Acoustic Investigation of Complex Seabeds

SPRINGER BRIEFS IN OCEANOGRAPHY
Jacques Yves Guigné
Philippe Blondel
Acoustic
Investigation of
Complex Seabeds
123
SpringerBriefs in Oceanography
More information about this series at http://www.springer.com/series/11754
Jacques Yves Guigné Philippe Blondel
•
Acoustic Investigation
of Complex Seabeds
123
Jacques Yves Guigné Philippe Blondel
Department of Physics Department of Physics
University of Bath University of Bath
Bath, Avon Bath, Avon
UK UK
and
Acoustic Zoom Ltd.

Paradise, NL
Canada
ISSN 2196-1212 ISSN 2196-1220 (electronic)

SpringerBriefs in Oceanography
ISBN 978-3-319-02578-0 ISBN 978-3-319-02579-7 (eBook)
DOI 10.1007/978-3-319-02579-7
Library of Congress Control Number: 2016963742
© The Author(s) 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
This book, and all this work, would not have
been possible without the love and support of
Anna during all these decades, and I would
like to dedicate the book to her.
For my grandchildren Éric, Raya, Mila and
Loïc, to inspire them to the world of seabed
acoustics.
Jacques Yves Guigné
For my parents: role models,

encouragements and always present when
needed. For Olga, source of inspiration and
constant support. And for Diego, to share
with him the wonders of science and the
discovery of underwater frontiers.
Philippe Blondel
Preface
Marine environments are increasingly important to human activities, with

increasing exploration of our “Blue Planet” and with the development of resource
extraction and offshore construction in often challenging environments. For these
activities, we need an accurate knowledge of the seabed and what lies immediately
beneath it. Are there methane reservoirs frozen below a particular Arctic seabed?
What is the “plumbing system” below a field of hydrothermal vents? Can resources
be safely extracted from a particular area, or is the seabed unstable? Will the
installation of gravity bases for a large offshore wind farm be possible, or will the
location of the different turbines be constrained by the local geology? Will drilling
work in this site, or are there hidden obstacles that are going to block it? Or is there
a risk of punch through, affecting economic viability as well as risks to humans,
assets and the environment? Human activities also affect the marine environments,
and knowing how they do so is now essential. How does resource extraction affect
the subsurface geology? How do marine habitats fare with particular activities (e.g.
trawl fishing) or how well do they recover after these activities cease (e.g. Grand
Banks fishing moratorium)? How successful is a particular carbon capture and
storage facility in keeping CO2 safely in the ground? If siting an offshore structure
at one place, will it be safer or riskier than siting it 15 m away? All these questions,
and many more, are routinely faced by marine scientists and geotechnical
engineers.
This book is for them and for anyone interested in the discovery of what lies
below the seabed. Through fact-packed chapters, we aim to illustrate the evolution
of the Acoustic Interrogation of Complex Seabeds (the title of this book) from its
initial concepts to the answer products now available on the market. First author
Jacques Yves Guigné did his Ph.D. at the University of Bath and graduated in 1986,
with some key ideas about how to improve existing geoacoustic tools. He then
founded Guigné International Ltd. to pursue further the prototyping and experi-
mentation of these thoughts in collaboration with his academic role at C-CORE
(Memorial University of Newfoundland). He later cofounded PanGeo Subsea Inc.,
to develop the prototypes into commercial sub-seabed exploration tools. He also
founded Acoustic Zoom Inc., to expand on the physics behind the sonar-based
vii
viii Preface
sciences to new applications for unprecedented earth imaging. Jacques’ work earned
him the Rayleigh Medal, the premier award from the Institute of Acoustics (in 2013),
and a D.Sc. from the University of Bath (in 2014). This book is focused on Jacques’
work and his achievements and in particular the development and use in the field
of the Acoustic Corer now marketed by PanGeo Subsea Inc. (Canada). Second
author Philippe Blondel joined the University of Bath in 1999, where he has spe-
cialised in sonar mapping, developing multistatic sonars and writing textbooks such
as the Handbook of Sidescan Sonar (Springer, 2009) and Bathymetry and its
Applications (InTech, 2011). He is teaching physics to undergraduate and post-
graduate students and he is also deputy director of the Centre for Space, Atmospheric
and Oceanic Science (CSAOS). Both authors started collaborating and exchanging
ideas after their first meeting in 2007, and Jacques Yves Guigné is now a visiting
professor at the University of Bath.
This double authorship offers a double perspective to this book. The insights
of the inventor of many devices for acoustic seabed interrogation (ASI) result from
several decades of hard work, in the laboratory and in the field. This book aims to
explain the thought processes but also the everyday use at sea and how it compares
with other technical approaches. Springer Briefs are meant for “experienced read-
ers”, and in this spirit, we assume known the basics of underwater acoustics, marine
geophysics and seismic prospection. The reader desirous to know more (or refresh
some concepts) will be invited where necessary to look at specific references known
in the field, such as Applied Geophysics (Cambridge University Press, 1990),
written by W.M. Telford, L.P. Geldart and R.E. Sheriff, and An Introduction to
Underwater Acoustics (Springer, 2009), written by X. Lurton. Both books have
seen several editions, a strong measure of their success. Other references will be
presented wherever felt necessary.
We hope that this short book will help appreciate the challenges of acoustic
seabed interrogation and how this can be successfully addressed in even the most
complex environments with a new instrument (the Acoustic Corer) presented in a
variety of situations.
Paradise, NL, Canada Jacques Yves Guigné

Bath, UK Philippe Blondel
October 2016
Acknowledgements
All books are collaborative endeavours, and the science presented in a book could
not be brought to a wider audience without the dedication and hard work of a team
working behind the scene to “make it happen”. We are very fortunate (and very
grateful) that Springer accepted our book proposal, and we would like to thank in
particular Janet Sterritt-Brunner for her initial encouragements. During the writing,
we interacted with Devi Ignasy and Karthik Raj Selvaraj, whose patience with the
delays coming from very busy professional lives (and the odd health issues along
the way) was always much appreciated.
The making of this book would not have been possible without a mutual
colleague, who also happens to be a mutual friend and who made us meet at the
Underwater Acoustics Conference in 2007. Both authors would like to acknowl-
edge the guiding figure of Prof. Emeritus Nicholas G. Pace, as supervisor (some-
times) and role model (always). Nick started working in the Department of Physics
at the University of Bath in 1970, and he stayed there until his retirement a few
years ago. He was the Ph.D. supervisor of Jacques Yves Guigné (1982–1986) and
the line manager of Philippe Blondel (1999–2000). Nick’s strong preferences for
experimental work, his wide and extensive knowledge of the entire field of
underwater acoustics (from what had been published to what had been tried and
never published) and his guidance in translating laboratory experiments into
sea-based products were defining influences for both of us.
Throughout the three decades of research and development that led to the
Acoustic Corer and derived answer products, many students, technical staff and
scientists worked with Jacques at the various companies and institutions that
teamed up with him to deliver on the sciences. Selected mention is to be given to
Dr. Chris Pike, Dr. Richard Charron, Dr. Sam Bromley, Dr. Ian McDermott, Gary
Dinn and Adam Gogacz for their work and direct participation. Special thanks to
PanGeo Subsea for their support in particular to Ms. Moya Cahill, cofounder of the
firm that took the acoustic interrogation ideas into commercial applications. Special
mention is also gratefully made to the Canada National Research Council’s
Industrial Research Assistance Program who provided grants to Guigné to pursue
ix
x Acknowledgements
the various experimental trials that shaped the concept for interrogating the seabed
with acoustics to commercial applications.
Except as indicated, all figures were produced by the authors in their respective
lines of work. When presenting work spanning several decades and a wealth of
scientific publications, technical and commercial reports and general presentations,
omissions or misattributions are always a possibility. The authors will therefore be
very grateful for notification of any error, which will be corrected as soon as
possible, either online or in the next editions.
Contents
1 Acoustic Coring—The Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 The Need for Accurate Assessments of Sub-seabed Sediments . . . 1
1.2 Importance to Present Maritime Operations . . . . . . . . . . . . . . . . . . 5
1.3 The Technology Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 Filling the Gaps—The Case for “Acoustic Coring” . . . . . . . . . . . . 13
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2 Imaging of the Near-Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.1 Capturing the Relevant Characteristics of Near-Surface
Sediments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Buried Objects and Benthic Habitats . . . . . . . . . . . . . . . . . . . . . . . 22
2.3 Significance and Value—Moving Toward an Answer Product . . . . 29
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3 Imaging into the Seabed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.1 Creating a Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Formulating the First Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.3 The Next Generation: The Acoustic Corer . . . . . . . . . . . . . . . . . . . 37
3.4 Data Processing—Coherent Summation and Flow Diagram . . . . . . 41
3.5 Processing Acoustic Corer Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.5.1 Data and Acquisition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.5.2 JYG-Cross Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.5.3 Synthetic Aperture—Rendering and Processing . . . . . . . . . 57
3.5.4 Combining Low-Frequency and High-Frequency
Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4 Acoustic Seabed Interrogation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.1 Using the Low-Frequency (LF) JYG-Cross Data . . . . . . . . . . . . . . 65
4.2 High-Frequency (HF) Identification of Non-specular Returns . . . . . 68
xi
xii Contents
4.3 Sizing and Identifying Individual Targets . . . . . . . . . . . . . . . . . . . . 70

4.4 Acoustic Textures and Substrate Variations . . . . . . . . . . . . . . . . . . 74
4.5 Combining with Other Measurements . . . . . . . . . . . . . . . . . . . . . . . 81
4.5.1 Traditional Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.5.2 Investigations of Sites with Complex Geology . . . . . . . . . . 82
4.5.3 Gassy Sediments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.6 Interpretations—Reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
5 The Future of Acoustic Seabed Interrogation . . . . . . . . . . . . . . . . . . . 93
5.1 Filling the Technology Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.2 Moving to New Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.3 New Environments—New Applications . . . . . . . . . . . . . . . . . . . . . 98
5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Abbreviations
AC Acoustic Corer
AGC Automatic Gain Control or Automatic Gain Correction
AMIE Acoustic Mapping and Interrogating Eye (portable, multistatic sonar)
ASI Acoustic Seabed Interrogator or Acoustic Sub-seabed Interrogator
(both concepts being equivalent in practice)
AUV Autonomous Underwater Vehicle
BH Borehole
CDP Common Depth Point (the common reflection point at depth on a
sub-seabed reflector)
CMP Common Mid Point (the seabed point halfway between source and
receiver, and shared by numerous source-receiver pairs)
CPT Cone-Penetration Testing
CPU Central Processing Unit
CT Computed Tomography
DRUMS Dynamically Responding Underwater Matrix Sonar (part of
Benthic-DRUMS™, developed by Guigné and co-workers in the
1990s, and DRUMS® = R200 parametric sonar)
FK Frequency (f) and wavenumber (k) reference framework, resulting
from Fourier transformations in time and in space
GIL Guigné International Ltd.
GLORIA Geological Long-Range Inclined Asdic (early British sidescan sonar)
GNU Type of permissive, free software license (derived from the recursive
acronym “GNU’s Not Unix”, used in Unix-operated software)
GPU Graphical Processing Unit
GSF Generic Sensor Format (a GNU open format, available
at https://www.ldeo.columbia.edu/res/pi/MB-System/formatdoc/gsf_
spec.pdf)
HF High-frequency
INS Inertial Navigation System
xiii
xiv Abbreviations
JYG-Cross Patented configuration of transducers, designed by Jacques Yves

Guigné (JYG)
LF Low-frequency
LFM Linear frequency modulation
MOU Mobile Offshore Unit
NMO Normal Move Out (compensation of the separation between acoustic
sources and receivers for a horizontal reflector)
QA Quality Assurance
ROV Remotely Operated Vehicle
SAS Synthetic Aperture Sonar (referring to either the instrument or the
mode of processing)
SAX Sediment Acoustics eXperiment (US research programmes known as
SA’99 and SAX’04)
SITAR Seafloor Imaging and Toxicity Assessment of Risks caused by buried
waste (EC research programme)
SPL Source Pressure Level
SVD Singular Value Decomposition (filtering technique)
UXO UneXploded Ordnance
Chapter 1
Acoustic Coring—The Rationale
Abstract The scope and ambition of marine geotechnical investigations have

greatly evolved over the last years. Offshore foundations, dredging operations of
harbours and channels, sub-seabed installations all place new demands on acquiring
more reliable knowledge on the composition of the seabed to address
cost-effectively the issues of buried geo-hazards, sediment property discontinuities
and trapped pollutants. This chapter shows the potential of physical sampling and
different remote sensing approaches, but also their limitations. The technology gap
can be addressed by designing an Acoustic Sub-seabed Interrogator (ASI), ideally
combining a range of low frequencies (to penetrate deeper into the seabed) and a
range of high frequencies (to enable higher vertical resolutions closer to the sur-
face). The use of a dense network of broadband receivers should allow capturing of
specular and non-specular returns, and bespoke data acquisition and processing
should give access to 3-D volumetric measurements. Its design should allow
deployments at all depths. The resulting acoustic corer should expand the horizontal
and vertical ranges of existing tools whilst preserving the best spatial resolutions
currently achievable.
! !
Keywords Acoustic coring Offshore construction Sub-seabed Geotechnics ! !
Borehole !
Cone-penetration testing !
Mobile Offshore Units Acoustics ! !
!
Spudcans Punch-through Blowout !
1.1 The Need for Accurate Assessments

of Sub-seabed Sediments
Industrial and environmental activities are steadily and increasingly turning toward
the marine environment, as technology developments enable greater access to its
resources. Traditional exploitation of oil and gas is now supplemented with
extraction of mineral deposits such as manganese or sulphide (e.g. Masuda et al.
2014). Renewable energies (tidal, wave and wind) and Carbon Capture and Storage
are expanding to adapt to climate changes and reduce the effects of CO2 in the
© The Author(s) 2017 1

J.Y. Guigné and P. Blondel, Acoustic Investigation of Complex Seabeds,
SpringerBriefs in Oceanography, DOI 10.1007/978-3-319-02579-7_1
2 1 Acoustic Coring—The Rationale
atmosphere (IPCC 2014). All these activities are associated with some type of
offshore construction, ranging from single piles to caissons of different sizes,
trenched pipelines or more substantial structures. Sustainable management of
marine habitats makes use of long-term structures in/on the seabed, like fish pens
and cages (e.g. Pilley 2008). Renewable-energy devices are installed singly or,
increasingly, as large arrays (e.g. Amoudry et al. 2009). Similarly, seafloor
observatories around the world rely on instrumented nodes and long cables con-
necting them back to shores (e.g. Favali et al. 2015). Successful siting and operation
rely on accurate and timely knowledge of the properties of the seabed, and of
relevant objects immediately below the seabed (e.g. boulders preventing piling or
drilling, gas pockets affecting the stability or safety of operations).
Advances in acoustic mapping, particularly in the last decades, have allowed
unprecedented access to seabeds all around the world, at depths down to the full
11 km of the Mariana Trench, and with resolutions varying from hundreds of
metres to centimetres (e.g. Blondel 2009; Blondel 2012). These maps provide
generic information about sites of interest. Further detailed investigations are
necessary to measure the physical and behavioural properties of the soil in those
places, assessing for example how easy it will be to drill or emplace structures (are
there any boulders big enough to stop or slow offshore work?), and how the bulk of
the seabed will respond to loads varying with time (for example through extraction
of the underlying gas reservoir, or processes such as scouring).
Knowing the properties of the immediate sub-surface has traditionally relied on
direct, physical sampling, for example using Cone-Penetration Testing (CPT) or
boreholes. The information provided is only valid for small areas, of the order of
square meters or generally less, at a maximum of 50 m below the surface (Harris
et al. 2008; Stark et al. 2014). Technical limitations mean deployments are currently
restricted to seabeds shallower than ca. 3000 m (Lunne 2012). Finally, these
properties are likely to change with distance. Detailed geophysical mapping
between test sites, as typified by the 10–50 m grid of North Sea surveys (e.g.
Semple and Rigden 1983; Ruffell et al. 1985; De Ruiter and Fox 1975), has proven
in general successful for a first examination of the broad horizontal uniformity of
soils, but limited in dealing with near-surface, sub-seabed geo-hazards. Difficulties
also exist in knowing what constitutes reality, also known as “the ground truth”. For
example, despite a test density of about 1 borehole or CPT every 800 m2, pile
driving in the North Sea Forties field revealed significant variations in soil prop-
erties that were not predicted by the borings (De Ruiter and Fox 1975). These
unexpected variations were apparently influenced by the assumption that anoma-
lous strength data in a weak zone were due to sample disturbance. In situ tests,
while providing some relief from the problems of sample disturbance, may be
affected by fabric-related discontinuities on a scale larger than that affected by the
test procedure (Marsland 1985). In cases like the exploitation of surface deposits,
limited ground sampling might also preclude full-scale assessments of the prof-
itability of extraction (e.g. Masuda et al. 2014). Exploitation of gas hydrates will
require similar levels of information (e.g. Hart et al. 2011).
1.1 The Need for Accurate Assessments of Sub-seabed Sediments 3
Can remote sensing techniques help? Geophysical sensing techniques used for
land-based work (Telford et al. 1990), such as electromagnetic (including radar) or
gravity can provide some information, but they are often limited at sea by their
resolution and/or use of operation in water. Seismic and acoustic techniques are
therefore mostly used.1 High-resolution acoustic profiles are used to trace the
continuity of acoustic interfaces (i.e. reflectors), based on their distinct physical
properties (such as bulk density, shear modulus, Young’s modulus or Poisson’s
ratio). These profiles are generally gathered from towed or propelled systems such
as Remotely Operated Vehicles (ROVs). Scientific literature however shows clearly
that, under certain conditions, acoustic reflectors do not correlate well with
boundaries identified by geotechnical tests (e.g. Mayer 1979; Mayer and LeBlanc
1983; Guigné 1986). This situation can be caused by interactions between the finite
bandwidth of an outgoing acoustic pulse and the soil stratigraphy, or by soils
characterized by a high degree of inhomogeneity (e.g. glacial tills or boulder lags).
Acoustic attenuation will also vary with frequency: in most cases, it is not known a
priori and it is affected by the geometric spreading as the sound waves travel
through the sediments. These conditions give anomalous reflections and distorted or
masked stratigraphic profiles. Their spatial extents will be limited by the beam-
widths of the transducers used. And the vertical resolution will consist in a trade-off
between higher frequencies (higher resolution, but also lesser penetration into the
seabed) and lower frequencies (lower resolution, but less attenuation, therefore
deeper imaging).
Whether acquired by direct physical sampling or by acoustic/seismic remote
sensing, the datasets acquired by engineers and geophysicists by necessity include
gaps, over which empirical correlations can prove tenuous. Even with much field
experience, there is always the risk that, for whatever reason, discontinuities exist
precisely in the region over which the data is to be interpolated (Fig. 1.1).
Correlating between datasets, from whatever origin, is not easy. Quantitative results
are directly tied to the level of calibration of the instruments, and how this was
tested in the field (Lunne 2012). Achieving agreement between, for example,
acoustic and penetrometer data is only achievable if the lateral extent and the
variability of the different sediments is known when planning in situ cone tests. The
ability to develop offshore resources in a safe and cost-effective manner is therefore,
unsurprisingly, based on the accuracy of the acquired sub-seabed information.
Significant losses, mostly economic but also of equipment or lives in the most
extreme cases, can result if the information is inaccurate.
1
Although there is no clear-cut and generally agreed definition, “seismic” techniques are generally
meant to encompass all uses of acoustic waves with frequencies below 1 kHz (e.g. Telford et al.
1990) whereas “acoustic” techniques extend above 1 kHz, up to hundreds of kHz or higher (e.g.
Lurton 2010).
Fig. 1.1 Top The standard piles used offshore approach five-meter diameters. Bottom A typical
buried boulder from the East Coast of Canada, wide enough to halt pile-driving if not detected in
time (PanGeo Subsea Inc. marketing archive, 2010)
1.2 Importance to Present Maritime Operations 5
1.2 Importance to Present Maritime Operations
Mobile Offshore Units are the traditional beneficiaries of detailed geotechnical

investigation, as their installation, operation and removal present particular risks
(Noble Denton 2013; UK Offshore Operators Association Ltd. 1997). Figure 1.2
shows how they are placed once towed onto the operation site. Accurate positioning
of the different platform legs requires good knowledge of the most stable places,
and the accurate identification of geohazards (to move away from them, or to
activate the necessary risk remediation measures) (Wong et al. 2010).
These mobile legs, called jack-ups, can have special footings (“spudcans”)
designed to increase their bearing area, thereby reducing the load requirements on
the sub-seabed immediately beneath each leg. They vary in size, the largest being
approximately 20 m in diameter.
Pre-loading is the most critical stage in placing a jack-up, as rapid penetration of
one or more legs may occur. A soil’s bearing capacity normally increases with
depth. But, when it is underlain by a weaker layer, there is a rapid reduction in soil
strength. As the spudcan reaches this interface, the weaker soil gives way and the
Fig. 1.2 Placement sequence of a Mobile Offshore Unit (Bennett and Associates 2005)
support of the leg moves downward faster than the jacking system can maintain
stability of the hull. This shifts the weights relative to the supports, thereby
increasing the required footing reaction needed to maintain equilibrium. This
continues until either the soil’s bearing capacity or the hull buoyancy (when it
enters the water) restores equilibrium. This phenomenon is referred to as “punch
through”. The risk of a punch through increases as jack-ups are being required to
work in deeper water and in locations where they are subject to greater environ-
mental loads. The consequences of an uncontrolled rapid jack-up leg penetration
can be extremely costly to the operator. It may result in structural problems:
– leg bending and/or damage to the leg-hull connection; failure of leg elements;
lost time and lost revenue due to downtime and repairs;
– excessive penetration, resulting in the operator discovering the jack-up legs are
not long enough for the location; and catastrophic events such as the collapse of
the rig (Fig. 1.3).
Spudcans can leave impressions on the seabed once the Mobile Offshore Unit
has been removed, particularly in locations with soft seabeds. If another unit is later
installed, these old spudcan impressions can induce horizontal forces on one or
more legs, as the spudcan tries to conform to the earlier impression. Because it is
Fig. 1.3 Failure of the AD19 jack-up in Saudi Arabia, in September 2002, was associated to
“punch through” by one of the legs, which subsequently collapsed under the load. Photography by
Bienen (2011)
Fig. 1.4 Jack-up rig toppling as a result of placing footing in an old spudcan footprint
disconnected from the other spudcans, this movement will bend the legs, causing
damage during pre-loading, or reducing allowable storm environment loads.
Identification of old spudcan impressions should therefore be an important part of
the site investigation. Typically, old spudcan footprints will have steep-sided
impressions with highly compacted soil at the base (Fig. 1.4). If still exposed, the
acoustic contrast with the surrounding surficial sediments will generally be enough
that they can be identified using multibeam echosounders or sidescan sonars.
However, if buried, these footprints are extremely difficult or impossible to map with
accepted practices (seabed penetration at the frequencies used not being high enough).
Even “pristine” seabeds are not without their problems. Hard ground patches
represent another type of localized anomalies, because they will have load-bearing
pressures different from their surroundings. Like spudcan impressions, they can be
detected with traditional tools if exposed at the surface, and if there is enough
acoustic impedance contrast with its surroundings, but are much more difficult, or
impossible, to detect if buried.
The upper boundaries of bedrock are another type of risk, because of rock
pinnacles and cavities (Figs. 1.5 and 1.6). Erratic variation in rock head elevation
presents another risk to jack-up rig placement, especially if buried (again, this is
because the foundation forces will not be balanced, inducing strong risks of sliding
or toppling). Jack-up legs reaching cavities might fail to reach a contact surface on
which they can rest. Once grouting starts, to consolidate the pile positions, large
amounts of grout can also be pumped and lost within the cavities, affecting overall
stability. This is a common occurrence during piling operations in the Gulf of
Arabia, as recently observed by the lead author.
In some cases, the upper layers of the seabed might also directly overlay gases or
fluids under pressure. Accidental breaching of the capping layers might create direct
environmental and technical problems, like gas flares, blowouts or the release of
pollutants into the water (e.g. OGP 2000). This problem will be increasingly likely
Fig. 1.5 Limestone formation exposed by erosion in Dukhan, Qatar. The structure shows different
layers and exposed cavities. Photography by Zitona [CC BY 2.0 (http://creativecommons.org/
licenses/by/2.0)], via Wikimedia Commons (https://commons.wikimedia.org/wiki/File:Zikreet_
rock_formations,_Dukhan,_Qatar.jpg)
as operations move toward previously exploited areas of the seabed, some of which
can be unevenly documented. For example, Lake Maracaibo in Venezuela has seen
more than a century of oil exploitation, with pipelines and other structures some-
times overlaying each other, some having been abandoned decades ago by now
untraceable operators. In other places, there is a risk of accidentally breaching into
gas or drilling fluid reservoirs left by previous operators. Similarly, the move
toward polar waters will increase risks of accessing gas hydrates close to the
surface, in waters made warmer by climate change.
Rapid expansion of the marine renewable energy sector is another driving factor
behind the need for accurate and competitive techniques of geophysical site inves-
tigation. A typical wind farm would consist in hundreds of turbines, pile-driving
several tens of metres into the seabed. Each foundation would be 40–60 m tall, 5-m
Fig. 1.6 Installation of offshore wind turbines requires specialist vessels and/or Mobile Offshore
Units, pile-driving at tens or hundreds of locations in a pattern made to maximise energy
generation. (Image credit Siemens, http://www.siemens.com/press/)
wide, weighing up to 530 tons (e.g. Sheringham Shoal, UK). Each pile would be
installed by Mobile Offshore Units needing stable positioning on the seabed, and its
position would need to follow an overall pattern designed to optimise energy gen-
eration from the farm. As tidal turbines progress from the prototype stage to actual
implementation, the same arguments will be made. In the case of renewable ener-
gies, the fact they are most often located close to shores and in areas of high
biodiversity add complexity to their installations: they need to take place at specific
times of years, or within specific timeframes, and the number of sites needs opti-
misation in respect to the ecological impacts of pile-driving (noise levels, dis-
placement or injury of specific animal species, etc.). Cost over-runs linked to
inaccurate geotechnical information will have the same detrimental effects as with
other offshore works, but will also run the risk of seriously affecting the project, as its
environmental cost will increase, potentially leading to cancellation.
1.3 The Technology Gap
The scope and ambition of marine geotechnical investigations have greatly evolved
over the last years. Offshore foundations, dredging operations of harbours and
channels, sub-seabed installations all place new demands on acquiring more reliable
knowledge on the composition of the seabed to address cost-effectively the issues of

