HV User Guidelines

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/320719343
Guidelines for the implementation of the H/V spectral ratio technique on

ambient vibrations measurements, processing and interpretation
Book · December 2004
CITATIONS READS
76 682
85 authors, including:
Anastasios Anastasiadis Kuvvet Atakan

Aristotle University of Thessaloniki University of Bergen
92 PUBLICATIONS 1,689 CITATIONS 102 PUBLICATIONS 1,899 CITATIONS
SEE PROFILE SEE PROFILE
Riccardo Mario Azzara Pierre-Yves Bard

National Institute of Geophysics and Volcanology Université Gustave Eiffel
95 PUBLICATIONS 1,659 CITATIONS 469 PUBLICATIONS 12,656 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
PRENOLIN View project
RETIS-Risk (2014-2015) View project
All content following this page was uploaded by Pierre-Yves Bard on 31 October 2017.
The user has requested enhancement of the downloaded file.

GUIDELINES FOR THE IMPLEMENTATION
OF THE H/V SPECTRAL RATIO
TECHNIQUE ON AMBIENT VIBRATIONS
MEASUREMENTS, PROCESSING AND
INTERPRETATION
SESAME European research project

WP12 – Deliverable D23.12
European Commission – Research General Directorate

Project No. EVG1-CT-2000-00026 SESAME
December 2004
SESAME H/V User Guidelines 17/03/05
SESAME: Site EffectS assessment using AMbient Excitations
European Commission, contract n° EVG1-CT-2000-00026
Co-ordinator: Pierre-Yves BARD

Administrative / Accounting assistance: Laurence BOURJOT
Project duration: 1st May 2001 to 31st October 2004
Project Web-site: http://sesame-fp5.obs.ujf-grenoble.fr/index.htm
Participating organisations:
Central Laboratory for Bridges and Roads – Paris, France
Centre of Technical Studies – Nice, France
Geophysical Institute Slovak Academy of Sciences – Bratislava, Slovakia
Institute of Earth and Space Sciences – Lisbon, Portugal
Institute of Engineering Seismology and Earthquake Engineering – Thessaloniki, Greece
National Centre for Scientific Research – Grenoble, France
National Institute of Geophysics and Volcanology – Roma, Italy
National Research Council – Milano, Italy
Polytechnic School of Zürich, Switzerland
Résonance Ingénieurs-Conseils SA – Geneva, Switzerland
University Joseph Fourier – Grenoble, France
University of Bergen, Norway
University of Liège, Belgium
University of Potsdam, Germany
List of participants:
Catello Acerra Giuseppe di Giulio Etor Querendez

Gerardo Aguacil Margaret Grandison Sandro Rao
Anastasios Anastasiadis Philippe Guéguen Julien Rey
Kuvvet Atakan Bertrand Guillier Gudrun Richter
Riccardo Azzara Ebrahim Haghshenas Johannes Rippberger
Pierre-Yves Bard Hans Havenith Mario la Rocca
Roberto Basili Jens Havskov Pedro Roquette
Etienne Bertrand Denis Jongmans Daniel Roten
Bruno Bettig Fortunat Kind Antonio Rovelli
Fabien Blarel Jörg Kirsch Gilberto Saccoroti
Sylvette Bonnefoy-Claudet Andreas Koehler Alekos Savvaidis
Paola Bordoni Martin Koller Frank Scherbaum
Antonio Borges Josef Kristek Estelle Schissele
Mathilde Bøttger Sørensen Miriam Kristekova Eva Spühler-Lanz
Laurence Bourjot Corinne Lacave Alberto Tento
Héloïse Cadet Alberto Marcellini Paula Teves-Costa
Fabrizio Cara Rosalba Maresca Nikos Theodulidis
Arrigo Caserta Bassilios Margaris Eirik Tvedt
Jean-Luc Chatelain Fabrizio Marra Terje Utheim
Cécile Cornou Peter Moczo Jean-François Vassiliades
Fabrice Cotton Bladimir Moreno Sylvain Vidal
Giovanna Cultrera Antonio Morrone Gisela Viegas
Rosastella Daminelli Jérôme Noir Daniel Vollmer
Petros Dimitriu Matthias Ohrnberger Marc Wathelet
François Dunand Jose Asheim Ojeda Jochen Woessner
Anne-Marie Duval Ivo Oprsal Katharina Wolff
Donat Fäh Marco Pagani Stratos Zacharopoulos
Lucia Fojtikova Areti Panou
Roberto de Franco Catarina Paz
2
FOREWORD
Site effects associated with local geological conditions constitute an important part of any
seismic hazard assessment. Many examples of catastrophic consequences of earthquakes
have demonstrated the importance of reliable analyses procedures and techniques in
earthquake hazard assessment and in earthquake risk mitigation strategies. Ambient
vibration recordings combined with the H/V spectral ratio technique have been proposed to
help in characterising local site effects. This document presents practical user guidelines and
software for the implementation of the H/V spectral ratio technique on ambient vibrations.
The H/V spectral ratio method is an experimental technique to evaluate some characteristics
of soft-sedimentary (soil) deposits. Due to its low-cost both for the survey and analysis, the
H/V technique has been frequently adopted in seismic microzonation investigations.
However, it should be pointed out that the H/V technique alone is not sufficient to
characterise the complexity of site effects and in particular the absolute values of seismic
amplification. The method has proven to be useful to estimate the fundamental period of soil
deposits. However, measurements and the analysis should be performed with caution. The
main recommended application of the H/V technique in microzonation studies is to map the
fundamental period of the site and help constrain the geological and geotechnical models
used for numerical computations. In addition, this technique is also useful in calibrating site
response studies at specific locations.
These practical guidelines recommend procedures for field experiment design, data
processing and interpretation of the results for the implementation of the H/V spectral ratio
technique using ambient vibrations. The recommendations given here are the result of a
consensus reached by the participants of the European research project SESAME (Contract.
No. EVG1-CT-2000-00026), and are based on comprehensive and detailed research work
conducted during three years.
It is highly recommended that prior to planning a measurement campaign on ambient

vibrations, a local geological survey, especially on Quaternary deposits, should be
performed. Interpretation of the H/V results will be greatly enhanced when combined with
geological, geophysical and geotechnical information.
In spite of its limitations, the H/V technique is a very useful tool for microzonation and site
response studies. This technique is most effective in estimating the natural frequency of soft
soil sites when there is a large impedance contrast with the underlying bedrock. The method
is especially recommended in areas of low and moderate seismicity, due to the lack of
significant earthquake recordings, as compared to high seismicity areas.
3
TABLE OF CONTENTS
INTRODUCTION................................................................................................................................................. 5
PART I: QUICK FIELD REFERENCE AND INTERPRETATION GUIDELINES .................................... 7
1. EXPERIMENTAL CONDITIONS + MEASUREMENT FIELD SHEET.............................................. 8
2. DIAGRAMS FOR INTERPRETATION OF H/V RESULTS................................................................ 10

2.1 CRITERIA FOR RELIABILITY OF RESULTS ............................................................................................... 10
2.2 MAIN PEAK TYPES ................................................................................................................................. 10
PART II: DETAILED TECHNICAL GUIDELINES1. TECHNICAL REQUIREMENTS .................... 15
1. TECHNICAL REQUIREMENTS ............................................................................................................ 16

1.1 INSTRUMENTATION ............................................................................................................................... 16
1.2 EXPERIMENTAL CONDITIONS ................................................................................................................ 17
2. DATA PROCESSING STANDARD: J-SESAME SOFTWARE ........................................................... 22
2.1 GENERAL DESIGN OF THE SOFTWARE .................................................................................................... 22
2.2 WINDOW SELECTION MODULE ............................................................................................................. 22
2.3 COMPUTING H/V SPECTRAL RATIO ....................................................................................................... 25
2.4 SHOWING OUTPUT RESULTS .................................................................................................................. 25
2.5 SETTING GRAPH PROPERTIES AND CREATING IMAGES OF THE OUTPUT RESULTS .................................... 26
3. INTERPRETATION OF RESULTS ........................................................................................................ 28
3.1 UNDERLYING ASSUMPTIONS ................................................................................................................. 28
3.2 CONDITIONS FOR RELIABILITY .............................................................................................................. 30
3.3 IDENTIFICATION OF F0 ........................................................................................................................... 31
3.3.1 Clear peak................................................................................................................................. 31
3.3.2 "Unclear" cases ........................................................................................................................ 32
3.4 INTERPRETATION OF F0 IN TERMS OF SITE CHARACTERISTICS ................................................................ 35
ACKNOWLEDGEMENTS................................................................................................................................ 36
REFERENCES.................................................................................................................................................... 37
APPENDIX A: H/V DATA EXAMPLES ......................................................................................................... 40

A.1 ILLUSTRATION OF THE MAIN PEAK TYPES .............................................................................................. 40
A.2 COMPARISON WITH STANDARD SPECTRAL RATIOS ................................................................................ 48
APPENDIX B: PHYSICAL EXPLANATIONS............................................................................................... 54
B.1 NATURE OF AMBIENT VIBRATION WAVEFIELD ...................................................................................... 54
B.2 LINKS BETWEEN WAVE TYPE AND H/V RATIO ....................................................................................... 56
B.3 CONSEQUENCES FOR THE INTERPRETATION OF H/V CURVES ................................................................ 59
4
INTRODUCTION
A significant part of damage observed in destructive earthquakes around the world is

associated with seismic wave amplification due to local site effects. Site response analysis is
therefore a fundamental part of assessing seismic hazard in earthquake prone areas. A
number of experiments are required to evaluate local site effects. Among the empirical
methods the H/V spectral ratios on ambient vibrations is probably one of the most common
approaches. The method, also called the „Nakamura technique“ (Nakamura, 1989), was first
introduced by Nogoshi and Igarashi (1971) based on the initial studies of Kanai and Tanaka
(1961). Since then, many investigators in different parts of the world have conducted a large
number of applications.
An important requirement for the implementation of the H/V method is a good knowledge of
engineering seismology combined with background information on local geological conditions
supported by geophysical and geotechnical data. The method is typically applied in
microzonation studies and in the investigation of the local response of specific sites. In the
present document, the application of the H/V technique in assessing local site effects due to
dynamic earthquake excitations, is the main focus, whereas other applications regarding the
static aspects are not considered.
In the framework of the European research project SESAME (Site Effects Assessment Using
Ambient Excitations: Contract No. EVG1-CT-2000-00026), the use of ambient vibrations in
understanding local site effects has been studied in detail. The present guidelines on the H/V
spectral ratio technique are the result of comprehensive and detailed analyses performed by
the SESAME participants during the last three years. In this respect, the guidelines represent
the state-of-the-art of the present knowledge of this method and its applications, and are
based on the consensus reached by a large group of participants. It reflects the synthesis of
a considerable amount of data collection and subsequent analysis and interpretations.
In general, due to the experimental character of the H/V method, the absolute values
obtained for a given site require careful examination. In this respect visual inspection of the
data both during data collection and processing is necessary. Especially during the
interpretation of the results there should be frequent interaction with regard to the choices of
the parameters for processing.
The guidelines presented here outline the recommendations that should be taken into
account in studies of local site effects using the H/V technique on ambient vibrations. The
recommendations given apply basically for the case where the method is used alone in
assessing the natural frequency of sites of interest and are therefore based on a rather strict
set of criteria. The recommended use of the H/V method is however, to combine several
other geophysical and geotechnical approaches with sufficient understanding of the local
geological conditions. In such a case, the interpretation of the H/V results can be improved
significantly in the light of the complementary data.
The guidelines are organised in two separate parts; the quick field reference and
interpretation guidelines (Part I) and detailed technical guidelines (Part II). Part I aims to
summarise the most critical factors that influence the data collection, analysis and
interpretation and provides schematic recommendations on the interpretation of results. Part
II includes a detailed description of the technical requirements, standard data processing and
the interpretation of results. Several examples of the criteria described in Part I and II are
given in Appendix A. In addition, some physical explanations of the results based on
theoretical considerations are given in Appendix B. In Part II, section 1, the results of the
5
experiments performed within the framework of the SESAME project are given in smaller
fonts to separate these from the recommendations and the explanations given in the
guidelines. The word „soil“ should be considered as a generic term used throughout the text
to refer to all kinds of deposits overlying bedrock without taking into account their specific
origin.
The processing software J-SESAME developed specifically for using in H/V technique, is
explained (provided on a separate CD accompanying the guidelines) in Part II. However, the
recommendations given in the guidelines are meant for general application of the method
with any other similar software. J-SESAME is provided as a tool for the easy implementation
of the recommendations outlined in this document. Regarding the processing of the data,
several options can be chosen, but the recommended processing options are provided as
defaults by the J-SESAME software.
6
PART I: QUICK FIELD REFERENCE AND INTERPRETATION

GUIDELINES
7
1. EXPERIMENTAL CONDITIONS + MEASUREMENT FIELD SHEET
Æ This sheet is only a quick field reference. It is highly recommended that the complete guidelines
be read before going out to perform the recordings. A field sheet is also provided on the next page.
This page, containing two identical sheets can be printed and be taken in the field.
Type of parameter Main recommendations

