Oxygen isotope values of water samples and foraminifera
The calculated equilibrium calcite isotope composition represents the
potential δ18O value of inorganic calcite precipitated in isotopic
equilibrium with the surrounding seawater. The offset found between the
equilibrium calcite value and the measured δ18O value of
foraminiferal tests is commonly described as a “vital effect” related to
differential isotopic uptake in carbonate organisms compared to equilibrium
conditions. The δ18O values of living foraminifera in our work area
during early summer were consistently lower than the calculated equilibrium
calcite values. Part of this offset may certainly result from mesoscale
oceanic variability, i.e. short-term changes of salinity and/or temperature
due to, e.g. sea-ice formation/melting or local vertical convection. The mean
offsets were -1.5 ± 1.3 ‰ in Neogloboquadrina pachyderma and -3.7 ± 1.7 ‰ in Turborotalita quinqueloba. Based on previously published results, the magnitude of the
vital effect in N. pachyderma appears to vary regionally. Bauch et
al. (1997) reported a consistent offset of -1.0 ‰ between
equilibrium calcite values and δ18O data of net-sampled N. pachyderma in the Nansen Basin. Volkmann and Mensch (2001) found an average
vital effect of -1.3 ‰ in the Laptev Sea for N. pachyderma and -1.6 and -1.3 ‰ in the Fram Strait for
N. pachyderma and for T. quinqueloba, respectively.
Plankton tows from various hydrographic regimes in the Nordic Seas revealed
vital offsets of -1.0 and -1.1 ‰ for N. pachyderma and
T. quinqueloba, respectively (Simstich et al., 2003). Significantly
smaller offsets were reported from the western subpolar North Atlantic,
calculated from shells collected with sediment traps (Jonkers et al., 2013).
Even studies conducted in the Fram Strait reveal slightly different values
(see Stangeew, 2001; Volkmann and Mensch, 2001). Figure 10 comprises results
on δ18ON.p. and equilibrium calcite values in the
upper water column reported from the Fram Strait.
(a) δ18O and (b) δ13C values of T. quinqueloba from the water column (red squares) and from the sediment
surface (green dashed line). The blue dots indicate (a) the equilibrium
calcite and (b) the δ13CDIC profile of the water column.
δ18O of N. pachyderma from the water column
(weighted average; black dots) with the range of equilibrium calcite values
in the upper 200 m (blue line: minimum, red line: maximum) along three
parallel E–W transects between 78∘50′ and 81∘50′ N in the
Fram Strait: (a) this study, (b) Stangeew, 2001; Bauch (unpublished data), (c)
Volkmann and Mensch, 2001.
In our study T. quinqueloba shows larger offsets between the
equilibrium calcite values and the measured δ18O values than
N. pachyderma (on average -3.7 and -1.5 ‰,
respectively). Earlier works (e.g. Fairbanks et al., 1980; Lončariċ
et al., 2006) also recorded a larger negative offset in spinose species
compared to nonspinose species. Moreover, symbiont-bearing species, like
T. quinqueloba, are known to be more depleted in 18O as a
consequence of higher CO2 fixation caused by photosynthesis (Bijma et
al., 1990; Spero et al., 1997). In N. pachyderma we found a clear
east–west difference in the magnitude of the vital effect along the transect,
similar to observations by Volkmann and Mensch (2001). In their study the
eastern and western part of the strait yielded significantly different
offsets, with highest deviations from the equilibrium calcite values in the
west. They concluded that ice coverage increases the magnitude of the vital
effect. In our samples in N. pachyderma the strongest disequilibrium
was indeed found at the two ice-covered stations (-4.0 and
-1.8 ‰, at 4 and 5∘ W, respectively) and at station 87
(-3.9 ‰, at 4∘ E). These results are also in line with
observations of Bauch et al. (1997) who found slightly increasing isotopic
differences between water and plankton samples with decreasing salinity and
temperature. Similar to these results, Volkmann and Mensch (2001) observed
greater vital offset in the cold and less saline waters of the western Fram
Strait. They concluded that unfavourable conditions here make the individuals
grow faster (i.e. increase their calcification rate). An increased
calcification rate decreases the δ18O of tests (McConnaughey, 1989)
and may thus increase the disequilibrium. While this hypothesis can explain
high offsets at increased calcification rates, the validity of the hypothesis
seems rather unlikely as unfavourable conditions generally lead to lower
metabolism and thus, decreased calcification rates. Moreover, lower
temperatures decrease metabolic rates in all organisms (Hemmingsen, 1960;
Gillooly et al., 2001). The abrupt increase in the offset close to the
sea-ice margin may rather be explained by increased primary production,
associated with the ice margin where ice melting increases stratification and
consequently the stability of the water column, which triggers phytoplankton
blooms (see Alexander, 1980; Carstens et al., 1997; Pados and Spielhagen,
2014). During biological production dissolved inorganic carbon is consumed.
