*4.2. SBE*

During the ASV developments, we considered the as SBE a "basic" sensor, since knowing water depth is crucial for safely navigating in shallow waters. Thus, we integrated a custom version of such instrument in the main electronic board of OpenSWAP. Our echosounder (μEcho) integrates a vertical incidence ultrasonic pinger, operating at a frequency of 200 kHz, a narrow (8◦, conical) beam width, a short time-gate pulse length (350 μsec). The ultrasonic signal, is digitally sampled in a selected time-window by an ArduinoDue ® board, and data (the echograms) are stored in SEGY-format files. Moreover, we implemented the procedure by Haynes et al., [14] to track the bottom reflection. First, an amplitude envelope of the echogram is computed by convolving the squared values of the original data trace with a rectangular window as wide as the source pulse:

$$a(t) = \int\_{-\infty}^{\infty} \left[ \mathbf{x}(\tau) \right]^2 w(t - \tau) d\tau \tag{1}$$

where *a(t)* is the amplitude envelope trace, *x*(*t*) is the recorded signal, is the rectangular window, and *L* is the window size.

$$w(t) = \begin{cases} \begin{array}{c} 1 \ t \in [0, L] \\ 0 \ elsewler \end{array} \end{cases}$$

Subsequently, a simple threshold-time delay algorithm was used to achieve the bottom reflected signal from the amplitude envelope trace. Conversion of travel-times into water-depth could be performed once the sound-speed is determined through specific estimates.

Acquisition of the entire echosounder sweep at each sounding point, rather than the simple depth value generally provided by echosounders, give us the opportunity of estimating the relative reflectivity of the sediment-water interface. Using the SeisPrho function RCCM [3], under the vertical incidence case and neglecting the e ffect of energy scattering due to the bottom roughness, we obtain an estimate of the relative reflection coe fficient (R) by using:

$$R = \left(A\_r/A\_s\right)z\tag{2}$$

where *Ar* and *As* are the amplitudes of the source and reflected signals, respectively, and *z* is the water depth.

In order to obtain an estimate of *R* from our finite-length echo-sounder pulse, we used, for *As* and *Ar*, the values:

$$A\_s = \sum\_{i=0}^{W} |\mathbf{x}\_i| \tag{3}$$

$$A\_I = \sum\_{i=B}^{B+W} |\propto\_i|\tag{4}$$

where *x(i)* is the digital sampled signal, *W* is the width of the source pulse, and *B* is the bottom detection time. Bottom reflectivity data obtained by these echograms are used to compile reflectivity maps, that could be diagnostic of geological processes.

Propagation and scattering of high frequency acoustic sound at or near the bottom is controlled by a number of factors, including biological, geological, biogeochemical and hydrodynamic processes operating at the benthic boundary layer [15]. However, experimental measurements sugges<sup>t</sup> that the single most important geotechnical property related to acoustic attenuation is the mean grain size of the insonified sediment [16–19]. In such a case, a combination of bathymetric and reflectivity maps could be used as an e ffective tool in analyzing geological processes acting at the sediment-water interface. Figure 9 reports an example of such procedure carried out along a river stream. We note a main erosional trough formed close to a bridge pillar, also marked by high reflectivity (red pattern in Figure 9d). We also note that areas with prevailing deposition are marked by lower reflectivity (blue pattern in Figure 9d) as along the internal part of the river meander in the northern sector of the study area. Ground-truthing reflection coe fficient estimates with sediment samples is mandatory to perform more accurate bottom classifications ([20] and references therein).

**Figure 9.** Echographic survey carried out along the Reno River close to an urban bridge (Bologna, Italy). (**a**) Configuration of the survey before start; (**b**) example of 200 kHz echogram in very shallow water; (**c**) morphobathymetric map highlighting the presence of a deep erosion close to the bridge pillar; (**d**) reflectivity map obtained using the same data showing sectors of prevalent erosion (red) and deposition (blue).

#### *4.3. Side-Scan Sonar Imaging*

We collected Side-Scan Sonar images along the Cavo Napoleonico artificial channel, in Northern Italy. The channel, which connects the Po and the Reno rivers in the Po plain, is oriented perpendicularly to the thrust and fold belt buried by alluvial sediments falling in the area which underwent the maximum superficial deformation during and after the Emilia 2012 earthquake sequence [21,22]. For this reason, it was chosen as an interesting site for geophysical surveys in search for co-seismic effects. Among other data collected with conventional methods, we carried out a side-scan sonar survey of this channel using a Starfish system mounted onboard of an OpenSWAP vehicle. We were searching for earthquake-related structures, such as fractures, fissures, sediment fluidization or slumps. Analysis of side-scan sonar images combined with high-resolution seismic reflection profiles (also collected with an OpenSWAP vehicle) suggests a correlation between the presence of "disturbances" at the channel floor (Figure 10) and the area o maximum deformation detected through satellite derived measures [23].

