200
Fischer et al.
Table 2). In addition, this year was characterized
by fluxes of the temperature-sensitive species G.
ruber (w) of almost zero (Fig. 1 b; Table 1). The
differences between measured SSTs and reconstructed temperatures derived from G. ruber were
always negative (4-year mean=-1.7°C; Table 2)
suggesting that some unknown effects may have
biased our calculations. As anticipated from the precipitation depth of L. injlata (SO m), this species
seems to provide reliable annual means of temperature which are 0.3-1.3°C (4-year mean=0.8°C)
lower than measured SSTs (Table 2). Considering
annual means of measured water temperatures, the
difference between the surface (0 m) and SO m is
about 1.4°C (Levitus and Boyer 1994). For
paleoceanographic purposes, temperature estimates based on species fluxes are more relevant.
Considering the species fluxes ofG. ruber (w) of
the fraction larger than ISO f!m, the water temperature averaged over a period of four years was significantly lower (=20. 7°C) as compared to the mean
value (=23.SoC; Table 2) and thus was about 1°C
lower than the measured SST (Table 2). Annual
mean values of water temperature measured at the
surface (0 m) and in a depth ono m differ by about
O.SOC (Levitus and Boyer 1994). Species fluxes of
juvenileL. injlata(1S0-S00 f!m)were obtained only
for 1989, a year of unusually high pteropod fluxes
in all size fractions (Kalberer et al. 1993). Applying these fluxes, the mean temperature value
amounts 22.7°C in contrast to the unweighted annual mean of21.SoC (Table 2). This flux-weighted
value of L. injlata calcifying in a water depth of
about SO m is unrealistically high when compared
to the measured SST (0 m) of 21.7°C. This is
mainly due to the unusually high numbers of this
species collected in summer 1989.
Carbon Isotopes, Organic Carbon Fluxes
and (PC IC0 2
The stable carbon isotope record of both species
is shown in Fig. 4. We obtained much higher values (sometimes> I.S%o) for the deeper living L.
injlata(SO m and below) than forGo ruber(0-30 m
depth). Part of this difference (about O.S%o) may
be attributed to an increased 12C removal at greater
depth, near the chlorophyll maximum. But differences as large as I.S%o can hardly be explained
by the different precipitation depths in 30 m and
around 50 m of G. ruber and L. injlata, respectively. In addition, the seasonal patterns are different and sometimes show contrary trends (e.g. in
summer 1989/90). G. ruber displays higher isotope
values during summer when chlorophyll biomass
(based on CZCS data) and carbon fluxes are highest (Fischer et al. 1996); minimum isotope values
occurred in winter.L. injlatarevealed less seasonal
changes (except in 1989) with variable maxima and
minima in the 4-year period. Here, for instance,
maximum values occurred in summer 1988 and
1991, but also in winter 1989 and 1990 (Fig. 4).
The stable carbon isotope composition of both
species is not very well understood. The Ol3C of
planktonic foraminifera should depend on several
factors: (1) the isotope composition ofthe ambient
water (ol3C LC0 2 ), (2) respiration and symbiont
photosynthesis ("vital effects"), (3) irradiance level
(Spero and Williams 1988; Spero and Lea 1993),
species size (Fairbanks et al. 1982; Ravelo and
Fairbanks 1995) and the CO} 2- content of ambient
sea water (Bijma et al. this volume). An overview
on the stable carbon isotope composition of
foraminifera is given by Multiza et al. (this volume);
a review of the sedimentary--8 l3 C signal is provided
by Berger and Vincent (1986). According to Spero
and Williams (1988), SSTs do not significantly affect the OJ3C record of planktonic foraminifera. The
Ol3C of aragonitic pteropods also depend on the
Ol3C composition of ambient sea water (e.g. Jasper and Deuser 1993), but, in contrastto planktonic
foraminifera, they are significantly affected by
ambient SSTs (Grossman and Ku 1986). Grossman
and Ku (1986) found that aragonitic molluscs decrease in 013C by 0.11 %0 per each degree centigrade of temperature increase; Kalberer et al.
(1993) obtained an almost similar negative temperature dependency of 0.10%0. This change may be
due to a temperature effect on growth rates,
whereby higher temperatures increase the magnitude of carbon isotope disequilibrium between DIC
and shell carbonate (Grossman and Ku 1986). Such
a temperature dependency may form the basis for
building a paleothermometer without the requirement of the isotopic composition of a paleo-ocean.
