34
Weferetal.
Similar changes resulting from the same processes also occur spatially, and leave a distinct imprint in the 8 15 N of sedimentary organic nitrogen,
e.g. with distance from the equatorial upwelling zone
(Altabet and Francois 1994), and with distance
offshore (Holmes et al. 1996, 1998). The latter
authors show that shelf and upper slope sediments
have distinctly lower 8 15 N values than further offshore in the Angola and Cape Basin. This observed
pattern reflects the high nitrate concentrations
nearshore (brought up by coastal upwelling) which
decrease to low and very low values further offshore.
Downcore 81~ analyses of bulk sediment, diatoms and foraminifers show a glacial minimum in
nutrient utilization in the equatorial Pacific (Farrell
etaI.1995), Southern Ocean (Shemesh et al. 1993),
and equatorial Atlantic (Altabet and Curry 1989)
pointing to lower to modestly increased biological
productivity compared to the modern ocean.
Francois et al. (1997) used the 81~ signal together
with other proxies for past nutrient supply and productivity to conclude that increased surface water
stratification south of the Antarctic Polar Front
(Atlantic Sector) contributed much to the lowering
of glacial atmospheric CO 2 ,
Despite the very promising results on nutrient
cycling in the ocean from nitrogen isotope studies,
many uncertainties remain. To estimate nitrate utilization in the past, the fractionation factorEu between nitrate and organic nitrogen (around 6 0 / 00 ),
the isotopic value ofthe source nitrate (3-7%0' e.g.
Wada et al. 1975) as well as alterations of the primary isotope signature due to degradation in the
water column and remineralization at the sedimentwater interface have to be taken into account.
Particle transformations in the water column combined with a transfer of nitrogen through higher
trophic levels and an accompanying isotope change
of 2-4% (e.g. Montoya 1994) may complicate
the interpretation. Surface sediments which are
characterized by an intensive remineralization of organic matter seem to be generally enriched by several per mil in 15N relative to surface water and
sinking particles (Montoya 1994; Altabet 1996;
Francois et al. 1997). In places, such changes may
be consistent (constant diagenetic offset) at a particular site and the downcore variations then would
record the changes in surface water nitrate utilization at least in a relative way. In contrast, in lowoxygen bottom water environments, the 8 15 N signature derived from the water column may be preserved without any offset during early diagenesis.
Other processes such as denitrification and nitrogen fixation may help determine the 8 15 N of
particulate matter. Denitrification mediated by bacteria under suboxic conditions results in an enrichment of15N in the subsurface nitrate (e.g. Montoya
1994; Schafer and Ittekkot 1993) which is then
reflected in the 8 15 N of sinking particles in the area
as for example shown for the Arabian Sea (Altabet
et al. 1995). Nitrogen fixation by cyanobacteria, on
the other hand, may be an important process in the
low-productive subtropical gyres and is responsible for extremely low values in the subtropical
North Pacific Gyre as demonstrated by Karl et al.
(1997). There, about one-half of the total annual
particulate nitrogen flux at 1,500m is supplied by
N2 fixation. In addition, the marine 8 15 N-signal in
near-shore environments may be altered due to the
contribution of some unknown fraction of terrestrial nitrogen with an unknown isotope signature.
An important aspect regarding the use of8 15 N
in paleoceanography is the prospect of reconstructing the overall intensity of denitrification in the
ocean as a function of geologic time. Changes in
available nitrogen (which is a limiting nutrient) have
been linked to changes in CO 2 partitioning between
ocean and atmosphere (McElroy 1983; Berger and
Keir 1984; Shaffer 1989). The idea is that increased
nitrate availability due to less water-column
denitrification during glacial periods would decrease
CO 2 in the ocean and atmosphere (Ganeshram et
al. 1995), analogous to the phosphate model of
Broecker (1982). If nitrate content of the ocean
changed in the manner envisioned, the 8 15 N in the
glacial ocean should have been lowered.
These fundamental questions await further research. New data from combined sediment-trap
and surface sediment studies in the Arabian Sea
show that the diagenetic change in 8 15 N across the
sediment-water interface is a constant offset (3 to
4 0 / 0 ) which depends on the intensity of oxic
remineralisation, so that the paleoflux of nitrogen
and the associated organic carbon can be calculated (Brummer, unpublished data). Moreover, tem-
Weferetal.
