Variability in Export Production Documented by Downward Fluxes and Species Composition
387
and (b) the range of primary production values for
the different areas over which the study sites span.
(2) Diatom sedimentation was markedly seasonal (except in the western equatorial Atlantic),
and enhanced fluxes at all sites corresponded in
time with the occurrence of upwelling.
(3) Diatoms were the main contributors to the
opal fraction (data are not available for radiolarians
in the western Atlantic), and their fluxes coincided
in time with total mass and opal fluxes.
(4) Whole diatom frustules containing well-preserved chloroplasts were present at all trap sites
(regardless of depth), suggesting that fast settling
through the water column may take place thus
minimizing dissolution and organic matter decay.
(5) Lithogenic, freshwater diatom and phytolith
fluxes document seasonal changes in Saharan dust
transport patterns, and the geographical extension
of the Saharan dust plume across the Atlantic.
(6) Accumulation rates of freshwater diatoms
and phytoliths in the sediments could be considered
as paleo indicators of the trade wind layer and the
winter dust plume.
(7) Marked trap-surface sediment discrepancies
in diatom assemblage composition was observed
at all sites (except off Cape Blanc) with the
sediments being enriched in robust taxa representative of the low productivity season of the surface
waters.
(8) Despite this considerable loss, typical diatom assemblages preserved in the sedimentary
record can be used as tracers of the corresponding hydrographic conditions of the surface waters.
Corollaries: Diatoms as Proxy of Export
Production and Paleoproductivity
Abundances of marine diatom valves in surface
and subsurface sediments have been extensively
used by paleoceanographers as a measure of the
diatom flux exported from the photic zone, and considered as a proxy for paleoproductivity. Data
based on the taxonomic composition of microfossil
assemblages are among the most useful methods
for interpreting past environmental conditions
(Sancetta 1989a). However, to make use of these
data we need to assess the accuracy with which
the fossil record reproduces the original living assemblage. To resolve this question, it is necessary
to examine the sequence from production in the
upper water column, through vertical settling, to
final burial in the sediments, at time scales of days
to months. The combination of sediment trap experiments and investigations of the sediment surface and downcore makes it possible to determine
the seasonal and interannual variations of assemblages, and to use downcore fluctuations in them
as indicators of paleoconditions for different
seasons over glacial-interglacial timescales (e.g.
Sancetta 1995).
Most aquatic systems experience both seasonal
and interannual variability in the timing, magnitude,
and composition of primary production, the variability being much lower in conservative systems such
as the central ocean gyres than in dynamic ones
like coastal upwelling systems and polar oceans
(e.g. Sancetta 1989b; Wefer 1989). Sediment trap
studies have demonstrated that particle sedimentation in the oceans is markedly seasonal and episodic, and can be related to biological processes in
the surface waters. Since diatoms account for more
than half of the total suspended amorphous silica
and are a substantial part of the primary production of the ocean (Lisitzin 1972; Nelson et al. 1995;
Treguer et al. 1995), time-series of diatom fluxes
in a variety of marine habitats (like the ones presented here) will help improve models of the
biogeochemical cycle of silicon (see goals of
OPALEO workshop described in Ragueneau et al.
subm), and contribute to the development of the
silicate pump model of Dugdale et al. (1995) and
Dugdale and Wilkerson (1998).
Although we have made an effort in
presenting flux estimates for radiolarians and
silicoflagellates in addition to diatoms, data on the
two former groups are still scarce, and interannual
variations for all three silica-bearing microorganisms are still lacking for many regions.
Silicoflagellates are less important in modem biogenic silica production (Treguer et al. 1995), and
fluxes of radiolarians do not show a strong relationship to primary production and/or organic carbon
fluxes (Boltovskoy et al. 1993, 1996). In contrast,
diatom fluxes are moderately well correlated with
total particle and organic carbon fluxes, the correlation being higher in the coastal domain than in the
387
and (b) the range of primary production values for
the different areas over which the study sites span.
(2) Diatom sedimentation was markedly seasonal (except in the western equatorial Atlantic),
and enhanced fluxes at all sites corresponded in
time with the occurrence of upwelling.
(3) Diatoms were the main contributors to the
opal fraction (data are not available for radiolarians
in the western Atlantic), and their fluxes coincided
in time with total mass and opal fluxes.
(4) Whole diatom frustules containing well-preserved chloroplasts were present at all trap sites
(regardless of depth), suggesting that fast settling
through the water column may take place thus
minimizing dissolution and organic matter decay.
(5) Lithogenic, freshwater diatom and phytolith
fluxes document seasonal changes in Saharan dust
transport patterns, and the geographical extension
of the Saharan dust plume across the Atlantic.
(6) Accumulation rates of freshwater diatoms
and phytoliths in the sediments could be considered
as paleo indicators of the trade wind layer and the
winter dust plume.
(7) Marked trap-surface sediment discrepancies
in diatom assemblage composition was observed
at all sites (except off Cape Blanc) with the
sediments being enriched in robust taxa representative of the low productivity season of the surface
waters.
(8) Despite this considerable loss, typical diatom assemblages preserved in the sedimentary
record can be used as tracers of the corresponding hydrographic conditions of the surface waters.
Corollaries: Diatoms as Proxy of Export
Production and Paleoproductivity
Abundances of marine diatom valves in surface
and subsurface sediments have been extensively
used by paleoceanographers as a measure of the
diatom flux exported from the photic zone, and considered as a proxy for paleoproductivity. Data
based on the taxonomic composition of microfossil
assemblages are among the most useful methods
for interpreting past environmental conditions
(Sancetta 1989a). However, to make use of these
data we need to assess the accuracy with which
the fossil record reproduces the original living assemblage. To resolve this question, it is necessary
to examine the sequence from production in the
upper water column, through vertical settling, to
final burial in the sediments, at time scales of days
to months. The combination of sediment trap experiments and investigations of the sediment surface and downcore makes it possible to determine
the seasonal and interannual variations of assemblages, and to use downcore fluctuations in them
as indicators of paleoconditions for different
seasons over glacial-interglacial timescales (e.g.
Sancetta 1995).
Most aquatic systems experience both seasonal
and interannual variability in the timing, magnitude,
and composition of primary production, the variability being much lower in conservative systems such
as the central ocean gyres than in dynamic ones
like coastal upwelling systems and polar oceans
(e.g. Sancetta 1989b; Wefer 1989). Sediment trap
studies have demonstrated that particle sedimentation in the oceans is markedly seasonal and episodic, and can be related to biological processes in
the surface waters. Since diatoms account for more
than half of the total suspended amorphous silica
and are a substantial part of the primary production of the ocean (Lisitzin 1972; Nelson et al. 1995;
Treguer et al. 1995), time-series of diatom fluxes
in a variety of marine habitats (like the ones presented here) will help improve models of the
biogeochemical cycle of silicon (see goals of
OPALEO workshop described in Ragueneau et al.
subm), and contribute to the development of the
silicate pump model of Dugdale et al. (1995) and
Dugdale and Wilkerson (1998).
Although we have made an effort in
presenting flux estimates for radiolarians and
silicoflagellates in addition to diatoms, data on the
two former groups are still scarce, and interannual
variations for all three silica-bearing microorganisms are still lacking for many regions.
Silicoflagellates are less important in modem biogenic silica production (Treguer et al. 1995), and
fluxes of radiolarians do not show a strong relationship to primary production and/or organic carbon
fluxes (Boltovskoy et al. 1993, 1996). In contrast,
diatom fluxes are moderately well correlated with
total particle and organic carbon fluxes, the correlation being higher in the coastal domain than in the