Wet phases in the Sahara/Sahel region and human migration patterns in North Africa

Abstract

The carbon isotopic composition of individual plant leaf waxes (a proxy for C(3) vs. C(4) vegetation) in a marine sediment core collected from beneath the plume of Sahara-derived dust in northwest Africa reveals three periods during the past 192,000 years when the central Sahara/Sahel contained C(3) plants (likely trees), indicating substantially wetter conditions than at present. Our data suggest that variability in the strength of Atlantic meridional overturning circulation (AMOC) is a main control on vegetation distribution in central North Africa, and we note expansions of C(3) vegetation during the African Humid Period (early Holocene) and within Marine Isotope Stage (MIS) 3 ( approximately 50-45 ka) and MIS 5 ( approximately 120-110 ka). The wet periods within MIS 3 and 5 coincide with major human migration events out of sub-Saharan Africa. Our results thus suggest that changes in AMOC influenced North African climate and, at times, contributed to amenable conditions in the central Sahara/Sahel, allowing humans to cross this otherwise inhospitable region.

Fig. 1.

Location of gravity core GeoB9528-3 offshore Guinea and modern vegetation zones of Northwest Africa (34). From north to south the main vegetation zones are Mediterranean (MED) (C₃ dominated), Mediterranean–Saharan transitional (MST) (mixed C₃ and C₄ plants), Sahara desert (C₄ dominated), Sahel-grass savanna (mixed C₃ and C₄ plants), and tropical rainforest (C₃ dominated). Three main wind systems influence the region of Northwest Africa, the Northeast trade winds (NETW), the Southeast trade winds (SETW), and the AEJ (also known as the Saharan Air Layer), and transport dust and plant leaf waxes from the African continent to the Atlantic. The NETW and the SETW converge at the Intertropical Convergence Zone (ITCZ; the meteorological equator). (A) During Northern Hemisphere summer, when the ITCZ is located at its most northerly position (surface expression at ≈20°N), the SETW are strongest and dust sourced in the Sahara and the Sahel is raised by easterly winds into the midaltitude flow (≈3 km) of the AEJ and transported beyond the continental margin between 10°N and 25°N. The AEJ is strongest during Northern Hemisphere summer when it is located at 10–12°N (36). (B) In Northern Hemisphere winter, the ITCZ migrates southwards (≈5°N) and the NETW are dominant. The AEJ is weaker in Northern Hemisphere winter and is located at ≈0–5°N (36). At this time, dust in the southern Sahara (the alluvial plains of Niger, Chad, and Faya Largeau) is uplifted by the low-altitude (500–1,500 m) NETW and deposited along wide areas off Africa between 2°N and 15°N (11), with the main axis of the dust plume located at ≈5°N. The Gulf of Guinea receives material sourced in the southern Sahara during Northern Hemisphere winter, which is transported over long distances and deposited offshore.

Fig. 2.

Geochemical records from GeoB9528-3. In traces c–e, the thick line represents the smoothed data (five-point running mean). In traces a, b, and d, the error bars represent the standard deviation of replicate analyses. All isotope data are reported in standard delta notation (‰) against the VPDB standard. The black bars at the top of the graph indicate documented human migrations out of Africa (23, 24). The vertical dashed lines indicate periods of major extinctions and turnovers of hominin populations at ≈75 and ≈45 ka (34). Wet intervals during the AHP, within MIS 3 and during MIS 5 are indicated by green shading. Arid intervals during MIS 6 and MIS 4 are indicated by yellow shading. The bar at the bottom indicates MIS 1–7. Trace a shows carbon isotope (δ¹³C) values of the C₂₉ n-alkane. On the right side of the graph, the estimated %C₄ plants is shown for the C₂₉ n-alkane, based on a binary mixing model assuming end-member values of −34.7‰ and −21.4‰ for C₃ and C₄ vegetation, respectively. On the left side of the graph, δ¹³C values of C₂₉ n-alkanes in surface sediments collected in the vicinity of GeoB9528-3 are shown. The red circle represents a δ¹³C value of −27.5‰ for site GIK16757-1 (8°58.60 N, 16°56.48 W) and the red triangle represents a δ¹³C value of −27‰ found at sites GIK16405-1 (12°25.37 N, 21°25.37 W) and GIK16408-2 (9°47.88 N, 21° 27.24W) (11). Trace b shows δ¹³C values of the C₃₁ n-alkane. On the right side of the graph, the estimated %C₄ plants is shown for the C₃₁ n-alkane, based on a binary mixing model assuming end-member values of −35.2‰ and −21.7‰ C₃ and C₄ vegetation, respectively. Trace c shows the δ¹³C of the benthic foraminifer C. wuellerstorfi. The precision of these measurements is ± 0.05‰ based on replicates of an internal limestone standard. Trace d shows alkenone (U₃₇^k') SST reconstruction for GeoB9528-3 (see SI Text for methods). Trace e shows the δ¹⁸O of the benthic foraminifer C. wuellerstorfi. The precision of these measurements is ± 0.07‰ based on replicates of an internal limestone standard. Trace f shows June insolation at 10°N (14).