B. Mušic1, B. Slapšak1, E. Farinetti2
1 Dept. of Geoinformatics, University of Ljubljana, Slovenia
2 Dept. of Archaeology, University of Leiden, Netherlands

Research in Tanagra in the framework of the Boiotia Project, co-directed by John Bintliff and Božidar Slapšak, is since 2002 part of the Late Roman Cities and Countryside project, an Interuniversity Attraction Poles programme (IUAP V/09). It is a survey oriented research of a complex urban site for which no foreseeable excavation can be envisaged. It applies a battery of non-desctructive methods, including archaeological, geophysical, surface morphology, and architectural survey.
The paper will concentrate on the potential of combined use of raster based GIS; 2D and 3D magnetic modelling; DEM, and of unsupervised classification for interpreting the results of geophysical prospection. 15 ha, about one half of the intra-muros area have been surveyed. Data sets acquired by geophysics are complementary in many details. Multivariate statistics is used to build up a site map in a single image, where archaeologically relevant units are discernable by spatially defined classes, based on the level of correlation between input data sets. The output is a composite image of differing, but normalised input data, and that could lead to partial automation of the interpretation procedure. 3D visualization is used for presentation of measured anomalies in physical fields, and of the final interpretation of the urban plan.