Estimation of amplitude noise reduction as a function of depth recorded by a deep vertical array (Northern Italy)

To design an efficient seismic monitoring infrastructure, the characterization of the background seismic noise level of each potential seismic station installation site is one of the most important data-quality metrics used to evaluate the suitability of such sites to host the seismic network. The background seismic noise can be generated by different sources such as, ocean waves (microseisms), atmospheric turbolences (strong wind and storms), and anthropogenic activities. Such disturbances are characterized by specific frequency bands, time-occurrence (diurnal and seasonal variation), and site location (close to populated area or to the coasts). Reducing the effect of these noise sources is one of the main challenges to face for designing seismic monitoring networks and, more specifically, when selecting the hosting site of a seismic stations. A solution to attenuate the seismic noise effect is obtained by deploying seismic stations in boreholes. The noise level reduction with depth has been observed and studied by different authors, however a general law estimating the sufficient depth to gain is still missing. In this study, we analyse the continuous seismic noise level at S. Potito-Cotignola gas storage in the Po Valley (Northern Italy) recorded from January 2019 to December 2021 by a broadband (BB) seismic station at surface and a vertical array composed by 6-short period 3-components seismometers installed at depth ranging between 35 to 285 m in borehole. We aim to characterize the seismic noise by computing the amplitude noise reduction in terms of dB as a function of depth for different frequencies and the SNR by selecting three seismic events, with different epicentral distance and magnitude. Our results show that the noise level decreases with depth following a logarithmic empirical trend and the lowest magnitude event records the maximum SNR difference between the deepest sensor and the one at the surface. The estimated empirical relationships can be used to help the design microseismic monitoring networks in similar geological settings.