Geophysical Survey

In 1993 Dr. Lewis Somers of Geoscan Research, http://www.geoscan-research.co.uk/ joined the Kerkenes to experiment with methods of geophysical survey. The excellent geomagnetic results led to larger scale survey led by Somers in 1995 and by the team in consultation with Somers in 1997 and 98. By the end of 2001 the entire 2.5km2, with the exception of the Byzantine remains on the Kale, of the city had been mapped. his achievement was made possible by the rapid development of ever more powerful desktop computers with gigabits of memory, a Windows version of Geoplot together with an array of new imaging software and advanced printers as well as through the dedication of the team and of the men employed to collect data. The survey was very largely co-ordinated by Nahide Aydin who also undertook the greater part of the image processing.

In 2000 Harald von der Osten Woldenburg http://www.lb.netic.de/hvdosten/ achieved excellent results with resistivity survey and carried out successful experiments with electrical conductivity. Since then the area of the resisitivty in the central part of the city has been gradually extended by an annual series of short spring seasons carried out while the ground is sufficiently damp. Although the results of the resisitivity survey usually display greater clarity than the magnetic, the method only works well under certain conditions and is impossible over much of the site because of the amount of stone on the surface. In level, stone-free areas resistivity survey is about four times slower that magnetic, and more than ten times slower on rough terrain.

Geomagnetic

Geoscan Research FM-36 fluxgate gradiometers are used to map subsurface features. Data is automatically collected in 20 x 20 m grid squares, each grid taking about 20 minutes to complete. Four readings per metre at one metre intervals, collected zigzag, yield results that bring out as much detail as higher data collection strategies. Although higher sample densities can result in more pleasing images, they do not give more information, are slower to collect and to process and much larger to archive. In 1997 two 40 x 40 m grids were very carefully surveyed at 8 readings per m. quarter metre intervals. The results, processed by Somers in a pre-release version of Geoplot 3, showed no more features than the sparser sample strategy. By 1999 it had become possible to survey about forty 20 x 20m grids per day with one machine. Using two machines increased the total to about 60 grids with two teams working an eight-hour day in the field. The new generation of Geoscan fluxgate gradiometers would make survey even more rapid. Efficient data management was crucial.

The geomagnetic results are so clear at Kerkenes for a number of reasons: the granitic bed-rock and soils, the single period of construction with no superimposition of buildings, proximity of wall tops to the modern surface and the destruction of the city by fire. Different types of construction yield a variety of results for a complex number of reasons that are still very poorly understood.

Electrical Conductivity

Harald von der Osten Woldenburg made trials with a Geonics EM 38 instrument. These proved to be very encouraging, although the machine was designed for geological rather than archaeological survey, the later requiring much denser sampling strategies as well as greater control over the quality of data collection.

Resistivity

Resistivity survey with a Geoscan RM 15 is carried out in exactly the same 20 x 20m grids as were laid out for the magnetic survey. On a good day it is possible to survey 9 grids in 8 hours, after which the data has to be downloaded. Although the wooden pegs rarely last from one season to another, being frequently burnt by shepherds for making tea, small amounts of lime placed around each peg can be located years later. Collection is at 0.50m traverse intervals with readings every 0.5m. Sharp differences in the level of contrast in the images of adjacent grids is the result of hydrological differences. They either indicate that the survey was conducted at different times or, in particularly damp areas, varions when the remote probes were moved.

Ground Penetrating Radar (GPR)

While the test trenches dug in 1997 were still open Somers experimented with GPR. The machine was dragged along the ground adjacent to the trenches so that results could be compared with the walls and surfaces exposed in the adjacent trenches. It was expected that GPR would provide evidence for the depth of buried features and the nature of surfaces. Results, for reasons that are still not understood, were disappointing. Further experimentation with GPR is planned for a future season.