buried geo-hazards, sediment property discontinuities and trapped pollutants being
released into the water during excavation operations. The two approaches seen so far,
namely physical sampling and acoustic imaging, have time and again shown their
potentials, but also their limitations, even if used in conjunction.
Diver- or camera-assisted sampling, like video observations, only show a limited
view of the seabed. They can easily be limited by water turbidity (for video) and
bottom currents (for diver sampling). Physical coring and sampling, e.g. with CPT,
goes deeper below the surface, but still only provides information over very limited
areas/volumes. These samples are generally collected at great costs, limiting their
numbers and sampling density. As an illustration, the European Wind Energy
Association recently stated (EWEA 2016) that the cost of geotechnical survey and
foundation construction is “great” at 21% of the total capital cost of an offshore wind
farm. Modern geotechnical surveys cost approximately2 USD 11.7 M for a typical
500-MW wind farm (Crown Estate 2010). This can be associated to individual costs of
typically USD 16 k for a CPT test in shallow waters (25 m), for an operation time of
4 h, compared to USD 117 k for a borehole in deeper water (50 m), for 14 h of
operation or more (Foley 2014). Current estimates are that at least USD 250 M will
need to be spent annually just for wind farm projects in the North Sea. There are of
course many other types of offshore projects planned, and the North Sea is but a small
part of the world’s seas. Market data suggests this will lead to a doubling or tripling of
the current levels of spending, notwithstanding a very likely shortage in the supply of
equipment, suitable vessels and appropriately trained personnel.
If boulders are found in some places, limiting for example the extent of bore-
holes, how do the measurements connect to the “real” state of sub-seabed condi-
tions? And how can all these measurements be integrated? Use of geostatistical
methods to optimise site investigations (e.g. by De Smith et al. 2006; Ditlevsen and
Madsen 2007) aimed for example at reducing the number of boreholes necessary
for wind farm installations, but models of foundation costs vs. site investigation
costs (Stuyts et al. 2010) concluded that even the best geostatistical techniques still
gave uncertain results. The risk of failure is also compounded by the often strong
disconnect between what is recorded in a physical sample, and what is measured in
geotechnical strength tests, e.g. CPTs. The mismatch in spatial and temporal scales
is reinforced by the provision of data at different resolutions, and at distinct loca-
tions. Figure 1.7 compares for example the information obtained by a borehole
(left), CPT sampling (middle) and what is actually found by acoustic imaging
(right). The borehole has a typical spatial footprint of 2–12 inches (ca. 5–30 cm)
and identifies the vertical sequence of sediment layers. Conversely, the CPT only
measures resistance and has no associated, quantifiable spatial footprint. It provides
the vertical geotechnical changes within each of these layers. But only the
2
The original quoted values of £7.5 M, £10 k and £75 k were converted to US dollars and are
therefore estimates. The orders of magnitude will remain, although individual costs will of course
depend on sites and marine environments.
1.3 The Technology Gap 11
Fig. 1.7 Comparison of information provided by a borehole (left), a CPT (middle) and wider-area
acoustic imaging (right). The borehole correctly identifies the different sediment layers. The CPT,
located a few metres away, detects slight changes in soil resistance. But only the acoustic imaging
reveals that the upper sand layer is actually thinning in just a few metres, implying a very strong
risk of “punch through”. (prepared by Ian McDermott, PanGeo Subsea Inc. marketing material,
2010)
wider-area acoustic imaging reveals that the upper sand layer is actually thinning
over the course of a few metres. This would definitely affect any offshore instal-
lation (as spectacularly shown in Fig. 1.3).
Does this mean acoustics should always be preferred? Not necessarily so.
Surface-mapping tools such as multibeam echosounders or sidescan sonars are
good at mapping larger areas, but acoustic returns correspond mostly to backscatter
from the immediate surface of the seabed (Fig. 1.8). Higher-density soils, in this
case from previous spudcan impressions, show as stronger acoustic returns, mod-
ulated by the local topography and the angle of imaging (see Blondel 2009 and
Fig. 2.1, for details of how backscatter is affected by the geometry). Conventional
sub-bottom profiling will be less affected, as it images these features from directly
above (normal incidence), therefore relying on specular energy. Traditional data
processing, or more sophisticated acoustic profiling acquisitions with long receiver
Existing Spudcan Impressions in the Seafloor
Fig. 1.8 Multibeam echosounder image of existing spudcan imprints on the seabed (Wong et al.
2010). Approximately 5 m in diameter, they are coded in blue. Note some of them show evidence
of scouring at their periphery, whereas smaller impressions on the seafloor are only visible because
of the purpose-defined colour scheme. The spudcan imprint highlighted with an arrow will be
analysed in more details later (Fig. 1.11)
offsets, render the data via migration/inversion protocols so as to discriminate

against the diffuse, non-specular energy, often treating it as noise. In many
instances, the very detail of the geological structure to be imaged is too small
(sub-wavelength scale) for a strong specular reflection to be built up laterally. The
textural character and detail of the sub-seabed is thus not always resolvable. Current
acoustic surveying approaches lack the ability to image the non-specular returns
that typify old buried spudcan impressions, patches of hard ground, punch through
potentials of thin, uneven sub-seabed formations, cavities, and rock pinnacles.
The technology gaps are therefore becoming clearer:
• There is a gap in scales between wide-area imaging techniques, such as seismics or
seabed mapping, and very localised ground sampling, such as boreholes or CPTs.
• There is a gap in resolutions between these two approaches. This can best be
addressed with another acoustic tool, which needs to be designed to provide
complementary information to existing techniques, if possible at lower costs, and
keeping the sub-seabed resolution as constant (and high) as possible (Fig. 1.9).
1.4 Filling the Gaps—The Case for “Acoustic Coring” 13
Fig. 1.9 The different approaches to geotechnical site investigation show a “technology gap”,
which can be filled by an “acoustic corer”, working in complement to existing measurements or as
a replacement if needed. (Image source Simmons and Company International—Information
Memorandum on PanGeo Subsea Inc., March 2012)
1.4 Filling the Gaps—The Case for “Acoustic Coring”
Acoustic “coring” should also make a fuller user of the rich information provided
by non-specular returns from complex sub-seabeds (Fig. 1.10). This requires sev-
eral receivers, emplaced where they will make the most use of the different com-
ponents of scattering: from the surface, from sediment layers, from heterogeneities,
from dipping large or small objects.
Sound source
Receivers
Fig. 1.10 Acoustic returns from the seabed and layers/objects below its surface make for a
complex set of information, which is best analysed using several receivers, and identifying both
specular and diffusive returns
The concept of acoustic coring, also called Acoustic Sub-Seabed Interrogator

(ASI), was introduced in Guigné (1986) and refined over the years (e.g. Guigné
et al. 1989; Guigné 2015), building on technological advances, progress in signal
processing techniques, and field experience all over the world. It addresses the
following concepts, presented in more detail in the following chapters:
– The ASI source will ideally be fixed on a stationary platform, allowing the
concurrent use of multiple data acquisition protocols. It will need to use a range
of low frequencies (to penetrate deeper into the seabed) and a range of high
frequencies (to enable higher vertical resolutions closer to the surface). Ideally,
it should use the interaction between these frequencies to increase the resolution
(suggesting the possibility of combining sources to form parametric arrays,
whose narrow beamwidths will enable higher horizontal vertical resolutions).
– The different receivers will need to cover a dense area, with horizontal
dimensions greater than 5 m and typically 12 m or more. These receivers will
need to be broadband, and they will use beam steering to focus on specific
reflectors. They will move in a controlled manner in the same plane as the
source transmitters for a wide range of possible emergent ray angles. They will
be organised as a phased array of hydrophones, to capture the time histories of
the returns and quantify beam spreading for particular reflectors. Focusing on an
emergent beam angle of interest is a powerful criterion of an ASI, allowing
particularly weak or distorted signals to be analyzed. The dynamic operation of
the transmitters and receivers will be controlled by a logical unit that uses
real-time processing.
– Data interrogation is done by first processing it into a volume whereby a
layer-by-layer (or depth slice where stratigraphy is obscured or lacking) data
analysis can be made to examine both the specular nature of stratigraphic layers
and the non-specular responses of discontinuous features such as boulders,
resulting in an “acoustic core” (Fig. 1.11). A sequential analysis of the time
histories (data traces) can be presented graphically. Bathymetry, layer thickness,
seismic velocities, attenuation, and other data such as the extent of seabed
inhomogeneity or internal scattering (attenuation) can also be emphasized to
allow for a thorough analysis of what truly characterizes the geotechnical nature
of a sub-seabed.
Typical acoustic cores quantify 3-D scattering in a cylinder-shaped volume, with
horizontal dimensions greater than 5 m (typically >12 m), extending down to 10 m
or more. How do they compare with other tools?
An acoustic core records variations of acoustic scattering, and strong intensities
correlate well with high CPT resistance values (Fig. 1.12). Figure 1.7, similarly,
showed strong correlation with CPT results, and the added benefits of imaging an
entire volume, sometimes highlighting unexpected discontinuities.
1.4 Filling the Gaps—The Case for “Acoustic Coring” 15
Fig. 1.11 “Acoustic corer” view of the jack-up spudscan highlighted with an arrow in Fig. 1.8.
Top Plan view, showing smaller circular imprints, interpreted here as reactions to the weight of the
CPT frame on the seafloor. Bottom Cross-section through the XZ profile, showing the depth
resolution and revealing differences between imprints A1 and A2, which appeared similar in the
plan view. (PanGeo Subsea Inc. Acoustic Corer Baltic 1 Survey Interpretive Report for EnBW
Ostsee Offshore GmbH February 2010)
Acoustic coring provides much more detail than would be achieved by con-
ventional seismic surveys, and the resolution does not degrade with distance from
the source, contrary to single-beam profiling. Compared to sidescan sonar or
multibeam echosounding, acoustic coring has the obvious advantage of penetrating
into the seabed to controllable depths. It also adds the benefit of multiple views
from the same objects. Using SAS rendering (see Chap. 3) and beamforming,
signals scattered by boulders and similar discontinuities can be analysed in the
directions of highest scatter, better capturing the details of complex seabeds.
Fig. 1.12 Comparison of an acoustic core (left) with two CPT cone resistances (right) and their
interpretation (middle). In this case, the sudden increase in core tip resistance comes from
alternating bands of melt water glacial (Mw Gc) sand and clay, contrasting with the overlaying
marine late glacial (Ma La) clay layer above, acoustically transparent. (PanGeo Subsea Inc.
Acoustic Corer Survey—Processing and Interpretation Report—Anholt Offshore Windfarm 2011)
The next chapters will present in more detail how this is achieved, first focusing
on acoustic measurements directly from the seabed and immediate sub-surface
(Chap. 2) and then extending the concept to deeper layers within the seabed
(Chap. 3).
References
Amoudry L, Bell PS, Black KS, Gatliff RW, Helsby R, Souza AJ, Thorne PD, Wolf J (2009) A
scoping study on: research into changes in sediment dynamics linked to marine renewable
energy installations. In: NERC Marine Renewable Energy Theme Action Plan Report, Natural
Environment Research Council, UK. Available via. http://www.nerc.ac.uk/research/funded/
programmes/mre/marine-scoping-study-2/. Accessed 17 Oct 2016
References 17
Bienen B (2011) Geotechnical considerations in the context of carbon storage in ocean spaces. In:
The Lloyd’s Register Educational Trust Research Collegium, Southampton. Available via.
http://www.southampton.ac.uk/assets/imported/transforms/content-block/UsefulDownloads_
Download/F85ED7735CDD48E3B1274272858EAF68/alecture12%20bienen.pdf. Accessed
17 Oct 2016
Bennett and Associates, LLC and Offshore Technology Development Inc. (2005) Jack-up units: a
technical primer for the offshore industry professional. Technical Report
Blondel Ph (2009) Handbook of sidescan sonar. Springer, Heidelberg
Blondel Ph (ed) (2012) Bathymetry and its applications. InTech Publishing, Rikeja. Available via.
http://www.intechopen.com/books/bathymetry-and-its-applications. Accessed 17 Oct 2016
Crown Estate (2010) A guide to an offshore wind farm. Available via. http://www.thecrownestate.
co.uk/media/5408/ei-a-guide-to-an-offshore-wind-farm.pdf. Accessed 17 Oct 2016
De Ruiter J, Fox DA (1975) Site investigation for North Sea Forties Field. In: Proceedings,
seventh annual, offshore technology conference, Houston, vol. 11. p 21–37. Available via.
https://www.geplus.co.uk/download?ac=1428270. Accessed 17 Oct 2016
De Smith M, Goodchild M, Longley P (2006) Geospatial analysis, a comprehensive guide. 3rd
edition. Matador, Leicester. Available via. http://www.spatialanalysisonline.com/. Accessed 17
Oct 2016
Ditlevsen O, Madsen HO (2007) Structural reliability methods. Technical University of Denmark.
Available via. http://chodor-projekt.net/wp-content/uploads/BiNSK/Literatura/Dilevsen,
Madsen,%20Structural%20Reliability%20Methods%20(2007).pdf. Accessed 17 Oct 2016
EWEA (European Wind Energy Association) (2016) The European offshore wind industry—key
trends and statistics 2015. Available via. http://www.ewea.org/fileadmin/files/library/
publications/statistics/EWEA-European-Offshore-Statistics-2015.pdf. Accessed 17 Oct 2016
Favali P, Beranzoli L, De Santis A (eds) (2015) Seafloor observatories—a new vision of the earth
from the Abyss. Springer Praxis, Chichester
Foley J (2014) Driving down the cost of offshore wind foundations. Available via. http://angles.
atkinsglobal.com/opinion/driving-down-the-cost-of-offshore-wind-foundations. Accessed 17
Oct 2016
Guigné JY (1986) The concept, design and experimental evaluation of “acoustic sub-seabed
interrogation” Ph.D. thesis, University of Bath, Bath
Guigné JY (2015) The future for “acoustics in the seabed”—the next chapter. J Ocean Tech
10(3):37–47
Guigné JY, Pace NG, Chin VH (1989) Dynamic extraction of sediment attenuation from
subbottom acoustic data. J Geophys Res 94(B5):5745–5755. doi:10.1029/JB094iB05p05745
Harris MM, Avera WE, Abelev A, Bentrem FW, Bibee LD (2008) Sensing shallow seafloor and
sediment properties, recent history. In: Proceeding of the IEEE Oceans’, 2008. doi:10.1109/
OCEANS.2008.5289424
Hart PE, Pohlman JW, Lorenson TD, Edwards BD (2011) Beaufort sea deep-water gas hydrate
recovery from a seafloor mound in a region of widespread BSR occurrence. In: Proceedings of
the 7th international conference on gas hydrates (ICGH 2011), 2011. Edinburgh
IPCC (2014) IPCC WGII AR5 summary for policymakers—climate change 2014: impacts,
adaptation, and vulnerability. Intergovernmental panel on climate change. Available via. http://
ipcc-wg2.gov/AR5/images/uploads/WG2AR5_SPM_FINAL.pdf. Accessed 17 Oct 2016
Lunne T (2012) The Fourth James K. Mitchell Lecture: the CPT in offshore soil investigations—a
historic perspective. Geomech Geoeng 7(2):75–101. doi:10.1080/17486025.2011.640712
Lurton X (2010) An introduction to underwater acoustics: principles and applications. Springer
Praxis, Chichester
Marsland A. (1985) The influence of geological processes and test procedures on measured
evaluated parameters. In: Advances in underwater technology and offshore engineering; 3,
Offshore site investigation. Graham and Trotman Ltd., London, pp 231–252
Masuda N, Okamoto N, Kawai T (2014) Sea-floor massive sulfide mining—its possibility and
difficulties to emerge as a future business. In: Drebenstedt C, Singhal R (eds) Mine planning and
equipment selection. Springer, Heidelberg, pp 105–112. doi:10.1007/978-3-319-02678-7_11
Mayer LA (1979) The origin of fine scale acoustic stratigraphy in deep-sea carbonates. J Geophys
Res 84(Bll):6177–6184. doi:10.1029/JB084iB11p06177
Mayer LA, Leblanc lR (1983) The chirp sonar: a new quantitative high-resolution profiling
system. In: Pace NG (ed) Proceedings institute of acoustics conference: acoustics and the
seabed. University of Bath, Bath
Noble Denton (2013) Seabed and sub-seabed data required for approvals of mobile offshore units,
0016/ND Rev 7. Available via. http://www2.dnvgl.com/nd-0016. Accessed 17 Oct 2016
OGP (2000) Flaring and venting in the oil and gas exploration and production industry: an
overview of purpose, quantities, issues, practices and trends, International Association of Oil
and Gas Producers Report No. 2.79/288. Available via. http://www.ogp.org.uk/pubs/288.pdf.
Accessed 17 Oct 2016
Pilley TVR (2008) Aquaculture and the environment, 2nd edn. Wiley, New York
Ruffell JP, Romeril PM, Jones KW (1985) A critique of arctic marine site investigation techniques.
In: Proceedings Arctic ‘85, civil engineering in the arctic offshore, San Francisco
Semple M, Rigden WJ (1983) Site investigation for Magnus. In: Proceedings of the fifteenth
annual offshore technology conference, Houston
Stark N, Hay AE, Trowse G (2014) Cost-effective geotechnical and sedimentological early site
assessment for ocean renewable energies. In: Proceedings of the IEEE Oceans’ 2014, St. John’s
Stuyts B, Vissers V, Cathie DN, Dorfeldt S (2010) Optimizing site investigations and pile design
for wind farms using geostatistical methods: a case study. In: Proceedings of the frontiers in
offshore geotechnics, Perth
Telford WA, Geldart LP, Sheriff RE (1990) Applied geophysics, 2nd edn. Cambridge University
Press, Cambridge
UK Offshore Operators Association Ltd. (1997) Guidelines for conduct of mobile drilling rig site
surveys v. 1.2. Available via. https://projects.dnvgl.com/iso-tc67-sc7-wg7/sname/Ref-02-The_
UKOOA_Surveying_and_Positioning_Commitee_Technical_notes_for_the_conduct_of_mobile_
drilling_rig_site_surveys_-_vol_1_of_2.pdf. Accessed 17 Oct 2016
Wong PC, Hau KW, Stonor RWP, Elfayoumi NF (2010) Feasibility of seafloor remediation for
multiple jack-up visits: RasGas Qatar North field. In: Proceedings of the offshore Middle East
Conference and Exhibition, Dohar
Chapter 2
Imaging of the Near-Surface
Abstract Acoustic sounding of the seabed and its immediate sub-surface produces
different types of returns, based on aspect, roughness and physical characteristics of
the different targets. This is presented within both seismic (low frequencies) and
sonar (high frequencies) contexts. Both authors describe their own studies in
controlled laboratory environments and at sea, contrasting larger-scale benthic
habitats with lower-scale buried targets. The key results are then compared with
existing research in acoustic Computed Tomography. Each example builds up to a
set of recommendations, about the source(s) and signals; about the receivers and
their spacing; about the hardware (constrained by the harsh realities of underwater
environments); and about the requirements of bespoke signal processing drawing
on a large range of independent techniques. The relative merits of standard pro-
jectors and parametric arrays are discussed and the important roles of acoustic
attenuation within the seabed and non-specular returns are introduced.
Keywords High-frequency ! Low-frequency ! Broadband ! Parametric arrays !

