380
Romeroetal.
Northern Africa, with the Saharan and Sahelian
regions, is the main source of windblown dust in
the northern hemisphere (Schlitz 1980; Swap et al.
1992). This dust is fed into the major wind systems
and transported outside the desert's margins in a
plume, across the equatorial Atlantic to the Caribbean and northeast coast of South America, in a
southwesterly direction toward the coast of Guinea,
and across the Mediterranean to Europe and the
Near East (prospero and Carlson 1972;Kalu 1979;
Schlitz 1980 and references therein; Swap et al.
1992).
The seasonal variation of the Saharan dust
layer is coupled to the annual migration of the
Intertropical Convergence Zone (lTCZ, Fig. 7;
Kalu 1979; Schlitz 1980). During northern summer,
when the ITCZ moves northward, the dust layer
is usually confmed to a latitudinal band between 15°
and 25° N (summer plume, Schlitz 1980), and affects our Cape Blanc site (Lange et al. 1998). As
the ITCZ moves to the south in winter, the dust
plume moves in a southwesterly direction and is
located between the equator and 15° N (winter
plume, Schlitz 1980), allowing dust particles to
reach our eastern and western equatorial sites (Fig.
7). At this time, the Cape Verde site is localized
across the main path of the African dust transport
in the NE trade winds.
Our trap data clearly show a decrease in
lithogenic, freshwater diatom and phytolith flux
values as a function of distance from the source
area (African continent). A sharp contrast is seen
between the northern (north of the equator) and
southern (south of the equator) sites, with lowest
values in the south and in the western Atlantic
(Table 2). Of all the sites studied, Cape Verde has
the largest contribution of freshwater diatoms
and phytoliths (Table 2). Aulacoseira granulata
(Plate I) andA. islandica dominated the freshwater assemblage in all areas. As with the flux data,
the relative contribution ofthe freshwater assemblage to the total diatom assemblage decreased with
increasing distance from the African continent.
The Walvis Ridge site, on the other hand, is
affected by the "berg" winds which are typical of
the entire Benguela region during fall and winter,
and may transport substantial quantities of sand and
dust out to sea up to 150 \an offshore between 18°
and 30° S (Shannon 1985). The only evidence found
for this transporting agent was the frequent occurrence of phytoliths in the WR2 trap; freshwater
diatoms were absent (Treppke et al. 1996b).
Surface Sediment Pattern
We know that the relationship between sediment
accumulation and measured rates of silica production and present-day primary productivity is not
straightforward (e.g. Nelson et al. 1995). The fraction of biogenic material that escapes dissolution in
the surface layers and during settling through the
water column finally reaches the seafloor where
dissolution continues until either the interstitial
waters become saturated or altered chemical composition of the particle surfaces prevents further
dissolution (see Nelson et al. 1995; Treguer et al.
1995, and references therein).
Diatom species composition in a particular area,
cell size and frustule thickness are important factors in determining whether siliceous sediments
accumulate below. In the eastern equatorial
Atlantic, the diatom assemblage of the biologically
most important season (summer upwelling), which
is characterized by small, lightly silicified diatoms
of the Nitzschia bicapitata group, is almost completely removed from the sedimentary record
(Lange et al. 1994; Treppke et al. 1996a). In this
area we are left with a sediment assemblage enriched in large and/or heavy silicified forms representative of the low-productivity season of the surface waters (Table 3). This conclusion can be extrapolated to the surface sediments of the Cape
Verde area (underlying trap CVI-2) and to the
western equatorial Atlantic (underneath traps
W A3 and W A 7) as well, where robust taxa (e.g.
Aulacoseira granulata, Azpeitia barronii,
A. nodulifera, Roperia tesselata, Rhizosolenia
bergonii; Plate I) are preferentially concentrated
in the sediments. In contrast, diatoms indicating persistent offshore spreading of the coastal upwelling
off Cape Blanc possess moderately to very robust
frustules readily preserved in the underlying
sediments. Here the sediment assemblage does
correspond with the trap assemblage and does
record the high-productivity season ofthe surface
waters (Lange et al. 1998).
