The Troodos Archaeological and Environmental Survey Project (TAESP) is investigating human activity in the north-central Troodos mountains of Cyprus during all time periods. The core of its methodology consists of intensive archaeological and geomorphological survey. The survey area includes fertile and well-watered valleys, drier plains, copper-bearing foothills, and the northern part of the Troodos Range itself. Other than some rescue excavation of tombs, no systematic archaeological work had been done in this area.
Our research is focused on the dynamic relationship between human society and the environment. We are documenting and analysing settlement patterns, land-use patterns and communication networks across the landscape at all time periods, and relating them to environmental factors such as physical landforms, soils and sediments, vegetation and water. A particular focus is the nature and development of resource exploitation, especially agriculture and metallurgical production. Related themes include the production of pottery and stone tools, forest resources and soil management.
Within our 164km² survey area we carried out fieldwalking transects in six intensive survey zones totalling 32km². These are the six areas mapped and discussed in the data and interpretation section, and were chosen to give a representative range of the topographic, environmental and cultural variation across the survey area. The rest of the area was sampled by means of a grid of 20 short transects and more purposive survey.
Transects are lines of survey units, traversed by fieldwalkers placed 5m apart. Where there were broad areas with high densities of artefacts and other features of interest, we also carried out contiguous block survey. These we termed Special Interest Areas (SIA, with identification numbers prefaced 'TS'). A particularly useful function of these was to allow us to investigate a settlement and its context as a single analytical entity. These methods were intended to give us a continuous view of human activity across the landscape, rather than limiting our analysis to a series of 'sites' or dots on the map.
A typical survey unit (SU) is 25m across and between 50 and 100m long. It was surveyed by a field team consisting of four or five fieldwalkers 5m apart. The walkers counted sherds, tiles, chipped stone, ground stone, slag, shotgun cartridges and other artefacts, recorded relevant features such as ground visibility, and made a representative collection of the sherds. This was done by each fieldwalker collecting one of every type of sherd, as distinguished in the field: one thin sigillata rim; one thick sigillata rim; one thick sigillata base; etc. By the end of the project we had collected 30.1% of the 99,676 sherds found in 1119 survey units.
Each survey unit sits within a geomorphological unit (GU), which describes environmental features such as soil type, slope and surface stability. No survey unit can cross the boundaries of a geomorphological unit, which gives it analytical consistency and integrity.
The collected artefacts were analysed and recorded by the artefact team. Once the data entry and digitising had been done, the database tables could be joined to the GIS to produce artefact distribution maps. Clearly, we had to assess the reliability of these density figures. A series of seeding experiments run by Robert Schon in 2003 showed that varying ground density and artefact distribution had a surprisingly low effect on sherd recovery rates. Fieldwalkers were apparently compensating for poor conditions by scanning the ground more carefully. On this basis we decided not to factor the visibility and background ratings into our density analyses.
With the pottery distribution maps, we were able to exploit our representative collection procedures. This involved combining the total figures for sherds counted with the proportion of sherds collected of a particular characteristic (e.g. Roman, or cooking ware). This resulted in an estimated number of sherds of that characteristic in the entire survey unit, not just the 30% that we collected. These figures were then divided by the square area of the unit, to give estimated densities of sherds per 10 x 10m square (expressed in this article as 'sherds per 100m²').
There were, of course, very clear focuses of human activity, many of them equivalent to the 'sites' of less intensive projects. These we termed Places of Special Interest (POSI, with identification numbers prefaced 'TP'). This is purely an administrative definition and carries no inherent cultural assumptions about meaning or function. For this reason, and because of their very varied nature, 'POSIs per km²' would be a meaningless statistic.
Depending on their character, POSIs were variously recorded, mapped, gridded and sampled. Full descriptions are available in the POSI database table, but this article focuses more on the transects and block survey to provide a continuous analysis of the landscape.
Similarly, our interdisciplinary team also investigated archaeometallurgy, architecture, botany, communication routes, documentary history, oral history and water distribution systems. In the final publication these will be integrated into the analysis, but are not the main focus of this article.