Computed Tomography
2.1 Capturing the Relevant Characteristics

of Near-Surface Sediments
The field examples in Chap. 1 highlight the important types of information pro-
vided by existing techniques, namely borehole, CPT, surface sampling and acoustic
mapping with multibeam and/or sidescan sonars. These examples also show their
limits in capturing sub-surface characteristics with a good enough resolution and a
wide enough area/volume coverage. Acoustic techniques seem a promising
approach (Sect. 1.4) but acoustic returns from the seabed and layers/objects below
its surface will be complex, even more so by combining measurements at different
frequencies, different horizontal resolutions and different vertical resolutions (with
frequency-dependent attenuation). “Traditional” (i.e. sonar) data will therefore

20 2 Imaging of the Near-Surface
require bespoke processing and interpretation techniques. The solutions chosen are
part of a rich field of possible approaches, detailed inter alia in Blondel (2009) and
Montereale Gavazzi et al. (2016), but not discussed here for the sake of brevity.
The physical and geological characteristics modulating acoustic returns from the
seabed are now pretty well understood, depending on the frequencies used and the
imaging mode (monostatic, where the same platform is used to transmit sound and
measure its backscatter along the same line of sight, as opposed to multi-angle, or
multistatic, where the sound source and receivers are physically decoupled).
Large-scale sonar approaches are highlighted in, for example, Blondel (2009) for
sidescan and Blondel (2012) for multibeam. Multi-static approaches to sonar
imaging of buried waste is also presented in Blondel and Caiti (2007). Research
from the last decades has identified the main factors in sediment acoustics, in
particular with the seminal SAX’99 and SAX’04 programmes in the US (e.g.
Thorsos et al. 2005). Local topography, roughness and physical make-up all
modulate the acoustic returns in predictable ways (Fig. 2.1). These principles are
used in habitat mapping (e.g. Kenny et al. 2003), in site investigations before
developments (e.g. OSIG 2014) and in similar activities.
Acoustic attenuation varies with sediment type and layer thickness. Not knowing
a priori values, it is often determined with the spectral ratio method. Because it uses
the geometric spreading of the sound waves as they travel through sediments, it
relies on accurate knowledge of the beam patterns at different frequencies. Guigné
et al. (1989a, b) showed the short and narrow-beam signals enabled by parametric
arrays had the right characteristics to investigate variations in different sand layers,
in a laboratory setting. This resulted in an adaptive determination procedure, further
validated in the field, which provides an exact model of sound velocities in
unknown seabeds.
Seabed surfaces are not always pristine geological environments. They are often
affected and reworked by marine life (e.g. coral reefs, burrows) and anthropogenic
activities (dumped objects or surface constructions, drill cuttings, trawling).
Large-scale structures such as carbonates or corals are easily identified at the sur-
face with sonar and conventional seismics (e.g. Hovland 2008). Pipe-lines, cables
or well-heads are most often proud above the surface but sometimes buried, through
sedimentation or as protection from trawling. Signatures of marine life can also be
much more subtle (Fig. 2.2) but still contribute significantly to acoustic returns
from different directions, especially if extending over large areas and at depth.
Although dumping is generally prohibited, 6.4 Mt of marine litter are deposited
every year (UNEP 2005). Legacy waste will also affect returns from the seabed and
immediate sub-surface, e.g. toxic buried waste (Blondel and Caiti 2007) or oil spills
(Parthiot et al. 2004; Medialdea et al. 2008). In areas of industrial activity, cuttings
(solid materials from the well bore) will generally smooth the seabed’s surface.
Their disposal is stringently regulated and like drilling fluids, they require proof that
vulnerable marine species or resources are not harmed (e.g. Storeng et al. 2009). In
other places, previous dredging or scouring around structures will change the
2.1 Capturing the Relevant Characteristics of Near-Surface Sediments 21
Fig. 2.1 Comparison of high- and low-frequency imaging at different angles. Top left, from
Blondel (2009): acoustic scattering will be influenced by the relative slope of the surface or object
of interest (with most of the energy scattering in the specular direction), by the roughness of the
target relative to the imaging wavelength (rougher surfaces presenting more facets likely to scatter
the energy in non-specular directions) and by the presence of heterogeneities (surface vs. volume
scattering). Top right, from Guigné et al. (2014): similarly, conventional seismic imaging will use
only the specular returns from horizons and discontinuities, if large enough relative to the imaging
wavelength(s). Bottom left, from Howey and Blondel (2008): differences between the scattering
expected at specific angles and the scattering actually measured can reveal buried targets. Bottom
right: the multi-angle technique designed by Guigné (1986) makes use of all possible returns,
revealing non-specular and diffuse scatterers
geotechnical characteristics of the surface (e.g. Wienberg and Bartholomä 2005).

Correctly understanding the exact characteristics of near-surface sediments can
therefore be challenging. It is also associated with very high costs. Clean-up costs
from marine litter are for example ranging in the millions of USD (UNEP 2005).
Benthic environmental surveys are now integral parts of assessments and con-
senting before offshore installations, and are estimated to cost around USD 0.75 M
for a typical 500-MW wind farm (Crown Estate 2010). It is therefore important to
“get it right”, achieving the best accuracy and the best repeatability.
Fig. 2.2 X-ray CT scan of a horizontal core from a coastal seabed near Venera Azzura (Italy). The
volume imaged (10 " 10 " 22 cm) matches the width of a borehole or the area of a
high-resolution sonar pixel. Shell pieces (yellow), small rocks (grey), animal burrows and water
pockets (both in red) penetrate the sediment (transparent) and contribute in different ways to
high-frequency volume scattering. From Pouliquen et al. (2001)
2.2 Buried Objects and Benthic Habitats
To image the fine details and stratigraphy of the immediate sub-surface, acoustic
instruments will require high frequencies, short pulses and ideally a broad band-
width. Parametric arrays are ideally suited because of their very fine beam patterns
and broad bandwidth signals (using the interaction between non-linear acoustic
waves, they produce a set of primary and secondary frequencies).
One example of this approach is the use of bistatic sonars for the detection and
identification of buried waste (Blondel and Caiti 2007). This was conducted as part of
the European project SITAR (“Seafloor Imaging and Toxicity Assessment of Risks
caused by buried waste”), and the approach was tested in the laboratory and validated
over a known dumpsite in the Stockholm Archipelago (Blondel and Pace 2009).
Conducted in shallow water and using a parametric array, decoupled from a
hydrophone chain further away (Fig. 2.3), these tests allowed the careful exploration
of the optimal geometries for imaging strong returns associated to man-made targets,
of the size of oil drums and often with metallic walls. In this case, the TOPAS-120
parametric sonar transmitted primary frequencies centred on 120 kHz and generated
secondary frequencies within the low frequency band 2–30 kHz. Repeat signals were
stacked to decrease the signal-to-noise ratio, enabling detection of milllimetric details
from up to tens of metres away. These experiments highlighted several important
issues. In shallow water, or close to the intended target, the need to transmit short
pulses directly limits the size of the scattering patch, and in some cases, sidelobes will
contribute significantly to the overall scattering strengths. Transmitter and receiver
acquisition need to be very accurately synchronised and positioned respective to each
other. These experiments also showed the necessity to move receivers away from the
strong reflectors. Other work (e.g. Schmidt et al. 1998) showed the role of recording
2.2 Buried Objects and Benthic Habitats 23
Fig. 2.3 Bistatic sonar setup: a the parametric array is positioned on an ROV, pointing at a target of
interest, and the signal is recorded on a chain of receiving hydrophones; b each signal will carry
distinct information, enabling reconstruction of scattering from below the surface (down to 10–20 cm)
and from inside the target, highlighting toxic waste even if buried. From Blondel and Pace (2009)
at very distinct locations, for example by positioning receivers on moving AUVs.

From a deployment perspective, the need to use several platforms at once makes this
method more expensive and potentially more time-consuming.
Fig. 2.4 Left: Details of the camera and 4 rows of transmitters (circles flush with the black face),
next to the receivers (white-tipped probes). Right: for deployments, the instrument was combined
with a frame grabber to directly sample seabeds of interest
Fig. 2.5 Left: Sampling strategy, with a sequential transmit/receive script executed in seconds for
each of the 40 positions, with hundreds of locally distributed soundings, giving a statistically
significant series of measurements in the region of interest. Right: example camera picture,
showing ground truth and the exact location of each area further sampled with acoustics (black
rectangles, encompassing approximately 12 " 30 cm on the seabed)
A related approach by Guigné and co-workers (Schwinghamer et al. 1996) had

been used slightly earlier to image the fabric and texture of benthic habitats,
applying it to the environmentally sensitive Grand Banks area offshore Eastern
Canada (Schwinghamer et al. 1996, 1998). In this case, the targets were much more
subtle, akin to those presented in Fig. 2.2. The challenge in this case was to acquire
information at high enough resolution (mm-sized voxels) to map potential habitats
over large areas. Designed by the lead author, Benthic-DRUMS™1 combined 4
rows of 10 independent, high-frequency, broadband parametric transducers, with
co-located receivers (Fig. 2.4, left). This complex acoustic instrument was posi-
tioned in the same frame as a grab sampler with a camera (Fig. 2.4, right). To cover
large areas of seabed in reasonable times, despite the varying water depths, a
“leap-frog” sampling strategy was adapted (Fig. 2.5, left). The Benthic-DRUMS
1
Dynamically Responding Underwater Matrix Sonar.
Fig. 2.6 Example data from a single transducer (from Guigné et al. 1991). Top left: dispersion test
of the water column, for frequencies of 120, 100, 80 and 60 kHz (from left to right). Top right:
acoustic returns from the seabed at the test site. Bottom left: frequency summation, highlighting
different fine-scale stratigraphy levels (B to K). Bottom right: instantaneous amplitudes for the 4
different frequencies
was deployed from a cable over the side of the ship, taking a picture of the “ground
truth” with accurate localisation of where the 40 detailed acoustic measurements
come from (Fig. 2.5, right) and offering the potential to sample the seabed in
locations of interest.
First tests in Hamilton Harbour, Ontario (Canada) showed the potential of the
frequencies used (60–120 kHz) to delineate fine seabed stratigraphy (Fig. 2.6). As
with the previous approach, stacking greatly increases the signal-to-noise ratio.
Here, though, stacking was also used with different secondary frequencies, as
recommended by Guigné et al. (1991). Hilbert transforms of the individual signals
provide their envelopes (Fig. 2.7), and the derivation of each waveform’s fractal
dimensions measures its irregularity and roughness, using a modified box-counting
method described in Schwinghamer et al. (1996). Combined with returns from 40
transducers at each sampling location, this technique provides cross-verification of
measurements over very small spatial scales (Fig. 2.5, right), and allows for rig-
orous statistical treatment of returns from each transducer and their variations.
The instrument and the signal processing methodology were tested extensively
during a 3-year experiment on the effects of otter trawling on benthic habitats on a
sandy-bottom ecosystem from the Grand Banks area offshore Newfoundland
(Schwinghamer et al. 1998). Each year, corridors 13 km long were trawled 12 times,
over widths of 120–150 m, each year from 1993 to 1995. Acoustic measurements
were taken before and after trawling, to document its effects and quantify sub-seabed
Fig. 2.7 Hilbert-transformed acoustic returns, for one transducer. They show the comparison
between untrawled control areas (left, plots A and C) and trawled areas (right, plots B and D). Five
depth zones are delimited each time (here, for the same intervals in time and for the upper 4.5 cm
of sediments). Signal levels vary between and within zones, giving access to a very fine
stratigraphy. From Schwinghamer et al. (1996)
changes (e.g. ecosystem recovery from year to year). The field data included sys-
tematic, high-frequency sidescan sonar surveys, to describe the surficial sediments.
In 1995, the RoxAnn™ acoustic bottom classification system was also used. These
two instruments were thoroughly compared with Benthic-DRUMS in 1994 and
1995. RoxAnn results, gathered with a footprint of 200 m2, were mixed and
sometimes contradicting the video evidence. Conversely, Benthic-DRUMS mea-
surements were consistent with the video evidence, with the sidescan sonar imagery
and with the many seabed samples (Schwinghamer et al. 1998), providing relevant
high-resolution benthic information over the first 4.5 cm of sediments over very fine
footprints. Typical results for a trawled area and a control area are shown in Fig. 2.8.
This approach was assessed independently and is now recommended by the UN
Food and Agriculture Organization (FAO) (Løkkeborg 2005) inter alia.
Similar challenges have been encountered in Computed Tomography, used for
example with X-ray imaging of the human body. This typically uses tens of thou-
sands of very closely spaced sensors, reconstructing 3-D and 2-D images from
thousands of different projections. But how easy would it be to translate this
approach to underwater environments, and buried objects? Acoustic Computed
Tomography has been tested by many authors, including Younis et al. (2002) who
investigated the imaging of shallow buried objects in a laboratory setting. Their study
used landmine-type objects buried 10–50 cm deep in homogeneous wet sand,
emplaced in an empty swimming pool (i.e. in air). 31 microphones emitted plane
Fig. 2.8 Fractal dimensions

(decreasing from white to
yellow, red and black) for the
40 transducers, sliced by
descending zones (1.6-cm
thick). They show the clear
differences between trawled
and untrawled (control) areas.
From Guigné and Pace (2007)
waves of frequencies 2000–3000 Hz. For each transmitted pulse, the reflected-
refracted signals are received by a line array located diametrically opposite the source
(Fig. 2.9), rotated at 1° intervals to cover a full horizontal circle. Pre-amplifiers
(with a fixed gain setting) were used before digitising (at 8 kHz) and multiplexing the
different signals. These were then used to provide individual time-series for each
sensor, combined into CT reconstructed images. Their measurements showed:
(1) the importance of coupling between sensors and the ground (arguably easier in
water than in air); (2) how pulse design must guide the design of the source array;
Fig. 2.9 Idealised view of the multi-transmitter, multi-receiver used by Younis et al. (2002) for
acoustic Computed Tomography of shallow buried objects in wet sand and air (not to scale)
(3) the need for a higher number of receivers to achieve good resolution (even when
the signal-to-noise ratio is good, which is less likely in complex sediments under
water) and (4) the need for bespoke signal processing techniques to get the most of
data, for example using adaptive interference cancellation.
These recommendations were carried out in independent studies later carried out
by Raytheon and Guigné International Ltd. (GIL), investigating ways to detect
targets buried in marine sediments (Raytheon/GIL 2004). A large pit (22 m long "
4 m wide " 7 m deep) was filled with relatively homogeneous marine sediments
typical of a beach or of shallow water environments, to a thickness of 3 m. These
sediments were carefully emplaced and made of 49% sand, 50% silt and 1%
gravel, with a mean grain size of 75 lm and a measured high attenuation of
0.53 dB/m/kHz. A large variety of targets were emplaced at selected positions
within the sediments, including an inert 81-mm mortar round (0.66 m long and
0.08 m in diameter). Like all other targets, it was supported with steel tubing (0.3 m
below) to prevent movements as the sediments compact and the targets settle under
their own weights. A DRUMS®-R200 parametric sonar, operating at 190 kHz,
transmitted short broadband impulses with modulation frequencies of 15, 20, 26
and 35 kHz. It was used to image the targets at accurately controlled locations and
tilt angles (Fig. 2.10) and distance to the seabed was monitored with a distinct depth
sounder.
Fig. 2.10 Schematics of the DRUMS-200 parametric sonar (light grey, top right of the frame) and
the ITC-6164 8-hydrophone array (dark grey, bottom right of the frame), with associated
electronics. Modified from Raytheon/GIL (2004)
Fig. 2.11 Example results, corresponding to the acoustic detection (left) of an inert mortar shell
(right) buried 50 cm deep. The 2-D slices at different (relative) depths show strong acoustic returns
as red, low or background returns as blue. The bottom slice shows additional (real) targets at two
of the corners. Adapted from Raytheon/GIL (2004)
Acoustic scattering from the water-sediment interface, from the individual tar-
gets (and supporting frames) and from the surrounding sediments was measured at
8 closely-spaced broadband hydrophones. Each received pulse included forward
scatter, reverberation and potential out-of-plane returns from strong reflectors (like
the targets). Signal processing included accurate positioning of the imaging and
receiving transducers relative to the simulated seabed, and beamforming (focused
at 1.0-m depth for this particular test). In this highly-controlled environment, it was
possible to detect most targets and in particular the mortar shell (Fig. 2.11). 2-D
slices of acoustic returns at different relative depths show the background return
from the surrounding sediments (top slice), and for each slice, at 1-cm interval,
the acoustic returns from the shell. Its shape is gradually revealed with depth, and it
is largest 4 cm after the first slice, consistent with its 8-cm diameter. Some of
the processing schemes used will be presented in Chap. 3 (“Imaging into the
sub-seabed”).
2.3 Significance and Value—Moving Toward

an Answer Product
Section 2.1 showed the importance of multi-aspect imaging, revealing more about
the sub-surface and the targets within, by using the variations in scattering at
different angles, and how this could be harnessed to detect subtle changes. The four
examples highlighted in Sect. 2.2 showed, respectively; (1) high-frequency
multi-aspect imaging of buried targets with a single parametric array and
multiple receivers on a vertical line array; (2) multi-aspect imaging of subtle

sub-seabed variations with multiple transmitters and multiple receivers organised
along a 2-D array; (3) acoustic Computed Tomography of large objects buried close
to the surface (in air) and (4) in water. Each of these approaches revealed much
more than traditional tools, illustrating how some of their key innovations can be
brought together in an answer product, consisting of carefully selected
hardware/sensors and a bespoke signal processing solution. These can be synthe-
sised in the following evidence-based recommendations:
Recommendation 1: The source(s) need to have a broad bandwidth combining
low and high frequencies, and they need to transmit high-amplitude, short signals.
The low frequencies are useful to detect discontinuities, and the high frequencies to
detect individual targets. The frequencies need to be tuned to the spatial wave-
lengths of each type of structure or target, but also to the acoustic penetration they
allow within the seabed. The signals transmitted need to be short, to improve image
resolution and distinction of the different arrivals. They need to have relatively high
amplitudes, to improve the signal-to-noise ratio as they get attenuated through the
sediments and back to the receivers close to the seabed.
Recommendation 2: The receivers need to cover different spatial scales but they
are necessarily limited in number. The spacing of the receivers needs to optimise
the possibilities of detecting individual returns from sub-seabed targets (disconti-
nuities or objects), with the potential aim of using beamforming to improve
localisation accuracy. Some of them can be above the potential targets of interest;
others will need to be far away, offering large scattering angles from the source(s) to
the target and forward to the receiver(s). The line of receivers will need to cover
large areas over the seabed, either by moving to cover a full circle (as in Younis
et al. 2002) or by being deployed in a larger pattern (e.g. as a circle or a spiral). The
total number of receivers is however limited and cannot be as high (typically tens of
thousands) as in X-ray Computed Tomography, because underwater use puts severe
constraints on the power available, multiplexing possibilities and data storage on
the subsea platform (or even on-board surface vessels).
Recommendation 3: The hardware needs to be relatively portable, for repeated
deployment in the harsh underwater environments. The relative positions of both
source(s) and receiver(s) need to be carefully controlled and synchronised accu-
rately enough to allow beamforming and other, more complex processing.
Recommendation 4: Bespoke signal processing will need to adapt to potentially
low signal-to-noise ratios, and it will need to distinguish between forward scatter
and out-of-plane returns. The resulting information will need to be presented as 2-D
horizontal slices and 3-D volume plots, relating the acoustic measurements to
parameters with a geophysical signification clear to the end-user. There should be
clear indications of the resolutions achievable each time, to allow confident iden-
tification of targets of interest.
The next chapter will show how this can be achieved in practice, introducing the
Acoustic Corer, the concept of the JYG-cross and the necessary signal processing
stages.
References 31
References

Blondel Ph (ed) (2012) Bathymetry and its applications. InTech Publishing, Rikeja. Available via:
http://www.intechopen.com/books/bathymetry-and-its-applications. Accessed 17 Oct 2016
Blondel Ph, Caiti A (eds) (2007) Buried waste in the seabed—acoustic imaging and bio-toxicity
(results from the European SITAR project). Springer-Praxis, Chichester
Blondel Ph, Pace NG (2009) Bistatic sonars: sea trials, laboratory experiments and future surveys.
Arch Acoust 34(1):3–17
Crown Estate (2010) A guide to an offshore wind farm. Available via: http://www.thecrownestate.
co.uk/media/5408/ei-a-guide-to-an-offshore-wind-farm.pdf. Accessed 17 Oct 2016
Guigné JY (1986) The concept, design and experimental evaluation of “acoustic sub-seabed
interrogation”. Ph.D. thesis, University of Bath, Bath, UK
Guigné JY, Pace NG (2007) An analytical acoustic framework to quantify the health of benthic
habitats. In: Proceedings 2nd Underwater Acoustic Measurements—Technologies and Results,
Crete
Guigné JY, Chin VH, Solomon SM (1989a) Acoustic attenuation measurements using parametric
arrays. Ultrasonics 27:229–301
Guigné JY, Pace NG, Chin VH (1989b) Dynamic extraction of sediment attenuation from
Guigné JY, Rukavina N, Hunt P, Ford JS (1991) An acoustic parametric array for measuring the
thickness and stratigraphy of contaminated sediments. J Great Lakes Res 17(1):120–131
Guigné JY, Stacey AJ, Clements C, Azad S, Pant A, Gogacz A, Hunt W, Pace NG (2014)
Acoustic zoom high-resolution seismic beamforming for imaging specular and non-specular
energy of deep oil and gas bearing geological formations. J Nat Gas Sci Eng 21:568–591.
doi:10.1016/j.jngse.2014.09.012
Hovland M (2008) Deep-water coral reefs—unique biodiversity hot-spots. Springer-Praxis,
Chichester
Howey R, Blondel Ph (2008) Bistatic scattering in sediments: comparison of model and scaled
tank experiments at 238 kHz. J Acoust Soc Am 123(5)2:3440
Kenny AJ, Cato I, Desprez M, Fader GB, Schuttenhelm RTE, Side J (2003) An overview of
seabed-mapping technologies in the context of marine habitat classification. ICES J Mar Sci
60:411–418
Løkkeborg S (2005) Impacts of trawling and scallop dredging on benthic habitats and
communities. FAO Fisheries Technical Paper 472. FAO, Rome. Available via http://www.
fao.org/docrep/008/y7135e/y7135e00.HTM. Accessed 18 Oct 2016
Medialdea T, Somoza L, León R, Farrán M, Ercilla G, Maestro A, Casas D, Llave E,
Hernández-Molina FJ, Fernández-Puga MC, Alonso B (2008) Multibeam backscatter as a tool
for sea-floor characterization and identification of oil spills in the Galicia Bank. Mar Geol 249
(1–2):93–107. doi:10.1016/j.margeo.2007.09.007
Montereale Gavazzi G, Madricardo F, Sigovini M, Janowski L, Kruss A, Blondel Ph, Foglini F
(2016) Evaluation of seabed mapping methods for fine-scale benthic habitat classification in
extremely shallow environments—application to the Venice Lagoon, Italy. Estuar Coast Mar
Sci 170:45–60. doi:10.1016/j.ecss.2015.12.014
OSIG (Offshore Site Investigation and Geotechnics Committee) (2014) Guidance notes for
the planning and execution of geophysical and geotechnical ground investigations for
offshore renewable energy developments. Cook M (ed.), Society for Underwater Technology.
Available via: http://www.sut.org/wp-content/uploads/2014/07/OSIG-Guidance-Notes-2014_
web.pdf. Accessed 18 Oct 2016
Parthiot F, de Nanteuil E, Merlin FX, Zerr B, Guedes Y, Lurton X, Augustin JM, Cervenka P,
Marchal J, Sessarego JP, Hansen RK (2004) Sonar detection and monitoring of sunken heavy
fuel oil on the seafloor. Proceedings of the Interspill 2004 Conference, Trondheim
Pouliquen E, Lyons AP, Pace NG, Michelozzi E, Muzzi L (2001) Backscattering from bioturbated
sediments at very high frequency. NATO SACLANTCEN Report SR-342. Available via:
http://www.dtic.mil/dtic/tr/fulltext/u2/a416953.pdf. Accessed 18 Oct 2016
Raytheon/GIL (2004) Technical demonstration of the DRUMS®-R200 sonar system. Technical
Report RPT-0440-900-015-B, Raytheon Integrated Defense Systems (USA) and Guigné
International Ltd. (Canada)
Schmidt H, Maguer A, Bovio E, Fox WLJ, LePage K, Pace NG, Guerrini P, Sletner PA,
Michelozzi E, Moran B, Grieve R (1998) GOATS’98: bistatic measurements of target
scattering using autonomous underwater vehicles. NATO SACLANTCEN Report SR-302.
Available via: www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA376854. Accessed 18 Oct 2016
Schwinghamer P, Guigné JY, Siu WC (1996) Quantifying the impact of trawling on benthic
habitat using high resolution acoustics and chaos theory. Can J Fish Aquat Sci (53)2:288–298
Schwinghamer P, Gordon DC Jr, Rowell TW, Prena J, McKeown DL, Sonnichsen G, Guigné JY
(1998) Effects of experimental otter trawling on surficial sediment properties of a sandy-bottom
ecosystem on the Grand Banks of Newfoundland. Conserv Biol 12(6):1215–1222
Storeng AB, Korneev O, Bambulyak A, Frantzen B, Lunde S, Novikov M, Olsen E, Shavikin A,
Storebø R, Sørgård T, Titov O (2009) Oil and gas activities—emission, operational and
accidental discharges. Barents-Portal (Joint Norwegian-Russian Environmental Status Report
for the Barents Sea). Available via http://barentsportal.com/barentsportal_v2.5/index.php/
en/barents-sea-status-report/background/human-activities/501-oil-and-gas-activities-emission-
operational-and-accidental-discharges. Accessed 21 Jan 2015
Thorsos EI, Williams KL, Tang DJ, Kargl SG (2005) SAX’04 overview. In: Pace NG, Blondel Ph
(eds) Boundary influences in high-frequency, shallow-water acoustics. University of Bath
Press, pp 3–10. Available via: http://opus.bath.ac.uk/8849/. Accessed 18 Oct 2016
UNEP (2005) Marine litter—an analytical overview. United Nations Environment Programme.
Available via: http://www.unep.org/regionalseas/marinelitter/publications/docs/anl_oview.pdf.
Wienberg C, Bartholomä A (2005) Acoustic seabed classification in a coastal environment (outer
Weser Estuary, German Bight)—a new approach to monitor dredging and dredge spoil
disposal. Cont Shelf Res 25(9):1143–1156
Younis WA, Stergiopoulos S, Havelock D, Grodski J (2002) Non-destructive imaging of shallow
buried objects using acoustic computed tomography. J Acoust Soc Am 111(5)1:2117–2127
Chapter 3
Imaging into the Seabed
Abstract Field and laboratory studies (presented in Chapter 2) showed the types of
acoustic returns from high-frequency and low-frequency sensors. These surveys have
also shown how data acquisition geometries and signal processing can be combined
to create a true Acoustic Seabed Interrogator (ASI), resulting in a set of 4 key
recommendations. Chapter 3 describes how this implemented, with the earliest ASI
prototype, developed by the lead author between 1990 and 1993. The procedures for
acquisition and data processing are presented along with field examples, and the main
data products are favourably compared with concurrent Cone-Penetrometer Testing
and borehole logs. Further developments led to a moving instrumentation platform,
blending low- and high-frequency sensors, and a generic flow diagram to process and
combine these measurements. The next generation Acoustic Corer™ is presented in
details, along with the several types of information available from its volumetric
datasets. This introduces the concept of JYG-Cross processing, and the use of
Synthetic Aperture for rendering and processing. The chapter also presents how
Quality Assurance can measure the quality of the measurements at all stages of
processing, in particular in shallow waters where multiple reflections are common.
Field examples are regularly used to illustrate these different achievements.
! !
Keywords Acoustic Seabed Interrogator (ASI) JYG-cross Synthetic Aperture !
! ! !
SAS Processing flow Quality Assurance Acoustic Corer™ Volumetric data ! !
! !
CMP gather Coherent summation Velocity focusing Chirp !
3.1 Creating a Prototype
The recommendations concluding Chap. 2 were first embodied with a prototype