Fischer et al.
Table 2). In addition, this year was characterized
by fluxes of the temperature-sensitive species G.
ruber (w) of almost zero (Fig. 1 b; Table 1). The
differences between measured SSTs and reconstructed temperatures derived from G. ruber were
always negative (4-year mean=-1.7°C; Table 2)
suggesting that some unknown effects may have
biased our calculations. As anticipated from the precipitation depth of L. injlata (SO m), this species
seems to provide reliable annual means of temperature which are 0.3-1.3°C (4-year mean=0.8°C)
lower than measured SSTs (Table 2). Considering
annual means of measured water temperatures, the
difference between the surface (0 m) and SO m is
about 1.4°C (Levitus and Boyer 1994). For
paleoceanographic purposes, temperature estimates based on species fluxes are more relevant.
Considering the species fluxes ofG. ruber (w) of
the fraction larger than ISO f!m, the water temperature averaged over a period of four years was significantly lower (=20. 7°C) as compared to the mean
value (=23.SoC; Table 2) and thus was about 1°C
lower than the measured SST (Table 2). Annual
mean values of water temperature measured at the
surface (0 m) and in a depth ono m differ by about
O.SOC (Levitus and Boyer 1994). Species fluxes of
juvenileL. injlata(1S0-S00 f!m)were obtained only
for 1989, a year of unusually high pteropod fluxes
in all size fractions (Kalberer et al. 1993). Applying these fluxes, the mean temperature value
amounts 22.7°C in contrast to the unweighted annual mean of21.SoC (Table 2). This flux-weighted
value of L. injlata calcifying in a water depth of
about SO m is unrealistically high when compared
to the measured SST (0 m) of 21.7°C. This is
mainly due to the unusually high numbers of this
species collected in summer 1989.
Carbon Isotopes, Organic Carbon Fluxes
and (PC IC0 2
The stable carbon isotope record of both species
is shown in Fig. 4. We obtained much higher values (sometimes> I.S%o) for the deeper living L.
injlata(SO m and below) than forGo ruber(0-30 m
depth). Part of this difference (about O.S%o) may
be attributed to an increased 12C removal at greater
depth, near the chlorophyll maximum. But differences as large as I.S%o can hardly be explained
by the different precipitation depths in 30 m and
around 50 m of G. ruber and L. injlata, respectively. In addition, the seasonal patterns are different and sometimes show contrary trends (e.g. in
summer 1989/90). G. ruber displays higher isotope
values during summer when chlorophyll biomass
(based on CZCS data) and carbon fluxes are highest (Fischer et al. 1996); minimum isotope values
occurred in winter.L. injlatarevealed less seasonal
changes (except in 1989) with variable maxima and
minima in the 4-year period. Here, for instance,
maximum values occurred in summer 1988 and
1991, but also in winter 1989 and 1990 (Fig. 4).
The stable carbon isotope composition of both
species is not very well understood. The Ol3C of
planktonic foraminifera should depend on several
factors: (1) the isotope composition ofthe ambient
water (ol3C LC0 2 ), (2) respiration and symbiont
photosynthesis ("vital effects"), (3) irradiance level
(Spero and Williams 1988; Spero and Lea 1993),
species size (Fairbanks et al. 1982; Ravelo and
Fairbanks 1995) and the CO} 2- content of ambient
sea water (Bijma et al. this volume). An overview
on the stable carbon isotope composition of
foraminifera is given by Multiza et al. (this volume);
a review of the sedimentary--8 l3 C signal is provided
by Berger and Vincent (1986). According to Spero
and Williams (1988), SSTs do not significantly affect the OJ3C record of planktonic foraminifera. The
Ol3C of aragonitic pteropods also depend on the
Ol3C composition of ambient sea water (e.g. Jasper and Deuser 1993), but, in contrastto planktonic
foraminifera, they are significantly affected by
ambient SSTs (Grossman and Ku 1986). Grossman
and Ku (1986) found that aragonitic molluscs decrease in 013C by 0.11 %0 per each degree centigrade of temperature increase; Kalberer et al.
(1993) obtained an almost similar negative temperature dependency of 0.10%0. This change may be
due to a temperature effect on growth rates,
whereby higher temperatures increase the magnitude of carbon isotope disequilibrium between DIC
and shell carbonate (Grossman and Ku 1986). Such
a temperature dependency may form the basis for
building a paleothermometer without the requirement of the isotopic composition of a paleo-ocean.