Benthic foraminiferal response to Late Glacial and Holocene sea level rise and rainfall variability off East Africa

Highlights

• Benthic foraminiferal assemblages and sediment elemental compositions from the East African continental slope are presented

• Stronger flow of southern-sourced AAIWto the study site during Heinrich Stadial 1

• Higher river runoff to the Indian Ocean, and hence more humid conditions in East Africa during the East African Humid Period

• Severe reduction in river runoff and hence arid conditions in East Africa between 8.5 and 8.1 kyr

• Redevelopment of the East African fringe reefs since 9 kyr

Abstract

Analogous to West- and North Africa, East Africa experienced more humid conditions between approximately 12 and 5 kyr BP, relative to today. While timing and extension of wet phases in the North and West are well constrained, this is not the case for the East African Humid Period. Here we present a record of benthic foraminiferal assemblages and sediment elemental compositions of a sediment core from the East African continental slope, in order to provide insight into the regional shallow Indian Ocean paleoceanography and East African climate history of the last 40 kyr. During glacial times, the dominance of a benthic foraminiferal assemblage characterized by Bulimina aculeata, suggests enhanced surface productivity and sustained flux of organic carbon to the sea floor. During Heinrich Stadial 1 (H1), the Nuttallides rugosus Assemblage indicates oligotrophic bottom water conditions and therefore implies a stronger flow of southern-sourced AAIW to the study site. During the East African Humid Period, the Saidovina karreriana Assemblage in combination with sedimentary C/N and Fe/Ca ratios suggest higher river runoff to the Indian Ocean, and hence more humid conditions in East Africa. Between 8.5 and 8.1 kyr, contemporaneous to the globally documented 8.2 kyr Event, a severe reduction in river deposits implies more arid conditions on the continent. Comparison of our marine data with terrestrial studies suggests that additional moisture from the Atlantic Ocean, delivered by an eastward migration of the Congo Air Boundary during that time period, could have contributed to East African rainfall. Since approximately 9 kyr, the gaining influence of the Millettiana millettii Assemblage indicates a redevelopment of the East African fringe reefs.

Introduction

The tropical Indian Ocean is a key region for late Pleistocene and Holocene paleoclimatic reconstructions, as it is governed not only by changes in strength of the Asian Monsoon, but also by asymmetric climatic variability of both hemispheres. Whereas Indian Ocean marine records south of 12°N document a link to the Southern Hemisphere (Jung et al., 2009, Naidu and Govil, 2010, Saraswat et al., 2013), continental records north of 20°S show Northern Hemisphere-controlled climate variability (Abell and Plug, 2000, Chase et al., 2011, Ivory et al., 2012, Truc et al., 2013). During the Glacial to Holocene transition, the African continent experienced fundamental hydrological changes with more humid conditions between ~ 12 and 5 kyr BP, relative to today. The African Humid Period led to fundamental ecological changes, such as the greening of the Saharan desert and subsequent expansion of Neolithic civilizations. Whereas millennial-scale changes of the West- and North African monsoon and fluctuations of the tropical rain belt are well understood (Collins et al., 2011, deMenocal et al., 2000, Zarriess et al., 2011, Zarriess and Mackensen, 2010), this is not the case for East African monsoon and rainfall variability. Especially the so-called “East African Humid Period” between 11 and 5 kyr (Tierney et al., 2011a) is a topic of current debate. On the continent, lake level studies suggest pluvial periods during the earliest Holocene (Garcin et al., 2009, Garcin et al., 2012, Gasse, 2000, Junginger et al., 2013, Junginger and Trauth, 2013), and geochemical data point to a complex moisture source history of East African rainfall (Costa et al., 2014, Konecky et al., 2011, Tierney et al., 2011b). In this study, we aim to reconstruct the runoff history of the equatorial East African continent into the Indian Ocean from marine sediments recording the environmental changes during the Glacial to Holocene transition. Therefore, we investigated the benthic foraminiferal assemblages as well as sediment elemental composition of a sediment core from the continental borderland at about 7°S off Tanzania, East Africa.

The last Glacial to Holocene evolution of the benthic foraminiferal faunal composition on the continental shelf and upper slope off East Africa has not been studied before. Earlier studies of benthic foraminiferal assemblages and their distribution in the Indian Ocean focused on the deep southeast Indian Ocean (Corliss, 1979, Corliss, 1983), the Northern Indian Ocean (Gupta and Srinivasan, 1990) including the Red Sea, Arabian Sea and Bay of Bengal, and the Somali Basin (Gupta, 1994, Gupta, 1997, Gupta and Satapathy, 2000, Gupta and Thomas, 1999), closest to our study site. In a recent study on the benthic foraminiferal fauna, De and Gupta (2010) cover a large part of the central Indian Ocean. While most studies concentrate on deep-water environments, Murgese and De Deckker, 2005, Murgese and De Deckker, 2007, Murgese et al. (2008), Gupta et al. (2011), Singh et al. (2012), and Sarkar and Gupta (2013) provide helpful insights into the Indian Ocean intermediate depths.