Acoustic Seabed Interrogator (ASI), developed by the lead author and refined
through field trials between 1990 and 1993. It was based on 16 broadband sparker
sources delivering high-amplitude pulses, delivered through a Huntec seismic
energy power supply set to either 480 Joules or 1080 Joules (more details about
sparkers can be found in Buogo and Cannelli 2002). These sources were arranged

34 3 Imaging into the Seabed
Fig. 3.1 First prototype of the Acoustic Seabed Interrogator (ASI™) during construction, taken
around 1990 by the lead author. The central outer structure contains the framework of 16 sparker
transmitters and the 12-m boom with equally-space hydrophones. Annotations show some of the
receivers (labelled Rx) and one of the transmitter (labelled Tx) in the central structure. Field
deployment of the completed ASI™ is presented in Fig. 3.2
in an octagonal polyethylene framework held by a circular aluminium structure

(Fig. 3.1) and positioned in two concentric circles around the centre of the frame.
A boom extending from this centre supported 12 equally-spaced receivers (Brüel
and Kjaer 8105 omnidirectional hydrophones, individually calibrated and with a flat
frequency response between 0.1 and 180 kHz). Each hydrophone signal was con-
ditioned before digitising with Brüel and Kjaer 2635 charge amplifiers with built-in
antialiasing filters (0.2 Hz low-cut and 30 kHz high-cut frequencies). The signals
were sampled at 100 kHz (i.e. every 10 ls). The receiver boom was rotated at 45°
angles during data collection and aligned with 4 transmitters to form 4 transmitter-
receiver rows called “beams” (with 12 measurements each). This directly addressed
Recommendation 1 (about the sources) and Recommendation 2 (about the recei-
vers) seen in Chap. 2. The geometry chosen would also ensure 3-D volumetric
information would be acquired over a minimum diameter of 10 m, with potential
penetrations below the seabed in excess of 10 m.
Engineering design ensured the whole setup was relatively portable (Fig. 3.2),
and a major experiment took place in the near-shore zone off Terrenceville,
Newfoundland and Labrador (Canada) in 1991 (Guigné et al. 1991; Inkpen et al.
3.1 Creating a Prototype 35
Fig. 3.2 Deployment of the prototype (ASI™) from a pier in Newfoundland and Labrador
(Canada), photographed by the author in 1991. The boom with the receivers is facing away from
the photographer. Note the relatively small crane necessary to lift and recover this working
prototype, showing how it can be deployed from a variety of surface vessels
1991; Pike 1994, 1998). Along with the accurate localisation and synchronisation of
receivers and transmitters, its design therefore directly addressed Recommendation 3
(about hardware) from Chap. 2.
3.2 Formulating the First Product
The final Recommendation 4 of Chap. 2 indicated the need for bespoke signal pro-
cessing, enabling the derivation of both 2-D and 3-D information about the seabed and
its sub-surface. This will be illustrated using drawings and results from the 1991
experiment (Guigné et al. 1991). The deployment of the Acoustic Seabed Interrogator
in a complex sedimentary setting was completed with Cone Penetrometer Tests
(CPT) and several boreholes around the centre of the frame (Fig. 3.3).
Four rotations of the boom, at 45° angles, provided linear “beams” of 12 series of
measurements, corresponding to the activation of each of the 16 transmitters in turn.
This aimed to create many folds that are the number of source/receiver combinations
Fig. 3.3 Seabed configuration of the ASI field test: the transmitters are kept in the same location and
the boom on which the receivers are fixed is rotated through 4 different angles (beams 1–4). CPT tests
and boreholes are used to validate the approach (see text for details). Redrawn from Guigné et al. (1991)
sampling the Common Depth Point (CDP).1 Low- and high-power soundings were
processed along two data flows, summarised in Fig. 3.4. These measurements were
carried out in two surveys at two power levels (480 and 1080 J) and fused to form
four distinct panels of 2-D traditional seismic results (shown in Figs. 3.5, 3.6, 3.7 and
3.8). The seismic processing follows the conventional steps presented in full details
in publications by Mueller (2005) and Scheidhauer et al. (2005).
The identification of different horizons for each beam, 45° apart and over dis-
tances of 10 m or more (depending on how strong the reflections are), allows the
definition of depth contours for each layer. Isopach maps can also illustrate the
thickness variations of each unit (Fig. 3.9). Their combination provides a full 3-D
volumetric representation of the actual seabed below the ASI frame (Fig. 3.10).
Cone Penetrometer Tests were also performed close to the centre of the frame
(Fig. 3.3), recording shear (S) and compressional (P) wave data at 1-m intervals
each time the CPT was pushed further into the seabed. They are supplemented with
borehole tests, recovering sediments to the same depths and interpreted traditionally
(Fig. 3.11). By providing individual depths of each horizon in 2-D slices along each
“beam”, i.e. 3-D measurements when combining the different beams, the ASI
provides much more information than CPT or boreholes, limited to 1-D profiles and
constrained by the exact location they were done at (see Fig. 3.12 for detailed
comparison of what can be measured with each instrument, and how they compare).
1
Definitions of common seismic terms are assumed known, as they are described in many text-
books (e.g. Telford et al. 1990) and on reputable internet sources (e.g. http://www.glossary.oilfield.
slb.com/Terms/.aspx).
3.3 The Next Generation: The Acoustic Corer 37
Fig. 3.4 Bespoke signal

processing of the ASI
measurements for each power
level. AGC is the abbreviation
of Automatic Gain Control,
increasing the visibility of
late-arriving events affected
by attenuation or wave
spreading. Redrawn from
Guigné et al. (1991), Inkpen
et al. (1991)
3.3 The Next Generation: The Acoustic Corer
These first trials of an Acoustic Seabed Interrogator were very encouraging: indi-
vidual horizons could be picked from their acoustic responses, up to 10 m away and
8–10 m below the surface. These horizons correlated well with the concurrent
“ground truth” from CPT and boreholes, and they extended this information well
beyond the point information these traditional means could provide. Desirable
engineering issues were to improve the portability of the system, making it easier to
Fig. 3.5 Brute stack data panel for data acquired in beam 1 (geometry presented in Fig. 3.3). The
seismic rendering shows scattering strengths positive toward the right. Layers of different
velocities, derived from normal-incidence returns from Semblance analysis plots, are labelled L I
to L VII. Distinctive sedimentary boundaries, called horizons, are picked from the locally highest
reflections and labelled H1–H7. Reproduced from Guigné et al. (1991) and Pike (1998)
deploy at sea. The capability of imaging larger volumes of sub-seabed, and of fully
using the multistatic geometry of transmitters and receivers to identify the different
types of returns from horizons and distinct objects, needed unambiguous associ-
ating to specific structures of interest to end-users. This was addressed through a
series of laboratory-based investigations and field experiments (summarised in
Guigné et al. 1991; Raytheon/GIL 2004, and many other references in this chapter
and others, authored or co-authored by Guigné), taking place over the period 1992–
2005.
The refinement of hardware and software data interrogation and analyses
resulted in a comprehensive solution, called the PanGeo Subsea Inc. Acoustic Corer
(Fig. 3.13). This sophisticated ASI consists of sonar hardware and data collection
scripts, advanced digital processing and interpretation protocols to acquire both the
specular and non-specular responses of the first 30 m of complex sub-seabeds. The
new approach also enables fusion with other available geotechnical and geological
datasets, addressing the link between new acoustic measurements and legacy or
third-party information.
Fig. 3.6 Brute stack data panel for data acquired in beam 2 (geometry presented in Fig. 3.3). Note
layers and horizons vary slightly from those identified in beam 1 (Fig. 3.5), either because of
geomorphological differences or because of intrinsic acoustic variations. Reproduced from Guigné
et al. (1991), Pike (1998)
Based on previous experience, the Acoustic Corer uses three distinct transmit-
ters, located on the mobile instrumentation platform on each boom (Fig. 3.14): a
low-frequency (LF) chirp transducer, a high-frequency (HF) chip transducer, and a
parametric transceiver.
The LF transducer is used for pseudo-seismic measurements with the JYG-Cross
(see Sect. 3.5) and generation of velocity profiles within the sub-seabed. It is based
on a Neptune 4108 A/B transducer, transmitting chirp pulses 4.5 ms long over a
frequency range of 1.5–6.5 kHz,. Laboratory calibrations show 3-dB beamwidths
decreasing from ca. 120° to ca. 25°, as frequency increases, resulting in a
Directivity Index of 6 dB. The transducer is operated with a 10% duty cycle and
delivers an effective2 Source Pressure Level (SPL) of 165 dB re. 1 lPa @ 1 m.
The HF transducer and the parametric transceiver are used for volumetric data
acquisition, gathering information about sub-seabed layers and structures at
2
The effective SPL is the difference between the peak source level and the combination of the
Directivity Index, i.e. how much sound is actually radiated in the direction of measurement, and
the duty cycle (expressed in dB: 10% * −20 dB), i.e. how much sound is actually radiated over a
unit time. It is therefore a good measure of the actual acoustic impact of an intermittent, directional
source.
Fig. 3.7 Brute stack data panel for data acquired in beam 3 (geometry presented in Fig. 3.3).
Horizons picked in this “beam” are compared with CPT and borehole data in Fig. 3.11.
Reproduced from Guigné et al. (1991), Pike (1998)
different resolutions and range of angles, which can then be combined. Based on
experience, the HF source is most effective in the upper 20 m of seabed, whereas
the parametric source can penetrate down to 30–40 m below the seabed. The HF
source is based on a Neptune 4108 C/D transducer, transmitting chirp pulses 4.5 ms
long over a frequency range of 4.5 kHz to 12.5 kHz. Laboratory calibrations show
3-dB beamwidths decreasing from 78° to 48° as frequency increases, resulting in a
Directivity Index of 5 dB. This transducer is also operated with a 10% duty cycle
and delivers an effective SPL of 169 dB re. 1 lPa @ 1 m. The Innomar parametric
transceiver transmits pulses 0.25 ms long, with a primary frequency of 100 kHz
and secondary frequencies set at 5, 6, 10, 12 and 15 kHz. Being a non-linear source,
it has a very narrow beamwidth (3.5° at 3 dB down with SPLs measured at 3 m
varying around 170 dB re. 1 lPa to 182 dB re. 1 lPa). It points vertically toward
the seabed, and its first-order sidelobes at less than ±6° are smaller by >12 dB,
meaning they contribute much less, if not at all, to the actual reflections at depth.
The receiver array comprises omnidirectional hydrophones with a broadband
response encompassing all frequencies used by the transmitters. Resulting signals
are sampled at very high frequencies, conditioned with bandpass antialising filters,
and used in the different processing tasks (presented in the next sections). By
rotating the boom at all angles over 360°, and moving the sonar platforms along
Fig. 3.8 Brute stack data panel for data acquired in beam 4 (geometry presented in Fig. 3.3). Note
that layers and horizons vary slightly but are broadly consistent with those identified in the
previous 3 beams. Reproduced from Guigné et al. (1991), Pike (1998)
each side of the boom, the hydrophones actually cover the entire circle 12 m in
diameter, potentially accessing a sub-seabed volume of 18,000 m3 (for a depth of
40 m).
3.4 Data Processing—Coherent Summation

and Flow Diagram
The main strengths of this approach to acoustic sub-seabed interrogation are: (1) the
very high number of points and geometries acquired each time; and (2) the coherent
summation of the waves measured for each combination of receiver/transmitter.
Back-projection consists in the collapse back in space and time of individual
measurements at each hydrophone to determine the location of the scatterer.
Coherent summation of the diffusively scattered wave fields also allows to increase
signal to noise ratios.
The three distinct datasets are used to perform seismic-type analyses (LF dataset)
and volumetric scattering analyses with synthetic aperture (HF and parametric
datasets). The focusing methodology relies on effective medium approximation (as
in pre-stack time migration) to image the discrete heterogeneities, where for each
Fig. 3.9 Depth contours and isopachs derived from depth cross-sections and from six horizons.
Reproduced from Guigné et al. (1991), Pike (1998)
voxel (small computational volume) and with appropriate weighting

(back-projection), the total scattered contribution is calculated. That is, if the voxel
non-trivially intersects the Fresnel volume (Cerveny and Soares 1992) of a scat-
terer, such as a boulder, the contribution would be high due to coherent summation.
On the other hand, if no scatterer (boulder) is present within the specified volume,
the total contribution would register values that are very low due to incoherent
summation. Moreover, because the size of the (synthetic) aperture is much larger
than the mean wavelength, the scattering 30 m deep or less occurs well within the
near-field of the source/receiver antennas, thus providing sufficient spatial resolu-
tion. The entire volume rendering/interrogating process then proceeds by
3.4 Data Processing—Coherent Summation and Flow Diagram 43
Fig. 3.10 Isopachs provide 3-D volumetric representations of the horizons picked up in each
“beam” (H1, corresponding to the seabed, to H7, nearly 10 m underground). Reproduced from
Guigné et al. (1991), Pike (1998)
Fig. 3.11 The locations of boreholes and CPT tests are shown superposed on bathymetric 0.1-m
contour lines for the 10-m diameter area of data acquisition (left). Fine-scale surface topography is
obtained by interpolating between beams (Fig. 3.3)
successively interrogating individual voxels. The answer product derived from

using such densely collected and processed data delivers a volumetric acoustic core
product as illustrated in Fig. 3.15.
Fig. 3.12 From left to right brute stack for beam 3, showing the different horizons picked from
the acoustic data alone; the corresponding normal incidence trace (replaced four times) for one of
the CPTs selected and their corresponding locations in the borehole (BH) samples (whose initial
interpretation log is given at extreme right). Dashed lines indicate the correlations between
horizons picked with each technique. Redrawn from Guigné et al. (1991), Pike (1998)
Fig. 3.13 Artist’s illustration of the PanGeo Subsea Acoustic Corer platform, deployed on the
seabed. The light-footprint tripod supports two antipodal booms (arms), rotating around the central
pivot, each with an independently moving instrumentation platform. The extent of the booms
allows imaging a 12-m diameter with multiple transmitters and receivers over a broad range of
frequencies. The tripod and booms are retractable when not in use, making for compact storage
(ca. 40 m3) and easier deployments at sea (with a 1500-m umbilical for deeper uses). Image
source: PanGeo Subsea Inc. marketing archive, 2011
3.4 Data Processing—Coherent Summation and Flow Diagram 45
Fig. 3.14 Close-up view of the moving instrumentation platform. It contains three sets of
transmitters: low-frequency (LF), high-frequency (HF) and parametric, and one hydrophone array.
This platform will move along the boom during operation, and as the boom rotates, cover the
entire 360° of a circle 12 m in diameter. Image source: PanGeo Subsea Inc. marketing archive,
2011
Figure 3.15 shows a striking, but typical, example of the Acoustic Corer data
after processing. The seabed surface has strong returns, and its topography can be
defined in great detail, allowing for example the detection of small depressions
(indicative of geological processes or previous industrial activity) or of small-scale
changes in terrain types (e.g. from sand to gravel or rock). Acoustic returns from
below the surface are clearly visible too. Only two “beams” are represented, at
cross-angles, and different horizons can be followed at several levels, and with
measurable variations in thickness. The acoustic returns also show distinct scat-
terers, which can be examined more closely to determine their exact nature. The
data processing is done in several stages, summed up in the flow diagram of
Fig. 3.16 and detailed in the following sections.
Fig. 3.15 Typical volumetric analyses using the PanGeo Subsea Acoustic Corer, in this case
corresponding to Statoil trials in the Norwegian Sea (Ormen Lange gas fields). The high returns
from the seabed do not mask the strong acoustic returns from individual horizons and from discrete
sources (potentially boulders or other obstacles) at several depths. The next steps of processing
will allow resolving these details. Image source: PanGeo Subsea Inc. marketing archive 2009.
Reproduced with permission
3.5 Processing Acoustic Corer Data
3.5.1 Data and Acquisition
The Acoustic Corer gathers acoustic data from the receivers (hydrophone array) and
from the three types of transmitters (LF chirp, HF chirp and parametric transducer).
This is completed with information from the platform itself, like the depth of
deployment, the tilt of the trip once fully deployed, the exact positions of the mobile
sonar platforms (“motor control can”, in Fig. 3.15). All signals are conditioned, in
particular to avoid aliasing, signal/noise ratios and general data quality are checked,
and the raw measurements are stored onto a dedicated server in GSF format
(Generic Sensor Format, a GNU open format available at https://www.ldeo.
columbia.edu/res/pi/MB-System/formatdoc/gsf_spec.pdf). The different data types
are then sent either for SAS (Synthetic Aperture Sonar) processing, in the case of
HF and parametric measurements, or to JYG-cross processing, in the case of LF
measurements (“processing” blocks, in green in Fig. 3.16). Both approaches are
explained in detail in the corresponding patents, respectively “Method for
3.5 Processing Acoustic Corer Data
Fig. 3.16 Generic flow diagram for the processing of acoustic measurements obtained by the PanGeo Subsea Acoustic Corer. See text for details of the
processing blocks in the legend
47
Accentuating Specular and Non-Specular Seismic Events from within Shallow

