The AHRB/C grant included a focus on developing innovative GIS -based analysis in the two areas of dating geophysical anomalies and colluviation modelling. A brief description of the work is provided here with initial references for the methodologies. The results for these two areas will be included within the final report. This work was carried out by Gary Lock and John Pouncett at the Institute of Archaeology,Oxford .
The challenge here is to try and date the complex of ditches and enclosures identified from geophysical survey. The methodology is based on Network Analysis, or more specifically Closest Facility Analysis in ArcGIS, and transports known dates from test-pits (established from ceramics and stratigraphy) around the network hence dating other ditches. If another date is encountered then the transportation stops. This work is reported in more detail in Lock and Pouncett (2007a and2007b) and Pouncett and Lock (2008)
This was tested on the geophysics of Sigwells West:
A similar approach has been used by eye, the colour-coded results are shown here:
The geophysics were digitised to create a vector network data-set and test-pits with dates established, the 'solution' is the shortest route to network end points.
Initial phasing is shown below:
Methods were developed to integrate stratigraphic and physical relationships which were known from the archaeology and needed to refine the model and produce a more nuanced result:
The robustness of the model was tested against 5 simulated data-sets which also allowed for individual anomalies to be investigated.
The methodology is being applied to the whole study area where appropriate geophysical coverage exists and will be reported in the final report.
Geomorphological processes have long been recognised as important within archaeology, not least for eroding evidence in upslope locations and covering evidence with colluvium downslope. It was established within the SCEP test-pits that several phases of colluviation have happened in the study area. These two figures show both regular and targeted test-pits with and without colluviation:
To understand and be able to analyse the topography the DEM was categorised into the following feature types:
Scale is important, by this we mean the number of cells within the DEM used in the classification, the window size as shown here where the different colours represent different topographic forms:
The methodology develops a concept of 'fuzzy classification' so that smaller windows (local topography) and larger windows (overall morphology) are both considered.
Colluvial zones can be identified where erosion and deposition are likely to take place:
Colluviation is more likely to occur on moderate slopes than steep or shallow ones and the depth of colluvium is likely to be greater on concave slopes than convex ones, hence surfaces could be weighted:
Phases of colluviation were identified from the test-pit stratigraphies and interpolation surfaces produced for each phase. These were then modelled against known colluviation (Locality 5 shown below) and evidence for sites from the geophysical survey
A flow model was also developed so that erosional and depositional zones could be estimated.
The methodology has been tested and seems to be robust. Results for the whole study area will be published in the final report, an interim report is Lock and Pouncett (Forthcoming).