This considerably increases pH and consequently the carbonate ion
concentration ([CO32-]) of the water (Chierchi and Franson, 2009).
Spero et al. (1997) showed that increasing seawater [CO32-] decreases
the 18O / 16O ratios of the shells of foraminifera and may thus
simultaneously increase the vital effect. Still, the effect of carbonate ion
concentrations alone cannot explain the high deviation from equilibrium
calcite found at the station at 4∘ E where no increased primary
production is expected. A possible reason for the increased vital effect at
the stations at 5∘ W, 4∘ W and 4∘ E might also be
a sampling during different ontogenetic stages. N. pachyderma is
known to reproduce on a synodic lunar cycle (Bijma et al., 1990; Schiebel and
Hemleben, 2005) and as these three stations were sampled in sequence in the
second half of the cruise, it is possible that in the respective samples
there were more specimens in early life stages compared to the stations
sampled 7–10 days before. Early ontogenetic stages are associated with higher
respiration and calcification rates (Hemleben et al., 1989). Rapidly growing
skeletons tend to show depletion in both 13C and 18O (McConnaughey,
1989), which could account for the increased vital effect observed at the
respective stations.
In contrast to N. pachyderma, the offsets found between equilibrium
calcite values and the δ18O values of T. quinqueloba do not
follow a clear trend along the transect and show great scatter (Fig. 9).
However, the low numbers of specimens found in the samples at most of the
stations did not allow us to determine δ18O over the whole water
column sampled. Moreover, as a consequence, lower numbers of tests (on
average ten) were used for stable isotope analysis than in N. pachyderma (25), which might also account for the scatter in both
δ13C and δ18O values in T. quinqueloba. We
therefore refrain from discussing the vital effect in T. quinqueloba.
Our analysis shows that recent specimens of planktic foraminifera from the
water column have a lower oxygen isotopic value than fossils on the sediment
surface (Figs. 5, 9). This is in agreement with a number of studies conducted
in different regions of the world (e.g. Duplessy et al., 1981; Schmidt and
Mulitza, 2002). Berger (1970) suggested in his hypothesis on intraspecific
selective dissolution that within one species preferentially the thin-shelled
individuals are dissolved during deposition. These tests are secreted during
the warmest period of the year and thus, their dissolution increases the
average δ18O value of the species in the coretop samples. Even
though the length of growing season of planktic foraminifera in the Fram
Strait is unknown, it has been shown that in the Nordic Seas the production
maximum of planktic foraminifera occurs during summer (Kohfeld et al., 1996;
Jonkers et al., 2010), with almost zero production during other seasons. This
means that the majority of the specimens calcifies the shells under similar
conditions. Accordingly, differences in the thickness of tests are not to be
expected. Therefore the hypothesis of Berger (1970) cannot explain the
isotopic differences between plankton and sediment surface samples in our
study area. Lateral transport of the shells during deposition is another
effect that could explain the discrepancies. However, mean transport
distances in the Fram Strait are only 25–50 km for N. pachyderma
and 50–100 km for T. quinqueloba (von Gyldenfeldt et al., 2000).
Even if we consider that specimens may also be carried a similar range during
their lifespan, these distances appear too short to transport isotopic
signatures from water masses with significantly different
temperature/salinity signatures into the sediments. The offset found in the
δ18O values between plankton and sediment surface samples can be
rather attributed to the age difference between living plankton and sediment
surface samples. Core top samples are assumed to represent modern conditions
in palaeoceanographic reconstructions. Nevertheless, depending on
sedimentation rates and bioturbation intensity, their average age can vary in
a great range (in the Fram Strait a few decades to 3 ky, on average 1 ky;
see Simstich et al., 2003) while net-sampled foraminifera reflect a snapshot
of actual modern conditions. Discrepancies found between isotopic composition
of shells collected on the sediment surface and in the water column may
therefore be related to changes in the oceanographic parameters between the
early summer of 2011 and average conditions during the period represented by
sediment surface samples. To explain the lower modern δ18O values,
the water mass in the calcification depth interval of the foraminifera must
have become warmer and/or the δ18Owater must have
decreased and thus, the salinity signature must have changed significantly.