Subsurface Rock Formations” (US Patent #2013/8,391,103, Guigné et al. 2013)
and “Discrete volumetric sonar method and apparatus for sub-seabed surveying”
(US Patent #2016/9030914: Guigné et al. 2015).
3.5.2 JYG-Cross Processing
The 1991 Terrenceville measurements (Sect. 3.2) showed the importance of mea-
suring scatterers from a multiplicity of angles (multi-fold3). The Mills Cross con-
figuration traditionally used in underwater acoustics (e.g. for multibeam sonars) and
in astronomy is not adapted to the hardware setup (co-linear booms) nor to the type
of data acquisition (seismic-like). Instead, a new configuration was designed,
named JYG-Cross (JYG being the initials of the lead author) and presented in
Fig. 3.17.
Using the LF measurements, the JYG-Cross plan-view geometry simulates two
2-D lines approximately 60° apart and 30° from the nearest tripod leg (Fig. 3.18).
Each line consists of 52 “shot” locations, with 52 co-linear receiver locations for a
total of 2704 data traces per line, meaning that 5808 traces are collected and
analysed each time. The sweeps are generated at radial increments of 10 cm;
similarly, the receiver platform is radially shifted at 10 cm increments.
This geometry of data acquisition mirrors conventional seismic data acquisition
protocols. An improvement is that the ASI’s very precise stationary location con-
trol, typically unattainable in sub-bottom profiling, allows for consideration to be
given to the application of predictive deconvolution operators and potential
applications of FK, Radon, or SVD filtering on the data to suppress repetitive
multiple echoes that can appear in the datasets.
Quality Assurance (QA) of the raw measurements is very important. Common
Mid Points (CMP), i.e. the points halfway between source and receiver and shared
by numerous source-receiver pairs, are displayed together (CMP gather). Each shot-
and CMP-gather is carefully analysed for anomalously high amplitude traces, as
well as for anomalous spectral content distribution. For example, if a trace pos-
sesses anomalously high amplitudes but the reflectivity is correct, a simple scaling
solves the issue; otherwise the trace is removed from the dataset. If the problem is
identified to be spectrally motivated, bandpass, notch filters, or velocity filters are
tested to determine if the issue is resolvable. A decision is made whether filtering is
the optimal solution or whether migrating alone may least disturb the final section.
Data migration consists in moving the space-time measurements to their correct
locations. For each of the two lines, various migration routines are considered and
3
In seismology, the fold measures how many times a particular point is measured, and is a direct
function of the number of hydrophones and their spacing. Seismic surveys typically use folds of
60–240 (2-D seismics) and 100–240 (3-D seismics).
3.5 Processing Acoustic Corer Data 49
Fig. 3.17 Generic JYG-Cross configuration, showing the source as a solid dot and the receiver
elements as solid triangles. The configuration is patented (#US2013/8,391,103) and ideally
designed to make the best use of the geometry of the Acoustic Corer
Fig. 3.18 The JYG-Cross

simulates two lines
approximately 30° and 60°
from each tripod, minimising
potential acoustic returns
from their frames. Sweeps are
generated along the red line
(using the sources on the
acoustic package from one
boom) and receivers are
stationed along the green line
(by moving the acoustic
package from the other
boom). Choosing to offset
their respective positions by
10 cm yields close to 5800
traces for each JYG-Cross
position
employed. Pre-stack migration is generally employed upfront along the regularly

distributed Common Mid Points, where each trace is assigned to its closest mid-
point bin and the trace headers (metadata) are updated to contain the CMP spatial
location information. Using the JYG-Cross, the CMP bins along each line are 5 cm
apart and span a total length of 5.1 m. The number of traces assigned to each CMP
Fig. 3.19 Left for the JYG-Cross configuration, the line of CMP bins is shown in yellow, with the
acquisition lines in green and red. Right CMP trace binning considers the angles to sources (red
dots) and receivers (triangles)
bin increases from the end of the CMP line toward the centre and drops off at the
same rate toward the other end. Maximal fold occurs at the central bin, where it
consists of 51 traces. The remapping/labelling of traces that fall within each bin is
commonly referred to as a CMP binning (Fig. 3.19).
The next step in the processing is the definition of a migration velocity model,
and Fig. 3.16 (light green boxes, centre top) shows this is a recursive process. It is
also the most crucial, and often poorly understood when examining imagery.
Taking input velocities at face value, i.e. without correcting them or checking them
in different ways (e.g. forward problem vs. inverse problem), it is possible to
“create” targets or misunderstand the extent and thickness of horizons. The adaptive
determination of acoustic velocities is based on the spectral ratio procedure of
Guigné et al. (1989a, b). It makes use of the short and narrow-beam signals made
possible by parametric sonars. More explanations of velocity analyses are presented
in Yilmaz (2001) and the full mathematical justification of the steps presented here
is available in Guigné and Gogacz (2015).
The outcome of this process is a stacking velocity model that is later converted
to an interval velocity model for depth conversion. For each line, the velocity
analysis is completed independently. This is carried out with simultaneous analyses
of semblance plots, multiplicity of constant velocity gathers (constant velocity
normal moveout (NMO)-corrected gathers) at selected CMPs, as well as a multi-
plicity of constant velocity stacks (constant velocity NMO corrected and stacked
CMP gathers). Based on semblance and constant velocity panels, a decision is made
on the compilation of a two-way travel-time and CMP-dependent stacking velocity
model.
Semblance analysis is used to identify the main reflections in seismic data. It is
done by recording normal incidence paths, when the waves are perpendicular to
discontinuities and take the shortest time to return. The resulting velocity model is a
first approximation. The technique is known to fail when the source-receiver offset
Fig. 3.20 Semblance analysis plot, using velocities derived from normal-incidence returns and
aiming to show the main scatterers. Velocities (in m/s) are represented on the horizontal axis and
depths (in m) on the vertical axis
is greater than the depth of the reflecting layers, which is obviously not a problem
here (as the maximum depth of 40 m is still much larger than the largest
source-receiver offset attainable with the JYG-Cross). Figure 3.20 presents an
example of a semblance analysis plot.
Because the acquisition reference datum is generally parallel to the seafloor and
the observed stratigraphic boundaries are sub-parallel to the acquisition datum, no
dip-dependent correction is required. Therefore, in the CMP gathers, the specular
reflections follow mostly symmetric hyperbolic trajectories symmetric about the
time-axis (Fig. 3.21).
Once the stacking velocity model for each line is obtained, it is used to ‘flatten’
the hyperbolic reflection trajectories in each CMP gather. The flattening is refer-
enced to the same datum with respect to the zero-offset trace, as dictated by the
following normal move-out (NMO) equation, compensating for the distance
between source and receiver:
DsNMO ð xÞ ¼ sð xÞ % sð0Þ ð3:1Þ
where sðxÞ is the two-way-travel-time to a specified depth point on a reflector, with

source to receiver offset of x units apart, and obeying Snell’s Law of reflection at the
interface.
The equation for sðxÞ can be obtained using the geometric relationship:
x2
s2 ð xÞ ¼ s2 ð0Þ þ ð3:2Þ
vðsð0ÞÞ2
where vðsð0ÞÞ is the two-way travel-time, dependent on the stacking velocity

model. Using the Binomial Theorem, to first-order accuracy, the move-out cor-
rection DsNMO ð xÞ can then be expressed as:
x2
DsNMO ð xÞ ' ð3:3Þ
2sð0Þvðsð0ÞÞ2
In a horizontally stratified medium and where the offset is smaller than the
reflector depth, the stacking velocity in Eq. (3.3) is well approximated by the RMS
velocity model.
Marine seismics traditionally takes place in deep waters and over large distances.
Seismic measurements in shallow waters will however be affected multiple
reflections, from the sea surface, the seabed, and strong reflectors below the seabed
(Fig. 3.22, top). Field experience with the Acoustic Corer showed it was more
visible at depths less than 20 m. Advanced processing filters such as predictive
deconvolution, FK, Radon, or SVD are used to suppress the overprint of the
multiples onto the primary reflectivity (Alessandrini and Gasperini 1989).
The morphology of the sea surface will be governed by wave action, which
follows a typical Pierson-Moskowitz spectrum (e.g. Lurton 2010) in response to
external forcing from the wind. Depending on the spacing of the waves relative to
the acoustic wavelengths, the multiple reflections will come either from individual
facets, oriented at different angles, or from a sea surface approximated as flat (for
the wavelengths in use). Given this variability of the multiples, predictive decon-
volution is the most suitable technique for their suppression. This can be applied to
Fig. 3.21 Top constant-velocity CMP gathers. Bottom constant-velocity stacks. The horizontal
axis shows receiver (trace) numbers and the vertical axis shows depths (in m) for each example
Fig. 3.22 In shallow water, seismic measurements will be affected by multiple reflections from
the sea surface, the seabed and any strong reflector within the seabed. Top typical example with the
Acoustic Corer to scale. Bottom example of multiple reflections as visible in raw traces (using the
same labels as above)
CMP-gathers or to the stacked data. The reflection mechanisms particular to

Acoustic Corer typically generate triplet of multiples, which can be observed in the
processed data (Fig. 3.22, bottom).
Fig. 3.23 Examples of predictive deconvolution for suppression of multiples. From the bottom up
(a) reflectivity, (b) impulse response, (c) trace. Traces (d–h) are obtained by an application of
predictive deconvolution using an operator length n (in ms) and a prediction lag a (also in ms)
The two parameters essential to the predictive deconvolution filter are the pre-
diction lag and the operator length. Figure 3.23 shows a sensitivity study of the
predictive deconvolution filter for the suppression of multiples. The optimal filter
design is an iterative process, where the optimal filter parameters selection may take
a number of trials.
Attenuation as the acoustic waves travel down (and then back up to the receiver)
needs to be corrected. This is done by Automatic Gain Correction (AGC), a
trace-by-trace process to balance amplitudes across the entire time record. To
eliminate spurious energy which might have been introduced during the processing,
the profiles are bandpass filtered with a zero-phase Butterworth filter. Comparison
between a raw profile and its AGC version is presented in Fig. 3.24. The “en-
hancement” of the different structures, by modifying their acoustic amplitudes, is
not a problem in seismic analysis as it is not used in their interpretation (only
stratigraphy and velocity information are noted to be relevant).
The stratigraphic boundaries are in most cases sub-parallel to the line of data
acquisition, and the scale of any ASI is generally very small compared to the local
geology. It is therefore justified to use a 1-D model (depth-dependent only) of
velocities to depth convert the entire processed data.
Generally, a small set of representative CMPs is identified, from which an
average rms velocity model is compiled. This is then converted to an interval
velocity model using Dix’s equation, commonly used in seismics (Dix 1955):
t2 v2 % t1 v2
Vint ¼ pffi2ffiffiffiffiffiffiffiffiffiffiffiffiffi1ffiffi ð3:4Þ
ðt2 % t1 Þ
Fig. 3.24 Comparison of a raw profile (left) and Automatic Gain Correction (right)
where Vint is the interval velocity, t is the travel time to each reflector (1 or 2) and
v is the rms velocity to each reflector (i.e. from the surface to reflector 1, then from
reflector 1 to reflector 2). This equation is more valid for flat, parallel layers, which
is generally the case at the local scales considered (and over the entire volume
covered by an ASI).
The last step in the interval velocity model compilation relates to the conversion
from interval velocity, as a function of two-way travel-time, to average (whole
column) velocity as a function of depth. The two-way travel-time to depth con-
version is computed via integration,
ZtVint ðsÞ
d ðt Þ ¼ ds ð3:5Þ
0
2
where t is the two-way travel-time, d is the depth as a function of t; and Vint is the
interval velocity model as a function t. The time-to-depth mapping of the profiles is
converted from two-way travel-time to depth.
At this stage, the JYG-Cross data processing is deemed complete, and the final
profiles are handed over to the interpreters. The multifold dataset is used to show
coherent stratigraphic layers within the region of interest, and to provide a velocity
profile for the site.
3.5.3 Synthetic Aperture—Rendering and Processing
Synthetic Aperture is a technique commonly used in radar remote sensing, and

increasingly with sonars as improvements in subsea technology start to make it
more accessible. The principles of acoustic Synthetic Aperture are presented in
recent textbooks (e.g. Lurton 2010; Blondel 2009) and their applications to sonar
imaging are summarised in the excellent review by Hansen (2011).
In a nutshell, successive measurements of acoustic scattering by a surface (or
within a volume) are combined coherently along a known track to increase the
along-track resolution. The positions of the imaging platform need to be known
along-track at accuracies higher than a fraction of the wavelength (Hansen 2011)
and this is a problem for vehicles in the water column (e.g. Autonomous
Underwater Vehicles), as this requires very good navigation between successive
measurements. This is traditionally done with INS (Inertial Navigation Systems),
whose clocks are synchronised with the acoustic sensors, collecting information at
relatively high rates (10–25 Hz, typically). For vehicles moving independently in
the water, these positions need to be regularly integrated to avoid drifts with
absolute positions, taken with GPS (at the surface) or underwater transponders (see
Blondel 2009, for a review of the different positioning techniques for sonar vehi-
cles). This is obviously not a problem with the Acoustic Corer, whose position is
known at deployment, and where the positions of the transmitters and receivers are
known accurately along the “booms” (mauve boxes in Fig. 3.16). The other sig-
nificant challenges in Synthetic Aperture (summarised in Hansen 2011) are vehicle
stability (roll, pitch, depth, yaw during successive measurements), the contributions
of the underlying bathymetry (roughness and slopes facing away/toward the
transmitters), and multipath environments in shallow waters. Vehicle stability is not
an issue with the Acoustic Corer, as the sonar packages follow the same
pre-determined paths on each arm, and the overall orientation of the tripod is known
from depth and tilt sensors. For Synthetic Aperture Sonars (Hansen 2011), the
underlying bathymetry needs to be determined independently and at scales com-
mensurate to the acoustic wavelengths used. Although the Acoustic Corer uses a
large range of frequencies (HF sensor and parametric transceiver being used for
Synthetic Aperture), each vertical measurement along the JY-Cross, i.e. at intervals
of 0.1 m, actually provides direct knowledge of the small-scale topography of the
seabed (see Fig. 3.15). Multiple reflections from the sea surface and seabed are also
well constrained, using predictive deconvolution results from the LF measurements
(e.g. Figs. 3.22 and 3.23).
The Acoustic Corer is therefore ideally suited, as its stable and fixed geometry
directly addresses the major challenges to Synthetic Aperture imaging. Figure 3.25
shows the typical sequence of signal transmission and reception when moving the
sonar package along each boom. Each time a chirp signal is transmitted, it is
received by each hydrophone along the same line, as well as by the hydrophones on
the other line of the JYG-Cross (Figs. 3.18 and 3.19). The Acoustic Corer volume
imaging is based on coherent summation of backscattered wavefields. As
previously indicated, the focusing methodology is based on straight-ray travel-time

computation through an effective medium approximation, much the same as in
pre-stack Kirchhoff time migration. To capture the discrete heterogeneities, the
entire volume interrogating process proceeds by successively interrogating indi-
vidual cells (corresponding to the volume centred on position (x, y, z) in Fig. 3.25).
In Synthetic Aperture (Hansen 2011), higher resolution performance and better
coverage rate require increasingly large processing capability and storage space,
potentially prohibitive on subsea platforms with limited power and space. As the
processing scheme chosen with the Acoustic Corer is highly parallelisable, a net-
work of Graphics Processing Units (GPUs) is used to perform the coherent
summation.
Coherent summation starts with the filtering of data to within the operating
bandwidth of the chirp transducers, followed with calculation of instantaneous
signal amplitude and phase at the sampling frequency desired, and linear frequency
chirp pulse compression (to distinguish features within the range). Pulse com-
pression works by compressing the energy of the signal transmitted into a shorter
duration, drastically improving signal/noise ratios, and its principles are detailed in
acoustics textbooks (e.g. Lurton 2010; Ainslie 2010).
The acoustic transmissions involve a unique configuration of low-frequency and
high-frequency chirps, linearly frequency-modulated (LFM), lasting 22 ms each.
Figure 3.26 (top) shows the original pulse transmitted into the water, unshaded (i.e.
constant amplitude range) or shaded (i.e. with amplitude modulation, to decrease
sidelobe levels). This type of signal is highly recommended for sediment profiling
Fig. 3.25 Successive platform locations Pi, Pm enable focusing of backscattered fields by
coherent summation (see text for details). Following the convention of Figs. 3.18 and 3.19, the
sources are represented in red and the receivers in green. A potential scatterer (black dot) is located
within a volume centred on position (x, y, z)
Fig. 3.26 Top linearly frequency-modulated (LFM) chirps transmitted by the Acoustic Corer
(arbitrary vertical units). The shaded signal uses amplitude modulation, to increase signal/noise
ratios. Bottom comparison of the unshaded and shaded pulses after compression, showing the
drastic improvement in the Mainlobe to Sidelobe Ratio (MSR)
(e.g. Lurton 2010), as LFM chirps compensate for the strong attenuations below the
seabed whilst maintaining a good time resolution.
Strong acoustic attenuation within the seabed affects returns from individual,
finite scatterers like boulders or from extended targets like sedimentary horizons
(twice, as the waves travel down to the targets and back to the receivers). Acoustic
returns from the top of the seabed will be comparatively much stronger. Multiple
reflections within the water column, especially in shallow environments (Fig. 3.22),
will also contribute undesirable echoes. To overcome these differences, the chirp is
segmented into 10 chirplets 4.52 ms long, with overlapping frequency bands. The
reconstruction of the desired (approximately) 25-ms chirps is obtained via

“stitching” together the chirplets with a 50% overlap. The pulse compression
consists in either match or mismatch filtering. Matched pulse compression consists
of cross-correlation of the raw data with the stitched pulse. This is described in
more details in US Patent #2016/9030914 (Guigné et al. 2015), along with the chirp
signal generations protocols defining a discrete volumetric sonar method. The
mismatched pulse compression consists of filtering of raw data with a function,
which after application to the pulse yields lower temporal side-lobes. The reduction
of side-lobes is obtained at the expense of widening of the main-lobe. Hence, a
compromise is established between main-lobe width and the relative amplitude of
the side-lobes (see Fig. 3.26).
The respective apertures hT and hR of the transmitter and receiver processing
cones constrain the transmitter and receiver locations considered in beamformer
summation (Fig. 3.27). Based on experimental observations of the beam patterns
of individual transmitters and receivers, they are found to be optimal between 30°
and 45°, for both transmitters and receivers. Shading, if applied, can be “hard” or
Gaussian. Hard shading uses a hard-edged aperture window, which allows a trace to
contribute to the summation if it coincides with the transmitter and receiver cones
(Fig. 3.27). The Gaussian shading adds roll-off Gaussian scaling toward the
boundaries of the aperture(s).
Chirp transmitter locations can be selected to grid the entire survey area (i.e. the
cylindrical volume below the Acoustic Corer) into a half-wavelength lattice.
Synchronised clocks (<10 ls deviations) are used to co-locate the distinct sets of
receivers for each transmission.
Trace summation is used for analytic or incoherent beamformer summation. The
analytic beam-forming summation algorithm, unlike conventional migration, con-
sists of first mapping the recorded signals to their individual analytic extensions.
The real time signal x(t) is sampled at regular intervals t = k ( Dt (k 2 ℵ; Dt being
the sampling interval). Its complex analytical signal xanalytical(k) at each sample is
Fig. 3.27 The respective

apertures of the transmitter
processing cone hT and
receiver processing cone hR
can be adapted during signal
analyses and beamformer
summation. Experiments and
field investigations show
they are optimal between
30° and 45°
defined as the sum of the initial signal (real part) and its Hilbert transform (imag-
inary part):
xanalytical ðk Þ ¼ xðkÞ þ j H fxðkÞg ð3:6Þ
The instantaneous amplitude A(k) of the signal and its phase U(k) are given by:
"
Aðk Þ ¼ xanalytical ðkÞ
ð3:7Þ
/ðk Þ ¼ argðxanalytical ðkÞÞ
Since A(k) and U(k) vary slowly relative to x(k), simple linear interpolation is
enough to estimate the analytic signal at any time between k ( Dt and
(k + 1) ( Dt.
The weighted summation is then performed along the travel path in the complex
domain. The resulting voxel value embodies the total target strength as well the
envelope information of each summation contributor.
The incoherent summation option entails beamformer summation performed on
the envelope of each recorded trace. This method of summation is particularly
useful when the backscattered signal strength is high, in relation to background
noise, and when the velocity model is not well constrained, as can sometimes be
observed in shallow water, complex sub-seabed conditions.
Regions of space associated with a real sound scatterer (e.g. boulder) will have a
proportionally stronger signal (sum of coherent signals) than regions with no
meaningful acoustic scattering. At the end of this process, the magnitude mean
voxel sum for each section of the 3-D image volumes is displayed, for QA and to
provide a larger-scale acoustic image of the sub-bottom within the volumes
bounded by the transducer locations. Field examples of many such features, and
their interpretations, will be presented at length in Chap. 4, along with explanations
of how they were processed.
3.5.4 Combining Low-Frequency and High-Frequency

Datasets
These two stages of analysis of the low-frequency data (Sect. 3.5.1. JYG-Cross
processing) and the high-frequency and parametric data (Sect. 3.5.2. Synthetic
Aperture –Rendering and processing) are the key elements of the entire processing
of acoustic measurements with the Acoustic Corer (Fig. 3.16). Regular visualisa-
tion and statistical analyses at the different processing stages enable strong Quality
Assurance, and these different data types can now be combined for interpretation
and reporting (pink boxes in Fig. 3.16), leading to the final “Answer Product”, i.e. a
full 3-D visualisation of the seabed, layers and horizons, presence and size of any
obstacle, and confidence levels in the non-destructive “acoustic coring”, down to
depths of up to 40 m.
3.6 Conclusions
Successful acoustic interrogation of seabed and sub-seabed structures requires a

trade-off between the volume covered and the spatial resolution achievable.
Deployments far from the seabed, and using low frequencies, can image the
sub-surface structure at appreciable horizontal and vertical ranges. The low fre-
quencies are less attenuated within the seabed, but they yield poor spatial resolution
of any heterogeneities (either laterally, e.g. for the extension of sediment layers, or
vertically, e.g. indicating if they are thinning or disappearing). Deployments in the
water column also mean lower accuracies in the position of the imaging platform,
preventing any sophisticated processing like Synthetic Aperture. The first Acoustic
Seabed Interrogator developed by Guigné and colleagues in the early 1990s
(Sect. 3.1) addressed these challenges by being placed very close to the seabed, and
using a combination of low and high powers. This first generation of ASI enabled
the acquisition of large amounts of accurate data, already going beyond what was
achieved with conventional tools such as CPT and borehole (Sect. 3.2). Industrial
and academic collaborations over the next decades, coupled with advances in
subsea technology and computing power, yielded a new product, the PanGeo
Subsea Inc. Acoustic Corer (Sect. 3.3). Combining low and high frequencies,
using transmitters with different beam patterns (from large to very narrow) and
systematically recording many multi-angle geometries, the Acoustic Corer intro-
duced a step change in acoustic interrogation.
The flexibility of acoustic interrogation modes, with low and high frequencies, is
strengthened by the possibility of selecting different acquisition protocols.
Transmitters are usually activated sequentially, and acoustic returns are measured at
all receivers. In some cases (e.g. noisy data), the same transmitter can be activated
repeatedly, allowing for stacking of the individual traces on each receiver and
significant improvements in the signal/noise ratio. The patented JYG-Cross con-
figuration of transmitters and receivers, with derived velocity profiles, can be used
for low-frequency, pseudo-seismic mapping of layers and heterogeneities over large
volumes (typically 12 m in diameter ( 40 m in depth). The high-frequency chirp
transmitters and parametric sonar data are processed using Synthetic Aperture sonar
algorithms to render a 3D data volume or “acoustic core” representing the acoustic
reflectivity of the sub-bottom. This data volume can be viewed in 3-D as
cross-sections or plan-view slices. Anomalies appear as regions of relatively high or
low intensity, and are located and identified. Depth-converted stratigraphic
boundaries, thinning beds or discontinuous layers, cavities and fractured zones can
all be mapped over the entire volume, offering full 3-D imaging of the seabed and
sub-seabed. How these datasets are combined and used in practice will be presented
in the following Chapter.
References 63
References
Ainslie M (2010) Principles of sonar performance modeling. Springer-Praxis, Chichester

