Acoustic-seismic coupling of broadband signals – support for on-site inspections under the Comprehensive Nuclear Test-Ban-Treaty

Abstract

In this PhD thesis the process of acoustic-seismic coupling and a method to reduce acoustically induced soil vibrations by applying an acoustic shielding to seismic sensors are investigated. The research is motivated by the verification of the Comprehensive Nuclear-Test-Ban Treaty: During on-site inspections sensitive seismic measurements can be performed to record seismic aftershocks created in the aftermath of a large, underground, human-caused explosion. This aims to precisely localise the hypocentre of that explosion to verify whether its origin was a nuclear or chemical one. However, these seismic measurements can be disturbed by other seismic signals in the inspection area and thus weak aftershock signals might go undetected. In this work disturbances caused by airborne sources are analysed: When sound waves hit the ground they excite soil vibrations which can mask weak aftershock signals. The findings of this work can be used in the development of new guidelines to improve the sensitivity of seismic on-site inspection measurements. For this measurements of sound pressure and the acoustically excited soil velocity, recorded with geophones places at the soil surface and in burying depths of up to 0.6 m, in flat terrain with sandy soil, are presented. The acoustic excitation was realised by broadband sound of jet-aircraft overflights, covering a broad range of incidence angles and due to the large distance arriving as plane waves, and of noise artificially produced by a speaker. By evaluating a multitude of overflight events it is shown that the acoustic-seismic coupling coefficient (i.e. the spectral ratio between soil velocity and sound pressure) only depends on the angle of incidence of the acoustic wave and the frequency. Thus, angle-dependent averaging of the coupling coefficient, obtained from the signals of several overflight events, can be performed which significantly improves the signal-to-noise ratio. While previous publications presented only pointwise measurements in this work results for a wide range of angles of incidence and frequencies are presented. In seismic spectra signal frequency bands of increased and decreased soil velocity are observed. These are caused by interfering seismic waves: The directly acoustically excited wave and waves which have been acoustically excited in a certain horizontal distance and which have been reflected within the ground before reaching the sensor. The seismic response to the broadband acoustic excitation with a range of incidence angles is used to obtain near-surface soil properties e.g. the P-wave velocity and the depth of the reflecting boundary. For this three theoretical models are introduced taking into account contributions of different numbers and types of the interfering waves. While sensors placed at the surface generally lead to the most reliable results, buried sensors are used to verify the models. Additionally, during several measurements an acoustic shielding is placed over some sensors. Thus, the sound pressure of the incident acoustic waves is reduced significantly. From the soil velocity recorded with the shielded geophones and the reflection characteristics the horizontal propagation range of acoustically induced seismic signals is estimated. It is shown that treating the process of acoustic-seismic coupling as a local effect is an insufficient approach. Finally, first suggestions for acoustic shieldings to reduce disturbing signals during sensitive seismic measurements are presented and the applicability during on-site inspections is discussed. An outlook for further research is given: Design and material of a suitable acoustic shielding should be investigated to fulfil requirements of high acoustic damping properties for signals of frequencies of a few Hz.