Historical Geographic Data of the Silk Road Zone II Versions EN1 Vol 3 (3) 2018
A dataset of traffic routes of the east-central Silk Road in the Tang dynasty (618 – 907 A.D.)
: 2018 - 05 - 13
: 2018 - 06 - 05
: 2018 - 08 - 20
485 4 0
Abstract & Keywords
Abstract: Traffic route is an important component of the Silk Road studies. Historically, traffic routes of the Tang dynasty importantly laid the framework of the Silk Road. Based on historical documents and archaeological achievements, we used historical geographical methods and GIS technologies to restore the traffic routes of the east-central section, through which to build a dataset of traffic routes of the east-central Silk Road in the Tang dynasty (618 – 907 A.D.). This dataset not only presents an overall characterization of the east-central section of the Silk Road, but also offers year-specific traffic information in this particular period. It provides significant and fundamental data for scientific research on the Silk Road of the Tang dynasty.
Keywords: Silk Road; traffic route; Tang dynasty
Dataset Profile
Chinese title唐代丝绸之路东中段交通线路数据集(618–907 年)
English titleA dataset of traffic routes of the east-central Silk Road in the Tang dynasty (618 – 907 A.D.)
Corresponding authorZhang Ping (zhangping029@126.com)
Data authorsXu Xueqiang, Zhang Ping
Time range618 – 907 A.D.
Geographical scopeShaanxi, Gansu, Qinghai, Ningxia, Xinjiang in China, Mongolia, and Kazakhstan
Data volume7.45 MB
Data format*.shp, *.dbf, *.prj, *.sbn, *.sbx, *.xml, *.shx
Data service system<http://www.sciencedb.cn/dataSet/handle/587>
Sources of fundingMajor projects of the National Social Science Fund (14ZDB031)
Dataset compositionThis dataset is composed of two shapefiles. One stores the point data of the Silk Road and the other stores the line data of the Silk Road. Among the files, *.shp file is used to store the geometry of geographic features; *.shx file is used to store the index of graphic features and attribute information; *.dbf file is used to store the dBase table of element attributes; *.prj file is used to store spatial reference information; *.sbn and *.sbx are spatial index files of geometry; *.xml file contains metadata in XML format.
1.   Introduction
Since Qin and Han dynasties, there have been political, economic and cultural exchanges between China and western countries. Land transportation was the primary way of such exchanges in an era when modern navigation technology was yet to be mature. Starting from Xi’an, the traffic road passed through central and western Asia, and terminated in Europe, shaping an important passage for east-west exchange. In the late 19th century, German geologist Richthofen named this road as the "Silk Road".
Scholars have divided the Silk Road into three sections: the eastern section from Chang 'an to Yangguan and Yumenguan, the central section from Yumenguan and Yangguan to Pamir, and the western section from Pamir to Europe. Previous efforts to restore the Silk Road traffic routes largely relied on extensive research of historical literature, through which general traffic directions were outlined according to the traffic points identified, sometimes supplemented by schematic expressions.1–3 With the development of remote sensing technology and GIS technology in recent years, as well as the increasing accessibility of ancient maps and archaeological findings, scholars were able to carry out visualization or thematic research on the Silk Road of each dynastic periods by using multi-source data.4–7 Some proposed to use GIS technology as a new method to restore the 2,000-year-old Silk Road traffic routes.8 Indeed, restoration of the Silk Road by using new technologies makes the spatio-temporal information more accurate and facilitates a comprehensive study of the involved regions .9 As no such data for the Silk Road have been published so far, this study takes the eastern and central traffic routes of the Silk Road of Tang dynasty (618 – 907) as research object (these traffic routes are collectively referred to as “the east-central Silk Road” hereinafter). Based on historical documents and archaeological achievements, we used historical geographical method and Geographic Information System (GIS) technology to build a year-scale dataset, in contradistinction to previous studies which used dynasty-scale and hence less accurate data for restoration. This study aims to reflect the features of the Silk Road of Tang dynasty as precisely as possible, and provides data for further academic research on the Silk Road.