Alessandrini B, Gasperini M (1989) The deconvolution of a marine seismic source: an iterative
approach. Geophysics 780–784
Buogo S, Cannelli GB (2002) Implosion of an underwater spark-generated bubble and acoustic
energy evaluation using the Rayleigh model. J Acoust Soc Am 111:2594–2600
Cerveny V, Soares JEP (1992) Fresnel volume ray tracing. Geophysics 57:902–915
Dix CH (1955) Seismic velocities from surface measurements. Geophysics 20(1):68–86
Guigné JY, Gogacz A (2015) Mapping of sub-seabed anomalies in seismically heterogeneous
marine soils. Proc Inst Acoust 37(1):166–174
Guigné JY, Chin VH, Solomon SM (1989a) Acoustic attenuation measurements using parametric
arrays. Ultrasonics 27:229–301
Guigné JY, Pace NG, Chin VH (1989b) Dynamic extraction of sediment attenuation from
Guigné JY, Rukavina N, Hunt P, Ford JS (1991) An acoustic parametric array for measuring the
thickness and stratigraphy of contaminated sediments. J Great Lakes Res 17(1):120–131
Guigné JY, Welford KY, Gogacz A, Clements C (2013) Method for accentuating specular and
non-specular seismic events from within shallow subsurface rock formations. US Patent
2013/8,391,103, 5 Mar 2013
Guigné JY, Pace NG, Gogacz A, Dinn G (2015) Discrete volumetric sonar method and apparatus
for sub-seabed surveying. US Patent 2016/9030914, 12 May 2015
Hansen RH (2011) Introduction to synthetic aperture sonar. In Kolev N (ed) Sonar systems.
InTech, Rijeka, p 3–28. Available via: http://www.intechopen.com/books/sonarsystems/
introduction-to-synthetic-aperture-sonar. Accessed 18 Oct 2016
Inkpen ST, English J, Pike CJ, Guigné, JY, Guzzwell J, Hunt P, Tuff K (1991) A three-dimensional
acoustic site surveying probe. In: Proceedings of OTC 91, 23rd annual offshore technology
conference 1:525–535
Lurton X (2010) An introduction to underwater acoustics: principles and applications. Springer
Praxis, Chichester
Mueller C (2005) The marine VHR 2.5-D seismic brute stack cube as a feasible tool for the low
budget investigation and research. Mar Geophys Res 26:197–206
Pike CJ (1998). High resolution acoustic investigations of sub-seabed soils: relationship of wavelet
transformed acoustic image to soil properties and some geotechnical parameters. Ph.D. thesis,
University of Wales, College of Bangor
Pike CJ (1994) Analysis of high resolution marine seismic data using the wavelet transform. In:
Foufoula-Georgiou E, Kumar P (eds) Wavelets in geophysics. Academic Press, Boca Raton,
pp 183–211
Raytheon/GIL (2004) Technical demonstration of the DRUMS®-R200 sonar system, Technical
report RPT-0440-900-015-B, Raytheon Integrated Defense Systems (USA) and Guigné
Scheidhauer M, Marillier F, Dupuy D (2005) Development of a system for 3D high-resolution
seismic reflection profiling on lakes. Mar Geophys Res 26:183–195
Telford WA, Geldart LP, Sheriff RE (1990) Applied Geophysics, 2nd edn. Cambridge University
Press, Cambridge
Yilmaz O (2001) Seismic data analysis. Society of Exploration Geophysicists, Tulsa
Chapter 4
Acoustic Seabed Interrogation
Abstract Chapter 3 presented the main characteristics of the Acoustic Corer™ and
Chapter 4 details how to use the low and high-frequency datasets. Backed up with
many field examples from commercial surveys in a variety of water depths and
environments, the chapter shows how the volumetric datasets produced by the
Acoustic Corer™ reveal high-resolution details (decimetric scales), for areas 12 m
in diameter and down to depths of 40 m. The examples show removal of strong
out-of-range reflections, identifications of non-specular returns, identifications and
sizing of individual targets (e.g. boulders). Acoustic textures of horizons and
substrate variations are presented for complex seabeds and compared with borehole
data and Cone-Penetrometer Testing, clearly illustrating how geotechnically rele-
vant information can be extracted at much larger scale without affecting the seabed.
The combination of acoustic corer measurements with traditional tools is illustrated
in sites with complex geology, including gassy sediments. These strands are
weaved into customer reports, showing synthetically and simply the geotechnical
information necessary to offshore activities and several examples are provided.
! ! !
Keywords JYG-Cross Synthetic Aperture SAS Specular Non-specular ! !
! ! !
Velocity focusing Reverberation Boulders UXO UneXploded Ordnance ! !
! ! !
CPT Borehole Stratigraphy Target identification Target sizing Ground! !
! !
truth Acoustic textures Gassy sediments Geohazards!
4.1 Using the Low-Frequency (LF) JYG-Cross Data
The first illustration of the improvements offered by multi-frequency, multi-aspect

Acoustic Seabed Interrogation (ASI) is already visible with the low-frequency
measurements from the PanGeo Subsea Inc. Acoustic Corer. The JYG-Cross
(Sect. 3.5.2) collects high-resolution multifold data along two near-orthogonal
lines. Close to 6000 traces are collected each time, and these “shot gathers” are
processed to fold the data in a spatially migrated record of 5 m from the 12 m data
collection, which provide synoptic views of the seabed along each line (Fig. 4.1).

66 4 Acoustic Seabed Interrogation
Fig. 4.1 Typical display of lines acquired with the JYG-Cross configuration, showing sub-seabed
structures and layers down to 40 m. See Sect. 3.5.2 for details of data acquisition and processing
Using the bespoke signal processing stages outlined in Sect. 3.5.2, the
JYG-Cross measurements allow the accurate determination of velocities within the
seabed, resolving dips and strikes of the different layers, with high signal-to-noise
ratios. Another added value of acoustic coring is the possibility of operating far
from “sensitive” assets, or in complex environments in which traditional borehole
or CPT measurements are not possible.
For example, an offshore platform at a site in the Danish North Sea needed
additional piles for reinforcement. The site was 60 m deep, but sub-seabed risks
meant that borehole coring close to the structure was not possible. The ASI option
was preferred for technical reasons, but the acoustic interpretations were potentially
limited by the presence of a caisson close to the investigation site (Fig. 4.2),
inducing strong out-of-range reflections in the acoustic measurements. This would
have been a problem for any type of acoustically based investigation, especially at
lower frequencies.
The Acoustic Corer acquired simultaneous high-frequency and low-frequency
data, the latter with the JYG-Cross (Fig. 4.3). Caisson reflections are indeed
introducing strong reflections in all traces, at all depths. This reflective interference
is known as sideswipe in seismic processing, and it can also be associated with
out-of-plane structures such as anticlines or faults. It was removed with pre-stack
F-K filtering, presenting a much cleaner series of traces, with a high signal to noise
ratio. The JYG-Cross lends itself to advanced signal processing analysis typically
used in seismic investigations but not seen in sonar and shallow sub-seabed
imaging applications.
4.1 Using the Low-Frequency (LF) JYG-Cross Data 67
Fig. 4.2 Drawings of an offshore caisson in the Danish North Sea, with the footprint location of
the PanGeo Subsea Acoustic Corer. The acoustically reflective structure of the caisson is most
likely to induce out-of-range returns in the Acoustic Corer measurements. Source PanGeo Subsea
marketing material (reproduced with permission) and Guigné et al. (2010)
Fig. 4.3 Left: raw traces acquired by the JYG-Cross are contaminated by out-of-range returns
from the neighbouring caisson. Right: pre-stack F-K filtering removes the caisson reflection and
presents a clearer view of the sub-seabed geology. From Guigné et al. (2010)
Interpretation of these traces, based on depth variations, strengths of the acoustic

returns, and high-frequency measurements, revealed there was a gravel layer at 5 m
depth, and a sloping stiff clay layer 20 m below the seafloor (Fig. 4.4). This had of
Fig. 4.4 Data after JYG-Cross processing and geological interpretation (Guigné et al. 2010)
course implications for the installation of reinforcement piles. If they had been
feasible, several borehole tests would have been necessary to reveal this more
complex structure, and they would not have revealed its higher complexity.
4.2 High-Frequency (HF) Identification

of Non-specular Returns
Low-frequency JYG-Cross measurements are complemented with high-frequency

measurements using two types of transducers: narrow-beam parametric sonar and
high-frequency transmitters (Sect. 3.3, Fig. 3.14). A typical acquisition would
include approximately 20,000 discrete data traces. Synthetic Aperture processing
and rendering (Sect. 3.5.2) can then be used to accentuate non-specular targets or
anomalies in the volumetric dataset.
In the North Sea investigation presented in Sect. 4.1, depth slices revealed a
series of anomalies at distinct depths. These anomalies appear as strongly reflective
and distinctive targets (Fig. 4.5, left). Systematic analyses of different depths lead to
the identification of all these anomalies in the 3-D acoustic core, and the determi-
nation of their respective vertical extents. Plotting them together (Fig. 4.5, right) can
then be used for geotechnical decisions. In this case, the anomaly detected 4.25 m
below the seafloor, with a diameter of 0.55 m, was the most problematic for the
client as it was positioned where the support caisson was originally to be placed. The
other anomalies identified are all characterised by their burial depths (to the nearest
decimetre after processing), their volumes and possible identifications (see Sect. 4.3
4.2 High-Frequency (HF) Identification of Non-specular Returns 69
Fig. 4.5 Left: depth-slice of Synthetic-Aperture processed HF measurements, 12 m in diameter,

showing a strong acoustic anomaly 4.25 m below the seabed. Scattering strengths are coded from
blue (lowest) to red (highest). Right: synthesis plot of all sizeable anomalies identified in the 3-D
acoustic core, with their respective depths below the seabed (measured down to 20 m)
for more details of how this is achieved in practice). These quantitative details will
then inform the different types of interventions envisaged.
These anomalies manifest themselves as non-specular returns, and their extents
(and shapes) are important parameters in defining what they are. This is why the
accurate determination of velocities within the seabed is paramount. Generally, little
or no refinement of the JYG-Cross (LF) stacking velocity model is required to achieve
an optimally focused post-stack time-migrated profile. In the case where refinement is
necessary, the analysis is carried out by post-stack time-migrating the profile at a
number of incrementally varied velocity models and selecting the velocities, which
collapse the migration ‘smiles’ and/or ‘frowns’. How does it work in practice?
By nature of being discrete, the boulders in the migrated data appear as
high-amplitude, spatially localized events in an otherwise ‘transparent’ background.
Figure 4.6 shows scattering from a single boulder, evolving as a function of time.
Fig. 4.6 Constant time slices through volumetric data collected in the Grand Banks area, offshore
Eastern Canada. The waves scattering from a boulder are seen as expanding circular events, at
different times of data acquisition (every 4 ms: assuming a nominal sound velocity in sand of
1828 m/s, this corresponds to a two-way travel time value of 7.3 cm, i.e. a depth separation of
3.65 cm between the slices shown). From Guigné and Gogacz (2015)
With diffraction focusing (mathematically justified in Guigné and Gogacz 2015),

the background sound speed is obtained from the recorded data by adjusting the
background sound speed (e.g. through inversion, or simple trial-and-error), so as to
maximize the amplitude (in the migrated data) of the event against the neighbouring
background.
For small boulders (with sizes smaller or equal to the imaging wavelength), their
signatures in the migrated data are expected to be high-intensity and localised in
space (area and also depth). However, if the boulder is sufficiently large, then the
point-scatterer assumption no longer holds and imaging proceeds in a manner
similar to imaging of reflectors where resolution of the boundary becomes focal.
Accuracy in determining the sound speed in the surrounding layers is essential, as
shown in Fig. 4.7.
With optimal velocity focusing, boulders and other anomalies reveal themselves
with focused maximal intensity. Detection of boulders (rock fragments >0.3 m) in
glacial and post-glacial soils requires utmost care in data processing as unconsol-
idated soils prove highly absorptive in the >4-kHz bandwidth regime, leading to
diminished penetration and ‘target’ resolution. Buried boulders must be treated as a
3-D problem (conventional sub-bottom profiling methods are too coarse). Absence
of laterally extensive seismic horizons requires diffraction/scattering focusing
methods to be employed for acoustic speed estimation, subsequently used in
imaging/detection.
4.3 Sizing and Identifying Individual Targets
The discontinuities identified in ASI measurements can be large, at least relative to

the diameters imaged, and they are then likely to correspond to discontinuous layers
of material with geotechnical properties distinct from the background. This is
important, as it can affect any future activity (e.g. drilling) and it is important to
identify what these discontinuities can be (e.g. gravel bed or rocky substrate).
Smaller targets need to be accurately located, sized and if possible identified
(cobbles? left-over equipment from previous activity?) (in some applications of the
Acoustic Corer, small targets have even been identified as unexploded ordnance, or
UXO). The choices of frequencies and geometries of acquisition made when
designing the Acoustic Corer are particularly useful for this purpose. At high
frequencies, the SAS acquisition and processing possess coherency and signal
strengths to detect boulders in highly reverberant gravelly sediments, something a
conventional sub-bottom profiler could not do or achieve. Boulders tend to scatter
diffusely into non-specular directions, and the multi-aspect views available with the
ASI method make the returns more coherent and part of the “true” seismic response,
raising potentially very weak signals above the background noise.
Based on experience, a simple way to size these targets is to look at time slices at
different depths and plot the acoustic intensity along their longest axis, broken
down into X and Y components (Fig. 4.8). The ranges at which the intensities fall
4.3 Sizing and Identifying Individual Targets 71
JFig. 4.7 Illustration of the effects of acoustic velocity (derived using the JYG-Cross measure-
ments) on the detection of non-specular returns. Plots on the left show depth slices, migrated and
using diffraction focusing with different sound speeds. Middle plots show time slices through the
volumetric data, centred on the anomaly detected. Plots on the right show vertical slices through
the migrated data. Top row: the sub-seabed velocity was correctly surmised, and the boulder is
crisply defined in all plots. Middle row: the acoustic velocity is artificially taken as too low,
resulting in blurring and decrease of intensities for the entire object. Bottom row: the velocity is
increased intentionally, showing the “breaking up” of the boulder in different sub-targets. Adapted
from Guigné and Gogacz (2015)
Fig. 4.8 Time slices of acoustic scattering at a specific depth allow identification of anomalies
(left). Their dimensions are found (right) by measuring distances at 3-dB from the maximum
intensity, giving basic information about their size at this particular depth (basic sizing of targets as
practiced with the Acoustic Corer measurements)
by more than 3 dB are taken as the respective lengths Dx and Dy, or rounded up to
avoid underestimation. With the decimetric accuracy of the measurements, and the
availability of time slices at different depths, this provides a full 3-D description of
the size and shape of each target.
Targets like boulders can be isolated or come in clusters. Ground truth has often
been used to validate the measurements made with the ASI method, showing good
agreement between acoustic data and actual sizes and shapes. Figure 4.9 (top row)
shows a close-up view of a time-slice (3 " 3 m) obtained by SAS rendering. Strong
targets (in red) are visible in the map, and the middle cluster scatters so much
energy that it saturates the measurements (elongated red shape with a white centre).
Time of arrivals, once processed, indicated a depth of 2.2 m. Ground sampling at
this depth recovered a cluster of boulders, in a background of till with sandy gravel.
4.3 Sizing and Identifying Individual Targets 73
Fig. 4.9 Comparison of boulders, isolated or in clusters, detected using the Acoustic Corer (left)
and visually (right). From PanGeo Subsea’s data archive—Kriegers Flak, Baltic Sea, 2010
Figure 4.9 (bottom row) shows several large anomalies, in a larger time-slice: these
boulders are visible on the seafloor, with the expected dimensions.
The 3-D extent of these anomalies can be followed vertically within the volumetric
data, giving more information about their geological background (and their
geotechnical implications). Figure 4.10 shows how a distinct anomaly (close to 10 m
below the seabed) can be identified. In this case, it is located in a marine late glacial clay
unit, with strong but small reflectors much deeper (roughly 10–20 m deep).
Detailed cross-sections within the volumetric data can provide a lot of useful
information. Figure 4.11 (top) shows for example a 3-m thick gravel/cobble layer
buried at a depth of 11.0 m and sandwiched between clay soil and sandy soil, both
of which are relatively acoustically transparent. The gravel is acoustically rever-
berant (sub-wavelength mean free path) and therefore it has a pixelated texture, as
shown in this cross-section. Typically, gravels appear as patches, layers, and dis-
continuous rough patterned slices. Gravel/cobble layers appear as regions of high
background acoustic intensity with localized areas of extremely high acoustic
intensity, which indicate cobble-sized particles, higher concentrations of gravel, or
small boulders, as illustrated in the sequence of horizontal sliced images in
Fig. 4.11 (bottom).
Fig. 4.10 Left: vertical section (10 m wide) with a distinct anomaly suggestive of a boulder.
Right: horizontal (time) slice corresponding to the depth (dbsf: depth below sea floor). This
particular boulder is 0.5 m in diameter (i.e. large enough to be of concern), with smaller boulders
around it. From PanGeo Subsea’s data archive—Kriegers Flak, Baltic Sea, 2010
Comparison between the co-located HF (chirp) dataset and the parametric pro-
filing can also provide additional measurements (Fig. 4.12). This example corre-
sponds to an area of the Baltic Sea with a complex geology (recurrence of boulder
clusters and layers). The same structures can be seen in the HF and parametric data.
They are interpreted as a mound of boulders and cobbles typical of a glacial
moraine deposit (the individual boulders merge to form the curved band in
Fig. 4.12, but individual horizontal slices at these depths reveal their sizes, loca-
tions and grading with depth and distance from the top of the mound). This is
further supported by CPT sampling, showing clear boundaries at the same depths.
4.4 Acoustic Textures and Substrate Variations
The coherence between echoes is maintained as the ASI platform is stationary.

Dunsiger et al. (1979) emphasized the importance of coherence in high-resolution
mapping. Small misalignments (in the order of 10–20 ms, assuming a velocity in
the sediment of 1600 m/s) between echoes destroy the fabric of the signal response.
The ASI’s coherent signals and density will provide the necessary precision and
data to statistically evaluate (in three dimensions) the homogeneity of sediment
properties and their distribution. There are many ways of quantifying textures in
acoustic images, using for examples second-order statistics or fractal measurements
(e.g. Blondel 1996; Montereale Gavazzi et al. 2016, and references therein).
Horizontal (time-migrated) slices obtained at different depths with the Acoustic
4.4 Acoustic Textures and Substrate Variations 75
Fig. 4.11 Top: vertical cross-section, 3 m thick and 12 m long, of a gravel/cobble layer 11.0 m
deep, sandwiched between clay soil and sandy soil. Bottom: successive horizontal slides going
down through the acoustic core (12 " 12 m) show cobble-sized particles, then higher
concentrations of gravel (12.00 m deep) and small boulders (e.g. 12.50 m deep). Dashed circles
indicate the 12-m diameter circle (dbsf: depth below sea floor)
Fig. 4.12 Example of complex geology, in this case a mound of boulders and cobbles (curved
hashed lines) typical of a glacial moraine deposit, as seen through the HF chirp and parametric
datasets. The two vertical slices are approximately 10 m wide and 15 m deep. The corresponding
CPT data shows clear spikes corresponding to these boundaries. Two different measurements had
been made, with one stopped by the densest layer of boulders
Corer provide direct information about the different types of sub-seabed soils. They
can be interpreted in conjunction with borehole or CPT data, if available.
The most easily recognized marine deposits are those containing gravels and
cobbles, since they produce a highly reverberant acoustic return. For example, a
cobble layer close to the surface of the seafloor acts as an extremely strong reflector.
Depending on the acoustic impedance contrast with the surrounding soil matrix,
boulders can act as extremely strong reflectors. The resulting size, shape, and
acoustic intensities of these types of anomalies are used to determine whether the
boulder is isolated or belongs to a cluster of boulders. Figure 4.13 shows examples
of boulders in three different sub-seabed soil types, namely a clay matrix (with low
acoustic returns and a smooth, homogeneous texture), a sand matrix (medium to
high acoustic returns from small areas, yielding a mottled texture) and a sandy
gravel matrix (showing a more heterogeneous texture, with lower acoustic returns
for the sand matrix and higher returns for the stronger densities of gravel). The
generic methodology presented in Sect. 4.3 permits the individual cobbles to be
isolated for further conformation.
In the course of a demonstration survey in 2009, the Acoustic Corer was used in
waters approximately 410 m deep, at the Statoil Omen Lange site in the Norwegian
Fig. 4.13 Examples of three HF chirp slices of boulders (identified with arrows and, when
necessary, circles) as detected in a clay sediment (left), a sand matrix (middle) and a sandy gravel
layer (right). Acoustic returns are coded from blue (lower) to red (higher). From PanGeo Subsea’s
data archive—Norwegian Fjord, 2011
Sea. Figure 4.14 shows a vertical slice (left), with horizontal slices (right) providing
accurate information about specific features. An erosion boundary at the base of the
till layer is detected as an irregular surface between 13 and 11 m deep (dashed red
line). It separates the till from an underlying clay soil, with a distinct acoustic
texture (bottom left horizontal slice). The individual plan views show reverberant
acoustic returns and distinct structures, indicating that the till contains considerable
amounts of cobbles and gravel. The erosion boundary is geotechnically significant,
given it is associated with the highest magnitudes of unit weight and undrained
shear strengths as recorded by cone penetrometer tests. The high shear strengths are
attributed to the increased coarse grain content. After acquiring and rendering the
SAS data into an acoustic core, it was noted that the erosion boundary could very
weakly and ambiguously be observed in previously acquired sub-bottom profiler
records. However, the subtle undulations in the boundary, as observed in the ver-
tical slice extracted out of the Acoustic Corer data, had been lost in the sub-bottom
profile owing to its scale, lack of spatial resolution and general smearing of
stratigraphic detail, inherent to profiles acquired with towed platforms.
These measurements can be interpreted as geological horizons and geotechni-
cally relevant features (Fig. 4.15), and comparison with borehole measurements
shows the higher volumetric coverage of the Acoustic Corer and its detection of
boulders and obstacles fortuitously just outside the borehole but still geotechnically
important.
Deployments of the Acoustic Corer have also sometimes revealed unexpected
features. Measurements at a site where geotechnical investigations had already taken
place showed for example the presence of debris associated to a failed CPT test
(Fig. 4.16). Horizontal slices taken very close to the seabed show significant changes
in just 15 cm. Figure 4.16 (left) shows strong acoustic returns corresponding to a
Fig. 4.14 Example of an Acoustic Corer slice after SAS rendering (left) and vertical slices at
different depths (right). Acoustic intensities are relative to a site-specific reference, therefore
showing acoustic contrast and providing better evidence of variations in sub-seabed soil types (see
text for details). Data taken at the Statoil Ormen Lange site in the Norwegian Sea, 2009. Presented
with permission from StatoilHydro to PanGeo Subsea
small topographic change (bottom right) and irregular structures (top left, within a
red circle). Based on their vertical and horizontal extents, they can be confidently
interpreted as debris (their downward continuation is visible in Fig. 4.16, right).
Acoustic returns from 0.35 m deep (right) also show a long, thin and highly
reflective structure (top right, within a red ellipse). Based on its location and its
dimensions, it is interpreted as a CPT rod lying close to the seabed, and left over
from the previous investigation. Other investigations (e.g. Raytheon/GIL 2004 and
unpublished reports) have also shown this approach could be used to successfully
detect and identify other types of buried targets (e.g. Unexploded Ordnance, UXO).
Fig. 4.15 Interpretation of the Acoustic Corer slice of Fig. 4.14 shows excellent comparison with
the geology derived from a borehole measurement. The intended position of a pile (dashed blue
lines in the interpretation) show the added benefits of the Acoustic Corer, with its larger coverage
and finer definition of geotechnically relevant structures (e.g. boulders, identified as red squares,
thickening boundaries at the 11–13 m horizon). Data taken at the Statoil Ormen Lange site in the
Norwegian Sea, 2009. Presented with permission from StatoilHydro to PanGeo Subsea
Other evidence of past activity is presented in Fig. 4.17. In this case, the first
horizontal slices (left) show generally strong returns, 30 dB higher than a circular
structure close to the centre of the Acoustic Corer’s field of view. This feature is
2 m in diameter, visible to 0.3–0.5 m deep with acoustic contrasts increasing to the
background level. The vertical slice (right) confirms its interpretation as a seafloor
depression. In this particular case, it was caused by a cone penetrometer retracted
because of fears of a punch-through to possible sub-surface gas. The Acoustic Corer
volumetric data showed this CPT measurement could have gone through without
risk, as the sub-seabed is relatively homogeneous and free of any geo-hazard. This
is confirmed with two vertical HF SAS profiles (Fig. 4.18, middle and right),
Fig. 4.16 Horizontal HF slices obtained with the Acoustic Corer at different depths very close to
the seabed (dbsf: depth below seafloor). Left: debris (within the red circle), associated with
previous geotechnical investigation. Right: long and thin structure (within the red ellipse)
corresponding to an abandoned CPT rod. The dashed white lines indicate the extent of the full
volumetric dataset, extending 40 m down. From PanGeoSubsea’s data archive—Baltic Sea
Fig. 4.17 Left: HF SAS image of the seabed, 12 m in diameter, with a pronounced indentation near
the centre (see text for details). Right: vertical slice showing it corresponds to a seafloor depression,
approximately 2 m wide and 1 m deep. From PanGeoSubsea’s data archive—Baltic Sea
compared to the aborted CPT measurements (Fig. 4.18, left). The ASI answer
product not only verified the validity of CPT soil profile results, but added value by
showing the lateral extent, depth variability, and complexity of the soil horizons.
Overall, correlation between the acoustic core data collected and the (limited) CPT
results were excellent.
4.5 Combining with Other Measurements 81
Fig. 4.18 The aborted CPT measurements (left) at the site of Fig. 4.17 can be compared to two
vertical HF SAS profiles (middle and right). The Acoustic Corer data verify that no geo-technical
hazard is present (although potential punch-through was originally envisaged), and they extended
the CPT measurements (ca. 5 m deep) to more than 40 m deep and 12 m across. From
PanGeoSubsea’s data archive—Baltic Sea
4.5 Combining with Other Measurements
4.5.1 Traditional Tools
The Acoustic Corer benefits from a large knowledge base acquired through multiple
deployments, including the notoriously challenging geological environments (e.g.
Lunne 2012) of the Grand Banks, Canada and the North Sea. The examples pre-
sented so far have mostly used CPT and borehole measurements either as validation
of the first surveys (Chap. 3) or as confirmation of volumetric datasets (Sect. 4.1).
Throughout, ASI measurements showed they provided high-accuracy measure-
ments over larger areas, sometimes extending deeper and sometimes revealing
unsuspected geo-hazards (e.g. boulders) or verifying that no hazard was present
even at larger ranges. But this does not mean that ASI precludes or replaces tra-
ditional CPT and borehole measurements, far from it.
Borehole data has long shown its importance in geotechnical sub-surface mea-
surements. Drilling parameters like drill bit load, torque, rate of penetration and fluid
pressure are first indicators of the geomechanical properties of the sub-seabed.
Recovery of soil samples enables accurate identification of the geology (e.g.
sand/gravel contents) and parameters most important to later activities (e.g.
mechanical resistance, fluid or gas content). CPT, including CPT-while-drilling
(CPTwd: Sachetto et al. 2004 in Lunne 2012), can access other parameters, e.g. cone
resistance, sleeve friction, pore pressure (which is also used to correct cone resistance
measurements) and in some cases lateral stress (although its pertinence is not always
accepted: see discussion in Lunne 2012), electrical resistivity (related to density), heat
flows and S- and P-wave velocities (see the open-access book by Brouwer 2007, for
more detailed information). Measurements at depth are generally hampered by the
prevailing hydrostatic pressure. Two thousand meters deep, cone resistance will be
close to 15 MPa and will need measuring with very small relative accuracies of
±35 kPa (e.g. Lunne 2012), severely constraining the sensor capabilities.
Borehole and CPTs have been in use for the last 50 years or so. CPT is now
widely accepted. Sensor calibration, data acquisition and processing are mostly
standardised now (Lunne 2012), making them workhorses of any site investigation.
However, as shown several times in this book, they are also spatially limited, and
the challenges and costs associated to uses in deeper seas mean they are increas-
ingly constrained. What the ASI’s volumetric imagery provides the CPT mea-
surements is a confirmation on the conditions that the probe encountered in
association with potential discontinuities and details on texture of the surrounding
soils. The CPTs are linear one dimensional measurements being made in three
dimensional real world in situ realities.
4.5.2 Investigations of Sites with Complex Geology
The following two sections will show examples taken from the Anholt Offshore
Wind Farm, in a collaboration between PanGeo Subsea and Dong Energy (www.
pangeosubsea.com/files/Anholt_Poster.pdf). Located in the Kattegat Strait offshore
Denmark, the Anholt site was a project undertaken by Dong Energy with 111 wind
turbines, installed by driving 6-m diameter steel monopole foundations into a
seabed to depths of 20 to 30 m. The total site investigated was approximately
20 km long and 5 km wide, in water depths of 15–19 m. Detailed geotechnical
investigation by experts, using CPT and boreholes, concluded that “the geological
architecture of the study area is very complex” (Leth and Novak 2010). Overall, the
complexity of the site is such that it proved impossible to reliably interpolate
between measurements and predict anomaly concentrations at the planned foun-
dation sites, even with the accompanying sub-bottom profiling. The Acoustic Corer
was used to survey 25 proposed installation sites to determine the stratigraphy of
the sub-seabed and to identify any geohazards that could pose a threat to foundation
installation, such as boulders, gassy soils and gravelly deposits.
Figure 4.19 shows two typical acoustic cores obtained on site. Volumetric ASI
information is combined with CPT measurements of the cone tip resistance (giving
information about the type of soil) to provide more complex interpretations.
Figure 4.19 (top) shows for example the presence of boulders, or clusters of
boulders, at shallow depths, and multiple reflections 14 m below the seabed.
Figure 4.19 (bottom) shows a close-up view of the first 10 m of seabed, with a
discontinuous layer approximately 6 m down, visible as a slight increase in CPT
cone resistance but averaged out during CPT interpretation. Overall, however, both
sites display good agreement between the very localised CPT data and the
larger-coverage acoustic cores, which reveal additional information relevant to
future offshore activities. Higher cone resistance generally corresponds to strong
reflections in the acoustic profiles, with a good correlation in simple uniform or
homogeneous geostrata.
Poorer correlations between acoustic cores and CPT measurements are visible in
more complex sites. Discrepancies primarily reflect the lack of spatial coverage by
the CPTs and, because of the non-linear soil interactions or influences commonly
present in mixed complex seabeds, that limits the reliability of geotechnical
investigations in such sites. Figure 4.20 shows a typical discrepancy (event “A”,
identified in the CPT interpretation around 9 m below the seabed). The acoustic
core shows the interface to be 1.25 m shallower than in the CPT data. This is due to
the placement of the CPT; because of its small footprint, it sampled a small atypical
area of stratigraphy and did not accurately represent the local trend.
This small footprint of CPT and borehole measurements constrains how much
information is available laterally, and in most cases it is not feasible, technically or
economically, to sample on a denser grid by repeating CPT deployments close to
the same positions. This is an important issue if the geology varies laterally, as
shown in Fig. 4.21. This particular vertical slice sports 6 different discrepancies
between Acoustic Corer data and CPT measurements, at regular depths throughout.
The discrepancies E and F are the most notable (around 10 m below the seabed)
and the important changes in acoustic reflectivities and depths were not picked up
by the CPT interpretation of the terrain as glacial lower till (Gl Gc).
4.5.3 Gassy Sediments
Offshore installations require accurate detection and mapping of “gassy sediments”,

