GNGTS 2013 - Atti del 32° Convegno Nazionale

Given this overall interesting geological setting, the area around the Malpasso site ( Fig. 1 ) is also particularly characterized by the presence of an emerging sandstone formation protruding towards the lake and is moreover interesting from the archeological point of view. Several historical reconstructions suggest indeed that within this area, and particularly in the Tuoro plain, the Trasimeno battle (217 a.C.) took place. This battle was one of the most important episodes of the II Punic war and it was fought between the army of the Carthaginian general Hannibal and two Roman legions under the command of the consul Caius Flaminius. New evidences related to historical level fluctuations of the lake and more careful reading of the sources and a critical analysis of previous studies have been significant in respect to a positive identification of this site as a most probable location of the battle (Brizzi and Gambini, 2008). The actual shore line leave only a short passage between the lake itself and the Malpasso (Fig. 1), while the same should be wider in ancient times to allow for the passage of huge armies. In this respect the findings of several dated ceramic materials has shown that for long periods in the Etruscan and Roman times lake levels were usually lower than the present (Cattuto et al. , 2011). With the aims of both reconstructing the geological setting of the area around the Malpasso site and to eventually find some localized remains of the battle we carried out several waterborne geophysical surveys on the area. Adopted methodologies were: magnetic surveys, seismic reflection Chirp Sonar surveys and Continuous Vertical Electric Soundings (CVES) profiles. Within this paper we report some preliminary results along an example profile (Fig. 1) for the last two methodologies underlining the main evidences observed and further planned processing on the data and data integration. Materials And Methods. Seismic reflection profiles were collected using a Benthos Chirp III system with 4 transducers mounted on board of a small catamaran towed by the boat. Data were collected using a pulse length of 5 ms, a frequency sweep from 2 to 7 kHz, and a digitally sampled at 16 kHz rate. Seismic data were stored in SEGY files, after deconvolution and instantaneous amplitudes computation. Data processing was carried out using SeisPrho (Gasperini and Stanghellini, 2009) and included static correction, filtering and automatic gain control. Vertical resolution of acoustic images is <10 cm; maximum penetration reached, about 10 m. CVES measurements were obtained by the use of an array of nine electrodes fixed on a floating cable dragged by a boat. The array has two current electrodes, in the cable part closest to the boat, followed by seven potential electrodes. The current electrodes were 16 m apart, while the six couples of potential electrodes in dipole-dipole configuration had exponentially increasing spacing from 0.5 to 16 m. The first potential electrode was 0.5 m from the farthest current electrode. This configuration allowed a maximum depth penetration similar to the seismic one. The towed cable floated on the lake surface thanks to appropriate plastic floaters and was kept stretched by a floating anchor fixed at its end. A multichannel georesistivimeter (Syscal-pro by Iris Instruments in Sysmar update) was used to simultaneously acquire the six potential measurements. The acquisitions were treated with a Laterally Constrained Inversion (LCI) process. The LCI approach was developed to invert vertical electrical sounding data acquired along a profile by Auken and Christiansen (2004), using a pseudo-2D layered parameterization of the investigated geological medium and has been implemented in Matlab environment (Sambuelli et al. , 2011). The inversion result of LCI is a set of 1D resistivity models, each of which corresponds to a sounding. All the soundings are inverted simultaneously by minimizing a common objective function, which contains the acquired data, the available a priori information and the constraints. Through the lateral constraints, information from one electro-stratigraphic section are then interconnected with the neighboring ones, producing a smoothly varying 2D section. Both surveys are accurately georeferenced by means of independent GPS acquisition to allow for data integration. 187 GNGTS 2013 S essione 3.3