Recommended minimum recording
Minimum expected f0 [Hz] duration [min]
Recording duration 0.2 30'
0.5 20'
1 10'
2 5'
5 3'
10 2'
Æ Microzonation: start with a large spacing (for example a 500
m grid) and, in case of lateral variation of the results, densify the
Measurement spacing grid point spacing, down to 250 m, for example.
Æ Single site response: never use a single measurement point
to derive an f0 value, make at least three measurement points.
Æ level the sensor as recommended by the manufacturer.
Recording parameters Æ fix the gain level at the maximum possible without signal
saturation.
Æ set the sensor down directly on the ground, whenever
In situ soil-sensor coupling possible.
Æ avoid setting the sensor on "soft grounds" (mud, ploughed
soil, tall grass, etc.), or soil saturated after rain.
Æ avoid plates from "soft" materials such as foam rubber,
cardboard, etc.
Artificial soil-sensor coupling Æ on steep slopes that do not allow correct sensor levelling,
install the sensor in a sand pile or in a container filled with sand.
Æ on snow or ice, install a metallic or wooden plate or a
container filled with sand to avoid sensor tilting due to local
melting.
Æ Avoid recording near structures such as buildings, trees, etc.
in case of wind blowing (faster than approx. 5 m/s). It may
Nearby structures strongly influence H/V results by introducing some low
frequencies in the curves
Æ Avoid measuring above underground structures such as car
parks, pipes, sewer lids, etc.
Æ Wind: Protect the sensor from the wind (faster than approx. 5
m/s). This only helps if there are no nearby structures.
Weather conditions Æ Rain: avoid measurements under heavy rain. Slight rain has
no noticeable influence.
Æ Temperature: check sensor and recorder manufacturer's
instructions.
Æ Meteorological perturbations: indicate on the field sheet
whether the measurements are performed during a low-pressure
meteorological event.
Æ Monochromatic sources: avoid measurements near
construction machines, industrial machines, pumps, generators,
Disturbances etc.
Æ Transients: In case of transients (steps, cars,...), increase the
recording duration to allow for enough windows for the analysis,
after transient removal.
8
1 SESAME HH/V User G uideliines 17/03/05 MEASUREMENT FIELD SHEET
9
1 SESAME HH/V User G uidelines 17/03/05
2. DIAGRAMS FOR INTERPRETATION OF H/V RESULTS
This section presents diagrams with criteria and recommendations to help in the result
interpretation for different cases. For detailed explanations of each case, see section 3 in
Part II of the guidelines. See Appendix A for illustrations. The definitions given in the table
below are valid for all section 2.
2.1 Criteria for reliability of results
Criteria for a reliable H/V curve • lw = window length

• nw = number of windows selected for the average H/V curve
i) f0 > 10 / lw
• nc = lw . nw. f0 = number of significant cycles
and • f = current frequency
ii) nc (f0) > 200 • fsensor = sensor cut-off frequency
• f0 = H/V peak frequency
and • σf = standard deviation of H/V peak frequency (f0 ± σf)
iii) σA(f)<2 for 0.5f0<f<2f0 if f0>0.5Hz • ε (f0) = threshold value for the stability condition σf < ε(f0)
or σA(f)<3 for 0.5f0<f<2f0 if f0<0.5Hz • A0 = H/V peak amplitude at frequency f0
• AH/V (f) = H/V curve amplitude at frequency f
• f- = frequency between f0/4 and f0 for which AH/V(f-) < A0/2
Criteria for a clear H/V peak • f+ = frequency between f0 and 4f0 for which AH/V(f+) < A0/2
(at least 5 out of 6 criteria fulfilled) • σA (f) = "standard deviation" of AH/V (f), σA (f) is the factor by
i) ∃ f- ∈ [f0/4, f0] | AH/V(f-) < A0/2 which the mean AH/V(f) curve should be multiplied or divided
• σlogH/V (f) = standard deviation of the logAH/V(f) curve, σlogH/V (f)
ii) ∃ f+ ∈ [f0, 4f0] | AH/V(f+) < A0/2 is an absolute value which should be added to or subtracted
iii) A0 > 2 from the mean logAH/V(f) curve
• θ (f0) = threshold value for the stability condition σA(f) < θ(f0)
iv) fpeak[AH/V(f) ± σA(f)] = f0 ±5% • Vs,av = average S-wave velocity of the total deposits
v) σf < ε(f0) • Vs,surf = S-wave velocity of the surface layer
• h = depth to bedrock
vi) σA(f0) < θ (f0) • hmin = lower-bound estimate of h
Threshold Values for σf and σA(f0)

Frequency range [Hz] < 0.2 0.2 – 0.5 0.5 – 1.0 1.0 – 2.0 > 2.0
ε (f0) [Hz] 0.25 f0 0.20 f0 0.15 f0 0.10 f0 0.05 f0
θ (f0) for σA (f0) 3.0 2.5 2.0 1.78 1.58
log θ (f0) for σlogH/V (f0) 0.48 0.40 0.30 0.25 0.20
2.2 Main peak types
This section is not an exhaustive list of the different types of H/V curves that might be
obtained, but it gives suggestions for the processing and interpretation of the most common
situations.
10
Sharp peaks and industrial origin see $3.3.2.d and appendix A
If industrial origin is proved, i.e., if

♦ raw spectra exhibit sharp peaks on all
components
♦ random decrement technique indicates no reliable f0
very low damping (< 5%)
♦ peaks get sharper with decreasing
smoothing
♦ Perform continuous recordings during

day and night
If industrial origin is not obvious ♦ Check the existence of 24 h / 7 day
plant within the area
Clear peak see $3.3.1 and appendix A
If industrial origin ( $3.3.2.d ) , i.e.

♦ the raw spectra exhibit sharp peaks on all three
components no reliable f0
♦ the random decrement technique indicates very
low damping (< 5%)
♦ peaks get sharper with decreasing smoothing
f0 reliable
♦ Likely sharp contrast at depth (A0 > 4-5)

If f0 > fsensor and no industrial origin ♦ f0 = fundamental frequency
♦ If h is known then VS,av ~ 4h.f0
♦ If VS,surf is known then hmin ~ VS,surf/4f0
11
Unclear low frequency peak see $3.3.2.b and appendix A
If rock site no reliable f0
If steady increase of H/V ratio with decreasing frequency

♦ Check H/V curves from individual windows and eliminate
windows giving spurious H/V curves
♦ Use longer time windows and/or more stringent window
selection criteria
♦ Use proportional bandwidth and less smoothing
If reprocessed H/V curve fulfils the clarity criteria f0 reliable
Perform additional measurements over a

If reprocessed H/V curve does not fulfil the clarity
longer time and/or during night and/or quiet
criteria
weather conditions and/or use earthquake
recordings using also a nearby rock site
Broad peak see $3.3.2.c and appendix A
Decrease the smoothing bandwidth
If bump peak is not stable no reliable f0

and/or AH/V(f) is very large
♦ If clearer peaks are observed in the vicinity

If bump peak is stable and and
AH/V(f) is rather low - if their related frequencies lie within the
frequency range of the broad peak
- if their related frequencies exhibit significant
variation from site to site
Then, examine the possibility of underground
lateral variation, especially slopes.
♦ Otherwise not recommended to extract
any information
12
Multiple peaks (multiplicity of maxima) see $3.3.2.c and Appendix A
Check no industrial origin of one of the peaks

Increase the smoothing bandwidth
If reprocessed H/V curve fulfils the clarity criteria f0 reliable
If reprocessed H/V curve does not fulfil the clarity Redo measurements over a longer time and/or
criteria during night
2 peaks cases (f1 > f0) see $3.3.2.d and appendix A
If industrial origin of f0 f0 not reliable and f1 reliable

AND
No industrial origin for f1 ♦ Likely sharp contrast at shallow depth
♦ f1 = fundamental frequency
♦ If h is known then VS,av ~ 4h.f1
♦ If VS,surf is known then hmin ~ VS,surf/4f1
If industrial origin of f1 f0 reliable and f1 not reliable

AND
No industrial origin for f0 ♦ Likely sharp contrast at rather large depth
♦ f0 = fundamental frequency
♦ If h is known then VS,av ~ 4h.fo
♦ If VS,surf is known then hmin ~ VS,surf/4fo
f0 and f1 reliable
If the geology of the site shows possibility of
having two large velocity contrasts at two ♦ Likely two large contrasts at shallow and
different scales large depth at two different scales
AND ♦ f0 = fundamental frequency
The clarity criteria are fulfilled for both f0 ♦ f1 = other natural frequency
and f1 ♦ If VS,surf is known then h1,min ~ Vs,surf/4f1
13
Flat H/V curve (meeting the reliability conditions) see $3.4 and appendix A
if soil deposits ♦ Likely absence of any sharp contrast

at depth
♦ Does not necessarily mean no
amplification
♦ Perform earthquake recordings on site
and nearby rock site
♦ Use of other geophysical techniques
♦ likely unweathered or only lightly

if rock weathered rock
♦ may be considered as a good
reference site
14
PART II: DETAILED TECHNICAL GUIDELINES
15
1. TECHNICAL REQUIREMENTS
It is important to understand which recording parameters influence data quality and reliability
as this can help speed up the recording process. H/V measurements in cities are conducted
within the following context:
– Anthropic noise is very high.
– It is quite rare to be able to get data on the soil per se. Most data will be obtained on
streets and sidewalks (i.e. asphalt, pavement, cement or concrete), and to a lesser extent
in parks (i.e. on grass or soil).
– Measurements are performed in an environment dominated by buildings of various
dimensions.
– Recordings are not always performed at the same time and under the same weather
conditions.
– The presence of underground structures (i.e. pipes,...) is often unknown.
The influence of various types of experimental parameter had to be tested on the results of
H/V curves both in frequency and amplitude. For each tested parameter, H/V data were compared
with a "reference situation". This comparison had to be made in an objective way, i.e. with the use of a
statistical method. The Student-t test was chosen as it is one of the most commonly used techniques
for testing an hypothesis on the basis of a difference between sample means. It determines a
probability that two populations are the same with respect to the variable tested. The t-test can be
performed knowing only the means, standard deviation, and number of data points. For further details,
or for users who would like to perform some comparison on their own, please refer to the SESAME
WP02 Controlled instrumentation specification, Final report. Further investigations would be welcome
in some cases using a common software process and processing parameters to compare records and
quantify their similarity.
This chapter is based on the following SESAME internal reports:

– WP02 Controlled instrumentation specification, Final report, Deliverable D01.02,
– WP02 H/V technique : experimental conditions, Final report, Deliverable D08.02.
1.1 Instrumentation
An instrument workshop was held during the SESAME project to investigate the influence of different
instruments in using the H/V technique with ambient vibration data. There were four major tasks
performed, which consisted of testing the digitisers and sensors, and of making simultaneous
recordings both outside in the free field and at the lab for comparisons.
Influence of the digitiser

In order to investigate the possible influence of the digitisers, several tests were performed to quantify
the experimental sensitivity, internal noise, stability with time and channel consistency. The influence
of various parameters was checked (warm up time, electronic noise, synchronism between channels,
difference of gain between channels, etc.):
Æ Most of the tested digitisers finally gave correct results.
Æ In general it was found that all digitisers need some warm up time. From 2 to 10 minutes is
usually sufficient for most instruments to assure that the baseline is more or less stable. Users
are encouraged to test instrument stability before use.
Æ For an optimal analysis of the H/V curve, it appears to be necessary to check the energy
density along the studied frequency band, to ensure that the energy is sufficient to allow
the extraction of the signal from the instrumental noise. Furthermore, deviations must be
the same on all components.
Æ Users should check synchronisation between channels: From numerical simulation, it
was demonstrated that no H/V modification occurs below 15 Hz. Over 15 Hz
modifications in the H/V results depend on the percentage of sample desynchronisation,
the minimum being obtained for a round number of desynchronised samples and the
maximum (up to 80% of H/V amplitude change) for 0.5 sample.
Æ Users should choose the same gain for all three channels. Small gain differences might
cause slight changes to the results.
16
Influence of the sensor

The influence of the type of sensor was tested by recording simultaneously with two sensors coupled
to the same digitiser. In total 17 sensors were tested. In general, signals are similar, as expected.
However,
Æ The accelerometers were not sensitive enough for frequencies lower than 1 Hz and give
very unstable H/V results. It is not recommended that H/V measurements be performed
using seismology accelerometers, as they are not sensitive enough for ambient vibration
levels.
Æ Stability is important. It is not recommended that H/V measurements be performed using
broadband seismometers (with natural period higher than 20 s), as they require a long
stabilisation time, without providing any further advantage. Users are encouraged to test
sensor stability before use.
Æ It is not recommended that sensors that have their natural frequency above the lowest
frequency of interest be used. If f0 is lower than1 Hz, while the sensor used is a high
frequency velocimeter, double check the results with the procedure indicated in 3.3.2.b.
1.2 Experimental conditions

As a general recommendation, it is suggested that before doing H/V measurements on the
field, the team should have a look at the available geological information about the study
area. In particular, the types of geological formations, the possible depth to the bedrock, and
possible 2D or 3D structures should be looked at.
An evaluation of the influence of experimental conditions on the stability and reproducibility of H/V
estimations from ambient vibrations was carried out during the SESAME project [2]. The results
obtained are based on 593 recordings used to test 60 experimental conditions that can be divided into
categories, as following:
– recording parameters,
– recording duration,
– measurement spacing,
– in-situ soil-sensor coupling,
– artificial soil-sensor coupling,
– sensor setting,
– nearby structures,
– weather conditions,
– disturbances.
Recording parameters
Æ As long as there is no signal saturation, results are equivalent irrespective of the gain
used. However, we suggest fixing the gain level at the maximum possible without signal
saturation. The only noticeable effect is a compression of the H/V curve when too high a
gain value is used implying too much saturation of the signal.
Æ A sampling rate of 50 Hz is sufficient, as the maximum frequency of engineering interest
is not higher than 25 Hz, although higher sampling rates do not influence H/V results.
Æ Length of cable to connect the sensor to the station does not influence H/V results at
least up to a length of 100 meters.
Recording duration
Æ In order for a measurement to be reliable, we recommend the following condition to be
fulfilled : f0 > 10 / lw. This condition is proposed so that, at the frequency of interest, there
be at least 10 significant cycles in each window (see Table 1).
Æ A large number of windows and of cycles is needed: we recommend that the total
number of significant cycles : nc = lw . nw .f0 be larger than 200 (which means, for
17
instance, that for a peak at 1 Hz, there be at least 20 windows of 10 s each; or, for a
peak at 0.5 Hz, 10 windows of 40 s each, or 20 windows of 20 s, but not 40 windows of
10 s). See Table 1 for other frequencies of interest.
As there might be some transients during the recording, these should be removed from the
signal for processing, and the total recording duration should be increased, in order to have
the above mentioned conditions fulfilled for "good quality" signal windows.
Table 1. Recommended recording duration.