2.   Data collection and processing
2.1   Data sources
This dataset includes data of traffic points and traffic lines of the east-central Silk Road in the Tang dynasty. Raw data were drawn from modern remote sensing images, traffic research results of Tang dynasty, cultural relics and census data, provincial maps of modern China, and historical maps. The data are composed of vector graphics and property sheets. Specifically, the sources can be divided into the following two categories.
(1) Data of traffic routes in the Tang dynasty
By extensively citing existing literature, Yan Gengwang restored the traffic routes nationwide in the Tang dynasty through schematic diagram. In the work, Yan used ancient names to record important traffic nodes,3 sometimes coupled with their current general positions (though not without doubts), and his work constitutes an important source for reconstructing the traffic dataset of the Silk Road of this period. While grounded in Yan’s work Transportation Maps in the Tang Dynasty, the study also used a large number of historical literature and research as a supplementary, such as modern remote sensing data (geospatial data cloud, http://www.gscloud.cn/), Atlas of Chinese Cultural Relics, Data of the Third National Cultural Relic Census, Atlas of Chinese History, Toponym in Ancient China,10–16 field work findings, and so on.
(2) Data of traffic attributes in the Tang dynasty
Traffic attribute data of the Tang dynasty mainly includes current administrative locations of the traffic points, and starting and ending time of the traffic points and traffic lines. The current administrative location data are derived from Atlas of Chinese Provinces published by Planet Press17 and the China Administrative Division Network (http://www.xzqh.org/), both of which provide updated information on the latest administrative division. The starting and ending time comes from various historical documents including General History of China's Administrative Division in the Tang Dynasty.18
2.2   Collection and processing methods
This dataset includes traffic point and traffic line data. Traffic point refers to the currently-known places where the Silk Road passed, which may not necessarily be official staging posts built during the Tang dynasty. Traffic points also include capital cities, military towns, ancient cities, fortresses, passes, lakes, etc. While the density of traffic points varies greatly from region to region, this study tries to restore the main artery of the east-central Silk Road in the Tang dynasty (618 – 907). For traffic arteries existing before the Tang dynasty, their traffic condition of year 618 is presented. For arteries established or abandoned during this period, this study presents the traffic condition of their establishment and abandonment year when applicable. The data generation process is shown in Figure 1.

Figure 1   Technical process of dataset generation
From the perspective of processing method, the acquired data falls into two categories: traffic graph data and attribute data. The traffic point and traffic line data were collected using a handheld GPS acquisition terminal and Google Earth (professional version) maps (on the GCS_WGS_1984 geographic coordinate system by default). EXCEL table was then used to store traffic attribute data such as the name, lifecycle and current administrative location of the traffic points and lines. ARCGIS property connection function was used to correlate the traffic graph data with the property data.
Data processing needed to solve two critical issues. The first was to correlate the traffic points and traffic lines extracted from historical documents with their current geographical locations. In order to retain the highest data accuracy, the practice of minimum precision was not used. Instead, we selected multiple scales to identify traffic point data of different types while abandoning practices that might reduce data precision. Data of lakes and passes were of the lowest precision, followed by cities and fortresses, whereas those of city gates and bridges possessed the highest accuracy. Indeed, as the traffic points were of different types, their current statuses varied as well. While many cities and fortresses of the period had their ruins preserved till this day, few staging posts remained; identification of the passes and lakes, on the other hand, could refer to their nearby relics. In general, location methods were selected based on the following two conditions into which the traffic points fall.
(1) Traffic points with ruins preserved. Although few cities, fortresses, passes and staging posts of the Tang dynasty remained intact, many of their ruins were nevertheless preserved. Though some of the cities were found to be built before Tang dynasty, they provided reliable references for locating the traffic points. With the help of ancient maps and image data, we could locate most of the sites, and only a few was located within a radius of four or five kilometers.
