**3. Results**

#### *3.1. ARP Results*

The geophysical surveys carried out by means of the ARP system permitted to simultaneously obtain three apparent resistivity maps, referring to the three depth intervals. Their measured global resistivity values ranged between 11 and 154 Ω·m, with di fferent responses between each depth map (Figure 4). First, the shallowest map (Figure 4a), referred to a 0–0.50 m layer, reports the apparent resistivity values ranging from 50 to 117 Ω·<sup>m</sup> with a data distribution showing maximum frequency at about 80 Ω·<sup>m</sup> but generally more shifted towards low than high resistivity values for less numerous classes (Figure 4d). The intermediate map, the 0–1.00 m layer (Figure 4b), presents apparent resistivity values ranging from 11 Ω·<sup>m</sup> to 56 Ω·<sup>m</sup> with maximum concentration around 25.5 Ω·<sup>m</sup> and a general symmetrical distribution (Figure 4). The 1.00 m map appears spatially more noisy than 0.50 m one but it is possible to note a good coherence between their resistivity patterns. The deepest map, 0–1.70 m layer (Figure 4c) shows apparent resistivity data ranging from 30 to 154 Ω·<sup>m</sup> with maximum concentration of around 60 Ω·m, significantly lower than the mean value (92 Ω·m) distributed with a lower spreading (Figure 4f).

**Figure 4.** Apparent Resistivity maps obtained with ARP at 0.5 m (**a**), 1.0 m (**b**) and 1.7 m (**c**); data frequency histograms referred to three maps above: 0.5 m (**d**), 1.0 m (**e**) and 1.7 m (**f**).

As is predictable, the results of the 1.70 m map are the smoothest obtained, with clear readability of resistivity patterns and anomalies at a bigger spatial wavelength and a lower presence of both spatially high-frequency signals and noise. Vertical gradients of resistivity ranges are probably related to soil layering properties, in particular, a possible compaction of deepest volumes due to the proximity to bedrock and a moisture gradient for the two shallowest layers due to superficial drying phenomena or different clay concentration with depth. In all the cases, the apparent resistivity patterns show no relation with the topography gradient that characterizes the investigated sector of the site. The result of tillage operations, carried out in the recent past, appear in the form of stripes east–west oriented and in some route traces, are recognizable as double parallel curves on the shallowest map (0.5 m), Figure 5a,b.

**Figure 5.** (**a**) Normalized apparent resistivity maps in gray color scale (0.5 m nominal depth); (**b**) traces of farm vehicles clearly readable in the ARP resistivity map referred to 0.5 m depth (dashed red lines).

Looking for archaeological purposes at apparent resistivity patterns, the most significant ones seem to be resistive areas (red tones), which are highlighted in the following maps. Nevertheless, some conductive patterns (blue tones) are also evidenced and reported as signals of potential interest, especially for the top and bottom layers.

In Figure 6, the resistivity map at 0.5 m nominal depth is reported with the indication, bordered in red, of some interesting resistive areas which have vertical continuity even in other two maps. The same areas are reported also at a 1.0 m depth (Figure 7) and 1.7 m (Figure 8).

Figure 9 reports the most interesting conductive areas: in particular, the most superficial map (0.5 m, Figure 9a) shows, in its southern part, some conductive regular patterns (indicated as **I1**), just inside the protected area where digs were concentrated, which could have archaeological relevance such as tombs or similar geometries filled with moist silty or clay soil. The 0.5 m map presents other conductive larger regions which have dimensions not compatible with archeological features and are most probably interpretable as superficial geological variations: this is the case of the **G1** region that is confirmed by deeper layers of the dataset and interpreted as a continuous geological body. Indeed, the map at 1.7 m shows wide conductive areas which could have importance to reconstruct paleoenvironment and paleo-landscape of the area (Figure 9b, **F3**, **G3** and **H3**); area **I3** of the same map shows that at higher depths, the conductive anomalies indicated before as **I1** inside the protected are less evident and probably their low e ffect could be due to the fact that the deeper map still includes all the layers between 0.0 m and 1.7 m.

**Figure 6.** Identification of the main resistive patterns in the ARP map at 0–0.5 m depth interval.

**Figure 7.** Identification of the main resistive patterns in the ARP map at 0–1 m depth interval.

**Figure 8.** Identification of the main resistive patterns in the ARP map at 0–1.7 m depth interval.

**Figure 9.** (**a**) Shallow conductive anomalies located in the southern part of the site of study (0–0.5 m interval depth); (**b**) wide conductive sectors at the northern part of the surveyed area (0–1.7 m interval depth).

#### *3.2. Vertical Analysis of ARP Anomalies and Comparison with GPR Data*

To explore the three-dimensional features of the ARP results, in the next paragraph, the data are analyzed by means of their vertical pseudo-sections in correspondence to the most important planimetric patterns. ARP pseudo-sections are also compared with GPR sections acquired over the same paths (Figure 10). Five vertical profiles have been selected over areas where coverage of both datasets was available: three are located at the north-western part of the ARP-surveyed area, intersecting three resistive anomalies A-B-C in Figures 6–8; the other two profiles are located at the eastern part of the ARP-surveyed area, just outside the protected area partially interested by archaeological excavations, in correspondence of the northern part of the D resistive anomalous pattern of Figures 6–8.

