HIGH-RESOLUTION MULTI-CHANNEL GPR SURVEY OF THE ROMAN TOWN FLAVIA SOLVA IN AUSTRIA W. Neubauer, S. Flöry, I. Trinks, A. Hinterleitner, K. Löcker, S. Seren The Roman town Flavia Solva, located in the Roman province Noricum (Styria, Austria) and founded in the early first cen- tury AD, has been investigated by geophysical prospection since 1998 through the ZAMG Archeo Prospections R  team. Aerial archaeology and excavations have revealed the rectangular street system of the town. The municipal area of the ancient town is located at the western bank of the river Mur and covers an extent of some 39 ha, partially destroyed by the modern town, roads and industrial zones. Over the past ten years, the Archeo Prospections R  team has conducted systematic geophys- ical prospection surveys at Flavia Solva, combining magnetic and GPR measurements in order to explore the extent and inner structure of this Roman town. As of 2009, an area of 5.43 ha has been surveyed with GPR and 13.84 ha with magnetometry. With the establishment of the Ludwig Boltzmann Institute for Archaeological Prospection and Virtual Archaeology (LBI ArchPro), the Roman town of Flavia Solva was selected as one of the case study areas for developing and testing a new large- scale integrated prospection approach. In the meantime, the ini- tial large-scale prospection approach has been copied by a group of amateur prospectors, covering the same area previously sur- veyed with high-resolution Caesium magnetometers, with dual sensor fluxgate magnetometers, proving that today comparable results can be achieved with commercially available instrumen- tation in a fraction of the time required in earlier years. As large- scale prospection projects can be very delicate to conduct in re- gard to the understanding and support of the generally numerous land owners involved, as well as regarding other stakeholders and public awareness, such duplicate activities are not at all help- ful in pursuing the objective of integrating non-invasive prospec- tion methods further into the archaeological process. However, despite such interventions, which may also be questioned from an ethical point of view, the main goal is to develop new tech- niques and methodology by empirically applying them to suit- able archaeological sites and landscapes, and to investigate the limitations of the respective hardware, software and logistics for their further improvement. The initial Earth resistance surveys conducted in 1998 lo- cated the forum and confirmed high resistivity contrasts, in- dicated favourable conditions for GPR measurements for the prospection of further details of the town structure. From 2000 onwards, GPR surveys have been carried out in Flavia Solva (Neubauer et al., 2001; Neubauer 2008). The primary objec- tive of the prospection work was a detailed survey of the north- eastern part of the municipal area. At the time, an area covering more than 4 ha has been recorded three-dimensionally, in detail, over the course of ten days of fieldwork. At that time, the survey was conducted using a PulseEKKO 1000 GPR system with 450 and 900 MHz antennae in a raster of 0.5 × 0.05 m, covering the greater parts of 10 insulae (domestic structures). The early GPR surveys from the year 2000 were repeated manually in 2007 and 2008 using the latest instrumentation (PulseEKKO Pro and Nog- gin Plus systems from Sensors & Software) and higher spatial resolution (0.25 × 0.05 m) . These repeated measurements doc- umented the progress made in the field of archaeological GPR applications (Neubauer 2008; Neubauer et al., 2009) and were presented at the 8 th International Conference on Archaeological Prospection held in Paris in 2009. In winter 2012, the same area was investigated again using a motorized multichannel MALÅ Imaging Radar Array (MIRA) in collaboration with the system manufacturer MALÅ Geo- science (Figure 1). The 16 channel 400 MHz antenna array using real-time kinematic GPS positioning was able to cover an area of 9 ha within 20 working hours at a resolution of 0.08 × 0.08 m. The objectives of this reinvestigation were the documenta- tion of the development in comparison with former prospection approaches and the more detailed imaging of the mapped struc- tures. The MIRA system covers a 128 cm wide swath for each driven track. In-line GPR trace sampling was set to 8 cm with a trace stacking factor of 4. The antenna array was placed in a front-mounted box on the hydraulic system of a small trac- tor. Power-supply and a field computer for data collection were installed on the vehicle. Accurate positioning of the GPR mea- surements is crucial and was implemented using real-time kine- matic GPS. The position information is transferred via radio link to the measurement vehicle, where the information is recorded together with the GPR data. The GPR and positioning data stream is processed using an adapted version of the APradar soft- ware. This software, which was also used for the previous sur- veys, allows the pre-processing, interpolation, coordinate system transformation and three-dimensional migration of the enormous amount of data within short time. The processing results are ex- ported as geo-referenced TIFF images, specific raster formats, or animations representing the generated 3D data volumes. For further data analysis and interpretation, the visualisations are im- ported into a GIS environment. On basis of these first GPR surveys, an archaeological and historical interpretation of the north-eastern part of the munici- pal area of Flavia Solva has been published (Groh et al., 2002) as an extension of the archaeological interpretation report by Archeo Prospections R  . With the new-high resolution GPR re- sults, many of the interpretations can be refined and open ques- tions can be addressed. The new survey reveals unprecedented details of the Roman buildings detected earlier. A vast Villa Urbana with a central peristyle court includ- ing water basins has earlier been postulated to be located in the north-eastern part of the municipal area, despite concerns de- rived from the high magnetic response of the "courtyard". Based on the combined interpretation of the magnetic and GPR survey data, several rooms with hypocausts were distinguished with a high probability. Now these interpretations have been verified by the results of the MIRA measurements, as the hypocausts be- came clearly visible in great detail in the GPR depth-slices gen- erated from the high-definition data volumes (Figure 2). As the "peristyle court" is hypocausted, it has to be interpreted as a large hall providing access to several rooms with further hypocausts