(2) Traffic points without ruins. Only a limited number of counties in the Tang dynasty had city walls, among which most had been destroyed. However, the names of these places remained. According to historical geographical findings (such as Atlas of Chinese History), especially historical documentation of place names (such as Dictionary of Chinese Ancient and Modern Place Names), counties in the Tang dynasty was accurately pinpointed to a current location of township level. As for staging posts with no ruins remained, most of them were located at or near county gates. Hence their location was identified by the location of respective counties.
As restoration of the traffic lines relied on the traffic points involved, the traffic lines with densely-distributed traffic points identified naturally had a higher accuracy. Considering temporal continuity of the Silk Road routes in history, we used data of the period with the densest distribution of known traffic points for restoring respective sections of the Silk Road in this period.8 In addition, when identifying the directions of traffic lines, we took into account topographical information, together with large-scale maps of the Republican period and road networks of modern times. It needs to be noted that the traffic lines in this dataset were mainly recorded in their historical terms, such as Baishuijian road and Huihu road. Those without historical terms were named after the places of origin and termination, such as Liangzhou—Guazhou road.
The second important issue was the timing of each traffic route. The artery of the Silk Road in the Tang dynasty was often controlled by the government. It consisted of road surface facilities, including pavilions, posts, towns, guards and capture facilities, and a set of management systems. The starting time of traffic data was recorded as the time when these official facilities and systems were established, while the ending time was recorded as the time when these official facilities and systems were abolished, either due to wars or regime secession, regardless of continued public usage afterwards. The starting and ending years of a traffic point were recorded as the year when the administrative unit was established and abolished, respectively. For example, Hezhou was established in 618 during the Tang dynasty but was taken over by Tubo in 742, so the starting and ending years of Hezhou were recorded as 618 and 742, respectively. For places, lakes and cities that existed before the Tang dynasty, their starting and ending years were streamlined to the duration of the Tang dynasty, in order to fit in with the designated period of this study.
3.   Sample description
The dataset contains two parts of data: traffic point and traffic line. The property table of traffic point Shapefile contains ten major data fields: "Name_CH" represents Chinese name of the traffic point; "Name_EN" means its English name; "Type" refers to traffic point type, involving data precision of multiple scales that can be used for data screening; "Province", "PL_City", "County" and "Town" represent the province, city, county and town where the traffic points were located, respectively; "Site" records traffic points with ruins preserved – the name of its current site was recorded if (ruins of) the traffic point remained, or a null value would be given if otherwise; “Beg_year” and “End_year” represent the year when the traffic point was established and abolished, respectively, which records information about its lifecycle. The property table of traffic line Shapefile mainly contains three data fields: "Name", "Beg_year" and "End_year". "Name" records the name of the traffic line, while "Beg_year" and "End_year" store information on their survival status. The dataset design allows users to retrieve traffic condition of the east-central Silk Road of any specific year from 618 to 907, which is useful for investigating relevant diachronic changes. Figure 2 shows the traffic routes of the east-central section of the Silk Road in years 618, 720, 770 and 800 during the Tang dynasty.





Figure 2   Traffic routes of the east-central section of the Silk Road in the Tang dynasty
4.   Quality control and assessment
We identified two major issues that might impact the data quality, that is, reliability of historical sources and errors caused by data acquisition methods. Accordingly, we adopted the following measures to control the quality of this dataset: firstly, we used a large amount of multiple sources to ensure that each data entry was supported by sufficient and reliable historical records, including historical literature, unearthed documents, field investigation data and research findings. In case of controversies, facts were re-examined according to historical records. Only sources deemed credible were added as references. Secondly, new approaches to ancient traffic restoration8 were adopted to restore the traffic routes more precisely. Thirdly, the vector data of Google Earth image were validated against field data, and geometric correction was made by using Landsat images. Finally, topological errors were checked and rectified under ARCGIS 10.2 environment. By means of considering historical facts, restoration design and technical methods, the above measures were used to guarantee source reliability, data accuracy and standardization.
