional overturning circulation (NAMOC) mediates the climate for the surrounding land and climatic connections between hemispheres [1
]. The sea surface conditions in the North Atlantic significantly influence the major modes of regional atmospheric circulation (meridional or zonal) [2
] causing warming or cooling episodes. Rapid changes in the NAMOC and on the North Atlantic surface during the last deglaciation have influenced changes of the climatic conditions (global warming, melt-water discharge from land ice, sea-ice distribution) [3
]. Overpeck et al. [4
] postulated an apparent synchroneity of the rapid climate events in the circum-North Atlantic during the last deglaciation, but called for new high-resolution studies to prove this.
We use the micropaleontological data from the sediment core AMK-340 (the Russian RV “Akademik Mstislav Keldysh” station 340), combined with the radiocarbon dated oxygen and carbon isotopic record, to estimate the sea subsurface summer temperature and to describe the paleoenvironmental changes during Termination I on the Reykjanes Ridge circa (ca.) 60° N. Late Quaternary paleoceanography of the subpolar to polar North Atlantic has been extensively documented, particularly using microfossils to provide sea surface estimates [5
], and references therein). Our aim is to get additional quantitative information on sea subsurface temperature and compare it with the global and regional paleoclimatic archives. The approach is a reconstruction of the paleotemperature based on the factor analysis of the radiolarian and planktic foraminiferal data in the same samples. The radiolarian distribution and paleotemperature estimates based on the radiolarians (old “graphical” paleotemperature method, and method of spline interpolations as a modification of the Q-mode analysis) for the core AMK-340 was presented by Matul [6
], and Matul and Yushina [7
]. This paper uses an improved modern and core radiolarian database, newly constructed modern planktic foraminiferal database, a standard paleotemperature method of the Q-mode analysis and transfer functions realized as a software PanTool Box [8
] and also involves unpublished data on the planktic and benthic foraminifera in the core.
The study area is located on the eastern margin of the Subpolar (Irminger) Gyre (Figure 1
). The western branch of the Irminger Current split from the warm North Atlantic Current influences the Reykjanes Ridge oceanography [9
]. Conversion of surface and warm to deep and cold waters by the Irminger Current, East Greenland Current, and also the Labrador Sea Water within the Irminger Gyre is unstable, exhibiting large interannual variability of the water flows at both the sea surface and deeper levels, thus complicating the local features of the NAMOC [10
]. The average summer sea surface temperature at the location of the core AMK-340 is ca. 12 °C [11
]. At the end of the last glacial period, before ca. 14 thousand years ago (ka), sea surface temperatures south of Iceland dropped to 0–2 °C [12
]; therefore, prominent temperature and environmental changes within the transition from the last glacial to the Holocene interglacial on the Reykjanes Ridge occurred.
2. Material and Methods
The gravity core AMK-340 was obtained from the central area of the Reykjanes Ridge, North Atlantic (58°30.6′ N, 31°31.2′ W, water depth of 1689 m; core length of 387 cm) during the 4th cruise of the Russian RV “Akademik Mstislav Keldysh” in 1982 (Figure 1
). The lithology of the core is composed of (1) 0 to 241 cm pelitic calcareous muds with CaCO3
content of 40–50%, (2) 241 to 307 cm pelitic weakly calcareous muds with CaCO3
content of 10–25%, (3) 307 to 387 cm pelitic muds with CaCO3
content of ≈10% alternated by the thin layers of the pelitic weakly siliceous muds. The lithology figure can be found in the publication of Matul [6
]. In the paleoceanographic interpretation (see Results and Discussion section) we use the published paleotemperature, isotopic and micropaleontological data on the cores RAPiD-10-1P (62.9755° N, 17.5895° W, 1237 m water depth), RAPiD-12-1K (62.09° N, 17.82° W, 1938 m water depth), and RAPiD-15-4P (62.293° N, 17.134° W, 2133 m water depth) (Figure 1
The chronology of the core has been established from four AMS (Accelerator Mass Spectrometry) 14
C-dated samples of the planktic foraminifera Neogloboquadrina (N.) pachyderma
left coiled or sinistral (sin.) shells in the Leibniz Laboratory for Radiometric Dating and Stable Isotope Research at the Christian-Albrechts-University of Kiel, Germany (Table 1
Radiocarbon ages were converted to calendar ones via the calibration program CALIB 7.1 using the MARINE13 scale (standard reservoir age correction R is 405 years with the regional ∆R of 85 ± 79 years) [16
]. The age–depth plot is presented in Figure 2
. We have no direct age dating of the core top, and tentatively assume the core top to be modern but nonetheless are aware of its older age. If we were to take into account the old radiocarbon data of the core sediments based on the total organic carbon [6
], the calendar age of the AMK-340 sample 10–15 cm could be 1730 years, thus, the core top should be younger than 1 ka. Core AMK-340 spans the last ca. 14.5 kyr, which is the Holocene and the end of the last glacial period, including most of the deglacial warming. In the Results and Discussion section we concentrate on the paleoceanography of the Termination I as there is no appropriate age control of the upper core part. To identify the above-mentioned paleoclimatic intervals, we have used the standard ages of the Greenland stadials and interstadials recognized in the core, using the INTIMATE event stratigraphy [17
]: start of the Bølling-Allerød (BA) warm interval is at ca. 14.7 ka, start of the Younger Dryas (YD) cold interval is at ca. 12.9 ka, and start of the Holocene warming is at ca. 11.7 ka. The average sedimentation rate of the core is ca. 26 cm·ky-1
. The planktic foraminiferal species N. pachyderma
(sin.) and Globigerina (G.) bulloides
tests were picked out for the oxygen and carbon isotopic analysis in the Marine Stable Isotope Lab (MASTIL) of the National Centre of Antarctic and Ocean Research, Vasco-da-Gama, Goa, India. The size range of the planktic foraminiferal tests chosen is >100 μm. The external precisions of δ18
O and δ13
C analysis are ±0.15‰ and ±0.09‰, respectively (1σ standard deviation) obtained by repeatedly running NBS-19 (the reference isotopic standard material according the National Bureau of Standards of USA) as the Standard (n
= 33). The δ18
O and δ13
C values are reported with respect to V-PDB (Vienna Pee Dee Belemnite).