Minimum number Minimum Recommended
Minimum value Minimum useful
f0 [Hz] of significant number of minimum record
for lw [s] signal duration [s]
cycles (nc) windows duration [min]
0.2 50 200 10 1000 30'
0.5 20 200 10 400 20'
1 10 200 10 200 10'
2 5 200 10 100 5'
5 5 200 10 40 3'
10 5 200 10 20 2'
Measurement spacing
Æ For a microzonation, it is recommended that a large spacing be initially adopted (for
example a 500 m grid) and, in case of lateral variation of the results, to densify the grid
point spacing, down to 250 m, for example.
Æ For a single site response study, one should never use a single measurement point to
derive an f0 value. It is recommended that at least three measurement points be used.
In situ soil-sensor coupling

In situ soil / sensor coupling should be handled with care. Concrete and asphalt provide good
results, whereas measuring on soft / irregular soils such as mud, grass, ploughed soil, ice,
gravel, uncompacted snow, etc., should be looked at with more attention.
Æ To guarantee a good soil / sensor coupling the sensor should be directly set up on the
ground, except in very special situations (steep slope, for example) for which an
interface might be necessary (see next section).
Æ Topping of asphalt or concrete does not affect H/V results in the main frequency band of
interest, although slight perturbations can be observed in the 7-8 Hz range, which do not
change the shape of the H/V curves. In the 0.2 – 20 Hz range, no artificial peaks are
observed. Tests should be performed on higher frequency sites in order to check the
influence of the asphalt thickness. See Figure 1 for a comparison with and without
asphalt, at the same site.
Æ Grass by itself does not affect H/V results, provided that the sensor is in good contact
with the ground and not, for example, placed unstably on the grass as can be the case
for tall grass folded under the sensor. In such cases, it is better to remove the high
grass before installing the sensor, or to dig a hole in order to install the sensor directly on
soil. Recording on grass when the wind is blowing can lead to completely perturbed
results below 1 Hz, as shown on Figure 2.
Æ Avoid setting the sensor on superficial layers of "soft" soils, such as mud, ploughed soil,
or artificial covers such as synthetic sport covering.
Æ Avoid recording on water saturated soils, for example after heavy rain.
18
Æ Avoid recording on superficial cohesionless gravel, as the sensor will not be correctly
coupled to the ground and strongly perturbed curves will be obtained. Try to find another
type of soil a few meters away, or remove the superficial gravel to find the firm ground
below, if possible.
Æ Recording on snow or ice can affect the results. In such cases, it is recommended that
the snow be compacted and the sensor installed on a metal or wood plate in order to
avoid sensor tilting due to local melting under the sensor legs. When recording in such
conditions, make sure that the temperature is within the equipment specifications given
by the manufacturer.
Figure 1. Comparison of the H/V curves obtained with and without asphalt, at the same site,
showing no significant effect of the asphalt layer.
Artificial soil-sensor coupling

When an artificial interface is needed between the ground and the sensor, it is highly
recommended that some tests be performed before doing the recordings in order to examine
a possible influence of the chosen interface.
Æ The use of a metal plate in-between the sensor and the ground does not modify the
results.
Æ In case of a steep slope that does not allow correct sensor levelling, the best solution is
to install the sensor on a pile of sand or in a plastic container filled with sand.
Æ In general, avoid "soft / non-cohesive" materials such as foam-rubber, cardboard, gravel
(whether in a container or not), etc., to help setting up the sensor. See Figure 3 for a
comparison with the sensor directly on the natural soil, or on a Styrofoam plate.
Sensor setting
Æ The sensor should be installed horizontally as recommended by the manufacturer.
Æ There is no need to bury the sensor, but it does not hurt if this is the case. It can be
useful however, to set-up the sensor in a hole (no need to fill it) about its own size in
order to get rid, for example, of the effect of a weak wind on grass. This would be
effective only if there are no structures nearby, such as buildings or trees that might also
induce some strong low frequency perturbations in the ground, due to the wind (see
below).
19
Æ Do not put any load on the sensor.
Figure 2. Comparison of the H/V curves obtained at the same site on grass with and without
wind (top), and in a hole, on asphalt (bottom) and again on grass with wind. This comparison
shows the strong effect of the wind combined with grass, whereas on asphalt or in a hole, the
wind has no significant effect (if far away from any structure).
20
Figure 3. Comparison of the H/V curves obtained with and without a Styrofoam plate under
the sensor, at the same site, showing a strong effect of the Styrofoam.
Nearby structures
Æ Users are advised that recording near structures such as buildings, trees, etc., may
influence the results: there is clear evidence that movements of the structures due to the
wind may introduce strong low frequency perturbations in the ground. Unfortunately, it is
not possible to quantify the minimum distance from the structure where the influence is
negligible, as this distance depends on too many external factors (structure type, wind
strength, soil type, etc.).
Æ Avoid measuring above underground structures such as car parks, pipes, sewer lids,
etc., these structures may significantly alter the amplitude of the vertical motion.
Weather conditions
Æ Wind probably has the most frequent influence and we suggest avoiding measurements
during windy days. Even a slight wind (approx. > 5 m/s) may strongly influence the H/V
results by introducing large perturbations at low frequencies (below 1 Hz) that are not
related to site effects. A consequence is that wind only perturbs low frequency sites.
Æ Measurements during heavy rain should be avoided, while slight rain has no noticeable
influence on H/V results.
Æ Extreme temperatures should be treated with care, following the manufacturer's
recommendations for the sensor and recorder; tests should be made by comparing night
/ day or sun / shadow measurements.
Æ Low pressure meteorological events generally raise the low frequency content and may
alter the H/V curve. If the measurements cannot be delayed until quieter weather
conditions, the occurrence of such events should be noted on the measurement field
sheet.
Disturbances
Æ No influence from high voltage cables has been noted.
Æ All kinds of short-duration local sources (footsteps, car, train,...) can disturb the results.
The distance of influence depends on the energy of the source, on the soil conditions,
etc., therefore it is not possible to give general minimum distance values. However, it
has generally been observed, for example, that ambient vibration sources with short
periods of high amplitude (e.g. fast highway traffic) influence H/V ratios if they are within
15-20 metres, but that more continuous sources (e.g. slow inner city traffic) only
influence H/V ratios when they are much closer. Our experience is that it is the
impulsiveness of the noise envelope that is crucial Therefore traffic is much less of a
problem in a city than it is close to a highway, for example. Users are anyway
encouraged to check recorded time series in the field when they perform measurements
in a noisy environment.
Æ Short-duration disturbances of the signal can be avoided during the H/V analysis by
using an anti-trigger window selection to remove the transients. A consequence is that
the time duration of the recordings should be increased in order to gather enough
duration of “quiet” signal (see sections 2 and 3), unless, for example, only one train has
perturbed the recordings.
Æ Avoid measurements near monochromatic sources such as construction machines,
industrial machines, pumps, etc.
Æ The recording team should not keep its car engine running during recordings.
21
2. DATA PROCESSING STANDARD: J-SESAME SOFTWARE
This chapter is based on the following SESAME internal reports:

– Multi-platform H/V processing software J-SESAME, Deliverable D09.03 [3],
– J-SESAME User Manual, Version 1.07 [4],
J-SESAME is a JAVA application for providing a user-friendly graphical interface for the H/V
spectral ratio technique, which is used in local site effect studies. The program uses the
functions of automatic window selection and H/V spectral ratio by executing external
commands. The automatic window selection and H/V process are standalone applications
developed in Fortran. J-SESAME is mainly a tool for organising the input data, executing
window selection and processing, and displaying the processing results. The software
operates in Unix, Linux, Macintosh and Windows environments.
Details concerning system requirements and installation procedure are given in the J-
SESAME user manual delivered with the software.
2.1 General design of the software

The general design of J-SESAME is based on a modular architecture. There are basically
four main modules:
– browsing module,
– window selection module,
– processing module,
– display module.
The main functionalities are integrated through a graphical user interface, which is part of the
browsing module. The display module is also tightly connected to the browsing module, as
there is close interaction between the two modules due to the integrated code development
in Java. Only two waveform data formats are accepted: GSE and SAF (SESAME ASCII
Format), see J-SESAME user manual for more details.
2.2 Window Selection Module

Windows can be selected automatically or manually. The manual window selection mode
can be used if the check-box labelled as “Manual window selection” (Figure 4) is active.
Besides manual selection directly from the screen, which is often the most reliable, but also
the most time consuming method, an automatic window selection module has been
introduced to enable the processing of large amounts of data. The objective is to keep the
most stationary parts of ambient vibrations, and to avoid the transients often associated with
specific sources (footsteps, close traffic). This objective is exactly the opposite of the usual
goal of seismologists who want to detect signals, and have developed specific "trigger"
algorithm to track the unusual transients. As a consequence, we have used here an
"antitrigger" algorithm, which is exactly the opposite: it detects transients but it tries to avoid
them.
The procedure to detect transients is based on a classical comparison between the short
term average "STA", i.e., the average level of signal amplitude over a short period of time,
denoted "tsta" (typically around 0.5 to 2.0 s), and the long term average "LTA", i.e., the
average level of signal amplitude over a much longer period of time, denoted "tlta" (typically
several tens of seconds).
22
In the present case, we want to select windows without any energetic transients: this means
that we want the ratio sta/lta to remain below a small threshold value smax (typically around
1.5 – 2) over a long enough duration. Simultaneously, we also want to avoid ambient
vibration windows with abnormally low amplitudes: we therefore also introduce a minimum
threshold smin, below which the signal should not fall during the selected ambient vibration
window.
There are also two other criteria that may be optionally used for the window selection:
♦ one may wish to avoid signal saturation – as saturation does affect the Fourier spectrum.
As gain and maximum signal amplitudes are not mandatory in the header of SAF and
GSE formats, the program looks for the maximum amplitude over the whole ambient
vibration recording, and automatically excludes windows during which the peak amplitude
exceeds 99.5 % of this maximum amplitude. By default, this option is turned on.
♦ in some cases, there might exist long transients (for instance related to heavy traffic,
trains, machines, …) during which the sta/lta will actually remain within the set limits, but
during which the ground motion may not be representative of real seismic ambient
vibrations. Another option was therefore introduced to avoid "noisy windows", during
which the lta value exceeds 80% of the peak lta value over the whole recording. By
default, this option is turned off.
Figure 4. This figure shows the graphical user interface of J-SESAME. Selected windows
are shown in green.
23
The program automatically looks for time windows satisfying the above criteria; when one
window is selected, the program looks for the next window, and allows two subsequent
windows to overlap by a specified amount "roverlap" (default value is 20%).
The window selection module has been written as an independent Fortran subroutine, for
which:
♦ The input parameters are the selection parameters (tsta = STA window length; tlta = LTA
window length; [smin, smax] = lower and upper allowed bounds for the sta/lta ratio; tlong
= ambient vibration window length over which the sta/lta should remain within the
bounds; yes/no (1/0) parameters for turning on or off the saturation and "noisy window"
options; overlapping percentage allowed for two subsequent windows)
♦ The output parameters are the name of the ambient vibration file, the start and end times
of each selected window, the recording status of each component : the main processing
module then reads the ambient vibration file, and performs the H/V computation over
each selected window.
Figure 5 shows an example where the same signal has been processed with and without the
automatic window selection (that is the transient removal). This example shows that the peak
on the H/V curve is much clearer when the transient removal is applied, and also that
standard deviation is lower, especially at low frequencies.
Figure 5. Signal (top) processed with (red curves bottom) or without (blue curves bottom) the
automatic window selection (selected windows are indicated by the red segments on the top
of the signal). The peak on the H/V curve is much clearer when the transient removal is
applied, and also the standard deviation is lower, especially at low frequencies.
24
The choice of the window selection parameters should be carefully optimised before any
automatic processing.
2.3 Computing H/V spectral ratio

The main processing module is developed in FORTRAN90. It performs H/V spectral ratio
computations and the other associated processing such as DC-offset removal, filtering,
smoothing, merging of horizontal components, etc., on the selected windows for individual
files or alternatively on several files as a batch process. The instrument response is assumed
to be removed by the user (in the case of identical components H/V ratios should not be
affected by the instrument response). Main functionalities of the processing module are
described below:
♦ FFT (including tapering)
♦ Smoothing with several options. The Konno & Ohmachi smoothing is recommended as it
accounts for the different number of points at low frequencies. Be careful with the use of
constant bandwidth smoothing, which is not suitable for low frequencies.
♦ Merging two horizontal components with several options. Geometric mean is
recommended.
♦ H/V Spectral ratio for each individual window
♦ Average of H/V ratios
♦ Standard deviation estimates of spectral ratios
Details of the different options are to be found in the Appendix of the J-SESAME User
Manual. The parameter settings for the above options are controlled through an input file.
The processing module is applied according to the selected nodes in the tree. If the selected
node is a site, then all the selected windows of the data-files collected for this site will be
used for computing the average H/V spectral ratio. Output for each window can be also
obtained by setting up the configuration parameters of the processing module. Batch
processing will be performed when several sites or data-files are selected.
2.4 Showing output results

