490
Bijmaetal.
ate could be used as a paleothermometer for the
ocean. The method is based on the fact that sea
water and calcite differ in their 18 0/ 16 0 ratios,
when in thermodynamic equilibrium, and that this
difference decreases with increasing temperature.
McCrea (1950) demonstrated in the laboratory that
synthetic calcium carbonate when precipitated in
equilibrium with sea water is enriched in 180 relative to sea water but less so with increasing temperature and thus behaves as predicted by Urey.
This empirical relationship was the first, so called,
paleotemperature relationship. Another milestone
were the papers by Epstein eta!' (1951,1953) who
were the first to show that marine invertebrates
grown in the natural environment deposit calcium
carbonate in equilibrium with sea water. Many new
paleotemperature equations have been developed
since this early pioneer work (for a review see
Bemis et a!. 1998) but the problem in determining
the "true" paleotemperature is not only one of choosing the correct equation but, more importantly, one
of choosing the correct value for 0 18 0 of sea water (OISO). For the geologic past, this amounts to
knowing paleosalinity and the relationship
between salinity and 0 18 0 w (e.g. Berger 1979;
Broecker and Van Donk 1970; Fairbanks et a!.
1992; Rozanski et al. 1993; Schrag et al. 1996;
Shackleton 1967,1974).
Beyond the glacial ice volume effect and the
effect ofprecipitationlevaporation, so called "vital
effects" were recognized. Contrary to Emiliani's
(1954) interpretation that the oxygen isotopic differences among planktic foraminifers reflect differences in depth habitat and hence differences
in precipitation temperature, Duplessy et al. (1970)
argued that the differential oxygen isotopic fractionation in benthic foraminifers from the same level in
a core requires a different explanation and they
suggested that metabolic CO, depleted in 180 is
incorporated into the shells of benthic and possibly
also planktic foraminifers. The hypothesis
that planktic foraminifers have different
ISO-fractionation factors was subsequently confirmed by Shackleton et a!. (1973) and later substantiated by numerous other studies (e.g. Curry and
Matthews 1981a; Duplessy et al. 1981; Fairbanks
et al. 1980; Shackleton 1974; Vergnaud-Grazzini
1976; Williams et a!. 1979). For instance, Buchardt
and Hansen (1977) note that benthic foraminifers
bearing symbiotic algae are depleted in 180 relative to symbiont barren species and Erez (1978b)
argues that with greater photosynthetic activity,
incorporation of light metabolic CO, results in
anomalously light carbon and oxygen isotopic compositions of hermatypic corals and benthic
foraminifers. However, when dealing with differences in 0 18 0 through time, vital effects have generally been assumed to be constant (or zero) and
were therefore neglected.
The interpretation ofthe carbon isotopic composition proved to be more complicated thano l8 0.
Due to the preferential fixation of I 'CO, by primary
producers and subsequent transport and decomposition of organic matter below the euphotic zone,
surface waters of the ocean are depleted in I'C
whereas the deeper ocean is I'C-enriched. Thus,
as a consequence of community production and
respiration, olJC depth profiles are related to the
oxygen concentration and inversely correlated with
the nutrient concentration (the stoichiometric proportionality constant with P0 4 is ca. 0.93 %0 per
mol,kg"l: Broecker and Peng 1982). Hence the
olJC of surface dwelling foraminifers is used as a
surface water fertility proxy and the difference in
olJC between planktics and benthics (~olJC) was
introduced as a measure of the strength ofthe biological pump (e.g. Broecker 1971; 1973; 1981).
Another source for variation ofo l3 C in dissolved
inorganic carbon (OIJCl:CO) is the fractionation
between carbon in atmospheric CO 2 and the total
dissolved CO, (LCO,) in surface ocean water. The
lJC/I'C ratio in atmospheric CO, is on average 9
%0 lower than that in surface ocean LCO, and the
extent of fractionation depends on temperature (e.g.
Broecker and Maier-Reimer 1992; Charles et al.
1993; Lynch-Stieglitz et a!. 1995). As a result, high
latitude surface water have a higher lJC/12C ratio
than low latitude surface water. Generally, this thermodynamic imprint is not very strong due to very
slow isotopic equilibration (Broecker and Peng
1982), however, if there is sufficient time
for gas exchange, a significant signal is produced
(Charles and Fairbanks 1990). A third source of
variation is the so called "Mackensen effect".