Romeroetal.
Northern Africa, with the Saharan and Sahelian
regions, is the main source of windblown dust in
the northern hemisphere (Schlitz 1980; Swap et al.
1992). This dust is fed into the major wind systems
and transported outside the desert's margins in a
plume, across the equatorial Atlantic to the Caribbean and northeast coast of South America, in a
southwesterly direction toward the coast of Guinea,
and across the Mediterranean to Europe and the
Near East (prospero and Carlson 1972;Kalu 1979;
Schlitz 1980 and references therein; Swap et al.
1992).
The seasonal variation of the Saharan dust
layer is coupled to the annual migration of the
Intertropical Convergence Zone (lTCZ, Fig. 7;
Kalu 1979; Schlitz 1980). During northern summer,
when the ITCZ moves northward, the dust layer
is usually confmed to a latitudinal band between 15°
and 25° N (summer plume, Schlitz 1980), and affects our Cape Blanc site (Lange et al. 1998). As
the ITCZ moves to the south in winter, the dust
plume moves in a southwesterly direction and is
located between the equator and 15° N (winter
plume, Schlitz 1980), allowing dust particles to
reach our eastern and western equatorial sites (Fig.
7). At this time, the Cape Verde site is localized
across the main path of the African dust transport
in the NE trade winds.
Our trap data clearly show a decrease in
lithogenic, freshwater diatom and phytolith flux
values as a function of distance from the source
area (African continent). A sharp contrast is seen
between the northern (north of the equator) and
southern (south of the equator) sites, with lowest
values in the south and in the western Atlantic
(Table 2). Of all the sites studied, Cape Verde has
the largest contribution of freshwater diatoms
and phytoliths (Table 2). Aulacoseira granulata
(Plate I) andA. islandica dominated the freshwater assemblage in all areas. As with the flux data,
the relative contribution ofthe freshwater assemblage to the total diatom assemblage decreased with
increasing distance from the African continent.
The Walvis Ridge site, on the other hand, is
affected by the "berg" winds which are typical of
the entire Benguela region during fall and winter,
and may transport substantial quantities of sand and
dust out to sea up to 150 \an offshore between 18°
and 30° S (Shannon 1985). The only evidence found
for this transporting agent was the frequent occurrence of phytoliths in the WR2 trap; freshwater
diatoms were absent (Treppke et al. 1996b).
Surface Sediment Pattern
We know that the relationship between sediment
accumulation and measured rates of silica production and present-day primary productivity is not
straightforward (e.g. Nelson et al. 1995). The fraction of biogenic material that escapes dissolution in
the surface layers and during settling through the
water column finally reaches the seafloor where
dissolution continues until either the interstitial
waters become saturated or altered chemical composition of the particle surfaces prevents further
dissolution (see Nelson et al. 1995; Treguer et al.
1995, and references therein).
Diatom species composition in a particular area,
cell size and frustule thickness are important factors in determining whether siliceous sediments
accumulate below. In the eastern equatorial
Atlantic, the diatom assemblage of the biologically
most important season (summer upwelling), which
is characterized by small, lightly silicified diatoms
of the Nitzschia bicapitata group, is almost completely removed from the sedimentary record
(Lange et al. 1994; Treppke et al. 1996a). In this
area we are left with a sediment assemblage enriched in large and/or heavy silicified forms representative of the low-productivity season of the surface waters (Table 3). This conclusion can be extrapolated to the surface sediments of the Cape
Verde area (underlying trap CVI-2) and to the
western equatorial Atlantic (underneath traps
W A3 and W A 7) as well, where robust taxa (e.g.
Aulacoseira granulata, Azpeitia barronii,
A. nodulifera, Roperia tesselata, Rhizosolenia
bergonii; Plate I) are preferentially concentrated
in the sediments. In contrast, diatoms indicating persistent offshore spreading of the coastal upwelling
off Cape Blanc possess moderately to very robust
frustules readily preserved in the underlying
sediments. Here the sediment assemblage does
correspond with the trap assemblage and does
record the high-productivity season ofthe surface
waters (Lange et al. 1998).