By pressing the <View Output> button (Figure 4) the user can navigate through three
dialogue boxes. The first dialogue box (Figure 6) shows the H/V spectral ratio of the merged
horizontal components, as well as the plus and minus one standard deviation curves. The
second dialogue box shows the H/V spectral ratio for each one of the NS and EW
components. The third one (Figure 7) shows the spectral ratio of the merged (H), NS and EW
horizontal components and the spectra of V, NS and EW for each individual window, only if
“output single window information” is selected in the configuration parameter of the
processing module.
25
Figure 6. H/V ratio for the merged horizontal components (mean in black, mean multiplied
σ
and divided by 10 (logH/V), in red and blue). The pink strip shows the frequency range where
the data has no significance, due to the sampling rate and the window length. The grey strip
represents the mean f0, plus and minus the standard deviation. It is calculated from the f0 of
each individual window.
2.5 Setting graph properties and creating images of the output results
For each graph shown (for example as in Figure 6 and 7) there is a small red box in the
upper right corner without any label. By clicking there the properties and scale of the graph
can be modified and images of the graph can be created. The button <Properties and
series> pops up a dialogue box where line type, width and colour can be changed for each
spectral curve. The button <Scales> pops up a dialogue box where the minimum, maximum
and scale type for each one of the vertical and horizontal axes can be modified. The button
<Save> allows an image of the graph to be created.
26
Figure 7. Result for individual windows: H/V ratios and spectra shown are derived from the
signal displayed in the red window.
27
3. INTERPRETATION OF RESULTS
Table 2. Definitions of the parameters used in this section.
• lw = window length
• nw = number of windows selected for the average H/V curve
• nc = lw . nw. f0 = number of significant cycles
• f = current frequency
• fsensor = sensor cut-off frequency
• f0 = H/V peak frequency
• σf = standard deviation of H/V peak frequency (f0 ± σf)
• ε (f0) = threshold value for the stability condition σf < ε(f0)
• A0 = H/V peak amplitude at frequency f0
• AH/V (f) = H/V curve amplitude at frequency f
• f- = frequency between f0/4 and f0 for which AH/V(f-) < A0/2
• f+ = frequency between f0 and 4f0 for which AH/V(f+) < A0/2
• σA (f) = standard deviation of AH/V (f), σA (f) is factor by which the mean AH/V(f) curve
should be multiplied or divided
• σlogH/V (f) = standard deviation of the logAH/V(f) curve, σlogH/V (f) is an absolute value
which should be added or subtracted to the mean logAH/V(f) curve
• θ (f0) = threshold value for the stability condition σA(f) < θ(f0)
• Vs,av = average S-wave velocity of the total deposits
• Vs,surf = S-wave velocity of the surface layer
• h = depth to bedrock
• hmin = lower-bound estimate of h
3.1 Underlying assumptions

As detailed in Appendix B, the main information looked for within the H/V ratio is the
fundamental natural frequency of the deposits, corresponding to the peak of the H/V curve.
While the reliability of its value will increase with the sharpness of the H/V peak , no
straightforward information can be directly linked to the H/V peak amplitude A0. This latter
value may however be considered as indicative of the impedance contrasts at the site under
study: large H/V peak values are generally associated with sharp velocity contrasts.
Figure 8 shows a comparison between the H/V ratio of ambient vibrations and the standard
spectral ratio of earthquakes using a reference site. The comparison is performed using all
the sites investigated in the framework of the SESAME project. The top part of the figure
compares the value of the fundamental natural frequency f0 derived using both methods. An
overall good agreement can be observed for the frequency values. The bottom part of Figure
8 compares the value of the peak amplitude A0. This comparison shows that it is not
scientifically justified to use A0 as the actual site amplification. However, there is a general
trend for the H/V peak amplitude to underestimate the actual site amplification. In other
words, the H/V peak amplitude could generally be considered as a lower bound of the actual
site amplification.
28
f0-H/V Ambient vibrations vs f0-SPR earthquakes
10
f0-H/V Ambient vibrations (Hz)
0
0 1 10
f0-SPR-earthquakes (Hz)
ANNECY BENEVENTO CATANIA CITTACASTELLO

COLFIORITO CORINTH EBRON EUROSEISTEST
FABRIANO GRENOBLE GUADELOUPE LOURDES
NICE PREDAPPIO ROVETTA TEHRAN
VERCHIANO VOLVI94 VOLVI97
A0-H/V Ambient vibrations vs A0-SPR earthquakes
8
7
6
A0-H/V ambient vibrations
5
4
3
2
1
0
0 1 2 3 4 5 6 7 8
A0-SPR-earthquakes
ANNECY BENEVENTO CATANIA
CITTAdiCASTELLO COLFIORITO CORINTH
EBRON EUROSEISTEST FABRIANO
GRENOBLE GUADELOUPE LOURDES
NICE PREDAPPIO ROVETTA
TEHRAN VERCHIANO VOLVI94
VOLVI97
Figure 8. Comparison between H/V ratio of ambient vibrations and standard spectral ratio of
earthquakes. Top: comparison of the frequencies f0, bottom: comparison of the amplitudes
A0.
29
The following interpretation guidelines are mainly linked with the clarity and stability of the
H/V peak frequency value. The "clarity" however is related at least partly to the H/V peak
amplitude (see below). While there are very clear situations where the risk of mistake is close
to zero, one may also face cases (more than 50% in total) where the interpretation is uneasy
and must call to some extent on "expert judgement": the following guidelines propose a
framework for such an expert judgement, trying to minimise the subjectivity – which,
however, can never be completely avoided.
3.2 Conditions for reliability

The first requirement, before any extraction of information and any interpretation, concerns
the reliability of the H/V curve. Reliability implies stability, i.e., the fact that actual H/V curve
obtained with the selected recordings, be representative of H/V curves that could be obtained
with other ambient vibration recordings and/or with other physically reasonable window
selection.
Such a requirement has several consequences :
i) In order for a peak to be significant, we recommend checking that the following condition
is fulfilled : f0 > 10 / lw. This condition is proposed so that, at the frequency of interest,
there be at least 10 significant cycles in each window (see Table 1). If the data allow –
but this is not mandatory –, it is always fruitful to check whether a more stringent
condition [ f0 > 20 / lw], can be fulfilled, which allows at least ten significant cycles for
frequencies half the peak frequency, and thus enhances the reliability of the whole peak.
ii) A large number of windows and of cycles is needed: we recommend that, when using
the automatic window selection with default parameters, the total number of significant
cycles : nc = lw . nw .f0 be larger than 200 (which means, for instance, for a peak at 1 Hz,
that there be at least 20 windows of 10 s each; or, for a peak at 0.5 Hz, 10 windows of
40 s each), see Table 1 for other frequencies of interest. In case no window selection is
considered (all transients are taken into account), we recommend, for safety, this
minimum nc number of cycles be raised around 2 times at low frequencies (i.e., up to
400), and up to 4 to 5 times at high frequencies, where transients are much more
frequent (i.e., up to 1000).
iii) An acceptably low level of scattering between all windows is needed. Large standard
deviation values often mean that ambient vibrations are strongly non-stationary and
undergo some kind of perturbations, which may significantly affect the physical meaning
of the H/V peak frequency. Therefore it is recommended that σA(f) be lower than a factor
of 2 (for f0 > 0.5 Hz), or a factor of 3 (for f0 < 0.5 Hz), over a frequency range at least
equal to [0.5f0, 2f0].
Therefore, in case one particular set of processing parameters does not lead to satisfactory
results in terms of stability, we recommend reprocessing the recordings with some other
processing parameters. As conditions for fulfilling items i , ii and iii above often lead to
opposite tuning for some parameters (see section 2.), it may be impossible in some cases:
the safest decision is then to go back to the site and perform new measurements of longer
duration and/or with more strictly controlled experimental conditions.
In addition, one must be very cautious if the H/V curve exhibits amplitude values very
different from 1 (i.e., larger than 10, or lower than 0.1) over a large frequency range (i.e.,
over two octaves): in such a case, it is very likely that the measurements are bad
(malfunction in the sensor or the recording system, very strong and close artificial ambient
vibration sources, for instance), and should be redone ! It is mandatory to check the original
time domain recordings first.
30
In the following interpretation guidelines, we assume that these reliability conditions are met;
if not, some reprocessing with other computational parameters should be attempted to try to
meet them, or some additional measurements made. If none of these two options leads to
satisfactory results, then the results should be considered with caution, and some reliability
warning should be issued in the final interpretation.
3.3 Identification of f0
3.3.1 Clear peak
The clear peak case is met when the H/V curve exhibits a "clear, single" H/V peak.
• The "clarity" concept may be related to several characteristics: the amplitude of the H/V
peak and its relative value with respect to the H/V value in other frequency bands, the
relative value of the standard deviation σA (f), and the standard deviation σf of f0
estimates from individual windows.
• the property "single" is related to the fact that in no other frequency band, does the H/V
amplitude exhibit another "clear" peak satisfying the same criteria.
We propose the following quantitative criteria for the "clarity"

Amplitude conditions:
i) there exists one frequency f-, lying between f0/4 and f0, such that A0 / AH/V (f-) > 2
ii) there exists another frequency f+, lying between f0 and 4.f0, such that A0 / AH/V (f+) > 2
iii) A0 > 2
Stability conditions:
iv) the peak should appear at the same frequency (within a percentage ± 5%) on the H/V
curves corresponding to mean + and – one standard deviation.
v) σf lower than a frequency dependent threshold ε(f), detailed in Table 3.
vi) σA (f0) lower than a frequency dependent threshold θ(f), also detailed in Table 3.
Table 3 gives the frequency dependent threshold values for the above given stability
conditions v) σf < ε(f), and vi) σA (f0) < log θ(f), or σlogH/V (f0) < θ(f).
Table 3. Threshold values for stability conditions.

Frequency range [Hz] < 0.2 0.2 – 0.5 0.5 – 1.0 1.0 – 2.0 > 2.0
ε (f0) [Hz] 0.25 f0 0.20 f0 0.15 f0 0.10 f0 0.05 f0
θ (f0) for σA (f0) 3.0 2.5 2.0 1.78 1.58
log θ (f0) for σlogH/V (f0) 0.48 0.40 0.30 0.25 0.20
For the property "single", we propose that none of the other local maxima of the H/V curve
fulfil all the above quantitative criteria for the "clarity".
If the H/V curves for a given site fulfil at least 5 out of these 6 criteria, then the f0 value can be
considered as a very reliable estimate of the fundamental frequency. If, in addition, the peak
amplitude A0 is larger than 4 to 5, one may be almost sure that there exists a sharp
discontinuity with a large velocity contrast at some depth.
However, one has, in any case, to perform the two following checks:
31
• the frequency f0 is consistent with the sensor cut-off frequency fsensor and sensitivity : if f0
is lower than1 Hz, while the sensor used is a high frequency velocimeter, check the
results with the procedure indicated in 3.3.2.b.
• this sharp peak does not have an industrial origin (cf. 3.3.2.a).
Window Number of Number of Frequency statistics from individual windows

length lw [s] windows nw significant cycles nc
f0 [Hz] σf [Hz] A0 σA(f0)
41 14 1561 2.72 0.11 4.4 1.2
Figure 9. Example of a clear H/V curve, that fulfils all the criteria for "reliability" and "clarity"
given in sections 3.2 and 3.3.1.
3.3.2 "Unclear" cases

3.3.2-a Sharp peaks and Industrial origin
It often occurs in urban environments that H/V curves exhibit local narrow peaks – or
troughs. In most cases, such peaks or troughs have an anthropic (usually industrial) origin,
related to some kind of machinery (turbine, generators, ...).
Such perturbations are recognised by two general characteristics
• They may exist over a significant area (in other terms, they can be seen up to distances
of several kilometres from their source).
• As the source is more or less "permanent" (at least within working hours), the original
(non smoothed) Fourier spectra should exhibit sharp peaks .
Several kinds of checks are therefore recommended :

