Innovative approach to investigate seafloor slow vertical deformation using bottom pressure measurements in a volcanic area: Campi Flegrei caldera (Italy) case of study

Caldera systems generally show complex ground deformation phenomena, mainly associated with variations at different timescales of the hydrothermal systems normally hosted in calderas, magma storage and migration in the volcano plumbing system. Sometimes volcanic calderas are totally or partially submerged, which makes their study challenging, in particular if we consider shallow water systems. Many studies demonstrate the efficiency of adopting Bottom Pressure Recorder sensors (BPRs) for monitoring the vertical displacement of volcanic areas in deep-water environments. Moreover, BPRs measure the pressure at the seafloor over time, making possible their use to measure rapid or gradual inflation and deflation events. Here we propose an original approach to investigate the seafloor deformation over time in shallow volcanic areas using these instruments, and we applied it at Campi Flegrei caldera (Italy), a high-risk volcano system with a significant portion submerged in the Bay of Pozzuoli, near Naples. In this study we consider the data of two BPRs installed at the seabed within the Multiparametric Elastic-Beacon Devices and Underwater Sensors Acquisition (MEDUSA) infrastructure of the INGV Osservatorio Vesuviano, and we transform pressure measurements in equivalent water level changes to obtain the vertical seafloor displacement. To do this, we need the mean density of the water column during the periods of analysis, so we indirectly calculate it through the use of two tide gauges, one placed in the Gulf of Pozzuoli, and one located outside the deformation area for reference. The final results are then compared with the data acquired by the GPSs installed on the top of MEDUSA buoys, deployed at the same sites of the BPRs. The good correlation obtained supports the reliability of these sensors in measuring the seafloor deformation in a shallow water environment with an unprecedented level of accuracy.