It has been shown indeed that due to increasing river discharges in the last
8 decades (e.g. Peterson et al., 2002) the freshwater budget of the Arctic
Ocean has significantly changed (Morison et al., 2012), which resulted in
increased freshwater export through the Fram Strait. Moreover, rising
temperatures have been documented for the last decades in the Arctic as well
(e.g. Zhang et al., 1998; Serreze et al., 2000; Spielhagen et al., 2011). The
mean offset found between the δ18O values of net-sampled
foraminifera and the tests from the sediment surface along the transect is
∼ 1.3 ‰. Assuming that the oxygen isotope composition of the
water remained constant over the time, this difference would correspond to a
change in water temperature of about 5 ∘C. Neglecting the two
extremely high offsets found at 4∘ W and 3∘ E, the mean
offset would decrease to ∼ 0.6 ‰, corresponding to a
temperature change of ∼ 2.4 ∘C. A temperature change of
2.4 ∘C is similar to the reconstructed temperature increase of
Atlantic Water during the last 200 years (Spielhagen et al., 2011). However,
a temperature change of 5 ∘C during the last millenium over the
whole Fram Strait area seems much too large and, clearly, water temperature
changes may not solely account for the differences found in the isotopic
composition between modern and fossil foraminifera. The results nevertheless
suggest the combined effect of temperature rise and
δ18Owater-change, possible dissolution and transport
effects during the last ∼ 1000 years.
Calcification depth
With currently available methods we cannot directly determine the actual
calcification depth of planktic foraminifera in the water column. Therefore
we assume that planktic foraminifera build their shells at the depth where
they are most abundant. The average depth of calcification (calculated from
the standing stock) of N. pachyderma in the Fram Strait lies between
70–150 m water depth. T. quinqueloba shows a similar calcification
range at 50–120 m water depth (Figs. 6 and 7). Both species show deepest
average calcification depth at the easternmost station. Our results are in
accordance with Simstich et al. (2003) who calculated an apparent
calcification depth for N. pachyderma of 70–130 m and 70–250 m
in the EGC and off Norway, respectively. From the Nansen Basin (eastern
Arctic Ocean), Bauch et al. (1997) reported a deeper average calcification
depth for N. pachyderma. However, in the northern regime of the
Nansen Basin, where the water column properties are similar to those in the
western Fram Strait, N. pachyderma prefers shallower waters than in
the southern Nansen Basin where the water column is strongly influenced by
the subsurface inflow of Atlantic Waters (Bauch et al., 1997). This trend
observed by Bauch et al. (1997) coincides with our results. The difference
found in calcification depths in the Nansen Basin and in the Fram Strait
might be caused by the different habitats that these locations represent. The
northern Nansen Basin is covered by sea ice throughout the year and thus
represents a different habitat for planktic foraminifera than the narrow Fram
Strait. Here, the interannual E–W variability in the position of the average
summer sea-ice margin is high and the ice-covered stations sampled in this
study might therefore be ice-free in another summer. It has been shown that
the depth habitat of planktic foraminifera in the Fram Strait in the early
summer is predominantly controlled by the position of the deep chlorophyll
maximum (Pados and Spielhagen, 2014). The permanent ice cover in the Nansen
Basin may alter the factors controlling the depth habitat of foraminifera and
may consequently cause a different depth habitat (and calcification depth)
than in the Fram Strait.
Calculating the vital effect from differences between water and plankton
samples at each depth level assumes that foraminifera calcified their tests
at the depth interval where they were caught. This might not be true, as
foraminifera are known to migrate in the water column during their life
cycle. Alternatively we may assume that the main encrustation process of
foraminifera indeed happened solely at the average calcification depth that
is derived in our study from the standing stock. When calculating the average
offset between water and foraminifera for the calcification depth only, a
vital effect of -0.9 ± 0.5 ‰ in N. pachyderma and
-3.1 ± 2.9 ‰ in T. quinqueloba is determined. These
vital effects are significantly smaller than those determined over the whole
water column, which are -1.5 ± 1.3 ‰ and
-3.7 ± 1.7 ‰ for N. pachyderma and T. quinqueloba, respectively. In general, we have to take into account that
both calcification-scenarios represent extreme cases and the actual vital
effect may be between these two extremes.