• Have a look at the raw spectra from each individual window : if they all exhibit a sharp
peak (often on the 3 components together), at this particular frequency, there is a 95 %
chance that this is anthropic "forced" ambient vibration and it should not be considered in
the interpretation.
32
• Another check consists of reprocessing with less and less smoothing: in the case of
industrial origin, the H/V peak should become sharper and sharper (while this is not the
case for a "site effect" peak linked with the soil characteristics). In particular, checks with
"linear", "box" smoothing with smaller and smaller bandwidth should result in "box-like"
peaks having exactly the same bandwidth as the smoothing.
• If other measurements have been performed in the same area, determine whether a
peak exists at the same frequencies with comparable sharpness (the amplitude of the
associated peak, even for a fixed smoothing parameter, may vary significantly from site
to site, being transformed sometimes into a trough).
• Another very effective check is to apply the random decrement technique (Dunand et al.,
2002) to the ambient vibration recordings in order to derive the "impulse response"
around the frequency of interest: if the corresponding damping is very low (say, below
5%), an anthropic origin may be assumed almost certainly, and the frequency should not
be considered in the interpretation.
• Whenever it is hard to reach a conclusion from the previous tests and this information is
important, it is often very instructive to perform continuous measurements (over 24 h :
day + night, or over one week: working days + week-end) to check whether these peaks
also exist during non-working hours. There are, however, many plants that work 24h a
day, 7 days a week: the test will not be conclusive in such a case, but it should be
possible to identify such a plant with a minimum knowledge of the local industrial activity.
If any of the proposed checks does suggest an industrial origin, then the identified frequency
should be completely discarded: it has no link with the subsurface structure.
Note : It may happen that the spurious frequency of industrial origin coincides with, or is not
far from, a real site frequency. The existence of such artefacts may then alter the estimation
of the actual site frequency f0; as much as possible, it is then preferable to perform
measurements outside working hours to avoid this spurious peak, or to apply severe band-
reject filters to the microtremor recordings in order to totally eliminate the artefact and its
effects.
3.3.2-b Unclear low frequency peaks [criterion i) and possibly ii) not fulfilled]
There exist a number of conditions where the H/V curve exhibits a fuzzy, unclear low
frequency peak (i.e., at frequencies lower than 1 Hz), or a broad peak that does not satisfy all
the criteria above, especially the amplitude criteria.
It may have several origins (nonexclusive of one another)
• a low frequency site with either moderate impedance contrast (lower than approx. 4) at
depth, or a velocity gradient, or a low level of low frequency ambient vibrations (for
instance in continental areas)
• wind blowing during recording time, especially in the case of non-optimal recording
conditions (for instance proximity of trees or buildings)
• measurements performed during a meteorological perturbation that may significantly
enhance the low frequency content and alter the H/V ratio.
• a bad soil-sensor coupling, for instance on very wet soils (after rain), or with grass, or
with a non-satisfactory plate in between the sensor and the soil
• low frequency artificial ambient vibration sources (such as heavy trucks / public works
machines) at close to intermediate distance (within a few hundred meters)
• inadequate smoothing parameters (smoothing with a constant bandwidth may
completely or partially erase low frequency peaks)
• inadequate sensor with very low sensitivity at low frequency.
Distinguishing between these various possibilities is not easy. The following tests / checks
may however help to decide whether the "unclear" low frequency peak is indeed a site
characteristic :
33
• Consider the geology of the site: if it is on rock, it is likely that the low frequency fuzzy
peak is an artefact; if it is on sedimentary deposits, there might exist low frequency
effects, due either to very soft surface layers, or to stiff but thick deposits. In general,
very soft layers (such as in Mexico City) result in clear peaks because of large
impedance contrasts (at least, approx. 4), while unclear H/V peaks at low frequency are
more likely for thick, stiff sedimentary deposits.
• Check the weather bulletins corresponding to the recording period and the measurement
field sheets.
• Check the low frequency asymptote (limit value of H/V ratio when frequency is close to
zero): if the asymptotic value is significantly larger than 2, some low frequency artefacts
due to wind , or traffic, or bad sensor, are likely.
• Check the cut-off frequency fsensor of the sensor used.
• Check whether or not a peak around f0 also appears on the mean ± one standard
deviation curves (item iv in 3.3.1). If not, reprocess the data with longer windows and/or
more stringent window selection criteria (in this case the standard deviation σA (f) is
probably too large and has to be reduced).
• Check the smoothing parameters and reprocess the data with i) proportional bandwidth
and ii) less smoothing: if this improves the clarity and stability of the low-frequency peak,
it is a hint that there are high chances it is due to site conditions ; however, if even with
this reprocessing, the criteria of 3.3.1 are not fulfilled, it is recommended that the site be
re-measured with longer recordings.
• Have a look at H/V curves from individual windows, at the corresponding H and V
Fourier spectra, and at the corresponding time histories. Some of them may be
eliminated, some other windows may be added (document the reasons for each
window), which may lead to a clearer low frequency peak satisfying all criteria of 3.3.1.
3.3.2-c Broad peak case or multiple peak case

In some cases, the H/V curve may exhibit a broad peak, or a multiplicity of local maxima,
none of which fulfils the above criteria i, ii and possibly v (3.3.1).
In each of these cases, the first check to perform is to change the smoothing parameters: in
case of a broad peak, decrease the smoothing bandwidth; in the multiple peak case,
increase the smoothing bandwidth.
In this latter case however, another mandatory check consists of investigating the possible
industrial origin of any one of these peaks (see 3.3.2.a).
It may sometimes happen that other "acceptable" processing parameters allow the broad or
multiple peaks to be transformed into a "clear" H/V curve according to the criteria of 3.3.1.
This is, however, rather rare.
• If the broad nature seems stable with a rather small standard deviation, then one may
consider the possible link with a sloping underground interface (see below 3.4 and
appendix B).
• With large smoothing, a multiple peak curve may always be transformed into a "broad
peak" or "plateau-like" curve. Smoothing parameters that are too large are nevertheless
not recommended. Since our experience taught us the scarcity of such cases, and their
links, very often, to unsatisfactory recordings, we recommend, in such cases, either that
the H/V results for the site be discarded, or that the measurements be repeated.
3.3.2-d Two peaks case

In some cases, the H/V curve may exhibit two peaks satisfying the above criteria (3.3.1); this
is, however, rather rare.
Theoretical and numerical investigations have shown that such a situation occurs for two
large impedance contrasts (say, around 4 minimum for each), at two different scales: one for
a thick structure, and the other one for a shallow structure. The two frequencies, f0 and f1
34
(with f0 < f1), may then be interpreted as characteristics at each scale, f0 being the
fundamental frequency.
In order to check whether this is actually the case, the following checks are recommended:
• check the geology of the site and the possibility of a) shallow, soft deposits, b) thick,
rather stiff sediments (or soft rock) and c) very hard underlying bedrock at depth.
• Reprocess the data with other smoothing parameters: the peaks should be stable and
withstand broader and narrower smoothing (around the recommended default values) ;
consider in particular the possibility that one of these peaks (generally at the higher
frequency) may have an industrial origin (cf. 3.3.2.a).
• The statistics on several hundred measured sites and a number of theoretical cases
show that the two contrasts should be at very different scales, which means that the two
frequencies f1 and f0 should be sufficiently different so that both peaks fulfil the clarity
criteria.
3.4 Interpretation of f0 in terms of site characteristics

After having analysed and checked the H/V curves as indicated in the previous section, the
next step is to interpret the H/V curve, and in particular the peak frequency(ies) f0 (and f1), in
terms of site characteristics.
• When f0 is clear (case 3.3.1) and does not have an industrial origin, then there is a
quasi-certitude that the site under study presents a large impedance contrast (at least,
approx. 4) at some depth, and is very likely to amplify the ground motion. f0 is the
fundamental frequency of the site; there is around 80% chance that the actual site
amplification for the Fourier spectra, around the fundamental frequency f0, is larger than
the H/V amplitude A0; amplification starts at f0, but may occur at higher frequencies even
though the H/V amplitude remains small. If the local thickness is known, the average S-
wave velocity of the surface layer may be estimated with the formula VS,av ≈ f0 . 4h ; if a
reliable estimate of the S-wave velocity VS,surf is available close to the surface, then a
lower bound estimate of the thickness may be obtained with the formula hmin ≈ VS,surf /
4.f0. This conclusion holds, obviously, even in the case of a rock site: in that case, the
existence of a clear H/V peak (generally at high frequencies) is proof of the existence of
significant weathering at rock surface.
• When there exist two clear frequencies f0 and f1 satisfying the criteria described in
3.3.2d, it is likely that a) the surface velocity is low, b) the deep bedrock is very hard, and
c) there exist two large impedance contrasts (at least, approx. 4) at two different scales,
so that the amplification should be significant over a broad frequency range starting at f0
and extending beyond f1. If the local total thickness is known, the average S-wave
velocity of the surface layers may be estimated with the formula VS,av ≈ f0 . 4h ; if a
reliable estimate of the S-wave velocity VS,surf is available close to the surface, then a
lower bound estimate of the thickness of the topmost layer may be obtained with the
formula h1,min ≈ VS,surf / 4.f1 .
• Frequencies of industrial origin associated with sharp peaks must be completely

discarded for an interpretation in terms of site characteristics.
• In the case of an "unclear" low frequency peak (f0 < 1 Hz), the safest attitude is to refrain
from deriving quantitative interpretations from the H/V curve. If the same observation is
consistent over several measurement sites in the same area, and if stiff sediments are
present at this site, then we recommend going back in the field and to perform additional
measurements – if possible during night time and/or under quiet weather conditions, with
a low frequency velocity sensor and over long periods of time. The low frequency might
35
then be extracted in a clearer manner. In any case, one must then keep in mind the
possibility, at such a site, of low frequency amplifications: the safest way to investigate
the reality of such low frequency amplification is to install temporary stations equipped
with broad band velocity sensors in continuous recording mode, including one on a
nearby "reference" (rock) site, and to evaluate the classical site to reference spectral
ratios from earthquake recordings (regional or teleseismic events are perfectly suited for
showing low frequency amplification).
• If the H/V curve exhibits a broad maximum, whatever the processing (see 3.3.2.c), then
it is not easy to extract any information for the maximum frequency of this broad peak,
since this estimate may significantly vary with different processing parameters
(smoothing type, bandwidth, etc.). However, it may be that such a broad peak is due to
the presence of an underground sloping interface between softer and harder layers: if
other recordings are available in the same area, one is then invited to check whether
clearer peaks are observed in the vicinity, and whether the associated frequencies a)
exhibit significant variations from site to site and b) lie within the frequency range of the
broad peak. If all these conditions are fulfilled, then there are large chances that the
underground structure of the site under study exhibits significant lateral variations, which
lead to significant 2D or 3D effects. Such "broad peak" or "plateau-like" H/V curves are
indeed observed on many valley edges.
• If the H/V curve exhibits amplitude values very different from 1 (i.e., up to 10, or as low
as 0.1) over a large frequency range, it is very likely that the measurements are bad
(malfunction in the sensor or the recording system, very strong and close artificial
ambient vibration sources), and should be redone!
• Finally, if the H/V curve is flat (i.e., has values lying between around 0.5 and 2.0, without
any clear peak), it is very likely that the local underground structure does not exhibit any
sharp impedance contrast (at least, approx. 4) at any depth. It does not necessarily
mean, however, that there is no site amplification; there are several examples of sites
with low frequency amplification and flat H/V curve (see the example of Tehran-ABM
site, in Appendix A.2). One should not be too hasty therefore in interpreting flat H/V
curve in terms of a no-amplification site: this can be done only for rock sites. However,
such situations of a flat curve on non-rock sites correspond to less that 5% of the total
number of sites studied, as can be seen on Figure 8, section 3.1.
ACKNOWLEDGEMENTS
The SESAME project has been supported by the European Commission – Research General
Directorate (Project No. EVG1-CT-2000-00026) and by the Swiss Federal Office for
Education and Science (OFES No. 00.0085-1). We would like to kindly thank Benoît Le Brun,
Hanan Nimry, Michael Schmitz and Bill Stephenson for their constructive review of the
guidelines text.
36
REFERENCES
All deliverables and scientific papers stemming from the SESAME project are available on
the project web-site at http://sesame-fp5.obs.ujf-grenoble.fr/index.htm
Atakan K., A-M. Duval, N. Theodulidis, P-Y. Bard and the SESAME-Team. 2004, On the
reliability of the H/V Spectral Ratio Technique, ICSDEE & ICEGE, Berkeley, CA, USA.
Atakan K., A.-M. Duval, N. Theodulidis, B. Guillier, J.-L. Chatelain, P.-Y. Bard and SESAME-
Team. 2004. The H/V spectral ratio technique: experimental conditions, data
processing and empirical reliability assessment. 13th world conference in Earthquake
Engineering, Vancouver, Paper # 2268.
Bard P.-Y., and SESAME participants, 2004. The SESAME project: an overview and main
results. Proceedings of the 13th World Conference in Earthquake Engineering,
Vancouver, August 2004, Paper # 2207.
Bard P.-Y., 1998. Microtremor measurements: A tool for site effect estimation ? Proceeding
of the Second International Symposium on the Effects of Surface Geology on Seismic
Motion. Yokohama, Japan, 3. pp. 1251-1279.
Bonnefoy-Claudet S., C. Cornou, J. Kristek, M. Ohrnberger, M Wathelet, P.-Y. Bard, D. Fäh,
P. Moczo, F. Cotton, 2004. Simulation of seismic ambient vibrations: I. H/V and array
techniques on canonical models. 13th world conference in Earthquake Engineering,
Vancouver, Paper # 1120.
Cornou, C., G. Di Giulio, M. Ohrnberger, J. Kristek, M. Wathelet, 2004. Simulated Vs
observed seismic ambient noise in the Colfiorito basin: site effect estimation and noise
wavefield characteristics. XXIXth European Seismological Commission, Session F3,
September 2004.
Cornou C., J. Kristek, S. Bonnefoy-Claudet, D. Fäh, P.-Y. Bard, P. Moczo, M. Ohrnberger, M
Wathelet, 2004. Simulation of seismic ambient vibrations: II. H/V and array techniques
for real sites. Proceedings of the 13th World Conference on Earthquake Engineering,
Vancouver, Canada, August 2004, Paper # 1130.
Cultrera G., R. Azzara, F. Cara, R. d’Anna, G. Di Giulio, M. S. Giammarinaro, G.
Passafiume, A. Rovelli and P. Vallone, 2004. Microtremor Measurements in Palermo,
Italy: a comparison with macroseismic intensity and earthquake ground motion.
Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver,
August 2004, Paper # 915.
Dubos, N., A. Souriau, C. Ponsolles, J.-F. Fels and G. Sénéchal, 2003. Etude des effets de
site dans la ville de Lourdes (Pyrénées, France) par la méthode des rapports
spectraux. Bull. Soc. Géol. Fr., 174-1, 33-44.
Dubos, N., 2003. Contribution à l'évaluation du risque sismique dans les Pyrénées Centrales.
Thèse de l'Université Paul Sabatier (Toulouse III), 210 pages.
Dunand F., P.-Y. Bard, J.-L. Chatelin, Ph. Guéguen, T. Vassail, M.N. Farsi. 2002. Damping
and frequency from randomec method applied to in-situ measurements of ambient
vibrations: Evidence for effective soil structure interaction. 12th European Conference
on Earthquake Engineering, London. Paper # 869.
Duval A.-M., J.-L. Chatelain, B.Guillier and SESAME Project WP02 Team, 2004. Influence of
experimental conditions on H/V determination using ambient vibrations (noise),
Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver,
August 2004, Paper # 306.
Duval A.-M., J.-L. Chatelain, B. Guillier and the SESAME WP02 Team. 2004, Influence of
experimental conditions on H/V determination using ambient vibrations (noise),
ICSDEE & ICEGE, Berkeley, CA, USA.
37
Kanai K. and T. Tanaka, 1961. On microtremors. VIII. Bulletin of the Earthquake Research
Institute, 39- 97-114.
Koller M. G., J.-L. Chatelain, B. Guillier, A.-M. Duval, K. Atakan, C. Lacave, P.-Y. Bard and
the SESAME participants, 2004. Practical user guidelines and software for the
implementation of the H/V ratio technique : measuring conditions, processing method
and results interpretation, 13th World Conference on Earthquake Engineering,
Vancouver, Paper # 3132.
Konno K. and T. Ohmachi, 1998. Ground motion characteristics estimated from spectral ratio
between horizontal and vertical components of microtremors. Bull. seism. Soc. Am., 88-
1, 228-241.
Kudo K., 1995. Practical estimates of site response. State-of-the- art report. Proceedings of
the fifth Internationnal Conference on Seismic Zonation. Nice, France.
Malischewsky, P., and F. Scherbaum, 2004;Love's formula and H/V ratio (ellipticity) of
Rayliegh waves, Wave motion, 40-1, 57-67..
Nakamura Y., 1989. A method for dynamic characteristics estimation of subsurface using
microtremor on the ground surface. Quaterly Report Railway Tech. Res. Inst., 30-1, 25-
30.
Nogoshi M. and T. Igarashi, 1971. On the amplitude characteristics of microtremor (part 2)
(in japanese with english abstract). Jour. Seism. Soc. Japan, 24, 26-40.
Panou A. A. , N. P. Theodulidis, P. M. Hatzidimitriou , A. S. Savvaidis and C. B.
Papazachos, Reliability of ambient noise horizontal-to-vertical spectral ratio in urban
environment: The case of Thessaloniki city (Northern Greece), PAGEOPH, 2005 (in
press).
Panou A. A. , N. Theodulidis , P. Hatzidimitriou , K. Stylianidis and C.B. Papazachos,
Ambient noise horizontal-to-vertical spectral ratio in site effects estimation and
correlation with seismic damage distribution in urban environment: The case of the city
of Thessaloniki (Northern Greece), Soil Dyn. & Earthq. Engin., 2004 (accepted for
publication).
SESAME European project, 2002. Controlled instrumentation specification, Deliverable
D01.02.
SESAME European project, 2003. Nature of noise wavefield. Deliverable D13.08.
SESAME European project, 2003. H/V technique : experimental conditions, Deliverable
D08.02
SESAME European project, 2003. Multi-platform H/V processing software J-SESAME.
Deliverable D09.03.
SESAME European project, 2004. J-SESAME User Manual, Version 1.08.
SESAME European project, 2004. Report on parameter studies. Deliverable D12.09.
SESAME European project, 2004. Overall comparisons for test sites. Deliverable D17.10.
Stephenson, W.R., 2003. Factors bounding prograde rayleigh-wave particle motion in a soft-
soil layer, Proceedings of the 2003 Pacific Conference on Earthquake Engineering, 13-
15 February 2003, Christchurch, New-Zealand, Paper # 56, 8 pages
Teves-Costa, P., L. Senos and C.S. Oliveira, 2004. Correlation between damage distribution
and soil behaviour estimated with ambient vibrations. Proceedings of the 13th World
Conference on Earthquake Engineering, Vancouver, August 2004, Paper # 1004.
Theodulidis N., G. Cultrera, A. Tento, D. Faeh, K. Atakan, P.-Y. Bard, A. Panou and the
SESAME-Team, 2004. Empirical evaluation of the horizontal-to-vertical spectral ratio
38
technique: results from the SESAME project. Proceedings of the 13th World
Conference on Earthquake Engineering, Vancouver, August 2004, Paper # 2323.
39
Appendix A: H/V Data Examples
A.1 Illustration of the main peak types
Following the practical user guidelines for the implementation of the H/V ratio technique, for
the seven peak types presented in Part I, section 2, respective data examples are given.
Each case is illustrated by a Figure along with some detailed explanation/comments with
respect to the proposed quality criteria. Figures below represent the ambient vibration
σ
average H/V ratio (thick red line) multiplied / divided by 10 (logH/V) (thin red line).
When users of the H/V spectral ratio method face one of the unclear examples presented
below, it is recommended that they refer to the reprocessing suggestions and guidelines for
interpretation that are given for the corresponding situation, either in Part I, section 2, or in
Part II, section 3.3.
40
1. Clear peak
Volvi1997 - LTST
10
6
Amplification
F0
0
1 10
Frequency (Hz)
Basin geometry: Elongated alluvial valley, width~5km, length~40km, Depth~200m
Site Information
LTST site depth to bedrock: 196m
Type of bedrock: Gneiss
Average shear wave velocity of deposits: 570m/s
Comments
Criteria for a reliable H/V curve are fulfilled, that is:
f0 > 10 / lw
nc (f0) > 200
σA(f) < log10(2)
Criteria for an ideal H/V peak are also fulfilled:

