Climatic windows for human migration out of Africa in the past 300,000 years

The study addresses when and along which routes Homo sapiens could have left Africa during the last 300,000 years, a question complicated by sparse fossils and limited ancient DNA. While a major expansion around ~65 thousand years ago (ka) is supported by genetic splits and many fossils, earlier potential excursions (e.g., ~210 ka in Greece, ~194–100 ka in the Levant, ≥85 ka in Arabia) and signals of gene flow into Neanderthals suggest multiple earlier opportunities. The authors propose that palaeoclimate—specifically rainfall and aridity—set climatic “windows” for viable dispersal via northern (Nile–Sinai) and southern (Bab-el-Mandeb) routes. They aim to reconstruct the temporal availability of these windows using high-resolution climate reconstructions combined with empirically grounded human precipitation tolerance, and to evaluate consistency with archaeological and genetic records.

Prior work relied on sparse empirical climate records and limited time-slice models, yielding qualitative scenarios of Sahara–Arabia humidity and migration corridors. Quantitative dispersal models have been calibrated to archaeology or genetics but sometimes require implausible assumptions (e.g., extreme niche shifts or dramatic increases in coastal migration speeds). Archaeological evidence indicates episodic human presence in Arabia and the Levant during interglacials, and genetics points to a principal expansion ~65 ka with possible earlier movements and gene flow to Neanderthals (>130 ka, possibly up to ~250 ka). However, the archaeological record is uneven, and genetics only captures exits that left surviving lineages. This study instead grounds dispersal feasibility in climate-derived precipitation/aridity thresholds relevant to hunter-gatherer subsistence.

Climate reconstructions: The authors used a statistics-based emulator of the HadCM3 GCM to reconstruct global terrestrial climate for the last 300,000 years at 1-kyr intervals, then downscaled and bias-corrected to ~0.5° (~55 km hex grid) using a two-step delta method. Step 1 applied a dynamic (time-varying) multiplicative delta using high-resolution HadAM3H simulations (1.25°×0.83°) from 21 ka to present, selecting analogue times by proximity in global CO2 to capture fine-scale spatial heterogeneities. Step 2 applied a classical multiplicative delta to match observed 1960–1990 precipitation. Land–sea configurations were reconstructed using present-day elevation and a Red Sea sea-level history, ensuring appropriate coastlines for each time slice. Decadal variability: For each millennium, decadal (10-year) mean maps (n=100 per millennium) were generated by scaling 30-year climatologies with decadal-to-30-year ratios derived from HadCM3B-M2.1 annual simulations for the last 21 ka. Temperature was reconstructed analogously (additive dynamic and classical deltas) to compute Köppen aridity index A = P/(T + 33). Routes and connectivity: Two dispersal routes were analyzed—northern (Nile–Sinai land bridge) and southern (Bab-el-Mandeb). The Nile delta was assumed crossable irrespective of local precipitation/aridity. For the southern route, the strait was assumed in principle crossable; sea level set the minimum necessary continuous water crossing distance (4–>20 km over time). For each decade in each millennium, and for a given precipitation threshold p, grid cells with precipitation ≥ p were marked habitable. A pathfinding approach on the hex grid tested for a connected path from northeast Africa (start at 32.6°E, 10.2°N) to outside Africa (λ > 65°E or φ > 37°N). A bisection method determined the minimal precipitation threshold that still allowed an exit path in that decade. An analogous analysis used the Köppen aridity index. Human tolerance threshold: Contemporary hunter-gatherer distribution data (excluding groups adjacent to permanent freshwater) were analyzed to infer a minimum precipitation tolerance, identifying a threshold ~90 mm/year below which no hunter-gatherers are recorded. This aligns with ecological constraints (minimum rainfall for grazer biomass and desert–xeric shrubland transition). An analogous aridity tolerance around 1.7 was inferred. Windows of opportunity were defined where the reconstructed climate exceeded these tolerances sufficiently to allow connectivity. Validation: Simulations were compared qualitatively to long-term humidity and temperature proxies (e.g., Dead Sea lake levels, leaf wax δD from Arabian margin, NW African humidity index; δ18O speleothem records in Israel), and quantitatively to pollen-based reconstructions at 6 ka and 21 ka. Time series at dated Arabian palaeolake sites were also compared, generally showing higher rainfall during the last interglacial than at most subsequent times.

