Fig 1		Fig 4		Fig 5		Fig 6
Fig 7		Fig 8		Fig 9		Fig 10
	Fig 11		Fig 12		Fig 13

Urban Survey

Introduction
The 2001 field season took place in three parts: a short spring season devoted to resistivity survey at a time when the ground retained sufficient moisture, a six week summer season that combined geophysical survey with the study of materials from previous excavations and a period of additional geomagnetic survey and pottery restoration in September. New initiatives in 2001 included the geophysical experiments carried out by Harald von der Osten-Woldenburg (Fig. 1), the great nephew of Hans Henning von der Osten who was responsible, together with F. H. Blackburn, for the creation of the first map of the city's walled limits in 1927. The progress map (Fig. 4) shows the areas already covered by the geophysical survey.

Figure 5 shows the Total Station used to set out the 20m grid for the geophysical survey, the view looks south-east with the north-eastern stretch of city wall at left and the Kale at right. Wooden pegs are hammered into the ground every 20m (Fig.6). Each peg is marked with white lime so that its position is not lost between seasons if shepherds burn the pegs to make tea. Communication between the station and the roving prism is by means of a walkie-talkie. Teams of men from the village (Fig. 7) are sufficiently practised to conduct gradiometer survey on the steepest of slopes without loss in the quality of data collection.

Remote Sensing
Harald von der Osten-Woldenburg kept the Kerkenes Project at the leading edge of experimentation and application of geophysical methods in archaeology through extensive trials with electromagnetic induction techniques using a Geonics EM 38 instrument (Fig. 1). The existing Kerkenes geomagnetic images are of such quality that they can be readily used as a control for this new method. Initial results are most encouraging, but the instrument can only be used where the terrain allows rapid movement at a constant speed, sometimes running backwards, without tripping or banging the instrument. Although the initial results (Figs. 9b and 10a) do not provide the same degree of clarity that is to be seen in the geomagnetic and resistivity imagery, (Figs 9a, 10b and 10c) they are impressive for a first try. Design of an instrument geared more towards the needs of rapid survey with high density sampling strategies, as required by archaeology, combined with ongoing refinements to the software are likely to make this a method that will have wide applications in the future.

Geomagnetic survey of the entire city at Kerkenes, using Geoscan fluxgate gradiometers, is now only one single season away from completion (Fig. 4). When finished the magnetometers will have been carried over more than two square kilometres of the site for something in the region of 36,000,000 paces. Combination of two-dimensional magnetic imagery with the three-dimensional Global Positioning System (GPS) survey that was successfully concluded in 2000, now allows the application of sophisticated computer programs to create simulations and perform analysis of the urban environment (Fig. 12). Virtual reality is just around the corner. ArcView is becoming the central platform for the Project and the July issue ArcNews contains a piece that brings aspects of the Kerkenes research design to a GIS audience.

A short spring season and experiments in the early days of the summer program, before the Kerkenes soils became baked so hard that probes made of the hardest available steel became bent within minutes, proved that electrical resistivity survey (Fig. 8) in carefully selected areas that are more or less free of stone rubble can produce excellent results (Fig. 10c). In some instances the resistance images reveal some structures with greater clarity than the magnetics. This seems to be particularly true of areas where there is greater overburden.

Figure 9 shows a 40 x 40m block at the northern end of the city surveyed using different geophysical methods. By July, however, the ground had become too hard for the insertion of steel probes for electrical resistivity survey. Both the gradiometer (a) and the electromagnetic induction (b) survey show a large hall and anteroom at left and a row of four cells at right. The gradiometer shows a pair of posts on either side of the central door between the hall and the anteroom, but neither method was able to detect the bases for wooden columns that would have been required to support a roof.

Figure 10 illustrates three geophysical survey methods each produce imagery that brings out different features. The location of this one hectare (100 x 100m) survey area is given on Figure 4. Electromagnetic induction (Fig. 10a), used with a high density sampling strategy for the first time at Kerkenes, reveals most of the features seen in the other two images. As a result of this experiment improvements are now being made to the machinery at the same time as the processing software is being developed. Geomagnetic survey with a fluxgate gradiometer (Fig. 10b) reveals many of the buildings in very considerable detail. In some instances the position of burnt door posts can be seen (e.g. grid 740/1720 not seen in Fig. 10c). The imagery is influenced by different degrees of burning and by the magnetic properties of various building materials in ways that are not yet fully understood. The underlying geology and geomorphology also has effects, particularly where bedrock is close to the ground surface. Processing the same data with different parameters does, however, bring out considerably more detail than it is possible to show on a single image. Electrical resistivity survey (Fig. 10c) highlights details with remarkable clarity where there is a favourable combination of conditions. Unlike with gradiometer imagery, heat induced alterations to magnetic properties have no effect, although fire alteration to mud-brick and other materials is significant. The background is strongly influenced by hydrology, a stream bed accounting for the broad sweep of white. The two parallel white lines that run diagonally across the picture indicate the path of a tractor track.

