Noise cross-correlation technique

The seismic noise recorded by two receivers can be used to define the impulsive response of the medium (Green function).

Several researchers demonstrated that cross-correlation of a diffuse wavefield, recorded at two points, is related to the time derivative of the medium Green's function. Considering a time interval long enough, a wavefield diffused in every direction will produce a seismic ray that will be recorded by the first receiver and, after a certain time interval, by the second one. So noise cross-correlation method is very suitable for urban areas, where active seismic experiments are prohibitive. Moreover, it allows to obtain seismic data in those areas where earthquake recordings are rarely available, because of their scarcity and the seismic waves attenuation.

This method has provided very good results over a broad range of distances, from a few hundred meters to several hundred kilometres.

Seismic noise experiments have been performed in Neapolitan urban area using Kinemetrics Quanterra Q330 stations with an accelerometer Episensor ES-T, in order to obtain V_s models vs depth. For example seismic noise has been recorded along a 445 m array at Scampia (in the northern zone of Napoli) (see map) for a time interval of 90 minutes.

Cross-correlation functions have been computed using a time window of 4 s in a frequency band of 0.5-4 Hz. These functions (vertical and radial components) have been processed by the FTAN analysis in order to extract an average group velocity dispersion curve of Rayleigh surface waves fundamental mode.

In the same area an active seismic experiment have been performed (see map). The average dispersion curves extracted by the two experiments have been joined during the Hedgehog non linear inversion. In this way a representative V_s model (100 m depth) has been obtained for the studied area.

Vibration analysis

Seismic vibrations measurements have been carried out in the urban area of Napoli with Q330 accelerometer seismic stations.

Signal analysis compared with the UNI9916 prescriptions

The signals analysis, after the numerical integration, has been performed with a 21 s time window over a minimum time interval of 50 minutes. In some sites of Posillipo and Ponticelli quarters, the measurements have been also done during the night.

For each investigated site the measured velocities have been compared with the more restrictive threshold values proposed by the UNI9916 (the minimum values for each edifice class) on the basis of the building typology of the destination use that characterizes the site.

Signal analysis for seismic site effect

The signal analysis has been performed with a 21 s time window. For each window the Fourier amplitude spectra and the H/V ratio (ratio between the mean of the spectral components in the horizontal plane and the vertical component). The H/V  ratio has been averaged over the whole registration (at least 50 minutes) and the average value and the standard deviation have been computed.