Carbon isotope values of DIC and foraminifera
The interpretation of the carbon isotope composition of foraminiferal shells
is quite complicated as several factors can influence the carbon isotope
incorporation. The gas exchange between sea and atmosphere, the biological
production, the community respiration and species-dependent incorporations of
carbon isotopes are the main processes that can affect the
13C / 12C ratio in calcite tests. A number of studies reported
on a consistent offset between δ13C of calcite shells and the
δ13CDIC measured within the water column (e.g. Bauch et
al., 2000; Volkmann and Mensch, 2001). According to Romanek et al. (1992) the
δ13C of inorganic calcite that precipitates in equilibrium with
seawater is 1 ‰ higher than δ13CDIC. In our
study area below 75 m water depth the δ13C values of N. pachyderma run relatively parallel to the δ13CDIC, but
with an average offset of -1.6 ± 0.7 ‰. This reveals a vital
effect of about -2.6 ‰. Kohfeld et al. (1996) reported from the
Northeast Water Polynya on the Greenland shelf a vital effect of
-1 ‰ while another study in the Nansen Basin (Bauch et al., 2000)
revealed a vital effect of -2 ‰. The discrepancies found here may
suggest the influence of oceanographic variability on the vital effect in
δ13C of N. pachyderma. The δ13C of T. quinqueloba shows a stronger vertical scatter with an average vital effect
of -4.6 ± 1.5 ‰. Again we emphasize that in the case of
T. quinqueloba the low amounts of calcite analysed might have also
influenced the results. Nevertheless, in the upper 75–100 m of the water
column for both species the δ13CDIC and the
δ13C of shells show an exactly reverse tendency (Figs. 5, 9): the
δ13CDIC is increasing towards sea surface while the
δ13C of the tests is decreasing. The high δ13CDIC values found close to the sea surface are assumed to be
caused by high primary production, resulting in enrichment in 13C (Fogel
and Cifuentes, 1993): as 12C is taken for photosynthesis, the water
becomes enriched in 13C. However, if no other processes would affect the
incorporation of carbon into the calcite shells, the tests should also show
the enrichment in 13C. One possible explanation for the deviation in the
upper ∼ 75 m could be the effect of high (near-surface) temperatures
on the carbon isotope incorporation of the tests. Laboratory (Bemis et al.,
2000) and field experiments (Jonkers et al., 2013) have shown that
foraminiferal δ13C linearly decreases with increasing temperatures.
However, in our data set the offsets measured between
δ13CDIC and δ13Cforaminifera have no
correlation with in situ water temperatures. Therefore this hypothesis cannot
explain in our case the greater vital effect found in near-surface waters.
Another explanation for the deviation might be an increased carbonate ion
concentration ([CO32-]) as a consequence of strong biological
production in the upper water column (Chierchi and Franson, 2009). Both
culturing (Spero et al., 1997) and field experiments (Bauch et al., 2002)
have shown that the carbon isotope composition of foraminifera is correlated
to the carbonate ion concentration of the water. The “carbonate ion effect”
(CIE) describes that increasing seawater [CO32-] causes depletion in
13C of the foraminiferal tests. The CIE could therefore explain our
observed low δ13C values of shells living in 13C-enriched
waters. A direct interpretation of this effect is not possible as during
cruise ARK-XXVI/1 the concentration of [CO32-] or the parameters
needed to calculate [CO32-] (e.g. pH and total alkalinity of the
water samples) were not determined. However, vertical profiles of
[CO32-] measured in the area (CARINA database, 2015) show in the
upper 500 m of the water column a quite uniform [CO32-]
distribution, with values of 100–120 µmol kg-1. Only at the
surface in the WSC (upper 50 m) values are higher (up to
160 µmol kg-1). Applying the observed effect on
Globigerina bulloides (-1.3 ‰ in δ13Cforaminifera / 100 µmol kg-1 in
[CO32-]; Spero et al., 1997) the range of about
50 µmol kg-1 in [CO32-] implies a potential effect of
-0.65 ‰ on the δ13C values of foraminifera, and thus
might explain the lower values found in the surface waters in the east.
However, we cannot see this difference between east and west in the offsets
measured between δ13CDIC and
δ13CN.p., which points to the fact that other processes
are responsible for the deviation found in near-surface waters as well.
Nevertheless, assuming that the vital effect in δ13C close to the
sea surface is influenced by increased carbonate ion concentrations induced
through high primary production, the smaller average vital effects reported
by Volkmann and Mensch (2001; -2.15 ‰) and Stangeew (2001;
-2 ‰) from the Fram Strait more than 10 years earlier may point to
an increase in bioproductivity during the last decades in the area. Data sets
of [CO32-] recorded between 1982 and 2002 in the Fram Strait (CARINA
database, 2015) however do not show respective changes, which may indicate a
significant shift only after 2002. We also have to consider that
bioproductivity may vary interannually and within the summer season.
As also discussed with respect to the offset in δ18O between
coretop and living foraminifera, the age of sediment surface samples can vary
in a great range (between modern to 3 ky, with an average of ∼ 1 ky;
Simstich et al., 2003). Accordingly, they may reflect significantly older
environments than the plankton samples. The negative offset in δ13C between the sediment and plankton samples may thus be explained by
the surface ocean Suess effect: during the last 100 years the carbon isotope
composition of the atmosphere has changed due to the increased anthropogenic
combustion of fossil carbon which is extremely negative in δ13C.
The δ13C values of the atmospheric CO2 have decreased by about
1.4 ‰ (Friedli et al., 1986; Francey et al., 1999) and the
concurrent shift in the stable carbon isotope composition of ocean surface
water is reflected in the decrease of δ13C of recent foraminiferal
shells (Bauch et al., 2000). The offset of roughly -1 ‰ in
δ13C between the sediment and plankton samples observed both in
this study and in that of Bauch et al. (2000) may therefore be explained by
the different ages of the carbonate in both sample sets and the developments
that have occurred in the last ca. 100 years.