GNGTS 2014 - Atti del 33° Convegno Nazionale

GNGTS 2014 S essione 3.2 193 productive units (namely, a wine press room), while the central body had a residential function. The villa also shows a southern area including a small paved court which probably acted as a taberna deversoria for housing merchants and voyagers travelling on the way to ancient Gallia. The building had a water supply line and a sewer system. The walls of the villa show socles in opus caementicium or dry rubble with pebbles and fragmentary tiles and bricks, while the elevations were basically in perishable materials, like sun-dried bricks, pisé or opus craticium . The roofs, sometimes covering wide surfaces, were made of large rectangular tiles and cover-tiles. The building complex described is the result of a long series of enlargements and modifications, starting during the Augustan period. Near the main construction, other contemporary buildings are known, probable dependencies to the villa . The characteristics and destination of these partially explored evidences are still under evaluation. The geophysical investigation programme started from the immediate surroundings of some of these structures, south of the ”trincea F” (north area). Moreover, to test the effectiveness of the methods in the given geological context, another area was selected for a blind test (south area) where some structures were already unearthed and later covered . Data acquisition and processing. Wemade the survey on two areas within the archaeological site: the North area and the south area (Fig. 3). In the north area, on May 2014, we made both a magnetic and GPR survey while in the south area only a GPR survey was performed on July 2014, after the processing of North area data which lead us to exclude the magnetic method. In the North area we used a Radar K2 from IDS with a 500 MHz IDS antenna. On the whole we made 31 profiles, each one about 32 m long. The profile interval was 0.4m while the trace interval was 0.035 m. The trace’s length was 50 ns and they were digitized with 512 samples. All the profiles show, on average, a quite broad band spectrum with a -3dB band ranging from 200 to 700 MHz. Raw radargrams showed a lot of clutters and there were no evidences of clear coherent reflections in the x-twt domain. The processing flow was very reduced and it consisted of: move start time, to set properly the zero time; dewow to remove low frequency effects and a non linear manual gain, increasing with time, to recover energy in the deeper radar section. Time slices, on the envelope of the Hilbert transform of the radargrams, have then been made summing up the samples over 1 ns window length. Fig. 3 – a) GPR timeslices on the south and north areas. b) Map of the vertical gradient of the TMF. Near the northern side of the area a metallic fence, that could not be removed, strongly influenced the measurements.