**4. Discussion**

The archaeological questions underlying the multidisciplinary study conducted in the Bronze Age embanked site of Fondo Paviani and described in this contribution can be summarized in the following points:

(a) Check the state of preservation of the perimetric system in different parts of the site;

(b) Verify the degree of information available by the combination of different geophysical methods measuring the differences in the electrical conductivity of the soils and therefore defining the geometry (size and depth) and the spatial distribution of the palaeo-channels that characterize this embanked system;

(c) Identify the stratigraphic/sedimentation relationships describing the reactivation sequence of older river systems with more recent ones;

(d) Define a multidisciplinary protocol applicable to the whole site, capable of providing clear information on the position of the structures of interest.

To answer the question of point (a), we started from the evidence made by the analysis of some aerial photographs relating to a chronological span of about 40 years, which have also guided the excavations and archaeological reconnaissance for over a decade. Over time, aerial photographs have shown a marked decrease in the contrast, which was previously visible between the traces of paleochannels in the northern perimeter of the site and the hosting system. The stratigraphic section drawn up following the excavation carried out in a part of the perimeter of the system, which is preserved today, constituted the first direct useful data to plan the field strategy for the geophysical measurements carried out with the use of the ERT and the FDEM methods.

The ERT sections acquired in different points of the embankment system of Fondo Paviani allow here to formulate some interesting considerations, made possible thanks to the comparison and calibration of the no-direct data with the direct ones provided by the stratigraphic section. In general, ERT measurements find a wide application and allows us to obtain excellent results in contexts and on systems where there is a marked gradient of the soil water content between the target and the hosting system.

The archaeological site of Fondo Paviani, characterized by the presence of the remains of a housing settlement consisting of a perimeter system, with a moat, characterized by river deposits, and a bright contrast to the hosting system, represents in this sense the ideal case-study for these types of non-invasive measurements. Thanks to the information obtained from the L1 section (Figures 7 and 8), with the identification of the response of the preserved rampart, detected by the stratigraphic section made in the same position (Figure 8), it was, in fact, possible to verify what was hypothesized by the analyses carried out on aerial images and by the LIDAR data. The ERT sections, mainly made in the northern sector of the embanked settlement, confirm the absence of the preserved rampart in this sector in the shallow subsoil, where one would expect to find the same, as in section L1.

Therefore, it can be reasonably said that in the sections corresponding to the ERT lines from L2-3 to L7 the highest and therefore most shallow system relating to the rampart has been removed due to the agricultural practices that have taken place in the last 40 years. This data would also be confirmed by the topographical data extracted by LIDAR (Figure 3), where mean lower altitudes are registered corresponding to the areas in which the ERT sections were acquired. However, from the comparison between the DTM data extracted from LIDAR and the ERT sections, it is possible to make further considerations on the possible correlation between the topographical surface data and the real buried deep geomorphological data. In particular, despite the apparent identical topographic information relating to the NE sector at the two ERT sections L1 and L2-3, parallel to each other, the geophysical data shows that the L1 section intercepts a deep conductive incision where instead the section L2-3 intersects a shallower system. From these real data, it is, therefore, possible to a ffirm that the topographical data and the overall view, albeit informative of the DTM data, do not always correspond to the real deep-buried structure. The ERT sections undoubtedly provided additional information of extreme interest regarding the geomorphology of the area. In particular, only the lines L1 and L7 seem to intercept some deep incisions characterized by conductive deposits for the whole investigation depth. In contrast, in the other ERT sections, the conductive structures, attributable to the presence of river deposits related to the perimeter system of the dam site, appear less in-depth.

In the areas of the settlement not investigated by geophysical measurements, the DTM analysis allowed us to monitor the state of preservation of the elevated structures (natural or anthropogenic) on a preliminary basis. In particular, in north-eastern and south-western corners of the settlement, an abrupt and regular lowering of the surfaces is visible. These anomalies, considering that they follow exactly agricultural field borders, can be interpreted as the outcome of a strong agricultural impact with possible partial destruction of the ancient landscape features.

Considering the point b), the ERT method, capable of providing data on the lateral and in-depth extension of the targets, could be integrated into vast areas by faster soil electrical conductivity measurements. These can be obtained thanks to the use of multi-frequency FDEM systems that allow the inversion of the data providing detailed information at di fferent depths. Multi-frequency FDEM measurements carried out in a test area, in which the L6 ERT section was acquired, thanks to the inversion of the data, demonstrate absolute consistency with the ERT measurements and, therefore, the possibility of integrating this di fferent kind of data. ERT combined with inverted FDEM data, thus confirming the possibility to quickly and precisely verify the extension of the structures only locally identified with ERT (Figure 9).

The integration between ERT and FDEM methods allows us to also answer question c positively. Undoubtedly, the two geophysical methods allow us to detect the di fferent targets in the subsoil and, therefore, to define the real stratigraphic di fferences not detectable from the aerial photo images, where the surface signal does not precisely define the nature and the stratigraphic relationship between the di fferent targets.

Finally, the integrated multidisciplinary approach adopted in this study, answers to point (d), highlighting how the information provided by every single method used, although fundamental, may be limited in the absence of a multidisciplinary approach as adopted here. The positive evidence obtained thanks to this type of integration allows us to outline future investigation strategies to be adopted to complete the study of this important embanked site.
