GNGTS 2022 - Atti del 40° Convegno Nazionale

GNGTS 2022 Sessione 2.1 215 Fig. 3 - Local empirical amplification model with respect to Swiss rock reference (Poggi et al. , 2011) estimated using the SSRh method for 1.2 Hz. Blue triangles represent stations of the seismic monitoring network where earthquake recordings are available, black circles are ambient noise measurements. Base maps source: Federal Office of Topography (map.geo.admin.ch) . 2011). The final model is consistent with some site response proxies such as the fundamental resonance frequency and geological data i.e. thickness of unconsolidated deposits derived from the gravity measurements. In the deep areas of the basin, the calculated amplification factors exceed 10 at the fundamental frequency (between 0.8 and 1.5 Hz) and remain high for the long frequency band. Even in the shallower parts characterized by sediments thickness less than 50 m, high amplification is expected. We can conclude that local seismic hazard is increased significantly in the Lucerne area. Moreover, in our study, we test the influence of the location of the stations on the results. Because we deployed more than one temporary station in the basin, we can calculate several realizations of the SSRh function. Based on the test performed on the temporary network, we observe the correlation between the goodness of the fit between earthquake- and noise- based spectral ratios and the similarity of the site condition between the tested station pair. In other words, if two stations have similar site response and are located on similar site conditions demonstrated by similar f 0 values, the noise-based spectral ratio provides a good estimate of the SSR amplification function. In case stations have much different site response, we can still estimate the relative amplification; however, the fit becomes worse, especially at the fundamental frequency. Another influencing factor that was studied was the effect of variability of ambient vibration wavefield. Depending on the time of the day when ambient vibration was measured, the resulting SSRh functions show some differences, however, a mean value over a longer period (e.g. 24 hours) provides a stable result. We recommend that the method is tested using a small temporary network before applying the technique to the whole basin in order to investigate the impact of described factors.