Using Ambient Vibration Array Techniques for Site Characterisation

Monday, Nov. 21st 2005 to Friday, Nov. 25th 2005
Université Joseph Fourier Grenoble, France

In recent years, the number of research papers dealing with ambient vibration analysis methods has increased considerably. Clearly, the interest in these methods originates from both the economical attractive cost benefit ratio and the straightforward data acquisition. Being a non-destructive passive technique, these methods also complement geotechnical and/or active geophysical methods within highly populated regions.

Within the SESAME project (''Site EffectS assessment using AMbient Excitation'', EU-Project EVG1-CT-2000-00026) detailed research has been accomplished to identify the capabilities and limitations of ambient vibration analysis techniques. Within this context a set of software tools has been developed to ease the processing and interpretation of ambient vibration wave field recordings.

The main findings of the SESAME project however show, that ambient vibration array analysis techniques have to be applied with care. It requires not only a careful measurement, but especially the interpretation and the inversion of analysis results need a (self-)critical review. From these conclusions it has been agreed among SESAME-partners, that it is highly necessary to distribute the software tools together with adequate training.

The main purpose of this course is to achieve the necessary understanding among the course participants for the problems related to these techniques. By doing so, we try to enable a correct usage of the software tools and to avoid misuse of these methods (''black-box usage''). The course fees are intended to support current and future improvements of the software package, distributed under an open source and free license.


organised by LGIT and IGUP
hosted by UJF-Grenoble
presented by Cécile Cornou, Matthias Ohrnberger, Marc Wathelet
supported by SESAME (EU-Project EVG1-CT-2000-00026) and Sismovalp (Interreg IIIB, Alpine Space)

Course outline

This is a preliminary program which may be subject to slight changes in the future. This document is also available as pdf.

