For centuries, a network of overland trade routes enabled cultural contacts between the Far East and the Mediterranean. Essentially the same route, which is known as the Silk Road, has been in use since the second millennium BCE, for trade in horses, precious stones, art pieces and the spice trade. The name of Silk Road, however, has stuck because of the lucrative silk trade with China around the start of the Christian era.

While geographical features explain the towns and cities that have arisen along the route, the disposition of passages through the mountainous stretches that separate urban centres, and which constitute the Silk Road, has been a subject of scholarly enquiry. Michael D Frachetti, C Evan Smith, Cynthia M Traub and Tim Williams, from Washington University, St Louis and University College, London describe in the journal, Nature, a simulation using satellite imaging, archaeology and computer analysis, which throws up most of the actual routes that have been used for human movement. 

The paper says that efforts to derive the origins of the network of inter-connections have generally sought “least effort” or “least cost” paths that “connect the dots” and link known habitation sites. In the present study, Frachetti and his colleagues reverse the approach, to discover pasture-driven travel paths within rugged terrain, as the setting for the growth of larger passages, rather than “ease of travel” between places that the passages connect.

Ease of travel being the motivation behind the evolution of movement pathways, has been the basis for many computer algorithms to create maps from geographical survey data or for planning new road or rail connections. The contour map that details the distribution of elevation over a region can be, for instance, the input to the Geographic Information System software, to throw up the path of water flow, leading to tributaries coming together and forming rivers. 
For a river flow application, the software would consider that tendency of water to flow under gravity as the driver. The software could, however, have other applications, with other motivating features, like seeking the shortest path or the quickest path, the least expensive path, the most profitable path, et al.

With respect to the emergence of the Silk Road, the authors say that the last 50 years’ data of the adaptation of nomadic people in the higher altitude regions of Asia suggests that “ease of travel” was not the factor that dictated mobility across mountains. “Variables that are fundamental to highland nomadic herding strategies, such as seasonal pasture quality and variation in annual mobility patterns, have not been reliably tested in relation to Silk Road geography,” they say. 

The team hence used the same GIS modelling tool but did not use the known Silk Road locations as the input, which had been used earlier, to look for factors that connect them. 

Instead, they used features like richness of pasture, which were relevant to nomadic people, and charted out paths of seasonal accumulation of herds of livestock, like little streams of water coming together to form larger flows.

The authors cite studies that show how the geography of Central Asia influenced the evolution of economies and human interaction in the region.  Archaeological studies of charred remains of plants, which point to livestock diet, for instance, indicate that the earliest farming societies exploited the plains and foothills in Turkmenistan till 3000 BCE. Desert land again restricted habitation to river valleys and drove agriculture and animal husbandry towards rain-fed highlands.

Herdsmen who roamed the higher reaches of the ranges in inner Asia are recognised as important agents of exchange between trade centres across Asia but the role of this mobility in shaping social centres and their communications has not been studied, the paper says. 

The team carried out computer studies of the flows of domesticated herd animals in an area that lies between elevations of 750 m and 4,500 m, within a broader area in highland inner Asia (see map). Modelling of population flows has been successfully used to simulate the seasonal wandering of Bronze Age herders between foothills and high altitude pastures in Kazakhstan, with the help of archaeological and environmental data, the paper says. 

The seasonal herding patterns in the study area were hence computer simulated, over 500 iterations, to derive reasonably persistent pathways that the herders used. And to assess whether these nomadic routes affected the connectivity at a continental scale, these derived pathways were compared with 258 historical Silk Road sites, within the area studied, to see if there was a correlation.

As the location of these sites had not been fed into the system when the pathways were generated, they embody an independent check to verify if the herder flow patterns compare with the actual, high altitude Silk Road, the paper says.

The results of the studies were first that the seasonal movements of sparse herding groups result in a network of interconnected paths that covers the area. When this modelling is iterated 500 times, which represents evolution over twenty generations, the network takes the form of nearly-continuous pathways. And along these pathways are found to lie, over 74 per cent of the historical Silk Road sites present in the region.

Finding a good proportion of known Silk Road sites on the map of pathways, generated based on pasture-driven herding groups, supports the idea that the longer routes through mountainous regions more likely grew along short-distance stretches of herding paths. Short-distance paths being rooted in the knowledge of local conditions and the availability of resources, food and water, would make for tree-like interconnection of the region and the complex of pathways that makes up the Silk Road.

The model, as opposed to routes between known sites based on “ease of travel” and “connecting the dots”, would also chart out a path along which to seek new, undiscovered Silk Road sites and other forms of Silk Road research and exploration, the paper says. 

The writer can be contacted at response@simplescience.in.