in particular for wind farm sites (e.g. Orange et al. 2005; Duck and Herbert 2006)
where very large (and expensive) wind turbines will be installed on monopole
foundations extending deep into the seabed. But the detection of gas with con-
ventional acoustic profiling is extremely difficult (Fig. 4.22). Gas in the sediment
acts as a strong reflective mirror due to impedance mismatches between the gas and
Fig. 4.19 Vertical profile slices taken at two separate sites and compared to geotechnical
information. Both sites display good agreement between the CPT data and the acoustic cores, with
the acoustic data displaying more events than are visible in the CPT data. The stratigraphic units
are identified with standard abbreviations. From PanGeo Subsea Inc. RPT-03131-1 AC Dong
Anholt Final Report Dec 2011, reproduced with permission
Fig. 4.20 Vertical profile slice showing a discrepancy between acoustic core and CPT data. Event
“A”, identified in the CPT interpretation around 9 m below the seabed, is 1.25 m shallower than
the corresponding increase in the CPT cone tip resistance. This is interpreted as coming from the
very localised emplacement of the CPT, which “missed” this discrepancy (from PanGeo Subsea
Inc. RPT-03131-1 AC Dong Anholt Final Report Dec 2011, reproduced with permission)
the surrounding sediments. Its presence thus creates a masking of the general
surrounding sediment character by imposing an acoustic blanking footprint.
This masking can be negated, to some extent, by imaging from different angles,
and using coherent sediment returns. The multi-aspect geometry of the SAS dataset
can better identify the presence of anomalous (e.g. “gassy”) returns in the horizontal
slices corresponding to specific depths. Depending on the quantity of gas, stratig-
raphy and anomalies may still be detected below gas layers owing to this
multi-aspect set of views. Figure 4.23 compares SAS-rendered HF chirp and
Fig. 4.21 Comparison of a vertical profile obtained with the Acoustic Corer, showing acoustic
variations in the sub-seabed down to 15 m and across 12 m laterally, with CPT measurements and
their interpretation. Note the high number of discrepancies that only the Acoustic Corer can reveal
(from PanGeo Subsea Inc. RPT-03131-1 AC Dong Anholt Final Report Dec 2011, reproduced
with permission)
parametric sonar data from a gas layer 6.20 m below the seabed. Both datasets
detect the spatial distribution and the thickness of this particular gas layer, with very
similar accuracies (but more details overall for SAS HF Chirp data). It is important
to use dissimilar acoustic transmission types like SAS HF chirp and parametric in a
co-located manner as their differentiating acoustics provide for a means to verify the
reliability of what we see in the imagery as being real.
4.6 Interpretations—Reporting
Site investigations are usually made of individual measurements using borehole and
CPT, complemented with a sizeable quantity of Acoustic Corer volumetric datasets.
In the case of the Anholt wind farm study, 25 acoustic cores were obtained,
showing stratigraphic variability and details of the “very complex geological
architecture” revealed by previous studies (Leth and Novak 2010). A large number
of individual hazards like buried boulders and cobble clusters were identified, all at
different locations, different depths and with different sizes. For each site, they are
4.6 Interpretations—Reporting 87
Fig. 4.22 Typical conventional sub-bottom profile with associated CPT response (green and
black vertical profiles). Notice the smeared masking character in the record caused by the presence
of gas in the sediment. Acoustic returns are coded in grey and coloured sub-horizontal lines
delineate specific horizons (from PanGeo Subsea Inc. RPT-03131-1 AC Dong Anholt Final Report
Dec 2011, reproduced with permission)
presented on synthetic plots like Fig. 4.5 (right), with object types and depths below
the seabed. In some places, lateral discontinuities and gas layers were also identified
and mapped in full 3-D. The measurements, their interpretations and their statistics
all need reporting in a synthetic and accessible way, and the typical flowchart is
illustrated in Fig. 4.24. Anomalies are identified on vertical and horizontal slices of
SAS-rendered HF chirp and parametric data, for complete volumes in the seabed
(25 of them, in the Anholt case). 58% of the anomalies were located in the Marine
Late Glacial Clay, which overlies the Melt Water Glacial Units, whilst 96% of the
anomalies were located in the uppermost 15 m of the soil column. This has clear
implications for offshore installations and any other activities.
At each site where an acoustic core was measured, end-users are provided with a
3-D identification of all anomalies specifying the location, burial depth, and
enclosing circle diameter of each. The geological interpretation yields an accurate,
3-D soil profile, which is supplemented with the number of boulders or other
obstacles in each layer, and the statistical distribution of their sizes (Fig. 4.25).
Fig. 4.23 Concurrent SAS-rendered HF chirp measurements (left) and parametric sonar data
(right) highlight the spread of the gas layer. The high spatial resolution allows mapping the area of
gas at different depths, hence the total volume, and assessing the surrounding geology (in this case,
a marine late glacial clay unit), and other potential geo-hazards (from PanGeo Subsea Inc.
RPT-03131-1 AC Dong Anholt Final Report Dec 2011, reproduced with permission)
4.6 Interpretations—Reporting 89
Fig. 4.24 Illustration of the interpretation and reporting flowchart for Acoustic Corer surveys,
using some of the data from the Anholt offshore wind farm investigation (from PanGeo Subsea
Inc. RPT-03131-1 AC Dong Anholt Final Report Dec 2011, reproduced with permission)
These individual interpretations can then be synthesised to establish the geo-

logical trend. The number of boulders identified is divided by the total volume
scanned for each geologic unit to determine the average boulder density. If par-
ticular geological units are found to have significantly more boulders per cubic
metre (e.g. Fig. 4.26), it is likely that other locations with the same geology (known
for example from borehole or CPT data, or from lower-resolution profiling) will
also have high numbers of boulders. It is therefore possible to extrapolate and
estimate the risk of geohazards present at unsurveyed installation sites, adding more
information to the general picture of the area.
Fig. 4.25 ASI investigations can be used to show the size distributions of anomalies, from very
small to very large (top) and how they are distributed with depth (bottom)
4.7 Conclusions 91
Fig. 4.26 The sub-seabed volumes surveyed with ASI can be used to identify the types of terrain
most likely to have boulders or other hazards. This information can then be extrapolated to make
informed assessments of how many other obstacles are likely to be found in other areas, where ASI
was not performed but where the geology is relatively well known. In this case, the Marine Late
Glacial Sand layers are likely to be the most affected
4.7 Conclusions
This chapter was much longer, as it presented a full ASI answer-product and
illustrated the different types of measurements with real field data in a large variety
of situations. The Acoustic Corer was presented with its different elements:
low-frequency JYG-Cross (Sect. 4.1), high-frequency identification of specular and
non-specular returns (Sect. 4.2), velocity focusing to capture the right form and
nature of the target, detection and sizing of anomalies (Sect. 4.3), acoustic textures
and substrate variations (Sect. 4.4), combination with other measurements like
borehole or CPT (Sect. 4.5) and presentation of the full analyses (Sect. 4.6). The
Acoustic Corer is now used all over the world, in a large variety of settings.
As subsea technology improves, and as signal processing techniques develop,
what can be the next developments? Can the imaging geometries be refined or
adapted to other challenges? Can the ASI platform be adapted to new environ-
ments? Can this acoustic sensing approach be translated into other domains of
remote sensing, e.g. radar or medical imaging? Current and future developments of
the ASI technique will be expounded in Chap. 5.
References
Blondel Ph (1996) Segmentation of the Mid-Atlantic Ridge south of the Azores, based on acoustic
classification of TOBI data. In: MacLeod CJ, Tyler P, Walker CL (eds) Tectonic, magmatic
and biological segmentation of mid-ocean ridges. Geological Society Special Publication, 118,
pp 17–28
Brouwer H (2007) In-situ soil testing. Available via http://www.conepenetration.com/online-book.
Duck RW , Herbert RA (2006) High-resolution shallow seismic identification of gas escape
features in the sediments of Loch Tay. Scotland: Tectonic and microbiological associations.
Sedimentology 53: 481–493
Dunsiger AD, Cochranen A, Vetter WJ (1979) Seabed characterization from broadband acoustic
echosounding with scattering models. IEEE J Oceanic Eng 4(3):94–106
Guigné JY, Gogacz A (2015) Mapping of sub-seabed anomalies in seismically heterogeneous
marine soils. Proc Inst Acoust 37(1):166–174
Guigné JY, Welford JK, McDermott IR (2010) Volumetric, multi-fold acoustic interrogations of
complex sub-seabeds. In: Proceedings of the 16th EAGE European Meeting of Environmental
and Engineering Geophysics on Near Surface 2010, Zurich. doi:10.3997/2214-4609.20144842
Leth JO, Novak B (2010) Late Quaternary geology of a potential wind-farm area in the Kattegat,
southern Scandinavia. Geol Surv Den Greenl Bull 20:31–34. Available via http://www.geus.
dk/DK/publications/geol-survey-dk-gl-bull/20/Documents/nr20_p031-034.pdf. Accessed 18
Oct 2016
Lunne T (2012) The Fourth James K. Mitchell Lecture: the CPT in offshore soil investigations—a
historic perspective. Geomech Geoeng 7(2):75–101. doi:10.1080/17486025.2011.640712
Montereale Gavazzi G, Madricardo F, Sigovini M, Janowski L, Kruss A, Blondel Ph, Foglini F
(2016) Evaluation of seabed mapping methods for fine-scale benthic habitat classification in
extremely shallow environments—application to the Venice Lagoon, Italy. Estuar Coast Mar
Sci 170:45–60. doi:10.1016/j.ecss.2015.12.014
Orange DL, Garcia-Garcia A, McConnell D, Lorenson T, Fortier G, Trincardi F, Can E (2005)
High-resolution surveys for geohazards and shallow gas: NW Adriatic (Italy) and Iskenderum
Bay (Turkey); Marine Geophysical Researches, 26: 247–266
Raytheon/GIL (2004) Technical demonstration of the DRUMS®-R200 sonar system. Technical
Report RPT-0440-900-015-B, Raytheon Integrated Defense Systems (USA) and Guigné
Chapter 5
The Future of Acoustic Seabed
Interrogation
Abstract The Acoustic Corer™ provides volumetric datasets of areas 12 m wide

and 40 m deep with decimetric resolutions. It fills the technology gap between
large-scale mapping by high-frequency sonars (with limited imaging capability
below the seabed) and small-scale profiling by low-frequency sensors (penetrating
deeper below the seabed at the expense of lateral resolution). Non-destructive, it
compares very favourably with traditional ground-truthing tools (e.g. Cone
Penetrometer Testing and boreholes). Chapter 5 shows current and future devel-
opments, moving to new subsea platforms (e.g. crawlers, Autonomous Underwater
Vehicles, drill-stems) and to new applications (e.g. mapping of deep-sea corals,
deep-sea mining, low-impact acoustic sensing). The adaptation of the Acoustic
Corer™ to terrestrial environments is detailed with the Acoustic Zoom™, able to
image areas wider than 10 square kilometres, down to kilometres below the ground.
Parallels between sonar and radar techniques also make it possible to envisage other
planetary applications, up to and including space, and some examples are provided.
! !
Keywords Bottom crawler Drill stem Autonomous Underwater Vehicle !
! ! !
AUV Deep-sea corals Deep-sea mining Acoustic Corer™ Acoustic Zoom™ !
5.1 Filling the Technology Gap
Seabed mapping gives large views of the seabed and its sub-surface, generally at
large resolutions and for limited depth ranges. Multibeam sonars and sidescan
sonars produce images of the seabed up to 30 km away from the imaging platform
(e.g. Blondel 2009), and more generally closer to a kilometre, with resolutions from
decametric to centimetric. These maps are usually restricted to the surface of the
seabed, although lower-frequency sonars (e.g. GLORIA working at 6.5 kHz) have
in some terrains been able to image sub-surface structures large enough to be
detected. Traditional sub-bottom profilers make use of these low frequencies to
provide information about the immediate sub-surface, but at resolutions generally
lower and only below the survey line. Parametric sonars, by their design, can