A0 (=6) > 2
∃ f- ∈ [f0/4, f0] | AH/V(f-) < A0/2
∃ f+ ∈ [f0, 4f0] | AH/V(f+) < A0/2
σf (=14%) < ε (f0) (=15%)
σA(f0) (=1.6) < θ (f0) (=2)
Interpretation : All criteria are fulfilled, the fundamental frequency of the site may be reliably
estimated at 0.7 Hz.
41
2. Unclear Low Frequency Peak
Citta di Castello - CD16

10
6
Amplification
F0
0
1 10
Frequency (Hz)
Basin geometry: Elliptical alluvium valley, width~10km, length~25km, depth~0.1km
Site Information
CD16 site is situated on soft alluvium sediments and silty clay.
Type of bedrock: Sandstone (Middle Miocene).
Comments
f0 > 10 / lw
nc (f0) > 200
σA(f) < log10(2)
In addition:
Although A0(=2.9) > 2, the peak cannot be qualified "clear" since the amplitude is not
decreasing rapidly on each side.
None f1 ∈ [f0/4, f0] | AH/V(f1) < A0/2
None f2 ∈ [f0, 4f0] | AH/V(f2) < A0/2
Interpretation : further tests should be performed as listed in section II-3.3.2-b
42
3. Two Peaks Cases (f1>f0)
Grenoble - pt50
10
8
Amplification
F0 F1
0
1 10
Frequency (Hz)
F1
Basin geometry: Y-shaped sedimentary valley, Depth~800m
Site Information
PT50 site is situated on late quaternary post-glacial deposits.
Type of bedrock: Jurassic marls and marly limestone.
Comments
f0 > 10 / lw
nc (f0) > 200
σA(f) < log10(2)
Interpretation :
For the low frequency peak, A0(=4.0) > 2 and ∃ f2 ∈ [f0, 4f0] | AH/V(f2) < A0/2
Although, strictly speaking, one cannot find f1 ∈ [f0/4, f0] | AH/V(f1) < A0/2, the general trend of
the curve, together with the known geology of the site, allow the meaning of the low
frequency peak to be assigned with confidence; another processing with more narrow band
smoothing would satisfy the criteria
For the second peak, all the criteria are fulfilled:

A1(=3.5)>2
∃ f1 ∈ [f1/4, f1] | AH/V(f1) < A1/2
∃ f2 ∈ [f1, 4f1] | AH/V(f2) < A1/2
This second peak around 13 Hz is certainly associated with a very shallow structure.
43
4. Broad Peak or Multiple Peaks
Volvi1997 - LPAS
10
6
Amplification
F1 F2 F3
0
1 10
Frequency (Hz)
Basin geometry: Elongated alluvial valley, width~5km, length~40km, Depth~200m
Site Information
LTST site depth to bedrock: ~180m
Type of bedrock: Gneiss
Average shear wave velocity of deposits: 570m/s
Comments
f0 > 10 / lw
nc (f0) > 200
σA(f) < log10(2)
Interpretation : All three peaks fulfil the criterion for amplitude, Ai>2. However, only the peaks
F2 and F3 fulfil all “clarity” criteria (3.3.1). The availability of other information (geology,
deposit thickness, geophysics) in that area allows us to identify f2 as the fundamental
frequency of the site. The location of this site close to a valley edge may explain the
presence of these two peaks with rather low amplitude, while another nearby site (LTST, see
above the "clear peak" example") exhibits a clear peak with larger amplitude: the latter is
located in the central, part of the graben.
44
5. Sharp Peaks and Industrial Origin

Tehran - TAR
10
6
Amplification
F0 F1
0
1 10
F requency (Hz)
Tehran - TAR
12
Smooth K-O: 10
Smooth K-O: 20
Smooth K-O: 40
Smooth K-O: 60
Smooth K-O: 80
8
Amplification
1 10
Frequency (Hz)
Basin geometry: Unknown
Site Information
TAR site is overlain with stiff soil (coarse grained alluvium).
Comments
f0 >10 / lw
nc (f0) > 200
σA(f) < log10(2)
The F1 local narrow peak has an industrial origin. This (H/V) spectral ratio peak is due to
manmade noise/machinery; the reprocessing with different smoothing parameters (bottom
right) shows it becomes narrower and narrower, with a larger and larger amplitude when the
b-value (Konno-Ohmachi smoothing approach) is increasing: this behaviour is typical of
industrial origin. Another confirmation is obtained form the fact this narrow peak occurs at the
same frequency in the Fourier spectra of all three components (Figure on bottom left).
45
6a. Flat H/V Ratio Curves [on sediments]
Tehran - ABM
10
6
Amplification
1 10
Frequency (Hz)
Basin geometry: Unknown
Site Information
ABM site is characterised by stiff soil (coarse grained alluvium) overlying bedrock at an
unknown depth.
Comments
f0 >10 / lw
nc (f0) > 200
σA(f) < log10(2)
Significant low frequency amplification (F<1.0 Hz) was found for the ABM sedimentary site
using earthquake data, which does not appear in the H/V ratio. This site is one of the few
examples of non-rock sites exhibiting a flat H/V curve though also exhibiting a significant low
frequency amplification (less that 5% of the total number of sites studied, as can be seen on
Figure 8, section 3.1)
Note: the peak around 1.3 Hz was shown to have an industrial origin.
46
6b. Flat H/V Ratio Curves [on rock]
Lourdes - ROC
10
6
Amplification
1 10
Frequency (Hz)
Basin geometry: Confluence of two valleys
Site Information
ROC site is situated on rock outcrop at the confluence of two valleys (reference site). (Dubos
et al., 2003; Dubos, 2003)
Comments
f0 >10 / lw
nc (f0) > 200
σA(f) < log10(2)
The H/V ratio is flat over the whole frequency range examined. As the available geological
information unambiguously indicates that it is a hard rock site, this flat H/V curve may be
interpreted as indicative of a good, non weathered reference site free of any amplification
even at high frequencies.
47
A.2 Comparison with standard spectral ratios
Following the Practical User Guidelines for the implementation of the H/V ratio technique,
data examples relevant to physical explanation are presented. Along with each figure some
explanation/comments with respect to the site information and its fundamental frequency are
given. Figures below represent the ambient vibration average H/V ratio (thick red line)
σ
multiplied / divided by 10 (logH/V) (thin red line) and the earthquake recordings average
σ
standard spectral ratio [SSR] (thick black line) multiplied / divided by 10 (logSSR) (thin black
line).
48
Volvi1997 - LTST
10
6
Amplification
0.1 1 10
Frequency (Hz)
Site: LTST (Mygdonian Basin, Northern Greece)

Surface Geology: Soft
Basin Geometry: Semi-elliptical 3D basin
Surface Topography: Flat
Site Description: Agricultural area
Comments: A clear and unique (H/V) spectral ratio peak is shown. Both approaches (SSR,
H/V) exhibit the same fundamental frequency (fo=0.7Hz). The LTST site is very well
documented with geotechnical/geophysical data and the 2D theoretical transfer function is in
good agreement with the experimental one.
49
Lourdes - AUZ
20
15
Amplification
10
0.1 1 10
Frequency (Hz)
Site: Auzon-Lourdes (France)

Surface Geology: Stiff and shallow sediments (Dubos et al, 2003; Dubos, 2003)
Basin Geometry: 3D Basin
Site Description: Urban
Comments: A clear and unique (H/V) spectral ratio peak is shown. Both approaches (SSR,
H/V) exhibit almost the same fundamental frequency (fo=5.0 Hz).
50
Predappio - SCA
12
Amplification
0.1 1 10
Frequency (Hz)
Site: Predappio (Italy)

Surface Geology: Soft (8m overlying rock)
Basin Geometry: Cylindrical
Comments: A clear and unique (H/V) spectral ratio peak is shown in high frequencies. Both
approaches (SSR, H/V) exhibit the same fundamental frequency (fo=7.0 Hz).
51
Tehran - ABM
6
4
Amplification
0.1 1 10
Frequency (Hz)
Site: Tehran (Iran)

Surface Geology: Stiff ABM site is characterised by stiff soil (coarse grain alluvium)
overlying bedrock at an unknown depth.
Basin Geometry: Unknown
Comments: No clear (H/V) spectral ratio peak appears in the low frequency band (f<1.0Hz),
although it is clear in the SSR technique. To the contrary, a sharp peak with an amplitude of
about 1.5 appears at about 1.3Hz. This (H/V) spectral ratio peak is due to manmade
noise/machinery since it appears in the Fourier spectra of all three components.
52
Katania - GENI
8
6
Amplification
0.1 1 10
Frequency (Hz)
Site: Catania (Italy)