Bijmaetal.
ate could be used as a paleothermometer for the
ocean. The method is based on the fact that sea
water and calcite differ in their 18 0/ 16 0 ratios,
when in thermodynamic equilibrium, and that this
difference decreases with increasing temperature.
McCrea (1950) demonstrated in the laboratory that
synthetic calcium carbonate when precipitated in
equilibrium with sea water is enriched in 180 relative to sea water but less so with increasing temperature and thus behaves as predicted by Urey.
This empirical relationship was the first, so called,
paleotemperature relationship. Another milestone
were the papers by Epstein eta!' (1951,1953) who
were the first to show that marine invertebrates
grown in the natural environment deposit calcium
carbonate in equilibrium with sea water. Many new
paleotemperature equations have been developed
since this early pioneer work (for a review see
Bemis et a!. 1998) but the problem in determining
the "true" paleotemperature is not only one of choosing the correct equation but, more importantly, one
of choosing the correct value for 0 18 0 of sea water (OISO). For the geologic past, this amounts to
knowing paleosalinity and the relationship
between salinity and 0 18 0 w (e.g. Berger 1979;
Broecker and Van Donk 1970; Fairbanks et a!.
1992; Rozanski et al. 1993; Schrag et al. 1996;
Shackleton 1967,1974).
Beyond the glacial ice volume effect and the
effect ofprecipitationlevaporation, so called "vital
effects" were recognized. Contrary to Emiliani's
(1954) interpretation that the oxygen isotopic differences among planktic foraminifers reflect differences in depth habitat and hence differences
in precipitation temperature, Duplessy et al. (1970)
argued that the differential oxygen isotopic fractionation in benthic foraminifers from the same level in
a core requires a different explanation and they
suggested that metabolic CO, depleted in 180 is
incorporated into the shells of benthic and possibly
also planktic foraminifers. The hypothesis
that planktic foraminifers have different
ISO-fractionation factors was subsequently confirmed by Shackleton et a!. (1973) and later substantiated by numerous other studies (e.g. Curry and
Matthews 1981a; Duplessy et al. 1981; Fairbanks
et al. 1980; Shackleton 1974; Vergnaud-Grazzini
1976; Williams et a!. 1979). For instance, Buchardt
and Hansen (1977) note that benthic foraminifers
bearing symbiotic algae are depleted in 180 relative to symbiont barren species and Erez (1978b)
argues that with greater photosynthetic activity,
incorporation of light metabolic CO, results in
anomalously light carbon and oxygen isotopic compositions of hermatypic corals and benthic
foraminifers. However, when dealing with differences in 0 18 0 through time, vital effects have generally been assumed to be constant (or zero) and
were therefore neglected.
The interpretation ofthe carbon isotopic composition proved to be more complicated thano l8 0.
Due to the preferential fixation of I 'CO, by primary
producers and subsequent transport and decomposition of organic matter below the euphotic zone,
surface waters of the ocean are depleted in I'C
whereas the deeper ocean is I'C-enriched. Thus,
as a consequence of community production and
respiration, olJC depth profiles are related to the
oxygen concentration and inversely correlated with
the nutrient concentration (the stoichiometric proportionality constant with P0 4 is ca. 0.93 %0 per
mol,kg"l: Broecker and Peng 1982). Hence the
olJC of surface dwelling foraminifers is used as a
surface water fertility proxy and the difference in
olJC between planktics and benthics (~olJC) was
introduced as a measure of the strength ofthe biological pump (e.g. Broecker 1971; 1973; 1981).
Another source for variation ofo l3 C in dissolved
inorganic carbon (OIJCl:CO) is the fractionation
between carbon in atmospheric CO 2 and the total
dissolved CO, (LCO,) in surface ocean water. The
lJC/I'C ratio in atmospheric CO, is on average 9
%0 lower than that in surface ocean LCO, and the
extent of fractionation depends on temperature (e.g.
Broecker and Maier-Reimer 1992; Charles et al.
1993; Lynch-Stieglitz et a!. 1995). As a result, high
latitude surface water have a higher lJC/12C ratio
than low latitude surface water. Generally, this thermodynamic imprint is not very strong due to very
slow isotopic equilibration (Broecker and Peng
1982), however, if there is sufficient time
for gas exchange, a significant signal is produced
(Charles and Fairbanks 1990). A third source of
variation is the so called "Mackensen effect".