The traditional indicators of productivity, organic matter and opal abundance, are strongly influenced by diagenesis, which complicates interpretations. To obtain additional clues to changes in productivity, we turn to benthic foraminifers. Benthic organisms live on the food provided by the export from overlying waters. The amount and nature of this export is reflected in the abundance and species composition of benthic foraminifers (Douglas and Woodruff, 1981; Woodruff, 1985; Lutze et al., 1986; Hermelin and Shimmield, 1990; Hermelin, 1992; Burke et al., 1993; Herguera and Berger, 1994; Loubere, 1994; Thomas et al., 1995). Unfortunately, the abundance of benthic foraminifers in many of the sites is largely controlled by carbonate preservation, as discussed above. We next explore the possibility that despite this handicap, species composition retains valuable information regarding trends in productivity.
The diversity of a benthic assemblage is sensitive to food supply: organic-rich environments tend to show reduced diversity but high abundance (Phleger and Soutar, 1973; Douglas and Woodruff, 1981; Hermelin and Shimmield, 1995). A simple index of diversity (which minimizes the effects from sample size) is the inverse of the percentage of the most abundant species. The distribution of this index shows minimal values for Sites 1081, 1083, and 1084. Maximum values are calculated for Sites 1076 and 1085 (Fig. 17) and in the other two pelagic sites (1086 and 1087; not shown in the figure). However, comparison with the ranking of sites in terms of productivity shows no clear relationship. In the cores south of the Walvis Ridge, there is a tendency for higher values to occur in sediments older than 3 Ma. This would agree with the proposition of lower productivity before the late Pliocene. However, there is no sign of recovery of diversity in the late Quaternary, for which we postulate lowered productivity. Thus, from these data, it appears that the late Quaternary productivity reduction applies only to opal, not to organic matter.
We plot the patterns for four taxa: Bulimina spp., Bolivina spp., Uvigerina spp., and Cibicidoides spp. The first two have long been recognized as forms abundant in high-productivity regions in coastal environments, such as off California (Uchio, 1960; Douglas and Woodruff, 1981). The presence of Uvigerina may indicate elevated levels of productivity (Lutze et al., 1986; Hermelin and Shimmield, 1990, 1995) and also has been tied to low-oxygen conditions in a pelagic setting (Burke et al., 1993). Different species within this genus apparently have different preferences (Berger et al., 1987). Cibicidoides is considered neutral with respect to productivity, at least with respect to its relative abundance (Berger and Herguera, 1992).
The genus Bulimina comprises several species (B. aculeata, B. exilis, B. marginata, B. mexicana, and B. truncana), which appear to have more or less similar distributional patterns. Maximum values occur at the Lüderitz site (1084) and in two of the Walvis sites (1082 and 1081), but not in the third (Site 1083). The high percentages appear within the Quaternary record (Fig. 18). There is a drop in the upper Quaternary section at Site 1081. High values also characterize many samples from the Congo Sites 1076 and 1077, but not from Site 1075. The patterns suggest that this taxon tracks high productivity in pulsing upwelling systems associated with strong oxygen minima. It prefers shallower areas within the set of Leg 175 sites; that is, areas with a strong influence from coastal upwelling. This agrees well with results from the upwelling area in the Arabian Sea (Hermelin and Shimmield, 1990). If this interpretation is correct, coastal upwelling effects within the Benguela Current system (Walvis and Lüderitz sites) increased greatly from the late Miocene to the Quaternary, with a possible maximum in late Pliocene to early Quaternary time.
The genus Bolivina is represented mainly by B. seminuda, B. pseudoplicata, and B. aenarensis. By far the highest abundances occur at the sites off Lobito (Angola), where Bolivina entirely dominates the assemblages (Fig. 19). These are environments of very high accumulation rates of terrigenous sediment, with a seasonal supply of organic matter from coastal upwelling. Furthermore, the dominance of Bolivina suggests low oxygen concentrations at the seafloor (Smith, 1963, 1964; Phleger and Soutar, 1973). High percentages of Bolivina also are noted in upper Quaternary sediments at Site 1081, where this genus replaces Bulimina as the dominant form in a pulse-like fashion. Substantial occurrences in the upper Miocene and lower Pliocene sections at Site 1082 are somewhat puzzling. Apparently, Bolivina also thrives here at the expense of Bulimina. The patterns suggest that this taxon prefers environments of moderate to high productivity in a coastal setting and does especially well where influx of terrigenous silts dominate the environment. Occurrences before and after the Bulimina-dominated late Neogene sequences (Sites 1081, 1082, and 1084) suggest that productivity went through a maximum centered between 2 and 3 Ma, with Bolivina occupying the flanks of that maximum.
The genus Uvigerina comprises the species U. auberiana, U. gallowayi, U. hispida, U. hispidocostata, and U. peregrina. Maxima occur at Site 1081 on the Walvis Ridge (Fig. 20). Among the Benguela set of sites, 1081 ranks low in productivity, but intermediate in opal deposition. It is at the edge of the Angolan oxygen minimum at a rather shallow depth (805 m). The high values for Uvigerina would seem to support the notion that this species prefers an environment of moderately elevated productivity in association with an oxygen minimum. If this interpretation is correct, oxygen deficiency is indicated for a period between 3 and 2 Ma on the Walvis Ridge and in the middle of the Quaternary at Site 1082 (Walvis Bay). In the Cape transect, high values occur in the lower Pliocene (Site 1085; other sites not shown because of spotty data) but also in the upper Pliocene (Site 1087). Thus, the distribution of Uvigerina would support the notion of an expanded oxygen minimum in the Cape Basin before the Gaussian termination of global warm-climate conditions.
The taxon Cibicidoides (also referred to as Cibicides and Planul-ina by various authors) comprises the species C. bradyi, C. pachyderma, and C. wuellerstorfi. These forms are of special interest because they serve as carriers of isotopic information for deep-water composition. They are thought to live mainly as epibenthos, avoiding interference from the isotopic composition of interstitial waters (Lutze and Thiel, 1989; review in Wefer and Berger, 1991). Cibicidoides is widespread and apparently well adapted to a wide variety of environments. It has substantial presence in the Congo Fan sites (1075, 1076, and 1077) and in the Benguela set (Sites 1081–1087; see Fig. 21). It tends to avoid the environment preferred by Bolivina (Sites 1078 and 1079). Maxima occur where the other abundant genera, which are tied to elevated productivity, are not entirely dominant. An interesting situation is shown in the Lüderitz site (1084), where Cibicidoides unexpectedly reaches high values within the high-productivity upper Pliocene section. Apparently, it alternates with Bulimina, the high-productivity taxon. This suggests that productivity varied considerably at Site 1084 right through the early Matuyama opal maximum. If so, it supports the notion of cyclic upwelling and mixing, as opposed to a continuous strong upwelling (as in the late Quaternary).
In summary, the abundance patterns of benthic foraminiferal species suggest that there is no simple ranking of species with respect to productivity, but that each responds to a mixture of environmental factors. We assume that such factors include sedimentation rate, quantity and quality of organic matter, oxygen supply, and substrate properties. The built-in negative correlation among abundant taxa also makes interpretations difficult. The present analysis by inspection suggests that the presumed productivity maximum associated with the Matuyama opal maximum is real and not just an artifact of diagenetic processes. Also, there are indications of alternating high- and low- productivity periods within and just before the maximum, between 3 and 2 Ma. The hypothesis of an expanded oxygen minimum before the Gaussian climate revolution is supported (albeit mildly).