Spatial Distribution of Archaeological Sites in Lakeshore of Chaohu Lake in China Based on GIS

: There are about 19 Neolithic Age sites and 120 Shang-Zhou dynasties sites in the western lakeshore of Chaohu Lake, Anhui Province, China. Based on sites data, topographic data, drainage maps, administrative maps and SPOT5 remote sensing data, spatial analysis methods were introduced into the archaeology of Chaohu Lake Basin with Geographic Information System (GIS). The spatial analysis methods include: 1) Point Density Analysis with density model to acquire the shift of the sites; 2) Distance Analysis to reveal the spatial structure of the sites; 3) 3D Analysis based on Digital Elevation Model to get micro-geomorphologic features of the sites; and 4) Buffer Analysis to discover the relationship between the sites and rivers. The results indicate that the archaeological sites spread from the western lakeshore of Chaohu Lake gradually to the northwest, later to the south, and then symmetrically spread. Controlled by productive forces and other factors, ancient people preferred to reside in those places near water, or in plain and fertile land, which resulted in linear and decentralized distribution of the sites, and presented such disciplines as river valley directivity, terrace directivity and soil directivity. This paper indicates that the distribution of archaeological sites was influenced earlier by natural elements and later by human elements in the study period. The research is important for the integration of GIS and archaeology.


Introduction
Archaeological site is the relic of human activities, which belongs to the concept of archaeology. From the historical, aesthetic, ethnic, or anthropological point of view, it is a human engineering with conspicuously universal value or an engineering united by nature and human being (Childe, 1990;Yan, 1997a). Most of the sites were ancient human settlements which were the outcome of development of human society in a certain historical stage. A site is a place where the ancient people lived, worked and took part in various activities and an integrated system composed of a variety of material elements and natural elements (Marziani and Citterio, 1999;Melanie, 2005). There were various natural factors af-tors affecting the choice of archaeological sites. At the same time, the sites were also affected by human factors such as productive forces, group economic life, family system, religious belief and so on. Therefore, archaeological sites had unique spatial and temporal distribution features (Yan, 1997b;Tang, 2004). Previous studies on sites mainly focused on the following aspects: 1) the relationship of laminated layers or individual features of sites (Sun, 1997;Ma, 2002;Wang et al., 2006;Zhang, 2007); 2) the reconstruction of production relations and natural environment (such as primitive agriculture or vegetation cover) of sites (Ohlendorf et al., 2003;Li and Zhou, 2007;Welmoed, 2008;Kohler-Schneider and Caneppele, 2009); and 3) spatial migration and distribution of ancient culture (Huang and Zhu, 2005;Zhu et al., 2007;Deng, 1997Deng, ). 2007Deng, 1997).
In the past, the research of archaeology of sites was only based on a simple literal description and summarization, but now with the development of remote sensing (RS) and Geographic Information System (GIS) technologies, it has rich contents, closely combining spatial database with attribute database of sites, and developing new research directions (Liu, 2007). The new directions include: 1) the features of remote sensing images of sites (Lu and Liu, 1992;Xiong and Liu, 2005;Gao et al., 2008); 2) the digitization of archeology data Zhang et al., 2007); 3) the visualization and management of spatial historic data (Li and Bi, 2004;Li, 2008); and 4) the restoration and reconstruction of sites by GIS spatial analysis methods (Liu, 2006;. The restoration and reconstruction of sites by GIS spatial analysis methods have drawn more and more attention for it can help us better understand the evolution of man-earth relationship and will be a research hotspot in archeology of sites (Liu, 2007). The objective of this study is not only to discuss the spatial distribution features of sites in the western lakeshore of Chaohu Lake, Anhui Province, China, but also to reconstruct the evolution of the sites in different periods, and to introduce the GIS spatial analysis methods into the archeology of sites in Chaohu Lake Basin.

Study area
As one of the five biggest lakes in China, Chaohu Lake, located in the central Anhui Province, is important for local environment and society. Feixi County is in the west of Chaohu Lake and crosses the Changjiang (Yangtze)-Huaihe River Basin. According to the identification made by some archaeological departments, there are about 19 Neolithic Age sites and 120 Shang-Zhou dynasties sites in this county (Fig. 1). There are three rivers flow through the county, i.e., the Fengle River, the Paihe River and the Tianhe River, and among them the former two flow into Chaohu Lake, which inflows into the Changjiang River, and the latter one flows into the Huaihe River (Gao et al., 2005).

Data sources
A database of Neolithic Age to Shang-Zhou dynasties sites was established based on Feixi County sites data, Feixi County topographic maps, drainage maps, administrative maps, SPOT5 (K/J:288-286/3) remote sensing data in 30th January 2004, etc. The sites data were acquired based on an investigation that was done by the Cultural Relic Management Institution of Feixi County in 1987. To compare spatial features, 97 modern settlements were selected at the scale of administrative village and investigated in the study area from June to August of 2006 ( Fig. 1). From the investigation, spatial features of modern settlements were obtained, such as the relation between settlements distribution and contours, etc.