Surface Geology: Soft
Basin Geometry: Unknown
Comments: No clear (H/V) spectral ratio peak appears in the frequency band [1.0 Hz, 4.0
Hz], although it is clear in the standard spectral ratio technique. The low frequency peak
(0.25 Hz) in the H/V curve is not reliable since the associated standard deviation is large, and
the amplitude does not decrease enough at lower frequencies.
53
APPENDIX B: Physical explanations
The interpretation of the H/V spectral ratio is intimately related to the composition of the
seismic wave-field responsible for the ambient vibrations, which in turn is dependent both on
the sources of these vibrations, and on the underground structure. It is also related to the
effects of the different kinds of seismic waves on the H/V ratio. The present section will
briefly summarise, for each of these issues, the status of knowledge and consensus reached
by the SESAME participants at the end of three years of intensive work and exchanges, also
benefiting from an abundant scientific literature (see Bonnefoy-Claudet et al., 2004, SESAME
deliverable D13.08, for a comprehensive review). It will conclude with a presentation of our
resulting "preferred" interpretations of the H/V peak, which form the basis of the interpretation
guidelines presented in section 3 of Part II.
One must however admit that our knowledge is still very incomplete and partial: we by no
means claim everything is known and clear, and much remains to be learnt! The reader is
thus strongly invited to consider this section simply as a snapshot of the views of the
SESAME participants, which, though based on three years of in-depth investigations, will
certainly evolve throughout the next decade.
B.1 Nature of ambient vibration wavefield

Understanding the physical nature and composition of the ambient seismic noise wavefield,
especially in urban areas, requires answering two sets of questions, which are not
independent of each other:
• What is the origin of the ambient vibrations (where and what are the sources)?
• What is the nature of the corresponding waves, i.e., body or surface waves? What is the
ratio of body and surface waves in the seismic noise wavefield? Within surface waves,
what is the ratio of Rayleigh and Love waves? Again within surface waves, what is the
ratio of fundamental mode and higher modes?
While there is a relative consensus on the first question, only few and partial answers were
proposed for the second set of questions, for which a lot of experimental and theoretical work
still lies ahead.
As known and taught for a long time in Japan, sources of ambient vibrations are usually
separated in two main categories: natural and human, which very often – and more
particularly within urban areas - correspond to different frequency bands:
• At low frequencies (f < fnh ≈ 1Hz), the origin is essentially natural, with a particular
emphasis on ocean waves, which emit their maximal energy around 0.2 Hz. These waves
can be very easily seen on islands and/or during oceanic storms. Higher frequencies
(around 0.5 Hz) are emitted along coastal areas due to the interaction between sea waves
and coasts. Some lower frequency waves (f << 0.1 Hz) are also associated with
atmospheric forcing, but this frequency range has very little interest for engineering
seismology. Higher frequencies (> 1 Hz) may also be associated with wind and water
flows.
• At high frequencies (f > fnh ≈ 1Hz), the origin is predominantly related to human activity
(traffic, machinery); the sources are mostly located at the surface of the earth (except
some sources like metros), and often exhibit a strong day/night and week / weekend
variability.
The 1 Hz limit for fnh is only indicative, and may vary from one city to another. Some specific
civil engineering works (highways, dams) involving very big engines and/or trucks may also
54
generate low frequency energy. Locally, this limit may be found by analysing the variations of
seismic noise amplitude between day and night, and between work and rest days as well.
Energetic low frequency sources are often distant (being located at the closest oceans), and
the energy is carried from the source to the site by surface waves guided in the earth's crust.
However, locally, these waves may (and actually often do) interact with the local structure
(especially deep basins). Their long wavelength induces a significant penetration depth, so
that the resulting local wavefield may be more complex: subsurface inhomogeneities, excited
by the long period crustal surface waves, may act as diffraction points and generate local
surface waves, and even possibly body waves. The energy at frequencies between 0.1 and
1.0 Hz decreases with increasing distance from oceans: extracting information from
microseisms is thus easier on islands (such as Japan) than in the heart of continental areas
(such as Kazakhstan).
High frequency waves generally correspond to much closer sources, which, most of the time,
are located very close to the surface: while the wavefield in the immediate vicinity (less than
a few hundred meters) includes both body and surface waves, at longer distances, surface
waves become predominant.
Besides this very qualitative information, only very little information is available on the
quantitative proportions between body and surface waves, and within the different kinds of
surface waves that may exist (Rayleigh / Love, fundamental / higher). The few available
results, reviewed in Bonnefoy-Claudet et al. (2004), report that low frequency microseisms (f
< fnh) predominantly consist of fundamental mode Rayleigh waves, while there is no real
consensus for higher frequencies (> 1 Hz). Different approaches were followed to reach
these results, including analysis of seismic noise amplitude at depth and array analysis to
measure the phase velocity.
The very few investigations on the relative proportion of Rayleigh and Love waves all agree
on more or less comparable amplitudes, with a slight trend towards a slightly higher energy
carried by Love waves (around 60% - 40%). In addition, there are a few reports about the
presence of higher surface wave modes from several very different sites (some very shallow,
other much thicker, some other with low velocity zone at depth).
The following Table summarises (with some simplification however) the above discussion.
Natural Human
Name Microseism Microtremor
Frequency 0.1 – fnh (0.5 Hz to 1 Hz) fnh (0.5 Hz to 1 Hz) – > 10 Hz
Origin Ocean Traffic / Industry / Human activity
Incident wavefield Surface waves Surface + body
Amplitude variability Related to oceanic storms Day/ Night, Week / week-end
Rayleigh / Love issue Incident wavefield Comparable amplitude – slight indication
predominantly Rayleigh that Love waves carry a little more energy
Fundamental / Higher Mainly Fundamental Possibility of higher modes at high
mode issue frequencies (at least for 2-layer case)
Further Comments Local wavefield may be diffe- Some monochromatic waves related to
rent from incident wavefield machines and engines. The proximity of
sources, as well as the short wavelength,
probably limits the quantitative importance
of waves generated by diffraction at depth
55
These results, although partial, do indicate that the seismic noise wavefield is indeed
complex, especially at "high" frequencies where the origin is human activity: when
interpreting the H/V ratio, one has therefore to consider the possible contributions to H/V
from both surface and body waves, including also higher modes of surface waves.
B.2 Links between wave type and H/V ratio

The wavefield being a mixture of Love, Rayleigh and body waves, the origin of the H/V peaks
(and troughs) is multiple: Rayleigh wave ellipticity, Airy phase of Love wave modes,
resonance of S and/or P body waves. This section begins with a brief outline of the effect of
each individual wave type on the H/V ratio, on the basis of simple wave propagation physics.
The issue of more complex wavefields is then addressed with a summary presentation of the
main learnings obtained from numerical simulation in complex media.
B.2.1 H/V and surface waves

B.2.1.1 H/V and Rayleigh waves
Rayleigh waves are characterised by an elliptical particle motion in the radial plane, and their
phase velocity. In horizontally stratified media where velocity varies with depth, both
characteristics exhibit frequency dependence. Stratification also gives rise to the existence of
distinct Rayleigh wave modes: while the fundamental mode exists at all frequencies, higher
modes appear only beyond some cut-off frequencies, the values of which increases with the
mode order. The H/V ratio of all modes of Rayleigh waves, which is a measure of the
ellipticity ratio, exhibits therefore a frequency dependence in stratified media.
There exist a rich literature on that topic, a list of which may be found in Kudo (1995), Bard
(1998), Stephenson (2003), Malischewsky and Scherbaum (2004) and Bonnefoy-Claudet
(2004). The main relevant results may be summarised as follows, considering first simple
one layer over half-space structures:
i) For intermediate to high S-wave velocity contrasts (i.e., larger than about 3), the H/V
ellipticity ratios of Rayleigh waves exhibit infinite peaks and/or zeros corresponding to
the vanishing of the vertical (respectively horizontal) component, and inversion of
rotation sense (from retrograde to prograde, or inversely). For low impedance contrasts,
because the rotation sense does not change with frequency, the ellipticity ratio only
exhibits maxima at some frequencies and minima at other frequencies with no zeroes or
infinities.
ii) Focusing first on the fundamental mode, the vanishing of the vertical component occurs
at a frequency fR which is very close (i.e., less than 5% different) to the fundamental
resonance frequency for S waves only if the S-wave velocity contrast exceeds a value of
4. For intermediate velocity contrasts (2.6 to 4), the ellipticity peak is still infinite and
occurs at a frequency that may be up to 50% higher than the S-wave fundamental
resonance frequency. For lower velocity contrasts, the infinite peak is replaced by a
broad maximum that has only low amplitude (less than 2-3), and occurs at a frequency
that may range between 0.5 to 1.5 times the S-wave fundamental frequency.
iii) The ellipticity ratio H/V of the fundamental mode may also exhibit not only a peak at fR,
but also a minimum (zero) at a higher frequency fz, corresponding to the vanishing of the
H component, and a second rotation sense inversion (from prograde to retrograde). A
few studies have been performed to investigate the variability of the ratio fz/fR; while
Konno and Ohmachi (1998) report a value around 2 for a limited set of velocity profiles,
Stephenson (2003) concludes that peak/trough structures with a frequency ratio around
2 witness both a high Poisson ratio in the surface soil, and a high impedance contrast to
the substrate. Some other studies for more complex velocity profiles report a
dependence of the ratio on the velocity gradient in the soft sediments.
iv) Higher modes exhibit also H/V peaks at higher frequencies corresponding to a vanishing
V component; some of these peaks, especially for high contrast structures, coincide with
56
the higher harmonics of S-wave resonance. However, for single layer structures, no
case is known, for which all existing Rayleigh wave modes exhibit simultaneously a peak
at the same frequency. In other words, for all frequencies for which several modes exist
simultaneously (i.e., generally beyond the S-wave fundamental frequency), there always
exist one Rayleigh wave mode (often the fundamental one) which may carry some
energy on the vertical component.
v) These results may generally be extrapolated to more complex horizontally layered
structures involving several layers or velocity gradients; one major difference however
concerns item iv) and the number of ellipticity peaks. Some sites may present a large
velocity contrast (i.e., exceeding at least 2.5) at different depths zk (however, given the
minimum threshold of 2.5, the number of such depths rarely exceeds 2…). In such
cases, the S-wave response will exhibit major amplifications at frequencies fk
corresponding to the fundamental frequencies of the layering located above these
depths with major discontinuity. In such cases, the available results show that all
Rayleigh wave modes existing at frequency fk do exhibit a common ellipticity peak (in
other words, the vertical component of all existing Rayleigh wave modes vanish at
frequencies fk). In such sites, a H/V curve with several sites is therefore consistent with
the surface wave interpretation. An example of this situation is a shallow very soft layer
resting on a thick, stiff unit underlain by very hard bedrock.
vi) No relation could be established between the S-wave amplification at the resonance
frequency, and the characteristics of the H/V infinite peak (for instance its width) or
maximum amplitude.
B.2.1.2 H/V and Love waves

Love waves carry energy only on the horizontal component. Their influence on the frequency
dependence of the H/V ratio can therefore come only from the frequency dependence of the
H component.
Different studies (see for instance Konno & Ohmachi, 1998) have shown that, at least for
high impedance contrast cases, Love waves do strengthen the H/V peak: all surface waves
carry their maximum energy for frequencies corresponding to group velocity minima ("Airy
phase"); for high impedance contrast layering, the group velocity minimum of the
fundamental Love mode occurs, like for the vanishing of V component, at a frequency fL
which is very close to the fundamental S-wave frequency.
Higher modes of Love waves may also have group velocity minima and associated Airy
phases at higher frequencies: this may result in other maxima if higher modes carry a
significant amount of energy.
B.2.2 H/V and body waves

As site amplifications occurring during actual earthquakes essentially involve incoming body
waves, it is obvious that the horizontal and vertical components of body waves are both
highly sensitive to site conditions. The main question to address here is the relation of the
H/V ratio and site conditions for body waves; a side question concerns the differences or
similarities between H/V ratios derived from earthquake recordings and H/V ratios derived
from ambient vibration recordings.
When considering, once again, a simple horizontally layered structure with one soft layer
over a half-space, and its response to obliquely incident plane waves, a striking result is the
fact that, whatever the incident wave type (P or SV or SH), the horizontal components
systematically exhibit resonant peaks at the S-wave resonance frequencies (even for P wave
incidence), while the vertical component always exhibit resonant peaks at the P-wave
resonance frequencies (even for S wave incidence). This result is valid when the impedance
contrast is large both for S and P waves, and comes from the conversion from P and SV
57
waves at the bedrock/layer interface, and their relatively small incidence angles within the
surface lower velocity layer.
This has very interesting consequences for the fundamental mode, since the S-wave
fundamental frequency is always significantly smaller than the P-wave fundamental
frequency (ratio equal to the S-P velocity ratio within the surface layer):
• As the fundamental frequency is only weakly dependent on subsurface topography – for
usual configurations –, this explains why the H/V ratio for a body wavefield should always
exhibit a peak around the fundamental S-wave frequency, for high impedance contrast
sites.
• In the case of horizontally stratified media, the H/V ratio should also exhibit peaks at the
S-wave harmonics, at least for all peaks that do not coincide with a lower order harmonic
of P-wave resonance.
• Finally, again for high impedance contrast, horizontally stratified media, the amplitude of
the first H/V peak is also expected to be somewhat correlated with the S wave
amplification.
These latter two items constitute the main differences from the surface wave case, where it is
not generally expected to have either harmonics, or any correlation between H/V peak
amplitude and actual amplification values. The presence or absence of harmonics at least for
a large impedance contrast, 1D structure, may thus be a good indicator of the composition of
the wavefield.
One should remain very cautious however in interpreting H/V ratios derived from earthquake
recordings beyond the fundamental S frequency, since this ratio is highly influenced by the
amplification of the vertical component, which cannot be neglected, especially in sites with
pronounced subsurface topography.
B.2.3 H/V and complex wavefields: results from numerical simulation