The study site is located in the western Indian Ocean on the upper continental slope off Tanzania, East Africa, at 07°08.30′S, 39°50.45′E in 446 m water depth (Fig. 1). The East African coast is framed by an almost continuous string of coral reefs, either fringing the coast and islands, or forming barrier or platform reefs offshore from Somalia to South Africa. Major rivers, such as the Rufiji, Limpopo and Zambezi, interrupt the continuous coral reefs alongside Kenya to Mozambique. These rivers deliver large amounts of sediments and nutrients, impeding coral reef growth (Arthurton, 2004) and providing ideal conditions for mangrove forests. The narrow continental shelf off Tanzania exhibits larger fringe reef structures around Pemba and Mafia islands, and patchy structures along the mainland (Obura et al., 2000). The Rufiji River is located close to the study site and its catchment comprises an area of 177.43 km². The Rufiji River delta is partially covered by one of Africa's largest mangrove forest of about 500 km² (Duvail and Hamerlynck, 2007). The region receives seasonal rainfall due to the passage of the Intertropical Convergence Zone (ITCZ) twice a year; the “long rains” during March–May and the “short rains” during October–December.

The ocean surface waters at 7°S are fed by the steady westward flow of the South Equatorial Current (SEC), splitting into the East African Coastal Current flowing northwards, and the Mozambique Current flowing southwards, when hitting the East African coast. The SEC marks the boundary between seasonal reversing surface circulation resulting from the monsoon cycle north of the equator, and perennial surface circulation without seasonal reversals south of the equator. For this reason, surface waters at 7°S do not experience monsoon-induced upwelling to any great extent, as seen off the Kenyan and Somalian coasts. Instead, surface waters are stratified year-round and characterized by low nutrient conditions (Birch et al., 2013, McClanahan, 1988). Well-oxygenized Subantarctic Mode Water (SAMW) and Antarctic Intermediate Water (AAIW) from the southeast and southwest Indian Ocean contribute largely to the thermocline and intermediate waters at 7°S (Fine et al., 2008) and bathe the core site today. High salinity waters formed from Red Sea outflow (Red Sea Water, RSW) spread southwards along the East African continental margin (You, 1998), which can be identified at roughly 600–1400 m water depth just below our study site (Birch et al., 2013; Fig. 2). As there is convincing evidence that Red Sea outflow was greatly reduced during the Last Glacial Maximum (LGM) due to lower sea-level (Rohling and Zachariasse, 1996), and the RSW settled deeper in the water column during the late deglaciation and early Holocene (Jung et al., 2001), we conclude that the RSW did not affect our study site during the last 40 kyr.

In the past three decades, based on the analyses of live (Rose Bengal stained) faunas, it has been emphasized that deep-sea benthic foraminiferal composition is controlled by a set of general environmental parameters, such as food supply and the depositional regime including sediment grain size and bottom current velocity, as well as physical and chemical properties of deep water masses including ventilation and carbonate corrosiveness (Fariduddin and Loubere, 1997, Gooday, 1988, Gooday, 2003, Loubere, 1991, Lutze and Coulbourn, 1984, Mackensen et al., 1985, Mackensen et al., 1995, Murray, 2001, Schmiedl and Mackensen, 1997, Van der Zwaan et al., 1999). It was also recognized that species prefer different living depths within the sediment and are categorized as epifaunal (living at the sediment surface), shallow infaunal (0–2 cm) and deep infaunal (> 4 cm) (Corliss, 1985, Corliss and Chen, 1988, Jorissen et al., 1992, Mackensen and Douglas, 1989). The balance between organic flux and therefore the availability of food and pore-water oxygen content are thought to control the microhabitat occupancy (Jorissen et al., 1995). Further, the species microhabitat is not static (Linke and Lutze, 1993, Mackensen et al., 2000); specimens move within the sediment in response to varying food supply and changes in oxygen concentrations within the sediment pore waters (Heinz et al., 2002, Moodley et al., 1998).

In paleoceanography the accumulation rate of total organic carbon (TOC) was often used as a productivity proxy (Knies et al., 1998, Mackensen et al., 1994, Müller and Suess, 1979, Sarnthein et al., 1988). Similarly, the benthic foraminiferal accumulation rate (BFAR) was used for ocean surface paleoproductivity reconstructions in open ocean deep-sea sediments (Herguera and Berger, 1991). The validity of such down-core reconstructions strongly depends on a reliable calibration with sediment-trap and sediment surface data, and the amount of lateral organic matter supply close to continental margins (Schmiedl and Mackensen, 1997). However, since refractory organic carbon is almost inert biologically, BFAR is supposed to dominantly reflect the labile particulate organic matter supply, whereas TOC fluxes include resistant and refractory portions (Loubere and Fariduddin, 2003). So in environments with substantial terrestrial organic carbon input, the reconstruction of ocean surface primary productivity might be biased. To differentiate between refractory and labile organic matter, hence to quantify the amount of terrigenous input, Guichard et al. (1999) compared BFAR and TOC flux rates to gain insights into the quality and thus the origin of organic matter. Given that sedimentary organic matter flux is positively correlated to BFAR, varying input of terrigenous material will affect this relationship.