Methods
Using GIS technology, the spatial distribution features of sites were studied. 1) Point Density Analysis was applied to researching the shift of the sites and estimate the spatial distribution of sites (Odland and Ellis, 1992). 2) Distance Analysis was used to reveal the spatial structure of the sites (Malczewski, 1999). 3) 3D Analysis based on Digital Elevation Model (DEM) and contours was done to get micro-geomorphologic features of the sites (Arnaud and Bernard, 1999). And 4) Buffer Analysis was used to discover the relationship between the sites and rivers (Peng et al., 2003).

Results and Analyses
There are some sites that maybe exist both in the Neo-

Spatio-temporal shift of sites
Density maps of sites in different periods were obtained by spatial analysis model in ArcGIS9.0 (Fig. 2).
From Fig. 2, it is found that there was a spatio-temporal shift of sites in the study area. At the Neolithic Age, the sites existed in a small region that was at the west of Chaohu Lake. In the Shang Dynasty, the distribution of the sites extended along the Paihe River and the Tianhe River and finally formed a spindle. When time came to the Western Zhou Dynasty, the sites vanished in the Paihe River and the Tianhe River, but appeared linearly in the Fengle River and extended along the river course. They experienced a booming period in terms of quantity and unit scale in the Eastern Zhou Dynasty, and spread in the west of Chaohu lakeshore. In sum, there was a migration path for archaeological sites. At first, the sites appeared near the west lakeshore of Chaohu Lake, then moved from east to northwest and south of Feixi County, and finally spread evenly in this region.
The pixels in darker color indicate the greater density of sites

Distribution features of sites in different periods
In order to access the relationship between the sites, Distance Analysis model was used to reveal spatial structure of the sites. By the analysis we obtained spatial distance maps of the sites in four periods (Fig. 3). Considering the low productivity and the poor capacity of transforming natural factors, ancient people must be influenced by such constraints as climate, hydrology, topography, soil and vegetation when selecting residen-tial areas, therefore spatial distance features of the sites were different in the four periods. For example, in the Neolithic Age and the Shang Dynasty, the sites were distributed obviously along rivers and in linear pattern; in the Western Zhou Dynasty, as the sites increased, most of them were distributed in the riversides, but some expanded to highland; and in the Eastern Zhou Dynasty, even aggregation-distribution sites group took form, which was constituted by several linear distribution.
The distance between isolines is 3.5km

Distribution directivity of sites
By further analysis, we can see that the sites distribution presented river valley directivity, terrace and hill directivity and soil directivity in general.
(1) River valley directivity. Based on 3D Analysis, we made four Triangulated Irregular Network (TIN) maps of the four periods (Fig. 4). From a macro perspective, the landform was high in the northwest and low in the southeast of the western lakeshore of Chaohu Lake. And there were many northwest-southeast river valleys with hills or low mountains on both sides. Ancient settlements were mainly distributed in the northwest-southeast river valleys, because the valleys gradually spread to southeast and had a good microclimate environment, which maybe was a leeward with a sunny exposure. This environment was helpful for the ancient agriculture.
We also analyzed the relations between sites and rivers by Buffer Analysis tool in ArcGIS9.0 (Fig. 5). Along riversides, we set a 1.5-km buffer strip and found that there were 113 sites in the strips in the western lake-shore of Chaohu Lake in Feixi County, accounting for 81.29% of the total 139 sites. Only 26 sites were out of the 1.5-km buffer strips but in the 2-km buffer strips. All those features reflected water source or river valley directivity and they were in accordance with ancient peo-ple′s idea of "choose settlement near water area" (Lu, 2000).
(2) Terrace and hill directivity. From a micro perspective, the sites were often distributed on the second terraces of streams, which were seldom overwhelmed by floods, or hills with sufficient water supply along stream valleys (Fig. 4). Those terraces and hills mainly spread in the area with an elevation between 20-40m. Therefore, ancient people could not only get water supply easily, but also avoid floods. Based on the comparison between the sites from the Neolithic Age to the Shang-Zhou dynasties and modern settlements (Table 1), it can be concluded that 55.40% of the sites were distributed in the area with an elevation of 20-40m, and only 2.88% in the area with an elevation above 60m, which indicates that ancient people primarily lived on terraces and hills  and seldom in high-elevation area. With the improvement in productive force, the distribution of modern settlements is no longer overly restricted by water-source or topography, and the distribution range is getting larger. However, there is still not too much extension in human activity in those areas with an elevation lower than 20m. The percentage of residence spots in the areas are enhanced from 23.02% to 23.71%, with an increasing range of only 0.69%. It shows that human activities were still restricted by natural conditions in lower elevation areas, such as flooding.
(3) Soil directivity. In the mountainous region in the central and western Feixi County, the soil parent materials were in most slopes and residual deposits, and the soil types were skeletal soil, purplish soil and yellow-cinnamon soil. In terraces and hills, the soil types were general paddy soil. In polder lands in riversides or lakeshores, the soil types were fluvo-aquic soils and paddy soils. Through field investigation, we found that the sites were usually located at fluvo-aquic soil areas and paddy soil areas. Those soils were good for agricultural production (Zhou and Huang, 2003).

