Skip to main content

A Silk Road in the Palaeolithic: Reconstructing Late Pleistocene Hominin Dispersals and Adaptations in Central Asia

Periodic Reporting for period 2 - PALAEOSILKROAD (A Silk Road in the Palaeolithic: Reconstructing Late Pleistocene Hominin Dispersals and Adaptations in Central Asia)

Reporting period: 2018-12-01 to 2020-05-31

The Silk Road was an ancient network of trade routes that crossed central Asia, constituting the primary link between the Middle East and East Asia for at least a millennium during late antiquity and the early Middle Ages. It is frequently cited as one of the most important and long-lasting conduits for bio-cultural exchange in human history and continues to play a crucial role in Asian geopolitics today. To avoid crossing the deserts and high mountains of arid central Asia – one of the driest regions in the world – travelers used the more temperate corridors through the mountain foothills (piedmonts). Recent research has shown that the movements of pastoral nomads during the Bronze and Iron Ages also reflected a connectivity of these same pathways through the piedmonts. The PALAEOSILKROAD project investigates the hypothesis that at least some of these corridors were used earlier still, by hunter-gatherers in the Stone Age. The extremity of the most arid phases of the last glacial cycle (the late Pleistocene, 110-11 500 years ago) makes such a strategy very likely. Because we know from genomic studies that these dispersals are in large part responsible for the origins and makeup of modern Asian peoples, uncovering their archaeological traces is of utmost importance.

There is now evidence of multiple encounters between the (so far) three different metapopulations occupying the region during the late Pleistocene, namely the Neanderthals, Denisovans, and modern humans. Where exactly these encounters took place is not clear, but the amount of time spent in spatiotemporal overlap between them is large enough that the entire territory of central Asia constitutes a possibility. Moreover, all current models of dispersal through the region suffer from the low quality and density of the input data, due to the lack of chronostratigraphic control and the extrapolation of paleoenvironmental data from regional or even global models, rather than from locally-derived proxies.

PALAEOSILKROAD aims to discover new Paleolithic sites in the foothills of the Karatau, Tian Shan, Dzhungarian Alatau, and Southern (Kazakh) Altai mountains in Kazakhstan, together known as the Inner Asian Mountain Corridor (Objective 1); and to use these new data to examine how humans adapted to the climatic fluctuations of the last glacial cycle (ca. 110-11 500 years ago) (Objective 2). To this end, we combine novel remote-sensing techniques (both satellite and ground-based) with traditional geomorphically-informed foot-survey and excavations, as well as paleoenvironmental reconstruction. Through field observations, mapping, and chronostratigraphy, we provide a framework for understanding the co-variation of human occupation with climatic phases. We also investigate biases in the archaeological record and their causes. This will help focus future efforts to discover new archaeological sites and interpret spatial and temporal gaps in the archaeological record of this crucial region.
In this first period, we have completed the first work package (WP 1, site prediction) and the initial phase of WP 2 (site mapping). We began by acquiring a digital lithological dataset for Central Asia that enabled us to restrict the remote sensing model to areas containing carbonate deposits. This allowed us to use exclusively topographic parameters for the model, which we tested in the field in the summer of 2018. After returning from the field, the model was re-calibrated to further restrict the areas considered to be of top priority. We tested this second iteration during the 2019 field season and we improved our results considerably, reducing the total target area from several thousand km2 to approximately 34 km2, demonstrating the viability of the model. The results are described in a manuscript currently under preparation for PloS One.

To date, we have surveyed a very large area (total length of GPS tracks: 27000 km), and recorded 95 caves and rockshelters (85 unknown, of which 28 contained sediments), as well as 31 lithic scatters from 16 localities. We have carried out 14 test excavations, of which 11 were in cave contexts. In total, the majority of excavated in situ finds are probably from the Holocene, with only 2 confirmed Paleolithic occupations and 2 Pleistocene sterile contexts documented. However, the depth of sediments in 3 cave sites is greater than what we have been able to excavate so far, and it is possible they contain Paleolithic occupations. Based on our geomorphic mapping and micromorphological analyses, we conclude preliminarily that a combination of structural and climatic factors contributed to a relative dearth of preserved Pleistocene cave sites. These preliminary results are summarized in an article currently in review at Quaternary International.

The second part of WP 2 (excavations) and the majority of WP 3 will continue into the second half of the project, concluding with a synthesis and model of dispersal and adaptation of human populations in Central Asia during the last glacial cycle (WPs 4 and 5).
Progress beyond the state of the art

At the moment, we have made significant progress on the methodological front. We have also discovered the first Paleolithic sites in two regions of Kazakhstan: the Jungarian Alatau (in Altyn Emel and the Ketmen range) and the Tarbagatai range.

We have created and field-tested a predictive model for the location of caves based on topography, which enables researchers to conduct fieldwork on larger areas much faster and with better results than was previously possible. This work is currently in preparation and will be submitted to PLoS One in the spring of 2020.

We have begun testing various combinations of non-destructive imaging techniques at the site-scale to speed up sub-surface survey in caves, and this will be reported in a separate paper (Tiago Attorre).

Expected results until the end of the project

Based on our progress so far, we expect to excavate and date at least two caves in the Karatau range, at least three loess sites in the southeast, several open-air spring sites in Altyn Emel and at least one cave site in the eastern part of the Junggar Alatau.

Following our studies of landscape evolution, we have already begun to understand the basic patterns of visibility and preservation of archaeological sites and by the end of the project we will have a comprehensive synthesis of the yield and biases of different regions and deposits.
Likewise, we have begun to gather data for a map of the distribution of raw materials and an assessment of their mechanical qualities (PhD student Abay Namen), which will be ready by the end of the project.

All the results will be integrated into a database in the open PaleoCore format ( and used, together with results from other databases, for improving computer models of dispersal through the region and generating hypotheses for future work (to be undertaken by postdoc (PD) 2).

Finally, we are planning to hold a workshop in Almaty to bring researchers working on Central Asia together to discuss the state of the art in the last 5 years. The results will be published in a special issue of a peer-reviewed journal. Additionally, based on our current research on the geological and paleoclimatic factors affecting the preservation of Pleistocene cave deposits, we are planning a second workshop on that topic.