**Figure 10.** Top: side-scan sonar image of the Cavo Napoleonico artificial channel close to the epicenter of the Emilia 2012 earthquake, showing slumps and gravitative failures affecting the channel-floor. Bottom: SBP profile collected in the same area, penetrating the first meters of alluvial sediments, showing paleo-channels and displacements reaching up to the surface.

#### *4.4. High-Resolution Imaging of the Subsurface Using the Embedded SBP*

A typical SBP, operating with magnetostrictive transducers at hundreds to thousands of volts, is not suitable for operating on board of any OpenSWAP vehicle, either for the heavy weight and for low efficiency in converting the DC electric power of batteries to suitable high-voltages. For this reason, we developed a lightweight chirped SBP system (μChirp) embedded in the OpenSWAP electronics (see above). We tested the potential of our μChirp in Lake Trasimeno, a shallow-water tectonic lake in Central Italy. The lake was investigated through conventional systems in the frame of a geological study carried out with different geophysical methods [24,25]. However, some sectors close to the northern shore, were too shallow to be accessible using conventional systems, and were surveyed using OpenSWAP. The problem in this case was imaging at the best resolution the first tens of meters stratigraphic sequence in very shallow water, where several sources of noise affect in general the data. Prior to the survey, for a quality control of data collected by μChirp, we performed a comparison with an industry standard chirp-sonar system, the Benthos-Teledyne Chirp III, mounted onboard of a small boat. Results of this benchmark are reported in Figure 11, where the shallowest part (<10 ms. TWT) of the Trasimeno sedimentary sequence is imaged with the two systems, showing similar vertical resolutions and penetrations.

**Figure 11.** Example of seismic reflection profiles along the same navigation line collected with two different systems in Lake Trasimeno. **Top**: Chirp III Teledyne-Benthos, with 4 Massa transducers; **Bottom**: μChirp with 4 Monacor transducers. Unconsolidated sediments are penetrated down to 15–20 ms. below the lake floor, by both systems, with high vertical resolutions (tens of cm), enabling a detailed imaging the sedimentary structures.

#### *4.5. Multibeam Echosounder Repeated Surveys*

In order to test the possibility of performing MBES surveys using an OpenSWAP vehicle we used a Klein HydroChart 3500 integrated echosounder/side-scan sonar system. The HydroChart 3500 is a professional bathymetric sonar with IHO hydrographic standard for shallow water operations that integrates the characteristics of a side-scan sonar with those of an interferometric multibeam. It is a portable system that includes a motion reference unit (MRU) as well as course and sound speed sensors located in the sonar head. Each echogram includes uncertainty on the estimate of the depth and angles of the beams used for ray-tracing. In this way, the sonar propagation uncertainty model is integrated into the data processing flow, to provide uncertainty estimates for individual depth measurements that can be used by third-party bathymetric postprocessing.

We performed repeated multibeam survey offshore Calabria (Southern Italy), in the Calabrian Arc accretionary wedge, one of the most tectonically active regions in the Mediterranean Sea [26–28]. For this purpose, we surveyed repeatedly some key areas in the nearshore, where interferences between coastal sediment transport and gravitative instability in the vicinity of active faults were observed. The surveys were carried out in the same areas at different time-intervals, ranging from a few days to several months, obtaining good results. An example is reported in Figure 12, where we observe that over 95% of bathymetric measures are coherently positioned within ±30 cm of differences (the normal accuracy gathered by navigation system), and discrepancies between the two surveys (Figure 12c) are probably related to short-term seafloor changes, such as sand ripple migration, which were captured by our 4D survey. This first test indicates that repeating bathymetric measures at regularly spaced time-intervals along the same acquisition lines could be an interesting tool to determine what type of natural process is active and what is its temporal scale. Such information would be crucial in areas highly prone to slumping, seismogenic, and tsunamigenic risks.

**Figure 12.** Bathymetric data collected using a Klein Hydrochart 3500 multibeam echosounder onboard of an OpenSWAP vehicle, including: (**a**) a first survey; (**b**) a second acquisition performed next day; and (**c**) the point-to-point difference between the two DTMs. Note that over 95% (colored dots) are within ±30 cm of difference. Red and blue undulations in c result from short-term seafloor changes, probably due to sand-ripple migration.
