Windows out of Africa: A 300,000-year chronology of climatically plausible human contact with Eurasia

Abstract

Whilst an African origin for Anatomically Modern Humans is well established 1 , the timings of their expansions into Eurasia are the subject to heated debate, due to the scarcity of fossils and the lack of suitably old ancient DNA 2 . Here, we estimate potential timings and routes out of Africa by deriving anthropologically and ecologically plausible precipitation requirements for human existence, and applying them to high-resolution palaeoclimate reconstructions for the past 300k years. We find that exit routes and timings previously suggested based on archaeological and genetic evidence coincide precisely with the presence of sufficiently wet corridors into Eurasia, while the gaps between the proposed exit timings co-occur with periods of insufficient rainfall. This demonstrates the key role that palaeoclimatic conditions played for out-of-Africa expansions. The challenging environmental conditions outside of Africa that occurred between windows of potential contact, coupled with the lack of a demographic rescue effect from migration and possible competition with other hominins, likely explain the demise of early colonists prior to the large-scale colonisation of the world beginning from ~65kya.

Population inter-connectivity over the past 120,000 years explains distribution and diversity of Central African hunter-gatherers

“…ENM performance in the hold-out data was extremely high (TSS=0.55 and AUC=0.87), correctly classifying presences and absences 81% and 74% of the time respectively. The main ecological factor rendering particular areas unsuitable for hunter-gatherer presence was precipitation seasonality ( BIO15 ) followed by the annual temperature range ( BIO7 )(Table S6), confirming what has been proposed for other African regions (11,12). Including predictors relating to Bantu populations ( Rural population density or Distance to populated places ) neither improved model fit (TSS=0.51 and AUC=0.85 for both models) nor significantly altered the relative contribution of the other variables.…”

“…Thus, the archaeological record of the region is limited and biased towards non-forest areas. Therefore, rather than to delimit the niche of CAHGs in the past (11), we use the archaeological record to test our model’s ability to predict the location and dating of known archaeological sites occupied by ancient hunter-gatherers (32). We compiled all published 14 C dates from 170 archaeological sites from the Middle Stone Age onwards (Data S2-S3; Fig.S6; Fig.…”

An empirical evaluation of bias correction methods for palaeoclimate simulations

“…Realistic reconstructions of global palaeoclimate are a key input for modelling many important long-term and largescale ecological processes (Eriksson et al, 2012;Timmermann and Friedrich, 2016;Leonardi et al, 2018;Zhu et al, 2018;Rangel et al, 2018;Beyer et al, 2020). In many of these applications, climatological normals with variables such as temperature and precipitation at quasi-equilibrium at different points in time represent the most relevant climatic inputs.…”

“…In many of these applications, climatological normals with variables such as temperature and precipitation at quasi-equilibrium at different points in time represent the most relevant climatic inputs. Simulations of these variables remain subject to substantial biases when compared to observational data despite advancements in how complex physical processes are represented in global climate models (Solomon et al, 2007;Ehret et al, 2012). Depending on the region of interest, these biases can be of the order of several degrees of temperature and tens of percents of precipitation, which can make the difference between markedly different vegetation types (Kottek et al, 2006).…”

“…Code and data availability. Code and datasets used in this analysis are available on the Open Science Framework: https://doi.org/10.17605/OSF.IO/8AXW9 (Beyer, 2020).…”