A climatic evaluation of the southern dispersal route during MIS 5e.

Samuel L. Nicholson^1,2, Rob Hosfield², Huw S. Groucutt^3,4,5, Alistair W. G. Pike⁶, Stephen J. Burns⁷, Albert Matter⁸, and Dominik Fleitmann^2,9

¹ Ecology and Evolutionary Biology, University of Reading, Reading, RG6 6LA, United Kingdom.

² Department of Archaeology, University of Reading, Reading, RG6 6AB, United Kingdom.

³ Extreme Events Research Group, Max Planck Institute for Chemical Ecology, 07745 Jena, Jena, Germany.

⁴ Department of Archaeology, Max Plank Institutes for Chemical Ecology, the Science of Human History, and Biogeochemsitry, 07745 Jena, Germany.

⁵ Institute of Prehistoric Archaeology, University of Cologne, Cologne, 5093, Germany.

⁶ Department of Archaeology, University of Southampton, Southampton, SO17 1BF, United Kingdom.

⁷ Department of Geosciences, University of Massachusetts, MA 01003-9297, United States of America.

⁸ Institute of Geophysics, University of Bern, CH-3012 Bern, Switzerland.

⁹ Department of Environmental Sciences, University of Basel, CH-4056 Basel, Switzerland.

Corresponding authors: Samuel L. Nicholson (sam.nicholson@reaeding.ac.uk)

Dominik Fleitmann (dominik.fleitmann@unibas.ch)

Highlights:

Abstract

Homo sapiens dispersals out of Africa are often linked to intensifications of the African Summer Monsoon and Indian Summer Monsoon. Current dispersal models advocate that dispersals along the “southern-route” into Arabia occurred during Glacial Termination-II (T-II), when reduced sea-level and Bab-al-Mandab width increased the likelihood of crossing. The precise phasing between sea-level and monsoon precipitation is thus key to assess the likelihood of a successful crossing or the behavioural and technological capacities that facilitated crossing. Based on a precisely-dated stalagmite record from Yemen we reveal a distinct phase-lag of several thousand years between sea-level rise and monsoon intensification. Pluvial conditions in Southern Arabia during MIS 5e lasted from ~127.7 to ~121.1 ka BP and occurred when sea-levels were already higher than at present. Based on our observations, we propose three models for the dispersal of H. sapiens which all have pertinent implications for our understanding of human technological and behavioural capacities during MIS 5e.

1 Introduction

Understanding how H. sapiens spread from Africa across the world is one of the most debated topics in human evolution (Mellars et al., 2013; Groucutt et al., 2015a; Bae et al., 2017). Two proposed main dispersal routes cross Arabia: a northern-route across the Sinai into the Levant and a southern-route from the Horn of Africa via the Strait of Bab-al-Mandab into Southern Arabia and beyond (Fig. 1). The accessibility of these entry points was spatiotemporally variable and related to major climatic changes across the Saharo-Arabian deserts. During interglacial periods, both the African and Indian Summer Monsoons (ASM and ISM, respectively) were much stronger, expanded northward and transformed the Saharo-Arabian deserts into green landscapes for a few millennia (Fleitmann et al., 2003b; Parton et al., 2015; Petraglia et al., 2015; Tierney et al., 2017; Nicholson et al., 2020). These pluvial periods, termed “Green Arabia Periods” and “South Arabian Humid Periods” (SAHPs) respectively, provided optimal periods for H. sapiens to disperse from sub-Saharan Africa into Eurasia (Fleitmann et al., 2011; Rosenberg et al., 2011; Larrasoaña et al., 2013; Nicholson et al., 2020). Over the last 130 ka BP, pluvial conditions in Southern Arabia with rainfall of more than 300 mm yr^-1 occurred during Marine Isotope Stages (MIS) 5 and 1, and lasted from ~128-121 ka BP (MIS 5e; SAHP 4), ~104-97 ka BP (MIS 5c; SAHP 3) and ~84-71 ka BP (MIS 5a; SAHP 2) and ~10.5 to 6.2 ka BP (SAHP 1) (Fleitmann et al., 2011; Nicholson et al., 2020). In addition, there is also some evidence for a period of enhanced rainfall between approximately 60 and 50 ka BP (the onset of MIS 3) (McLaren et al., 2009; Parton et al., 2013, 2018), though the nature and timing of this period remains uncertain.