9:00-9:30 Reception and welcome  
  • Technical issues and program overview
9:30-10:30 Physical background of ambient vibrations Lecture
  • Overview of the physical background of ambient vibration studies.
  • Basic assumptions and state of the art about the nature of the ambient vibration wave field.
  • The linkage between subsurface structure and wave field propagation properties.
  • The use of ambient vibration array analysis for observation and recovery of site characteristics (site response).
10:30-10:45 Coffee break
10:45-12:15 Single station measurement, H/V Lecture
  • Interpretations for horizontal to vertical spectral ratio technique (H/V).
  • Relationships between Rayleigh ellipticity and SH transfer function.
  • Effects of experimental conditions.
12:15-14:00 Lunch break
14:00-15:30 Signal database software (GEOPSY) Exercises
  • In depth description of the signal database software, GEOPSY.
  • Training with a step by step tutorial to create and manage databases.
  • Use of the standard tools for data visualisation, manipulation and standard signal processing.
15:30-15:45 Coffee break
15:45-17:15 Single station measurement, H/V Exercises
  • Geopsy tools for H/V interpretation.
9:00-10:30 Basic array processing concepts (frequency wavenumber, f-k) Lecture
  • Overview array processing methods.
  • The intuitive shift-and-sum technique in time domain.
  • Transposition into the frequency wavenumber (f-k) domain.
10:30-10:45 Coffee break
10:45-12:15 Array geometry and f-k response Lecture
  • Discrete spatial sampling of the continuous seismic wave field( by small groups of seismic stations - arrays) : concepts of resolution and aliasing.
  • Resolution and aliasing of simple linear array layouts.
  • Generalisation to bi-dimensional array settings.
  • Relation between array geometry and resulting estimation capabilities: rules of thumb and the array response function.
12:15-14:00 Lunch break
14:00-15:30 Array geometry and f-k response Exercises
  • Design of array geometry with BUILD_ARRAY tool: optimum usage and functionalities.
  • Building experience on how modifications in array layout affect the theoretical array performances.
15:30-15:45 Coffee break
15.30-17:15 Conventional f-k processing Exercises
  • Application of the conventional f-k algorithm to estimate the dispersion curves from the synthetic signals of the data base example created in the previous exercise
  • From the discussion of the analysis results, new insight into dependencies between the observability of dispersion curves, array geometries and signal spectra (high pass filter effect and H/V).
9:00-10:30 High resolution f-k (Capon's method) Lecture and exercise
  • Background of a widely used method, the high resolution f-k method after Capon (1969).
  • Application of the method to the synthetic data set, previously mentioned.
  • Comparison with the conventional f-k method.
10:30-10:45 Coffee break
10:45-12:15 Spatial autocorrelation method (SPAC) Lecture and exercise
  • Background of the spatial autocorrelation technique (SPAC) introduced by Aki (1957).
  • Ongoing developments around Aki's original work.
  • Application of the method to the synthetic data set, previously mentioned.
  • Discussion of differences with f-k techniques.
12:15-14:00 Lunch break
14:00-15:30 Array analysis of real data sets Exercises
  • Investigation of a real data set with the goal to derive the site dispersion curve.
15:30-15:45 Coffee break
15:45-17:15 Dispersion curve (DC) inversion Lecture
  • Short introduction to the fundamentals of inversion theory.
  • Direct search methods based on a random sampling, and particularly the Neighbourhood algorithm.
  • Specific issues solved for the inversion dispersion curves: robustness of the forward computation, non-uniqueness of the final solution.
20:00-... Workshop Dinner
9:00-10:30 DC inversion Exercises
  • Inversion of theoretical dispersion curve in order to identify critical issues.
  • Model parameterisation and parameter range (vP, vS), a natural way to introduce prior information and to reduce the non-uniqueness.
10:30-10:45 Coffee break
10:45-12:15 DC inversion Exercises
  • Inversion the analysis results which have been obtained from f-k (standard and high-resolution) and SPAC computations the day before.
  • Combining the results of arrays with distinct apertures and geometries: decision on the usable frequency range for the inversion target.
12:15-14:00 Lunch break
14:00-15:30 Discussion of inversion results Moderated discussion
  • Comparison of the inversion results obtained by the course participants with the true underlying model.
  • Discussion of the causes and the significance of apparent discrepancies between obtained inversion results.
15:30-15:45 Coffee break
15:45-17:15 ... Array analysis and DC inversion of test data sets I Exercises
  • Blind test experiment performed among the course attendees on a synthetic case.
  • Choice of the array layouts, analyse the selected records with f-k and SPAC techniques and finally derive the underlying velocity model.
9:00-10:30 Discussion of results from data sets Moderated discussion
  • Five minute presentations of the individual results for both synthetic and real data examples.
  • General discussion of experience gained from the different data sets (synthetic and real cases).
10:30-10:45 Coffee break
10:45-12:15 Course summary Lecture
  • Summary of the various steps involved for the task of ambient vibration array analysis and site characterisation by the inversion of one-dimensional earth models.
  • Overview of ambient vibration array techniques around the world.
  • Recall of current issues and limitations of using ambient vibrations to site characterisation.
12:15-14:00 Lunch break
14:00-... Departure of Participants Open discussion

Suggested readings

A paper copy of these publications will be sent to the attendees before the beginning of the course.

All deliverables and reports of the SESAME project are available here (especially, D18.06, D19.06, D24.13 may be of interest).

Registration and Fees

To offer the best attention to each of the course attendees, the maximum number of participants is fixed to 15. A minimum of 8 registrations is required for the organisation of this cession.

Online registration is open until Thrusday October 15th 2005. If the number of applicants is greater than 15, the organisers reserve their rights to select the participants. They will be informed of their selections on Monday October 19th.

The fees include lectures, practical exercises, lunches, workshop dinner and coffee breaks. One computer per participant will be made available during the exercises. Profits will be devoted to the development and improvement of the SESARRAY softwares.

Organisation Number of participants Price
(euros, excluding VAT)
SESAME partners 1 1000
  2 1500
  n > 2 1000+n*500
Sismovalp 1 1000
  2 1500
  n > 2 1000+n*500
Others 1 2500
  2 4000
  n > 2 4000+(n-2)*1000

Special discounts may be offered if really needed, please contact Pierre-Yves Bard (pierre-yves.bard at


Marc Wathelet marc at
Matthias Ohrnberger mao at

Photo gallery

You can find all photos taken during this first edition here (thanks to Sylvette Bonnefoy, Mehmet Ergin, Fabrizio Cara and Carlo Cauzzi).