Since the actual composition of the seismic noise wavefield is mostly unknown, a series of
numerical simulations have been performed to investigate the origin of an H/V peak. A wide
variety of subsurface structures have been studied (1D, 2D and 3D), but only "local" sources
have been considered (i.e., less than about 10 km away from the receiver): the "microseism"
case has not yet been investigated (incoming crustal surface waves generated by oceanic
waves). These comprehensive analyses are presented and discussed in various reports
(SESAME deliverables D12.09 and D17.10, Bonnefoy-Claudet, 2004), and only the main
conclusions are summarised below:
• for 1D structures, a single H/V peak is obtained only when the predominant noise sources
are located within the surface layer and within limited distances from the site (less than
around 20 -30 times the layer thickness); the wavefield then consists of a mixture of
Rayleigh, Love and body waves. For more distant dominant noise sources (a case not
often met in urban sites, but possible in the countryside), and/or deep sources (i.e.,
located at depth below the interface associated with the main impedance contrast), the
wavefield includes a large proportion of body or head waves, especially at high
frequencies, and H/V curves exhibit several peaks associated with fundamental and
harmonics of S-wave resonance. In all cases, the frequency of the fundamental peak
exhibits a very good agreement (within ± 20% at most) with the actual fundamental S-
wave frequency of the layered structure.
• For laterally varying structures, the wavefield associated with "local" noise sources is
more complex, since it also includes additional waves diffracted from the lateral
heterogeneities. The composition of this diffracted wavefield depends on the location of
the site of interest with respect to the lateral heterogeneities: away from them (for
instance, in the flat central parts of valleys and basins), it mainly consists of surface waves
(fundamental or higher mode depending on the frequency); close to them it also includes
58
a significant portion of body waves. For all the investigated simulation cases (all with a
large impedance contrast), H/V curves exhibit clear (sharp) peaks in the "flat" parts (i.e.,
those with only gentle underground interface slopes), and broader and generally lower
maxima at sites with rapidly varying thickness (for instance at valley edges). The
frequency of the sharp peaks ("flat" parts) agrees within ± 20% with the resonance
frequency of the structure (possibly different from the 1D value for deeply embanked
valleys), while the bandwidth of the broad peaks (sites with steep underground interfaces)
is generally indicative of the fundamental frequency variations between the shallowest and
deepest sections; however, the amplitude of this broad maximum is often too small to
allow a clear identification, at least in the absence of any additional geological information
and/or dense geographic coverage of ambient vibration measurement points.
Additional comment : H/V and surface topography

All the above discussion and results implicitly addressed mainly sedimentary / alluvial sites
within valleys or plains, and are therefore valid only for horizontal free surfaces. As frequency
dependent site amplification has been repeatedly observed also on top of rocky hills, several
attempts have been performed to investigate whether H/V ratios from ambient vibrations also
exhibit a peak in the same frequency range. These attempts have generally been successful,
but no theoretical interpretation and no numerical simulation have been proposed to explain
this apparent success.
The surface topography issue was not addressed within the SESAME project, and we will
therefore only indicate here that surface waves are certainly affected by non flat free
surfaces, resulting in a diffracted wavefield (as for non horizontal underground layering)
including both surface and body waves. Even when considering surface weathering, the
impedance contrast between surface and depth usually remains limited; there is therefore
only a small probability that the H/V peak comes from a vanishing of the V component, and
we consider more likely an explanation in terms of body waves, involving both diffraction and
focusing. This is however only a suggestion, and the topic would deserve thorough further
investigations, especially as the "surface topography" amplification effect itself is far from
being properly understood.
B.3 Consequences for the interpretation of H/V curves

The above mentioned results exhibit a few consistent, robust characteristics which have to
be kept in mind when interpreting H/V curves and peak(s):
• The results are clear and simple in case of horizontally layered structures with large
impedance contrasts (> 4-5).
• The results become more and more fuzzy a) for decreasing contrasts and b) for
increasing underground interface slopes.
• Theoretical and numerical results are by far more numerous and easier to interpret for
local, anthropic like sources, i.e., essentially, above 1 Hz.
• Clear urban situations with non stationary, spatially distributed local sources are generally
associated with one single H/V peak.
• In the Rayleigh wave interpretation, the H/V peak should be associated with a (local)
trough in the Fourier spectra of the vertical component, while Love or body wave
interpretations should be associated with a (local) peak in the Fourier spectra of the
horizontal component
This has the following direct practical consequences for H/V investigations:
• One should always gather the available geological and geotechnical information, looking
in particular for a priori rough estimations of impedance contrasts (keeping in mind that
large impedance contrasts are generally associated with either very young unconsolidated
deposits, or very hard bedrock), and indications as to the lateral variability of underground
structures.
59
• Low-frequency peaks (i.e., < 1 Hz) are often less easy a) to detect and b) to interpret than
high-frequency peaks. Additional measurements in the vicinity of the site of interest often
help to find a consistent H/V peak.
• One should never forget to have a look at the original Fourier spectra of the horizontal and
vertical components, especially when the H/V maximum is not very clear.
One must also bear in mind that real ambient vibration recordings also include a number of
"spurious" sources (such as wind, or industrial harmonic machinery) which may affect the
estimation of the H/V curve, and downgrade the possibility of interpretations: while clear and
sharp "natural" peaks generally remain visible, fuzzy maxima may completely disappear.
B.3.1 1D media
When the available geological information allows the deposit to be considered as horizontally
layered with a smooth and flat interface with the underlying bedrock, at least locally, then the
wavefield composition and interpretation of H/V ratio can be seen as follows:
B.3.1.1 High frequencies (human origin, f > fnh , fnh ≈ 1 Hz)

• If the site is located in an urban environment, the "noise" sources are essentially local and
superficial; the wavefield predominantly consists of surface waves (Rayleigh and Love)
with however a slight proportion of body waves. The H/V curve should exhibit one single
peak, at a frequency that is within ± 20% of the S-wave resonant frequency of the site.
The amplitude of that peak should not be interpreted in terms of amplification values.
• If the site is located in the countryside and the predominant sources are distant, the
wavefield also includes resonating head waves and the H/V curve may exhibit several
peaks corresponding to the various harmonics
B.3.1.2 Low frequencies (oceanic origin, f < fnh , fnh ≈ 1 Hz)

Unless there exist energetic low frequency sources (big trucks or engines) close to the site
under consideration, the seismic noise wavefield is most probably caused by oceanic activity.
This may be checked through continuous measurements and an analysis of the daily /
weekly amplitude variations.
The local geological structure certainly does not exist over the whole distance between the
site and the ocean: therefore the low frequency crustal surface waves carrying the noise
energy certainly undergo some conversion (mode conversion : Rayleigh / Love, fundamental
/ higher) and/or type conversion (surface to body waves) along laterally varying substructures
at some distance from the site. For instance, for a deep inland basin, incoming low frequency
crustal waves have a penetration depth of at least 1 km, and will be diffracted along basin
edges: in such a case, the sources of seismic noise may be seen as a collection of point
sources located along the basin / bedrock interface, re-emitting the same energy envelope
spectrum as different waves: local surface waves and body waves. The associated H/V ratio
should then be somewhat similar to the H/V ratio derived from earthquake recordings, in the
low frequency range only of course (if available, site to reference rock spectral ratios should
then be also comparable for ambient vibration and earthquake recordings).
Low frequency H/V peaks are associated either with extremely soft soil (e.g., Mexico City
clays) with a thickness of several tens of meters, or with "normal" soil deposits having a very
large thickness (several hundred meters at least): the former case is obviously associated
with a large impedance contrast, while this can occur in the latter case only if the bedrock is
very hard (the confining pressure at large depth automatically induces a stiffness increase).
B.3.2 2D / 3D structures
We consider here sites under which at least one of the interfaces with significant impedance
contrast exhibits steep slopes (i.e., larger than around 10°, this value being however only
60
indicative). Such sites are therefore either "transition" zones between areas with more or less
horizontal layering, or deeply embanked valleys and basins having a large thickness to width
ratio (typically, larger than 0.2).
B.3.2.1 Transition zones

If such transition zones are expected from the available geological / geotechnical /
geophysical information, it is always recommended that measurements be made on each
part of the transition zone, and to have a rather dense measurement mesh (the "density"
being related to the predominant wavelength, i.e., to the frequency: the higher the frequency
the smaller the mesh size.
High frequencies (human origin, f > fnh , fnh ≈ 1 Hz, possibly varying from site to site)
Numerical simulations have consistently shown that, for local surface sources, the H/V curve
at such "transition" sites exhibit broader and lower maxima, which may be hard to identify:
surface waves cannot develop with one single "pure" mode, nor can resonance of body
waves occur.
Low frequencies (oceanic origin, f < fnh , fnh ≈ 1 Hz, possibly varying from site to site)
No simulation is available for incoming crustal surface waves; the local diffraction
phenomena already mentioned for this case let us think that the wavefield should include a
significant proportion of body waves generated at depth, so that the H/V ratio should reflect
at least partly the differential amplification between H and V components, and be a more
reliable indicator of the site frequency than for high frequency local surface sources. This
interpretation is not consensual and should be taken with caution.
B.3.2.2 Deeply embanked valleys and basins

The presence of such deep structures should be at least guessed from the geological map
and a minimum knowledge of the geological history of the area (ancient glacier valleys filled
with lacustrine deposits are a typical example of such structures). However, this kind of
structure may also be met in the case of buried canyons, which may remain unknown in the
absence of detailed geophysical surveys.
When associated with large impedance contrasts, the seismic response of these structures
to incoming body waves exhibits a global 2D (or 3D) resonance pattern characterised by
similar resonance frequencies at all sites, whatever the local thickness, and mode shapes
with both a strong spatial dependence, and a strong polarisation. As a consequence, a)
resonance frequencies may be different on the two horizontal components (for instance, in a
valley, the components parallel - // - and perpendicular - ⊥ - to the valley axis), and b) the
amplitude of the corresponding H/V peak may undergo significant variations from site to site,
due to variations in both H and V components (for instance, a given mode may include a
node in the ⊥ component at one site, and a node on the V component at another site).
Considering the importance of diffraction phenomena of steep interfaces, and the fact that
these resonance modes are the eigen-solutions of the wave equation in such structures, it is
logical to conclude that such modes will exist whatever the excitation wavefield, provided
there is enough energy in the corresponding frequency range. This theoretical consideration
is supported by the few numerical simulations performed on such structures (Cornou et al.,
2004 – SESAME deliverables D12.09 and D17.10).
B.3.2.3 Common recommendations for 2D/3D structures

In any case, when other information sources (geology, geophysics, …) suggest a 2D or 3D
structure, three kinds of actions / processing are strongly recommended to check these
possibilities:
• One should never limit one's investigations to one single measurement, but should
have a wide and dense spatial coverage, starting with a mesh size comparable with
61
the expected thickness of the softer cover. The mesh size may be reduced in a
second step in case of strong lateral discontinuities, indicating either faults or very
steep underground slopes
• One should also investigate the differences between differently polarised horizontal
components (i.e., applying different rotation angles) ; for instance, along valley edges,
clear differences may appear between the ⊥ and // components.
62
View publication stats

HV User Guidelines

Uploaded by

Copyright:

Available Formats

HV User Guidelines

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

HV User Guidelines

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Guidelines for the implementation of the H/V spectral ratio technique on

Book · December 2004

Anastasios Anastasiadis Kuvvet Atakan

SEE PROFILE SEE PROFILE

Riccardo Mario Azzara Pierre-Yves Bard

SEE PROFILE SEE PROFILE

PRENOLIN View project

RETIS-Risk (2014-2015) View project

The user has requested enhancement of the downloaded file.

SESAME European research project

European Commission – Research General Directorate

SESAME: Site EffectS assessment using AMbient Excitations

European Commission, contract n° EVG1-CT-2000-00026

Co-ordinator: Pierre-Yves BARD

Catello Acerra Giuseppe di Giulio Etor Querendez

It is highly recommended that prior to planning a measurement campaign on ambient

PART I: QUICK FIELD REFERENCE AND INTERPRETATION GUIDELINES .................................... 7

1. EXPERIMENTAL CONDITIONS + MEASUREMENT FIELD SHEET.............................................. 8

2. DIAGRAMS FOR INTERPRETATION OF H/V RESULTS................................................................ 10

PART II: DETAILED TECHNICAL GUIDELINES1. TECHNICAL REQUIREMENTS .................... 15

1. TECHNICAL REQUIREMENTS ............................................................................................................ 16

APPENDIX A: H/V DATA EXAMPLES ......................................................................................................... 40

A significant part of damage observed in destructive earthquakes around the world is

PART I: QUICK FIELD REFERENCE AND INTERPRETATION

1. EXPERIMENTAL CONDITIONS + MEASUREMENT FIELD SHEET

Type of parameter Main recommendations

2. DIAGRAMS FOR INTERPRETATION OF H/V RESULTS

2.1 Criteria for reliability of results

Criteria for a reliable H/V curve • lw = window length

Threshold Values for σf and σA(f0)

2.2 Main peak types

Sharp peaks and industrial origin see $3.3.2.d and appendix A

If industrial origin is proved, i.e., if

♦ Perform continuous recordings during

Clear peak see $3.3.1 and appendix A

If industrial origin ( $3.3.2.d ) , i.e.

♦ Likely sharp contrast at depth (A0 > 4-5)

Unclear low frequency peak see $3.3.2.b and appendix A

If rock site no reliable f0

If steady increase of H/V ratio with decreasing frequency

If reprocessed H/V curve fulfils the clarity criteria f0 reliable

Perform additional measurements over a

Broad peak see $3.3.2.c and appendix A

Decrease the smoothing bandwidth

If bump peak is not stable no reliable f0

♦ If clearer peaks are observed in the vicinity

Multiple peaks (multiplicity of maxima) see $3.3.2.c and Appendix A

Check no industrial origin of one of the peaks

If reprocessed H/V curve fulfils the clarity criteria f0 reliable

2 peaks cases (f1 > f0) see $3.3.2.d and appendix A

If industrial origin of f0 f0 not reliable and f1 reliable

If industrial origin of f1 f0 reliable and f1 not reliable

if soil deposits ♦ Likely absence of any sharp contrast

♦ likely unweathered or only lightly

PART II: DETAILED TECHNICAL GUIDELINES

This chapter is based on the following SESAME internal reports:

Influence of the digitiser

Influence of the sensor

1.2 Experimental conditions