Figure 11 showing the results of the gradiometer survey over a twelve hectare (300 x 400m) area at the centre of the city (Figure 4 gives the location), provides an overview and details of subsurface remains at the same time. This particular image combines data from 2001 with results from earlier seasons at a single level of data processing

Aspects of the Urban Infrastructure
Space only permits brief mention and selective illustration of some of the more significant developments. A single twelve hectare (300 x 400m) image (Fig. 11) of sub-surface remains within the central portion of the city may serve as an example of the results that can be replicated over the greater part of the very large area that has been surveyed (Fig. 4). Urban blocks and streets were laid out according to a conceptual scheme that was sufficiently flexible to allow for distortion to rectilinear planning where topography made such deviation expedient. Areas that were not enclosed within urban blocks in the initial scheme of things were, if our interpretation of the horizontal stratigraphy is correct, gradually filled by more random building construction. In Figure 11 the large block to the left of centre, revealed in 1995, can now be seen in its wider urban environment. At centre right a less regular block can be seen, the south-west corner of which is defined by the meandering street that clings to the contour. This same block contains a variety of buildings which at first sight appear to be recorded with exceptional clarity, but which pose particular problems of interpretation. The block would appear to be dominated by a large rectangular space, the south-eastern third of which is divided from the larger section by a pair of strong signals. It is very possible that this feature is yet another of the large columned halls that are a particular architectural characteristic of the city, and that the two high spots were created when the wooden posts of double doors were burnt (a phenomena seen elsewhere at Kerkenes). Resistivity survey might reveal stone column bases which would verify this interpretation. Above the north-west corner of this putative hall are two parallel rows of rectangular rooms, each some 6 x 4m, with a corridor between. The less distinct structure to the left may be of similar form. It is somewhat exceptional to find such a large number of cells, which were presumably for storage, within a single complex. Single rows of similar cells are not, however uncommon; one row being visible at the centre of the image and another in the block left of centre. Also clearly identifiable are a number of the more usual two-roomed buildings one of which, in the block on the left, was partially excavated in 1996.

Another new development of particular significance has been the recognition of two buildings that appear as though they might be megarons (Fig. 10, especially bottom left of 10c). A megaron is a type of building, characterised by a pitched roof, open porch and central hearth, that has clear cultural connections with Phrygia in western Anatolia. The two structures seen on Figure 10, both of which lie in areas that were apparently outside any of the original urban blocks, are substantial buildings each measuring approximately 10 by 12m. The existence of these megarons may perhaps represent further evidence for degrees of Anatolianisation during the development and growth of the city, but before its destruction around 547BC.

Within yet another area of the central portion of the city, not illustrated here for want of space, magnetic survey has led to the identification of a large open area that was clearly a public place. One possible function for this feature could very well have been as a market place for which no other good candidate exists within the city walls. Situated to the north-east of the Büyük Göl (Big Pond), the largest of the artificial reservoirs within the city, this large level area occupies a sheltered position in the central sector of the site. Several of the major streets run tangentially to this space which must have been devoid of buildings for some special purpose. New imagery of the Büyük Göl itself, obtained when it was completely dry in September, has revealed the presence of substantial stone walls within the artificial banks on all four sides. Unlike the Sülüklü Göl (Leech Pond) on the high southern ridge, also fully surveyed when completely dry, the Büyük Göl does not seem to have an elaborate and carefully constructed stone lining. Although part of an overall, planned, scheme of urban water management, the various pools and reservoirs within the city evidently varied in status and had differing functions.

Geomagnetic images of the steep central slopes between the high southern ridge and the lower area of the city, also obtained in 2001 but omitted here, have delineated the two connecting streets, that to the east turning sharply round the head of a small stream so as to maintain a vehicular gradient, together with crescents of small blocks following the contours above. The extreme north facing slopes above the head of the western stream are, not surprisingly, devoid of structures. This expanse, the only large empty area within the entire city, difficult to negotiate, exposed to icy winds and far distant from the nearest gate, would have afforded the least desirable urban real estate.

Away from the site, analysis of the urban dynamics of the site using Geographical Information Systems (GIS) continues in the capable hands of Nahide Aydin, who recently completed the Masters Program in Archaeometry at METU. Scott Branting is also proposing to write a doctoral thesis on transportation and GIS at Kerkenes at the University of Buffalo.