208
Wolffetal.
One method encountered in recent publications
(Duplessy et al. 1991; Rostek et al. 1993) is based
on the oxygen isotopic composition of calcite shells
of planktonic foraminifera, which depends on the
temperature and isotopic composition ofthe ambient water mass. The isotopic composition of the
water is again empirically related to salinity. The
most severe problems with the oxygen-isotope
method result from uncertainties in the isotope-salinity relationship and uncertainties in the estimated
temperature.
Other methods are based on the ecological preferences of foraminiferal assemblages, relating relative abundances of planktonic foraminiferal species
to oceanographic parameters. The most critical
point here is the isolation of one desired parameter
(salinity) which does not account for complex dependency ofthe assemblages on various other factors such as temperature, nutrients, seasonality and
thermocline structure, anyone of which may be of
greater significance in determining the overall
faunal composition.
The aim of this paper is to summarize recent
approaches for paleosalinity reconstruction and to
evaluate the pros and cons, flaws and perspectives
of the various methods. This will be pursued by
investigation of core-top and downcore data for
each method.
Salinity Reconstructions Using Oxygen
Isotopes
The Paleotemperature Equation
Several attempts have been made in the last decade to estimate paleo salinities from the oxygen isotope compositions of calcite tests of planktonic
foraminifera (Fillon and Williams 1984; Broecker
1989; Spero and Williams 1990; Broecker 1990;
Sikes and Keigwin 1996). Recently, more quantitative approaches were made to reconstruct
paleosalinities down core (Duplessy et al. 1992;
Rostek et al. 1993; Maslin et al. 1995; Wang et al.
1995; Kallel et al. 1997) and on the time slice of
the last glacial maximum (Duplessy et al. 1991,
1996).
The oxygen isotope composition recorded in the
tests of planktonic foraminifera is determined
mainly by the temperature and the isotopic composition ofthe water mass in which they calcify. To a
lesser extent it also depends on the carbonate ion
concentration as shown by Spero et al. (1997). This
effect will be neglected here.
The ISO/ 16 0-ratio of the precipitated calcite is
greater than the ISO/ 16 0-ratio of the sea-water.
This difference decreases with increasing temperatures of the water at the time of calcite precipitation (Berger and Gardner 1975). Thus, measuring
the isotopic composition of planktonic foraminifera
and estimating the isotopic composition ofthe ambient sea-water enables us to compute temperatures for this water mass. This method has been
used to reconstruct sea-surface temperatures since
Emiliani applied it to Pleistocene sediments in 1955
(e.g. Emiliani 1955; Vincent and Shackleton 1980;
Van Campo et al. 1990; Keigwin 1996). The approach is based on the theoretically and empirically
derived equation (Urey 1947; Epstein et al. 1953):
where T is the temperature in °C, 8 18 0 C is the isotopic composition of the calcite test in %0 relative
to the PDB standard and 8 1S O W is the isotopic composition of the ambient sea-water in %0 (PDB).
Fig. I illustrates the dependency of ~8180 (8 IS O C -
8 IS O W ) on the temperature as determined in
culture studies by Erez and Luz (1983).
Reconstructing the Oxygen Isotope
Composition of Sea-Water
Returning to equation (1), however, it is obvious that
if the water temperature is known from an independent method the oxygen-isotope ratio of seawater can easily be calculated from 8 1s O measurements offoraminiferal tests. Solving equation (1)
for 8 1S O W gives
Computing the oxygen-isotope ratio of seawater can be regarded as the first step towards sa-
Wolffetal.
One method encountered in recent publications
(Duplessy et al. 1991; Rostek et al. 1993) is based
on the oxygen isotopic composition of calcite shells
of planktonic foraminifera, which depends on the
temperature and isotopic composition ofthe ambient water mass. The isotopic composition of the
water is again empirically related to salinity. The
most severe problems with the oxygen-isotope
method result from uncertainties in the isotope-salinity relationship and uncertainties in the estimated
temperature.
Other methods are based on the ecological preferences of foraminiferal assemblages, relating relative abundances of planktonic foraminiferal species
to oceanographic parameters. The most critical
point here is the isolation of one desired parameter
(salinity) which does not account for complex dependency ofthe assemblages on various other factors such as temperature, nutrients, seasonality and
thermocline structure, anyone of which may be of
greater significance in determining the overall
faunal composition.
The aim of this paper is to summarize recent
approaches for paleosalinity reconstruction and to
evaluate the pros and cons, flaws and perspectives
of the various methods. This will be pursued by
investigation of core-top and downcore data for
each method.
Salinity Reconstructions Using Oxygen
Isotopes
The Paleotemperature Equation
Several attempts have been made in the last decade to estimate paleo salinities from the oxygen isotope compositions of calcite tests of planktonic
foraminifera (Fillon and Williams 1984; Broecker
1989; Spero and Williams 1990; Broecker 1990;
Sikes and Keigwin 1996). Recently, more quantitative approaches were made to reconstruct
paleosalinities down core (Duplessy et al. 1992;
Rostek et al. 1993; Maslin et al. 1995; Wang et al.
1995; Kallel et al. 1997) and on the time slice of
the last glacial maximum (Duplessy et al. 1991,
1996).
The oxygen isotope composition recorded in the
tests of planktonic foraminifera is determined
mainly by the temperature and the isotopic composition ofthe water mass in which they calcify. To a
lesser extent it also depends on the carbonate ion
concentration as shown by Spero et al. (1997). This
effect will be neglected here.
The ISO/ 16 0-ratio of the precipitated calcite is
greater than the ISO/ 16 0-ratio of the sea-water.
This difference decreases with increasing temperatures of the water at the time of calcite precipitation (Berger and Gardner 1975). Thus, measuring
the isotopic composition of planktonic foraminifera
and estimating the isotopic composition ofthe ambient sea-water enables us to compute temperatures for this water mass. This method has been
used to reconstruct sea-surface temperatures since
Emiliani applied it to Pleistocene sediments in 1955
(e.g. Emiliani 1955; Vincent and Shackleton 1980;
Van Campo et al. 1990; Keigwin 1996). The approach is based on the theoretically and empirically
derived equation (Urey 1947; Epstein et al. 1953):
where T is the temperature in °C, 8 18 0 C is the isotopic composition of the calcite test in %0 relative
to the PDB standard and 8 1S O W is the isotopic composition of the ambient sea-water in %0 (PDB).
Fig. I illustrates the dependency of ~8180 (8 IS O C -
8 IS O W ) on the temperature as determined in
culture studies by Erez and Luz (1983).
Reconstructing the Oxygen Isotope
Composition of Sea-Water
Returning to equation (1), however, it is obvious that
if the water temperature is known from an independent method the oxygen-isotope ratio of seawater can easily be calculated from 8 1s O measurements offoraminiferal tests. Solving equation (1)
for 8 1S O W gives
Computing the oxygen-isotope ratio of seawater can be regarded as the first step towards sa-