We studied polycystine radiolarians (siliceous) and benthic and planktic foraminifera (calcareous) microorganisms in 37 sediment samples of 2-cm thickness each throughout the core, with a time resolution as high as 133 years.
For the radiolarian analysis, air-dried sediment samples of 1–2 g were boiled in a solution of 30% hydrogen peroxide and sodium pyrophosphate, and the carbonates were removed by adding a solution of 10% hydrochloric acid. The residue was washed through a 50 μm sieve. Then we subsampled a known aliquot with a pipette in order to later obtain the quantitative estimates and settled it on the cover glass and mounted on the slide in Canada balsam. As a rule, at least 250–300 radiolarians tests were counted under the transmitted light microscope at the ×300–600 magnification. Percentages of the radiolarian species and total radiolarian abundances as test·g−1 of bulk dry sediment were calculated.
For the foraminiferal analysis, air-dried sediment samples of available weight were washed through a 100 μm sieve. The dried fraction of >100 μm was thoroughly mixed and divided until the remaining split contained at least 300 tests. We identified and counted all components (mineral grains, planktic and benthic foraminifera) in this split, and, additionally, examined the whole fraction of >100 μm to find rare species that might be omitted from the split. The “foraminiferal number” (the number of shells larger than 100 μm per 1 g of dry sediment) and percentages of all planktic and benthic foraminiferal species were then calculated.
We found 25 benthic foraminiferal species in the core. The detailed study of the benthic microfauna from the core AMK-340 will be published later [18
]. In the current paper, we present data on two benthic species that can be good indicators of the organic matter fluxes to the bottom: Cassidulina (C.) teretis
and Globocassidulina (G.) subglobosa
. C. teretis
feeds on the bacteria from the soft, enriched by the organic matter sediments [19
], G. subglobosa
marks high spring phytodetrital fluxes [20
]; therefore, we can use data on their quantity changes as possible indirect indication of the bioproductivity variations.
The paleotemperature was reconstructed by a method of Q-mode factor analysis and transfer functions. The PaleoTool Box software with its latest PC-version [8
] provides the platform to make a statistical analysis of the micropaleontological data.
Transfer functions which allow paleotemperature estimates are based on the treatment of the modern micropaleontological datasets. The North Atlantic reference datasets used here consist of (1) 36 polycystine radiolarian species in 91 surface sediment samples from the area between 40 and 73° N [21
], and (2) 23 planktic foraminiferal species in 237 surface sediment samples from the area between 40 and 80° N as a compilation from the Atlantic Ocean database of the Shirshov Institute of Oceanology, Moscow, Russia (134 stations from [22
]) and World Ocean database (103 stations from [23
]. Stations are presented in Figure 3
. In the down-core record, we found no microfossil assemblages that are similar to the modern tropical/subtropical ones in the North Atlantic south of ca. 40° N. Therefore, we used stations from the areas north of this latitude. A temperature at the subsurface depth of 100 m is chosen as a basic parameter to be reconstructed for the interpretation because this depth can be assumed to represent the median level of the prevailed habitat of radiolarians [24
] and planktic foraminifera [25
]. In the North Atlantic, the highest abundances of the marine microzooplankton were found during May–September [26
]; therefore, we analyze the summer temperature (July-August-September as the standard season in The World Ocean Atlas 2013 [11
]). Most samples in our modern datasets on the radiolarians and planktic foraminiferas in the surface sediments were obtained from the end of 1950s to 1960–1970s, and the closest time interval with the averaged temperatures in the World Ocean Atlas 2013 [11
] which we use is 1955–1964.
A comprehensive description of the PanTool Box application can be found in publications of Zielinski [28
] and Zielinski et al. [29
]. In the construction of the modern micropaleontological datasets, we follow recommendations of Imbrie and Kipp [30
] that every microfossil species must have abundances >1–2% in one sample and occurs at least in 10 samples. Every radiolarian and planktic foraminiferal species from the modern datasets were found in the core. Geographical range of samples and species distribution from the modern datasets covers environments which occurred in the area of study during the Late Quaternary so that we do not expect the non-analogue situation. As Zielinski et al. [29
] argued, the method of the Q-mode factor analysis works better than the modern analogues method with the regional database (in particular, in the cold-water areas), which is our case.