Discussion
In general, the factors influencing site distribution could be divided into natural factors and human factors. The leading factor varied as all the factors functioned differently in different periods.

Natural factors
In ancient times, production capability was low and men were strongly dependent on natural environment and resources, so the distribution of the sites indicated ancient men′s adaptation to natural factors, such as climate, hydrology, topography, soil, vegetation, etc., which were very important in forming regional features and spatial differences of the sites (Yang, 2005).
(1) Climate. Climate factor had a very important influence on the general distribution of the earlier sites; later, with the improvement of productive force, its influence had dropped. In the Neolithic Age, men would live near lakeshores or riversides for water. The phenomena changed in the late Shang Dynasty and the early Western Zhou Dynasty, during this period, the weather was cold and dry in Chaohu Lake area and the soil degenerated severely (Jia and Dai, 2006;Zhu and Zhang, 2003). Men would move to the place where it was warm, humid, soil-fertile and water-accessible, like the area along the Fengle River. There was a good micro-climate, as it faced river and was located against hills. Therefore many sites appeared in the Fengle River basin while sites in other places were few. While in the Eastern Zhou Dynasty, as temperature rose and production capability improved, so did men′s capacity of adapting to natural environment, sites began to evenly spread.
(2) River. River always remained one of the main factors affecting sites distribution. In the study area, the sites would be built near water bodies, such as rivers or Chaohu Lake. Humid micro-climate and the fertile land in riverside and lakeshore not only provided favorable natural environment for the population there, but also ensured them abundant food, and against floods. Meanwhile, rich aquacultural products from rivers such as fishes and shrimps were good supplements for food (Chen, 2004). The results also showed that many sites were distributed along the two biggest rivers in Feixi County, the Fengle River and the Paihe River, from the Neolithic Age to the Shang-Zhou dynasties.
(3) Landform. The sites in the study area mostly lied at the second terrace or piedmont of hills, where the altitude was from 10m to 40m, especially between 20-40m. Some sites might also lie at mounds about 4-7m high. Ancient men choosing those sites could not only get water sources but also defend assaults from animals or other clans. Based on the field investigation in the study area, it was found that there were entrenchment water courses surrounding the sites in general, which were probably city moats in those periods. Qian (2002) and Tang (2004) stated that in the sites in the northern China, residences were protected by building city walls, and in the southern China, by digging entrenchment. That viewpoint was confirmed in the Chaohu Lake area.

Human factors
As enhancement of human production capability and capacities of transforming nature, human factors would play an increasingly important role in affecting sites distribution.
(1) Level of productive force. Relations of production are decided by productive force, which is the basic law in human society. In agriculture society, the self-sufficient economy resulted from low productivity and led to men′s interdependence and collectivity mode of production. In the Neolithic Age, men could get few foods in low productive conditions and could not afford big population, so the number of sites was limited. But when it came to the Eastern Zhou Dynasty, with the invention of ironware farm implements, the level of productive force improved greatly, and human being could get more foods. Sites boomed from about 40 sites in the Western Zhou Dynasty to about 100 sites in the Eastern Zhou Dynasty.
(2) Economic structure. With the development of productivity, economic structure in the study area changed from fishing and hunting economy to agricultural economy. The development of agriculture led to the altering of housing patterns. Numbers of settlements increased and sites expanded. To fight against drought, ancient men invented irrigation agriculture to maintain high production. Land-use patterns also changed. As a result, the sites were not controlled exactly by water system and expanded to some places that had fertile soil but not good natural water resources, such as high-altitude areas. As a result, the sites in the Eastern Zhou Dynasty featured as symmetrically spreading.

Conclusions
Based on GIS techlology, and considering natural factors and human factors, we studied the spatio-temporal distribution of the sites in the western lakeshore of Chaohu Lake and the factors influencing their distribution. Field verification of the site features from remote sensing images were carried out in the study area from June to August of 2006. Spatial features of the sites were obtained during field investigation and the primary results were certificated.
The study results show that the archaeological sites spread from the west lakeshore of Chaohu Lake gradually to the northwest, later to the south, and finally symmetrically spread. The distribution of the sites was influenced by natural and human factors. Natural factors played important roles in early period and human factors in late period. Influenced by both natural and human factors, the sites in the western lakeshore of Chaohu Lake appeared at riverside or highland. And the sites scattered in linear and decentralized shapes, and presented disciplines such as river valley directivity, terrace and hill directivity, soil directivity, etc.
The application of GIS to studying the archaeological sites in the lakeshore of Chaohu Lake is an example as combination of GIS and regional archaeology research. GIS is an important tool for identifying the relationships between the natural factors or human factors and sites, and field investigation is also necessary to prove the primary results from GIS spatial analysis tool.
To study sites, we could also set from site domain to analyze the natural resources around sites, and then restore how ancient human beings accessed to resources and carried out their daily activities. By those investigation and analysis, we could better understand ancient man-land relationship, which will be the next step of this research.