Similar changes resulting from the same processes also occur spatially, and leave a distinct imprint in the 8 15 N of sedimentary organic nitrogen,
e.g. with distance from the equatorial upwelling zone
(Altabet and Francois 1994), and with distance
offshore (Holmes et al. 1996, 1998). The latter
authors show that shelf and upper slope sediments
have distinctly lower 8 15 N values than further offshore in the Angola and Cape Basin. This observed
pattern reflects the high nitrate concentrations
nearshore (brought up by coastal upwelling) which
decrease to low and very low values further offshore.
Downcore 81~ analyses of bulk sediment, diatoms and foraminifers show a glacial minimum in
nutrient utilization in the equatorial Pacific (Farrell
etaI.1995), Southern Ocean (Shemesh et al. 1993),
and equatorial Atlantic (Altabet and Curry 1989)
pointing to lower to modestly increased biological
productivity compared to the modern ocean.
Francois et al. (1997) used the 81~ signal together
with other proxies for past nutrient supply and productivity to conclude that increased surface water
stratification south of the Antarctic Polar Front
(Atlantic Sector) contributed much to the lowering
of glacial atmospheric CO 2 ,
Despite the very promising results on nutrient
cycling in the ocean from nitrogen isotope studies,
many uncertainties remain. To estimate nitrate utilization in the past, the fractionation factorEu between nitrate and organic nitrogen (around 6 0 / 00 ),
the isotopic value ofthe source nitrate (3-7%0' e.g.
Wada et al. 1975) as well as alterations of the primary isotope signature due to degradation in the
water column and remineralization at the sedimentwater interface have to be taken into account.
Particle transformations in the water column combined with a transfer of nitrogen through higher
trophic levels and an accompanying isotope change
of 2-4% (e.g. Montoya 1994) may complicate
the interpretation. Surface sediments which are
characterized by an intensive remineralization of organic matter seem to be generally enriched by several per mil in 15N relative to surface water and
sinking particles (Montoya 1994; Altabet 1996;
Francois et al. 1997). In places, such changes may
be consistent (constant diagenetic offset) at a particular site and the downcore variations then would
record the changes in surface water nitrate utilization at least in a relative way. In contrast, in lowoxygen bottom water environments, the 8 15 N signature derived from the water column may be preserved without any offset during early diagenesis.
Other processes such as denitrification and nitrogen fixation may help determine the 8 15 N of
particulate matter. Denitrification mediated by bacteria under suboxic conditions results in an enrichment of15N in the subsurface nitrate (e.g. Montoya
1994; Schafer and Ittekkot 1993) which is then
reflected in the 8 15 N of sinking particles in the area
as for example shown for the Arabian Sea (Altabet
et al. 1995). Nitrogen fixation by cyanobacteria, on
the other hand, may be an important process in the
low-productive subtropical gyres and is responsible for extremely low values in the subtropical
North Pacific Gyre as demonstrated by Karl et al.
(1997). There, about one-half of the total annual
particulate nitrogen flux at 1,500m is supplied by
N2 fixation. In addition, the marine 8 15 N-signal in
near-shore environments may be altered due to the
contribution of some unknown fraction of terrestrial nitrogen with an unknown isotope signature.
An important aspect regarding the use of8 15 N
in paleoceanography is the prospect of reconstructing the overall intensity of denitrification in the
ocean as a function of geologic time. Changes in
available nitrogen (which is a limiting nutrient) have
been linked to changes in CO 2 partitioning between
ocean and atmosphere (McElroy 1983; Berger and
Keir 1984; Shaffer 1989). The idea is that increased
nitrate availability due to less water-column
denitrification during glacial periods would decrease
CO 2 in the ocean and atmosphere (Ganeshram et
al. 1995), analogous to the phosphate model of
Broecker (1982). If nitrate content of the ocean
changed in the manner envisioned, the 8 15 N in the
glacial ocean should have been lowered.
These fundamental questions await further research. New data from combined sediment-trap
and surface sediment studies in the Arabian Sea
show that the diagenetic change in 8 15 N across the
sediment-water interface is a constant offset (3 to
4 0 / 0 ) which depends on the intensity of oxic
remineralisation, so that the paleoflux of nitrogen
and the associated organic carbon can be calculated (Brummer, unpublished data). Moreover, tem-