- Precipitation tolerance required for exits fluctuated with palaeoclimate; wetter corridors periodically connected Africa and Eurasia along both northern and southern routes. - Empirical hunter-gatherer threshold: ~90 mm/year precipitation (below which no modern hunter-gatherers are recorded), consistent with ecological minima for grazer biomass and vegetation transitions; analogous aridity threshold ~1.7. - Northern route windows (assuming 90 mm/y tolerance): Several intervals between ~246–200 ka; a reopening near ~130 ka with intermittent feasibility until ~96 ka; additional opportunities around ~78 ka and ~67 ka; route largely closed thereafter until the wet Holocene. - Southern route windows (assuming crossing of Bab-el-Mandeb feasible): Extended favorable intervals before the last interglacial at ~275–242 ka, ~230–221 ka, and ~182–145 ka; during ~135–115 ka (last interglacial) with generally high sea levels except near ~135 ka; a long favorable window ~65–30 ka; additional connections just after the Last Glacial Maximum and during the mid-Holocene, consistent with Eurasian backflow into Africa. - Sea-level control on southern route: Minimum continuous water crossing distance across Bab-el-Mandeb varied from ~4 km (low sea level) to >20 km (high sea level), modulating crossing difficulty. - Sensitivity to tolerance: For northern route, tolerances >110 mm/y only allow shorter-than-millennial favorable intervals; >130 mm/y renders exits very challenging and restricted to unusually wet decades. Southern route is more permissive: up to 200 mm/y feasible during last interglacial; up to ~130 mm/y feasible during favorable intervals between ~65–55 ka. - Consistency with evidence: Climatic windows align with archaeological finds indicating early excursions into Eurasia and with genetic signals, including Homo sapiens to Neanderthal gene flow (>130 ka) and a major expansion ~65 ka. - Mechanisms for early failures: Frequent interruptions between windows likely isolated early colonists on the Arabian Peninsula, making them vulnerable to local extinction without continued demographic influx; additional barriers include Taurus–Zagros topography and competition with Neanderthals (north) and possibly Denisovans (east). The prolonged wet window ~65–30 ka, coupled with shorter crossing distances (~4 km) at times of low sea level, likely facilitated sustained demographic influx and successful global dispersal.

The findings indicate that rainfall and aridity set temporal windows in which dispersal corridors across North Africa and into Southwest Asia were viable, constraining when Homo sapiens could realistically leave Africa. By anchoring feasibility to an empirically derived human precipitation tolerance (~90 mm/y) and identifying connected paths, the study demonstrates that major proposed dispersal episodes—both earlier excursions and the main ~65 ka expansion—coincide with climatically permissive periods. This supports palaeoclimate as a key driver or gatekeeper of human expansions. The results also explain why earlier waves may have failed to establish persistent populations beyond Africa: periodic arid interruptions curtailed connectivity and demographic reinforcement, while topographic barriers and competition with other hominins further limited spread. The extended favorable window along the southern route from ~65–30 ka provided conditions for sustained demographic influx and expansion, consistent with genetic and archaeological signals. The study refines debates on the southern route by quantifying the role of sea level in crossing difficulty, while noting that seafaring capability remains unresolved. Overall, the work integrates high-resolution climate reconstructions with ecological-human tolerances to illuminate the timing and geography of out-of-Africa dispersal opportunities.

This study reconstructs high-resolution, decadal-scale climatic feasibility for Homo sapiens dispersals out of Africa over the last 300,000 years, identifying episodic windows of suitable rainfall and aridity along northern and southern routes. Using an empirically grounded precipitation tolerance (~90 mm/y), the inferred windows closely align with archaeological and genetic evidence for both early excursions and the major ~65 ka expansion. The prolonged favorable period ~65–30 ka, especially via the southern route when sea levels reduced crossing distances, likely facilitated the successful, large-scale dispersal and subsequent colonization of Eurasia. The approach highlights palaeoclimate as a primary constraint on dispersal timing and pathways. Future work should integrate these climatic windows with explicit demographic dispersal models, refine human ecological tolerances for arid regions, better constrain sea-level impacts on crossing feasibility, and synthesize archaeological and genetic data to resolve the roles of route choices and interactions with other hominins.

- Human tolerance proxies: Thresholds derived from modern hunter-gatherers may not perfectly reflect early modern human tolerances, given different technologies (e.g., water storage/transport), subsistence, and regional hydrological contexts. - Spatial sampling bias: Ethnographic hunter-gatherer data are unevenly distributed geographically and climatically, potentially biasing inferred thresholds. - Climate reconstruction assumptions: Delta methods assume present-day correction structures apply to the past (partially mitigated with dynamic delta); uncertainties in emulator extrapolation beyond calibrated periods and in bias corrections remain. - Temporal resolution: Feasibility assessed at decadal scales; shorter-term variability (storms, dry spells) could significantly affect water availability and mobility. - Hydrology simplifications: Local water sources (lakes, rivers, springs) and microrefugia were not explicitly modeled; Nile delta assumed crossable at all times; this may misestimate local feasibility. - Sea crossing uncertainty: Bab-el-Mandeb crossing was assumed in principle possible; actual seafaring capabilities and usage are not established archaeologically. - Validation limits: Proxy records are sparse and indirect; comparisons were qualitative for long records and limited to two time slices for pollen-based reconstructions. - No explicit demography: The study infers climatic feasibility but does not simulate population dynamics, dispersal speeds, or interactions with other hominins, which affect realized migrations.