Ground seismic noise can be used to identify:

1. the resonance frequencies of buildings

2. the resonance frequencies of subsoils

in a passive, non-intrusive, fast and cheap way.

Seismic noise exists everywhere on the Earth surface. It mainly consists in surface waves, which are the elastic waves produced by the constructive interference of the P and S waves in the layers near the surface. Seismic noise is mostly produced by wind and sea waves. Also industries and vehicle traffic locally generate seismic noise, although essentially at high frequencies (some Hz), which are quickly attenuated. In areas without local noise sources, in absence of wind and on flat rocky basements, the seismic noise spectrum is as shown in figure 1, where the blue and the green curves are respectively the 'minimum' and 'maximum' standard models of the United States Geological Survey for the seismic 'ground' noise.

Figure 1. The power spectra (in terms of velocity, vertical component) of the maximum and minimum standard seismic noise models of the United States Geological Survey (J. Peterson, Observations ond modelling of background seismic noise, Open-file report 93-322, USGS, 1993).

The noise spectrum decreases at high frequencies and has two peaks at 0.14 and 0.07 Hz, probably due to large ocean waves, which are only slightly attenuated even at thousands of kilometers from the ocean due to the wave-guide like properties of surface waves.

Local effects due to anthropic and natural sources add to this general trend. For example, if in the vicinity of an industrial plant with a proper frequency of 7 Hz, this will show up in the noise spectrum as a peak. In this case the noise source is of active type.

Much more interesting is the fact that seismic ground noise acts as an excitation function for the specific resonances of both buildings and subsoil, much like a white light illuminates objects exciting the wavelength of their own color. For example, if a building has resonance frequencies at 6 and 100 Hz, the ground noise will excite these frequencies making them clearly visible on the noise spectrum. In exactly the same way, seismic noise will excite the resonance frequencies of the subsoil, which will show up in the spectrum. For example, if the subsoil has proper frequencies of 0.8 and 20 Hz, these will appear in the spectrum as in figure 2.

Something similar happens during an earthquake, which can be thought as an episode of extremely high noise, with amplitudes up to 10¹⁰ times higher than those of the ground noise. In this case, if the soil resonance frequency is the same as that of a building on that soil, a coupled resonance will be induced. The latter, which greatly increases the amplitude of the stresses on the building, is called seismic amplification.

Seismic amplification is the first cause of earthquake damage, more important than the size of the earthquake itself. A notable example of this in recent years has been the comparatively modest (M=6.6) earthquake which stroke Central Mexico on September 19, 1985. It produced only light damage in the epicentral area, but caused the collapse of 400 buildings and the damage of many more in Mexico City, 240 km from the epicenter. The following seismic noise analyses showed that Mexico City is built on a sedimentary basin with a proper resonance frequency of about 1 Hz, which is the same as that of 10 storey buildings, which were the ones most affected by the earthquake. Similarly, the December 28, 1908, M=7.2, N-E Sicily earthquake destroyed 95% of the buildings in Messina but left practically intact all those built on solid rock, on which no seismic amplification occurs.

In conclusion, ground seismic noise can be profitably used to identify the resonance frequencies of buildings and subsoils in a passive, non-intrusive, fast and cheap way. This type of analysis is getting more and more popular.