Geoscience Reference
In-Depth Information
psu [Lea et al., 1999]. Results from culturing G. ruber are
not yet available. Ferguson et al. [2008] provided evidence
from sediment traps that high salinity in the eastern Medi-
terranean Sea was associated with elevated foraminiferal
Mg/Ca values, although high Mg/Ca calcite overgrowths
were later suggested as the cause [Hoogakker et al., 2009].
Sediment core-top data suggest that Mg/Ca may depend on
several factors in addition to temperature, including salinity,
but that the effects are strongest only at salinities >36.5 psu
[Arbuszewski et al., 2010]. Because modern annual mean
salinity in the vicinity of Orca Basin is 35.5 psu [Levitus
and Boyer, 1994], and increased freshwater would further
lower the salinity, the salinity effect on Mg/Ca-SST is likely
minimal (as discussed by Williams et al. [2010]).
On the other hand, one concern is that Mississippi River
water, which is known to have a low dissolved Mg/Ca value
of about 0.49 mol mol
1
compared to the Gulf of Mexico
value of about 5.1 mol mol
1
, might lower foraminiferal Mg/
Ca values. Nevertheless, a simple box model calculation
shows that the concentrations of both Mg and Ca are too low
in Mississippi River water (425 and 870
M, respectively)
[Briggs and Ficke, 1978] to affect seawater (53 and 10.3 mM,
respectively) or foraminiferal Mg/Ca values. Calculations
show that a 25% dilution of surface seawater would only
decrease seawater Mg/Ca by <3%, which is within measure-
ment error [Flower et al., 2004]. Therefore low-Mg/Ca melt-
water is unlikely to have affected our SST determinations
based on G. ruber.
Two commonly used equations for conversion of Mg/Ca
values to temperature [Dekens et al., 2002; Anand et al.,
2003] yield different absolute temperatures but identical
relative changes. Only the latter study supplied specific
equations for white and pink G. ruber (Mg/Ca = 0.449
exp(0.09
SST) and Mg/Ca = 0.381
exp(0.09
SST,
respectively). We have chosen to use the white G. ruber
equation for internal consistency and because core-top Mg/Ca
values yield SST values of 25.4°C [LoDico et al., 2006;
Richey et al., 2007, 2009], equivalent to mean annual SST
in the northern Gulf of Mexico [Earth System Research
Laboratory, 1994]. Paired Mg/Ca SST and
μ
Figure 6.
Comparison of polar ice core records for the 8
-
23 ka
interval. NGRIP
δ
18
O
[EPICA Community Members et al., 2006] are compared to the ice
volume-corrected
δ
18
O[Rasmussen et al., 2006] and EDML
δ
18
Osw record from EN32-PC6 [Flower et al.,
2004] updated to the MARINE09
14
C calibration data set [Reimer et
al., 2009] using Calib 6.0 (http://intcal.qub.ac.uk/calib/). YD, B/A,
Greenland stadial 2a (GS-2a), Antarctic Cold Reversal (ACR), and
AIM 1 are labeled on the ice core records. Modern Gulf of Mexico
δ
18
O values were
δ
18
Osw is indicated by horizontal line.
18
Osw values based on a paleotemperature
equation (SST(°C) = 14.9
used to calculate
δ
w)) [Bemis et al.,
1998] that is appropriate for G. ruber [Thunell et al., 1999].
Adding 0.27
‰
yields values on the Vienna standard mean
ocean water scale [Hut, 1987].
Paired Mg/Ca SST and
4.8(
δ
c
δ
al. [2002] for three reasons: (1) for consistency with our
other Gulf of Mexico Mg/Ca work (2) because early Holo-
cene Mg/Ca SST from this core (~26°C) is close to modern
mean annual SST and (3) because early Holocene
δ
18
O data from core EN32-PC6
[Flower et al., 2004], updated to the MARINE09 timescale
[Reimer et al., 2009] are compared to polar ice core records
(Figure 6). The original Mg/Ca data [Flower et al., 2004]
have here been calibrated to SST using the white G. ruber
equation from Anand et al. [2003] instead of from Dekens et
δ
18
Osw is
close to modern values [Fairbanks et al., 1992; Schmidt et
al., 1999]. Calculated
δ
18
Osw values have also been cor-
rected for changing ice volume effects using deglacial sea
level records [Fairbanks, 1989; Bard et al., 1990] and a
Search WWH ::
Custom Search