**Figure 10.** Location of Ground Penetrating Radar profiles.

In Figures 11a–c and 12a,b the comparison of radargrams and resistivity pseudo-sections is reported. As also observed in Figure 11a, looking at pseudo-sections, it is possible to notice the difference of apparent resistivity ranges at different depths: mean-resistive data on the top layer; a more conductive layer at middle-depth in which relatively resistive points vertically correspond with the top layer ones; the most resistive layer is the deeper one in which vertical correspondence of relatively high resistivity data with the two top layers is confirmed. This behavior leads to the consideration that the resistive bodies should have a certain vertical continuity from bottom to top, even in the intermediate layer where a probable higher humidity of soil partially influences apparent resistivity ranges shifting all values towards more conductive ones. The three nominal depths of ARP raw data are indicated with the black horizontal dotted lines on pseudo-sections.

**Figure 11.** Comparison of ARP pseudo-sections with GPR profiles: (**a**) TI040009; (**b**) TI040017; (**c**) TI060005. Four times vertical amplification.

**Figure 12.** Comparison of ARP pseudo-sections with GPR profiles: (**a**) TI040062; (**b**) TI100073. Four times vertical amplification.

In Figure 11a, the northern profile (**TI040009**) is reported. ARP pseudo-section presents two resistive patterns belonging to regions **B** and **C** in Figures 6–8. The bigger one, belonging to region **B**, is located at progressive horizontal coordinates range of 65–90 m and corresponds to parts of the radargram with an intermediate amplitude signal. Moving to the right in the radargram, it is possible to notice the signal weakening in correspondence of conductive areas where microwaves are more absorbed. From 135 m to 140 m, the second resistive anomaly appears as a section of the top linear resistive pattern in region **C**; at the same coordinates, GPR data present the most noisy traces.

In Figure 11b, the ARP profile and the radargram acquired along the **TI040017** line in the south-western side of the study area are represented. The profile crosses two wide resistive anomalies, which are identified as **A** and **C** in Figures 6–8. In correspondence of both resistive regions, the radar profile is characterized by the strong amplitudes of the signals. Moving from west to east, a strong reflective region is found from 62 m to 75 m, with a globally complex shape due to the superimposition of multiple diffraction signals. This anomalous region, at depth from 0.5 m to 2.0 m, is spatially related to the first resistive body of the **TI040017** pseudo-section, clearly detected also in ARP maps. The second significant radar-reflective region is located from 90 m to 110 m. This signal pattern is related to targets with a decreasing depth from left to right approximately ranging from 1.5 m to 0.5 m. The same trend is observable in the resistivity pseudo-section where the top layer resistive region is right-shifted compared to that of the bottom layer.

Figure 11c reports the comparison between the **TI060005** radar section and the resistivity pseudo-section at the same position. This line is parallel to the **TI040017** radar profile, previously presented (Figure 11b) and is located about 10–12 m south of this. As Figure 10 shows, this profile intersects the same resistive areas ( **A** and **C** in Figures 6–8) close to their southern border. The pseudo-section shows the same resistive volumes but with slightly lower resistivity values and wider spatial extents. The GPR profile is characterized by the presence of di ffuse intermediate signals with two regions where signal amplitude increases: the first one is particularly strong and concentrated between the 60 m and 80 m progressive coordinates, in correspondence with the western resistive body. The second region, located between 125 and 150 m, presents weaker GRP signals than the first one but is also weaker than the corresponding second anomaly in the **TI040017** profile. A preferential direction of radar signal is less evident in this case with regards to the **TI040017** GPR profile. The depth ranges are the same as in the **TI040017** profile for both regions.

The last comparisons of ARP and GPR data, Figure 12a,b, mainly concern the most resistive and widest area of ARP maps which was indicated with **D** in Figures 6–8 in the southern part of the prospection. This highly apparent resistivity region is partially located in the restricted archaeological area in the south, and in the bordering part of the free northern area. The two comparative profiles are placed in the free northern area, the first one over the northern limit of the wide resistive pattern and the second one crossing the central higher resistivity part.

In Figure 12a, the first profile, **TI040062**, is plotted. The pseudo-section is characterized by a large resistive region at its bottom layer, ranging from 15 m to 75 m progressive horizontal coordinate, with the highest apparent resistivity values concentrated between 45 m and 75 m. On the top layer, resistive values are right-shifted and less extended, located between 60 m and 80 m. the GPR profile is in very good agreemen<sup>t</sup> with ARP pseudo-section, presenting an intense group of signals between 60 and 80 m with a preferential tilted direction climbing up towards the right. These reflected signals come from variable depths ranging between 0.5 m and 1.8 m.

Figure 12b reports the **TI100073** comparison profile, which crosses the **D** in Figures 6–8 at its highest values position. This profile is almost parallel and east-shifted of about half length from the previous one, **TI040062**. Because of this, anomalous patterns linked to the same body, which are still present, are located at di fferent horizontal coordinates. The deeper ARP anomaly presents the highest resistivity values of all the datasets, saturating the color-scale at 150 Ω·<sup>m</sup> between 5 m and 45 m, but resistive area can be also considered between 0 m and 55 m. The top resistive area is a bit wider than the range from 20 m to 40 m. Even for this profile, GPR data are in good agreemen<sup>t</sup> with the ARP pseudo-section: in fact, a large region with clear sub-horizontal GPR signals is present between 5 m and 40 m, at an approximative depth of 1-2 m.