The southern dispersal route involves a maritime crossing of the Bab-al-Mandab Strait. However, its current width of approximately ~26 km represents a significant challenge to dispersal and was more likely traversable at times of lower sea-level, especially if sea-faring technologies were limited. One proposed timing for earlyH. sapiens dispersals is Glacial Termination-II (T-II), between 136-129 ka BP, when sea-levels, although rapidly rising, were lower than today and the width of the Bab-al-Mandab Strait (BaM_width) was reduced to a few kilometres (Armitage et al., 2011). From a palaeoclimatic perspective, a dispersal was most likely to have occurred at times of increased precipitation and biomass across Arabia. However, during T-II, several lines of evidence point to a phase-lag of several thousand years between sea-level rise and northward migration of the tropical rainbelt due to colder northern-hemisphere temperatures related to Heinrich Stadial (HS) 11 between 135 and 130 ka BP (Cheng et al., 2009; Böhm et al., 2015; Häuselmann et al., 2015; Marino et al., 2015). In other words, arid conditions may have prevailed in Arabia during T-II, forming a biogeographical barrier to widespread dispersals despite low sea-levels. Thus, establishing the precise phasing between sea-level change and ASM/ISM intensification during the MIS 6-5e transition from records close to the Bab-al-Mandab Strait could be one critical factor for understanding accessibility of the southern-dispersal route. ASM and ISM records with precise and accurate chronologies are an important prerequisite to reveal such a phasing. Here, we present a precisely-dated and highly-resolved speleothem-based climate record from Mukalla Cave in Yemen, covering MIS 5e (SAHP 4: Nicholson et al., 2020). Precise Uranium-series (²³⁰Th) ages allow us to evaluate the temporal phasing between ASM/ISM rainfall and sea-level change at a possible point of entry into southern Arabia.

2 Environmental Settings, materials and methods

Stalagmite Y99 was collected from Mukalla Cave (14°55’02”N; 48°35’23” E; ~1500 masl; Fig. 1) in southern Yemen, where climate is strongly governed by the ASM and ISM respectively. At present, both Mukalla Cave and the Bab-al-Mandab Strait are located at the northern and north-eastern margins of the ASM and ISM, with rainfall averaging <150 mm yr^-1 (Fleitmann et al., 2011). Stalagmite Y99 extends back to 1.1 million years and was deposited in 17 punctuated growth intervals identified through ²³⁰Th and Uranium-lead dating, with Growth Interval-I (GI-I) being the youngest and dated to MIS 5e (Nicholson et al., 2020). Previously analysis of Y99 was focussed on the broad timing and climatic conditions (bulked δ¹⁸O and δ¹³C isotope analysis) of SAHPs over the last 1.1 million-years (Fleitmann et al., 2011; Nicholson et al., 2020). Here, we provide a more high-resolution and focussed study of the timing of SAHP 4 compared to sea-level fluctuation. We used the StalAge algorithm to produce a robust age-model for SAHP 4 from previously collected ²³⁰Th ages. This was then used to provide δ¹⁸O_ca (ASM rainfall) and δ¹³C_ca records at <100 years resolution and can be accurately compared to sea-level records.

The width of the Bab-al-Mandab Strait (BaM_width) was reconstructed using bathymetry data and the Red Sea relative sea-level (RSL) curve. The RSL has been constructed using marine core δ¹⁸O_{G. ruber} records from the Red Sea (Siddall et al., 2003; Rohling et al., 2009), whereas the chronology of the RSL time-series is based on correlations with Mediterranean Core LC21 and the revised ²³⁰Th chronology of the Soreq Cave record for periods younger than 150 ka BP (Rohling et al., 2009; Grant et al., 2012, 2014). Using local sea-level records that exploit a basin isolation effect means that our assessment is unaffected by isostatic effects, allowing us to compare regional climates with sea-level variations that control the sill depth and the width of the Strait. We used the freely available QGIS software package and 15 arc-second (~450 m between 12-14^oN) interval elevation and bathymetry data (GEBCO Compilation Group, 2020) in combination with the RSL to estimate BaM_width over the last 150 kyrs (extended methods).

3 Results and discussion