5.   Usage notes
This dataset can be directly used to characterize the overall situation of the east-central Silk Road in the Tang dynasty, and a dynamic data display can be enabled by using the function of "time slider". Apart from that, this dataset can also be used for secondary data extraction, where traffic condition of the east-central Silk Road can be retrieved by any designated year within the temporal coverage of the Tang dynasty. Providing the most reliable and comprehensive data of the east-central Silk Road in the Tang dynasty hitherto, this dataset is fundamental for regional studies on the Silk Road in this period. Since the source data are of different precision levels, three precision levels have been preserved for each data entry. Accordingly, users are suggested to select an appropriate scale according to their specific research demands.
Ferdinand von R. China, Band 1. Berlin: Dietrich Reimer Verlag, 1877.
Zhou W & Ding J. The Silk Road Dictionary. Xi’an: Shaanxi People’s Press, 2006.
Yan G. Transportation Maps in the Tang Dynasty (Volumes 1 & 2). Shanghai: Shanghai Ancient Books Press, 2007.
Bi J & Wang X. Researches and application of geo-spatial technology to the history of the Silk Road. Arid Land Geography 30(2007): 954 – 961.
Wang X. A study of historical geographic information system of the Silk Road (Xinjiang section). Master’s Dissertation, Northwest University, 2015.
Du X. Study on the information extraction and historical route changes of the Desert Silk Road (Xinjiang section) from Han to Tang Dynasty. Master’s Dissertation, Xinjiang Normal University, 2015.
The Silk Road geographic information system project of Fudan University,.available at: <http://silkroad.fudan.edu.cn/>.
Zhang P. GIS technology and a new method of reconstructing the 2000-Year-Old Silk Road. Trends of Recent Researches on the History of China 2017(2): 57 – 61.
Zhang P. Historical geographic information platform for the Silk Road: Design, concept and application. Journal of Yunnan University(Social Sciences Edition) 16(2017): 50 – 54.
National Cultural Heritage Administration. Atlas of Chinese Cultural Relics: Shaanxi Volume. Xi’an: Xi’an Map Press, 1998.
National Cultural Heritage Administration. Atlas of Chinese Cultural Relics: Gansu Volume. Beijing: Mapping Press, 2011.
National Cultural Heritage Administration. Atlas of Chinese Cultural Relics: Qinghai Volume. Beijing: China Map Press, 1996.
National Cultural Heritage Administration. Atlas of Chinese Cultural Relics: Ningxia Volume. Beijing: Heritage Press, 2010.
National Cultural Heritage Administration. Atlas of Chinese Cultural Relics: Xinjiang Volume. Beijing: Heritage Press, 2012.
Tan Q. Atlas of Chinese History. Beijing: China Map Press, 1982.
Dai J. Encyclopedia of Toponyms in Ancient and Modern China. Shanghai: Shanghai dictionary press. 2005.
Planet Map Press. Atlas of Chinese Provinces. Beijing: Planet Map Press, 2016.
Zhou Z & Guo S. General History of Administrative Division in China (Volume of Tang Dynasty). Shanghai: Fudan University Press, 2017.
Data citation
1. Xu X & Zhang P. A dataset of traffic routes of the east-central Silk Road in the Tang dynasty (618 – 907 A.D.), Science Data Bank. DOI: 10.11922/sciencedb.587
Article and author information
How to cite this article
Xu X & Zhang P. A dataset of traffic routes of the east-central Silk Road in the Tang dynasty (618 – 907 A.D.). China Scientific Data 3(2018). DOI: 10.11922/csdata.2018.0015.zh
Xu Xueqiang
historical research, data collection and dataset generation.
PhD, mainly engaged in historical geography and GIS research.
Zhang Ping
data collection and data structure guidance.
PhD, Professor, mainly engaged in historical geography and historical GIS research.
National Social Science Fund (14ZDB031)
Publication records
Published: Aug. 20, 2018 ( VersionsEN1
Released: June 5, 2018 ( VersionsZH2
Published: Aug. 20, 2018 ( VersionsZH4
Updated: Aug. 20, 2018 ( VersionsZH5