94 5 The Future of Acoustic Seabed Interrogation
provide similar penetration into the seabed, with much higher spatial resolution, but
again they are limited to recording only information directly at the nadir of the
surveying platform. Seismic surveying (e.g. Telford et al. 1990) provides
large-scale information about the sub-surface, to large depths but usually at lower
resolutions and missing totally the immediate sub-surface (first few meters).
Acoustic profiling of complex seabeds is therefore limited until one proceeds to
in situ sampling, using boreholes and cone penetrometer tests (CPTs). These
approaches are however extremely restricted spatially (lateral coverage of a few
tens of centimetres at most) and although they can in some cases reach down to
50 m below the seabed (e.g. Lunne 2012), they are bound to “miss” anything not
directly sampled or be stopped if faced with hard obstacles such as boulders. These
traditional approaches, although tried, tested and fully validated, still leave marine
scientists with clear gaps (Fig. 1.9). How does one sample the immediate
sub-seabed, in particular at depths relevant to offshore installations? Deeper mea-
surements typically do not have enough spatial resolution, whereas high-resolution
measurements typically do not reach deep enough. Direct sampling will not have
the width of coverage necessary, and acoustic sampling is limited to high fre-
quencies (high resolutions but high attenuation, i.e. shallow sub-seabed depths) or
low frequencies (low attenuation, but also low resolutions).
Acoustic interrogation of the seabed (ASI) directly addresses these different gaps,
as shown in Chap. 4 with examples from PanGeo’s Acoustic Corer imagery. The
combination of low and high frequencies, the use of highly-directive parametric
sonars and multistatic geometries (in particular with the JYG-Cross), SAS and chirp
Fig. 5.1 The Acoustic Corer now fills the gap originally identified in Fig. 1.9. Benefiting from
several decades of research and development, it enables accurate knowledge of the seabed
properties (geology, buried targets, acoustic velocities and therefore densities) at decimetric res-
olution, for volumetric coverage typically of 12 m in diameter " 40 m in depth
5.1 Filling the Technology Gap 95
capabilities, all offer a high, decimetric resolution for volumetric, wide-area coverage
typically at decimetric scale, for areas 12 m wide and 40 m deep. It offers a direct
pathway between traditional seabed mapping and highly localised sampling (Fig. 5.
1). The processing steps outlined in Fig. 3.16 make it possible to map the sub-seabed
with different types of information: acoustic velocities and attenuations can be
computed throughout the volume (i.e. at different depths, but also horizontally across
the ASI volume), individual returns can be identified as specular or non-specular
(Sect. 4.3) and they can be immediately sized and interpreted (Sect. 4.4).
5.2 Moving to New Platforms
Site investigations often require large-scale surveys, and the Acoustic Corer can be
deployed multiple times to cover different areas/volumes (Sects. 4.5 and 4.6).
Repeated measurements to systematically cover a large portion of seabed would
require a succession of deployments down to the bottom, followed by a succession
of recoveries, rapidly becoming time-consuming (and therefore also expensive).
The exact positioning of each deployment relative to the other ones would also need
careful preparation. So wouldn’t it be easier to actually stay on the seabed, and
image large areas in one single deployment?
This is where existing tools like deep-sea crawlers can bring their unique
potential. They were identified early on as “a major requirement of future pro-
grammes of research on the continental slope” (Rogers et al. 2002). Crawlers
provide stable platforms travelling along accurately known paths on the seabed,
over distances of kilometres or more. Using a high-grade Inertial Navigation
System (INS), their position can be synchronised with the other sensors they would
carry (e.g. acoustic sensors). In a single deployment, a version of the Acoustic
Corer mounted on a crawler could therefore image very large volumes, combining
different scans to provide a full view of the surface and sub-surface (Fig. 5.2, top).
For this purpose, a line array of acoustic receivers and transmitters can be
mounted on the carriage assembly, in a direction transverse to the direction of
motion (Fig. 5.2, top). The INS is advantageous in that it captures changes in
position during scanning. A position-determining transponder would be attached on
the carrier assembly (Fig. 5.2, bottom), and several navigational transposers would
be disposed onto the seabed, at precisely located and known spaces forming a wide
localization network placed on the seabed. The acquisition routines would then
communicate with these different transponders to get accurate positions of all
transmitters/receivers in the line array. The imaging of sub-seabed formations
would follow the ASI protocols, and the acoustic energy they scatter would be
detected along a predetermined line parallel to the length of the swath. The carriage
assembly would then move to a selected distance, transverse to the length of the
swath. This would be repeated until the desired area of seabed is totally covered.
The acoustic energy measured al all selected geodetic positions would then be
stacked coherently, and beam-steered to each of a plurality of depths and positions
Fig. 5.2 Application of a large 3-dimensional volumetric seabed ASI mapping, which would rely
on a bottom-crawling vehicle or other ROV (top). Combining successive scans made with
accurately-positioned sensors deployed from the crawler (bottom) allow access to potentially very
large volumetric datasets. The drawings are in support of the US patent 2015/9,030,914 by Guigné
et al., and reproduced in Guigné (2015)
along the length of the swath. These would then be stitched together to generate a
wide volumetric dataset, extending typically 50 m into the seabed. The stable
platform afforded by the carriage assembly enables deployment of sensors at
2–3.5 m from the seabed. As a starting point, the broad-beam high-intensity
acoustic transmitter would comprise three LFM chirp projectors, oriented at suitable
angles to fully project into the sub-bottom. With 4.5–14.5 kHz bandwidth, each
projector could typically provide a uniform 5-m wide swath of acoustic energy.
This cellular operation (as shown in Fig. 5.2, top) would therefore combine vol-
umes 5 m long " 5 m wide " 30–50 m deep to create very large acoustic cores.
Offshore geotechnical engineers are trained to rely on borehole and/or cone pen-
etrometer data for marine geotechnical, shallow sub-bottom formation assessment.
Acquisition of these samples is expensive and limited in its spatial representation of
what the true sub-bottom really is. In addition, the samples recovered may have been
altered or biased during their collection. Soft sediments become compressed during
core extraction and, depending on the extraction conditions, they may also fluidize
some of the other materials. This would yield incorrect stratigraphic interpretations. In
5.2 Moving to New Platforms 97
addition, there may be disruptive blockages in physical cores, coming from large
particulates or the inclusion of very dense fragments. This would yield misinterpre-
tation as bedrock formations (which do not exist in reality at the depth sampled).
Without a multiplicity of cores taken in close proximity, the exact lateral extents and
the true nature of these boundaries may not be known.
Earlier chapters showed how important it was to verify the extent, horizontal and
vertical, of these layers and targets. Complex sub-seabeds can contain soft sedi-
mentary lenses, boulder/cobble erratics, glacial tills, hard pans, fluidized discon-
tinuous mud layers, gas hydrates, gas-charged sediments and periglacial frozen soil
features inter alia. A version of the Acoustic Corer can be emplaced onto a
drill-stem (Fig. 5.3), replacing the usual tripod with the physical barrel used in
drilling or the rod that pushes the cone penetrometer. The approach requires pen-
etration by at least one meter into the seabed, to allow the acoustic arms to rotate in
a precise, stable manner to carry out signal transmission and reception. Like the
Acoustic Corer described in the previous chapters, this allows the evaluation of a
large volume of the seabed even in very soft sediments. This approach maximizes
the use of typically one core drill or CPT to evaluate the lateral extent of layers in a
specific position of the seabed. By cross-correlating with the core sample and/or
data from the cone, it allows checking that acoustic iterations between the two
sources of information produce a final consistent calibrated geophysical and
geotechnical interpretation of both the vertical and spatial distribution conditions of
the seabed. If blockages have occurred in the physical sampling these are imme-
diately noted and corrected through the acoustic imagery results. In approaching the
Fig. 5.3 Drill-stem Acoustic Corer (numbered labels refer to sub-elements of the system, as
explained in the patent). Image attribution: Guigné (2010) (US patent 2010:7,715,274)
collection of acoustic data in this collocated manner there is less obstruction of the
transmitted energy by any tripod legs and feet, and hence less is transmitted in the
water column that could later have bounced back into the data set being collected.
5.3 New Environments—New Applications
The applications presented so far rely on similar concepts throughout: the trans-
mission of bespoke signals covering a wide band of low and high frequencies, their
reception with a dense, configurable array and the capture of specular and
non-specular returns, coupled with advanced signal processing. These ideas form
the base of an unconventional 3D/4D seismic imaging system called the Acoustic
Zoom and validated in complex settings on land. The full details of this approach to
acoustic interrogation of complex sub-surface features, e.g. oil and gas reservoirs,
are given in Guigné et al. (2014), to which the reader is referred for the more
mathematical justifications.
Acoustic Zoom uses higher frequencies (typically 50–200 Hz) than conventional
seismics, enabling higher resolution. They are transmitted into the ground with a
high-fidelity vibroseis source (Fig. 5.4, top left), deployed as an acoustic lens along
Fig. 5.4 Research for the Acoustic Corer (used at sea) has led to the development of Acoustic
Zoom (used on land). A low-impact source like a vibroseis truck (top left) is used to elicit
scattering from geological layers and potential heterogeneities (bottom left). The acoustic returns,
from specular and non-specular directions, are measured over a large array of up to 4000 receivers
in 16 lines (right). Sources and receivers can be scattered over 12 km2 or more depending on the
application. Image source: Acoustic Zoom marketing material, also used in Guigné et al. (2014)
5.3 New Environments—New Applications 99
a pattern specific to the target under investigation. The seismic source is tuned to
deliver higher frequency energy sweeps, in a stacked mode, than conventionally
seen or possible in petroleum exploration (e.g. Telford et al. 1990). The receiver
array consists in a number of “arms” in a star “hub and spoke” configuration
(Fig. 5.4, right), whose shape is optimised for the depth and extent of the
sub-surface target of interest (e.g. oil reservoir). A typically configured receiver
array has sixteen spokes at 22.5° increments, with each spoke having discrete
vertical-component sensors (3-component acquisition can also be considered) at
sub-wavelength spacing, over a total spoke length of about 2000 m (120 wave-
lengths are required, spanning 50–200 Hz). 4000 receivers were distributed to form
the acoustic lens. The aperture thus created is essentially circular, approximately
4 km in diameter. Its resulting size means that the array’s near-field extends to
thousands of meters in depth, well beyond the zones generally of interest.
Acoustic Zoom uses the discrimination/steering power of the broadband vertical
component-sensor receiving array, purposely arranged with element spacing at
half-wavelengths across the frequencies of interest, to achieve both high vertical
and high lateral resolution (similar techniques are used in radio-astronomy arrays).
This results in the creation of steerable pencil beams for any volume of interest in
the ground (Fig. 5.4, bottom left). The number of independent beams, and thus the
size of the spot or voxel corresponding to an independent beam are controlled by
the aperture of the array.
To mitigate signal loss due to frequency-dependent attenuation, the sweep is both
high-frequency biased and non-linearly modulated. Coherent stacking is then
employed to boost the final signal-to-noise ratio. With the source and receiver arrays
stationary, the effectiveness of coherent stacking rests solely on the repeatability or
high-fidelity of the source signals at the higher frequencies being propagated (Fig. 5.5,
top). The resultant seismic data can then be focused with the sharpening of discrete
half-degree beams (formed on reception), providing significant discrimination against
energy arriving from internal earth volume reverberations, generally viewed in con-
ventional seismic as coherent noise. Capturing non-specular data is especially
important in unmasking the finer internal structure that often controls the fluid mi-
gration within the reservoir, and the provision of a 0.5° beam (Fig. 5.5, bottom) is
extremely important in distinguishing between these different returns..
The choices made with Acoustic Zoom induce several key departures from
conventional seismic imaging. Typical receiving arrays constrain the volume of
investigation to lie (well) inside the bounds of their footprint area, whereas the
smaller size of the Acoustic Zoom array can access larger lateral ranges and depths.
This is stemming from its wide-angle steering ability, similar to arrays used in
radio-astronomy. A full range of beam-forming and steering can be established for
clusters of sensors in the array leading to further discrimination and opening the
potential separation in subsequent attribute analyses (as visible in Fig. 5.6). Each
image of a particular volume in the ground can be built from a wide range of
backscatter angles, approaching ±45°. Multiple “views” will reveal more about
textures of formations, contrary to migration-based processing with final rendering
of only coherent reflection-based returns, filtering out the discontinuities or “noise”.
Fig. 5.5 Top The ambient noise shows a flat power spectral density (at 1.5 s Two-Way Travel time
in the earth). The effect of the vertical stacking is dramatic as it lowers the noise floor: one can note
the power spectra of varying folds from 1, 200, 300, 400, and 512 with the resultant exponentially
decaying power spectral density function, meaning the transmitted signal has useable energy all the
way to 170 Hz. Bottom Over 16 lines, the large array forming the acoustic lens enhances the
dynamic range, with side lobes 15 dB below the main lobe. Combining 120 wavelengths distributed
over 4000 receivers also results in a highly directional main lobe, 0.5° wide at −3 dB
The Acoustic Zoom concept can be adapted to marine operations. Developments

started in August 2013 and included the design of a novel acoustic source, similar
to a vibroseis source on land. Three resonant, in-phase programmable tubes form a
broadband array transmitting bespoke signals at frequencies from below 50 to
300 Hz. Each unit radiates 190 dB re. 1 lPa @ 1 m, directly into the ground, which
can be favorably compared to traditional marine seismic sources (often tens of dB
louder), and weighs 450 kg in air (with batteries and frame). Figure 5.7 illustrates
Fig. 5.6 Results from Acoustic Zoom measurements are very similar to those obtained with the
Acoustic Corer. In this example, 3-D seismic measurements were available to check Acoustic
Zoom interpretations, at least at comparable accuracies (Acoustic Zoom detected smaller features
and identified heterogeneities missed during traditional seismic surveys). Image source: Acoustic
Zoom marketing material, also used in Guigné et al. (2014)
the intended deployment of the marine version of Acoustic Zoom, with imaging at
different positions and receiving over a similar spoke array. Benefiting from the
processing experience acquired with both Acoustic Corer and Acoustic Zoom, it
provides very high-resolution details about potential reservoirs.
Whereas Acoustic Zoom (on land and at sea) aims at imaging very large vol-
umes, it is sometimes desirable to image much smaller targets. The lessons learned
with these tools have been used to propose a new concept of “Acoustic Mapping
and Interrogating Eye” (AMIE). This aims to be a low-cost, highly portable system,
which can be deployed from subsea vehicles like Autonomous Underwater
Vehicles (AUVs) or Remotely Operated Vehicles (ROVs). AMIE (Fig. 5.8) consists
in a half-dodecahedron array with 10 vertices. The centre of the array and each of
the 10 nodes supports parametric sonars, tuneable between 50 and 500 kHz
(Blondel et al. unpublished document 2010).
The parametric sonars have nominal 2° beamwidths. They are used sequentially to
transmit acoustic waves, whose reflections are measured by the other 10 sonars, used
as receivers. Because of the geometry, they are equidistant and at similar angles from
the focal point. The scattered signals provides a multi-aspect view of reflection and
attenuation within the object under investigation (a coral mound in Fig. 5.7, but other
objects like buried waste can also be studied) in a coherent spatial and temporal
manner, with 100 signals for each sequence. By sampling at MHz rates the signals on
each transducer, this will provide acoustic snapshots with millimetre resolutions in
volumes of several cubic metres. Interrogating pulses will sweep through available
Fig. 5.7 The extension of Acoustic Zoom into the marine realm uses low-impact seismic sources
working at higher frequencies and directing the largest part of their energy directly into the seabed
(no water column spreading or reflections, contrary to shallow-towed seismics). Acoustic Zoom
marketing material, reproduced in Coley (2015)
Fig. 5.8 Physical setup of the Acoustic Mapping and Interrogating Eye (AMIE) used for
high-resolution mapping of complex seabed structures like deep-sea corals. AMIE uses parametric
sonars (located on each of the 10 nodes of the array and activated in sequence) and high-frequency
multistatic processing of their measurements (at broadband hydrophones located on each of the
other nodes). The use of a Remotely-Operated Vehicle (ROV) allows for “on-the-fly”
measurements of coral reefs and more detailed investigations when needed. It also constrains
the size, weight and power budget of the entire AMIE setup. From Blondel et al., unpublished
document (2010)
frequencies, accessing different physical scales and resolutions. Outgoing waveforms

and pulse-repetition rates can then be matched to investigate surface scattering
(e.g. coral branching densities or buried waste structural integrity) and volume scat-
tering (e.g. coral growth rates or buried waste internal content). Feasibility was
demonstrated in laboratory studies by Blondel et al. (2006, 2010), for different coral
types, and in Blondel and Caiti (2007) for buried waste.
5.4 Conclusions
The key concepts of Acoustic Interrogation, as introduced by Guigné (1986) and as

presented in this book, are not limited to seabeds. The validation of the Acoustic
Zoom showed how it could successfully and dramatically augment traditional
seismic studies to fully exploit the reflected and scattered wavefields to produce an
augmented conventional migrated volume with a rich and unique complementary
“sediment texture’ volume for large underground reservoirs. The development of
the Acoustic Mapping and Interrogating Eye, complemented with laboratory
studies, showed it was possible to image and characterise small marine targets
above ground. The analogy of radar with seismo-acoustic techniques makes some
of these processing techniques useable in space (e.g. Sava et al. 2015), in particular
for asteroid mapping and mining (Blondel 2016). Current plans for advanced
exploration of the Moon and Mars rely on the accurate identification of sub-surface
resources like oxygen or frozen water, meaning that smaller, space-adapted versions
of the Acoustic Corer and its derivatives could one day be used beyond the Earth.
This would be a highly fitting extension to the voyage of scientific discovery and
technical innovation started 30 years ago. Closer to home, and closer in time, what
will be the immediate developments of Acoustic Seabed Interrogation techniques?
First, the domains of application are slightly shifting based on economical and
geopolitical changes. Offshore site investigations will continue to be dominated by
exploration and by developments of marine-based petroleum reserves. What has
changed over the past two decades is that, along with oil and gas exploration, vast
wind farm projects have entered the scene worldwide and in a dominant manner.
Unconventional petroleum and mining activities such as gas hydrates, heavy
minerals and marine mining will also become important offshore resources,
expanding in economic importance. It is almost certain that resources from the
Arctic and other offshore regions will be extracted before the end of the next
decade. In later years, as developments move from the continental shelf to its
deeper slopes, comprehensive marine geotechnical information will be required and
depended on to mitigate installation risks.
In all these investigations, the challenge to marine geotechnical engineers is to
develop the more cost-effective approaches to characterise the soils upon which
structures can be built, through which boreholes can be drilled, and from which
minerals will be mined. Not only do these approaches need to be economical, they
also need to be increasingly accurate. Practices in 2016 with respect to determining

seabed loads remain conservative, with large margins to accommodate uncertainties
and errors. Such structural design loads, if they are vastly overestimated and if the
foundation capacity is underestimated, would combine and lead to massive struc-
tures and foundations that are so expensive that it would be uneconomical to
develop the reservoir or wind farm installation. Thus the influence of marine
geotechniques is not limited to improving the economics of site characterization
methods, although this is an important goal; it also has a direct input to the selection
and design of the most appropriate structure from which to develop the field.
Advances have been made over the last decade in site characterization. The
in situ methods, particularly the cone penetration tests described by Robertson
(1986) and the self-bored pressuremeter tests by Jefferies et al. (1987), have
improved significantly in their ability to determine the state of the soil and provide
the property values required for design. Programs for curve fitting and interrogating
data will continue to evolve such that the engineer can have multi-facetted, fused
and cross-correlated data delivered in whatever form specified, whether it be state
parameter, specific engineering properties, a pile design, or indeed an entire
foundation design. The adoption and growth of ROVs, AUVs, and bottom crawlers
suggest that it may soon be possible to do all of this remotely, especially in
exploring and developing resources in extreme environments. For example,
geotechnical site investigations in polar regions are being planned to be carried out
by engineers using customized remote vehicles for surveying route selections,
detecting, identifying, and removing geo-hazards and then safely burying cables
systems (Dinn 2012). Rapid volumetric site characterization and geo-hazard
delineation will be essential for quality-controlled foundation designing without
incurring cost-prohibitive, conventional sampling programs. The use of large
(e.g. deep-sea) vessels for soil sampling and test boring is generally extremely high.
The current trend is to lessen the reliance of drill holes and undisturbed samples to
in situ tests with no samples, where the cost per meter of a cone test is typically
about half that of a drilled hole. Acoustic Seabed Interrogation is therefore ideally
placed to fill this characterisation gap.
Data acquisition is nothing without adequate processing, and the methods pio-
neered by Guigné (1986) and later collaborations are now commercialised by
PanGeo Subsea and by Acoustic Zoom Inc. since 2010, and routinely built upon.
The combination of JYG-Cross processing, chirp and SAS rendering, and the
combination of low and high frequencies in multiple-view presentations of the
acoustic data create a high computational burden. ASI’s answer was to use dedi-
cated clusters of GPUs (Graphical Processing Units; e.g. Yuen et al. 2013) and
CPUs (Central Processing Units). Trends are now seen in the application of expert
systems to site characterization and design, such as automatic curve fitting methods
for conventional in situ tests, innovative methods of interrogating the seabed by
acoustics, and real-time graphic parallel processors that can produce answers
quickly to engineering decisions that affect the economics and safety of offshore
operations. Stuyts et al. (2010) introduced the use of geo-statistics to help eliminate
5.4 Conclusions 105
the need for a borehole at every wind farm foundation given that the input into the
statistics carries spatial information of the substrata boundaries in a reliable, rep-
resentative manner. This trend will reduce the requirement to look at a soil or touch
it or break it or shake it. That will all be done indirectly. Design alternatives will be
provided that could be economically predicted and conceptually constructed and
optimized through expert computer aided design software to eliminate decision
biases and subjective judgments on the part of the installation engineer.
Acoustic Seabed Interrogation is now providing answer products that can pene-
trate 50 m or deeper (100 m in theory) into the seabed, and identify small-scale
structures at a scale below 10 cm, from horizons to discontinuities. Deployment costs
are current transitioning from thousands of US$/meter deep to hundreds of US$.
The use of new platforms and the extension to deeper and more varied (and
challenging) environments make the future of acoustic interrogation of complex
seabeds even more interesting and important in the years to come.
References
Blondel Ph, Caiti A (eds) (2007) Buried Waste in the Seabed—Acoustic Imaging and Bio-toxicity
(Results from the European SITAR project). Springer-Praxis, Chichester
Blondel Ph, Huvenne VAI, Hühnerbach V (2006) Multi-frequency acoustics of deep-water coral
habitats and textural characterisation. Proc. 8th European Conference on Underwater Acoustics
(ECUA):379–384
Blondel Ph, Munday J, Nash D (2010) Multi-angle ultrasonic imaging of corals. Proc. 8th
European Conference on Underwater Acoustics (ECUA):123–130
Blondel Ph (2009) Handbook of Sidescan Sonar. Springer, Heidelberg
Blondel Ph (2016) From Deep Sea to Outer Space—Tools for De-Risking Mining Operations. In:
Graps A. (ed) Proc. Asteroid Science Intersections with In-Space Mine Engineering, p. 39.
Available via: http://europlanet-scinet.fi/index.php?id=390. Accessed 18 Oct 2016
Coley K (2015) Acoustic Zoom: the future of offshore exploration, Mar Tech Rep:32–37
Dinn G (2012) Field experience with a new sub-bottom investigation tool: Acoustic 3-D imaging
of the sub-seabed. Proc, IEEE Oceans 2012
Guigné JY (1986). The concept, design and experimental evaluation of “acoustic sub-seabed
interrogation”. PhD thesis, University of Bath, Bath, UK
Guigné JY (2010) Wide-area seabed analysis, US Patent 7,715,274, 11 May 2010
Guigné JY, Stacey AJ, Clements C, Azad S, Pant A, Gogacz A, Hunt W, Pace NG (2014)
Acoustic Zoom high-resolution seismic beamforming for imaging specular and non-specular
energy of deep oil and gas bearing geological formations. J. Nat. Gas Sci. Eng. 21:568–591.
doi:10.1016/j.jngse.2014.09.012
Guigné JY (2015) The future for “acoustics in the seabed”—The next chapter. J. Ocean Tech.
10(3):37–47
Jefferies MG, Crooks JHA, Becker DE, Hill PR (1987) Independence of geostatic stress from
over-consolidation in some Beaufort Sea clays. Can Geotech J 24:342–356
Lunne T (2012) The Fourth James K. Mitchell Lecture: The CPT in offshore soil investigations—a
historic perspective. Geomechanics and Geoengineering 7(2):75–101. doi:10.1080/17486025.
2011.640712
Robertson PK (1986) In situ testing and its implication to foundation engineering. Can Geotech J
23:573–594
Rogers A, Billett D, Berger W, Flach E, Freiwald A, Gage J, Hebbeln D, Heip C, Pfannkuche O,

Ramirez-Lodra E, Medlin L, Sibuet M, Soetaert K, Tendal O, Vanreuseul A,
Wlodarska-Kowalczuk M (2002) Life at the edge: achieving prediction from environmental
variability and biological variety. In: Wefer G, Billett D, Hebbeln D, Jorgensen BB, Schlüter
M, Van Weering TCE (eds) Ocean Margin Systems. Springer, Heidelberg, pp 387–404
Sava P, Ittharat D, Grimm R, Stillman D (2015) Radio reflection imaging of asteroid and comet
interiors I: Acquisition and imaging theory. Adv Space Res 55:2149–2165
Stuyts B, Vissers V, Cathie DN, Dorfeldt S (2010) Optimizing Site Investigations and Pile Design
for Wind Farms Using Geostatistical Methods: A case study. In: Proc. Frontiers in Offshore
Geotechnics, Perth
Telford WA, Geldart LP, Sheriff RE (1990) Applied Geophysics, 2nd edn. Cambridge University
Press, Cambridge
Yuen DA, Wang L, Chi X, Johnsson L, Ge W, Shi Y (eds) (2013) GPU solutions to multi-scale
problems in science and engineering. Lecture Notes in Earth System Sciences, Springer,
Heidelberg
Index
A H
Acoustic attenuation, 3, 20, 59 High-frequency (HF), 21, 22, 26, 29, 45, 58,
Acoustic Corer, 13, 15, 38, 39, 45, 46, 49, 54, 61, 68, 102
57, 58, 62, 65, 71, 77–79, 81, 85, 90, 95, Horizon, 2, 14, 19, 20, 36, 38, 43, 46, 59, 71,
97, 103 76, 78, 79, 86, 105
Acoustic (or seismic) velocity, 14, 71
Acoustic Seabed Interrogation (ASI), 65, I
103–105 Inertial Navigation System (INS), 57, 95
Acoustic Zoom, 98, 99, 101, 103, 104 Interrogation, 14, 38, 62, 94, 98, 105
Automatic Gain Correction (AGC), 55
J
B JYG-Cross, 39, 48, 49, 57, 62, 65, 67, 94, 104
Beamforming, 15, 28, 30
Bedrock, 6, 96 L
Borehole, 2, 10, 11, 19, 35, 36, 62 LFM, 58, 59, 95
Boulder, 1–3, 10, 14, 15, 42, 46, 59, 61, Low-frequency (LF), 20, 39, 61, 65, 90
69, 71–73, 75, 76, 82, 89, 97
M
C Migration, 11, 41, 48, 58, 99
Chirp, 40, 46, 58–60 Mobile Offshore Units (MOU), 4, 7
CMP, 48–50
Cobbles, 71, 75–77 N
Computed Tomography (CT), NMO, 50, 52
26, 27, 29 Non-specular, 11, 13, 20, 48, 68, 90, 98
Coring, 9, 13, 15, 65
CPT, 2, 9–11, 14, 19, 36, 37, 40, 65, 75, O
79, 81–83, 85 Offshore (construction, foundations), 1, 9
D P
Drill/drilling, 1, 2, 7, 21, 71, 81, 97 Parametric (array, sonar), 20, 22, 23, 28, 39,
46, 62, 68, 75, 86, 101, 102
F Positioning, 7, 22, 57, 95
F-K, 67 Punch-through, 80, 81
G R
Gas/Gassy sediments, 1, 7, 85–87, 103 Rendering, 15, 42, 61, 68, 78, 104

SpringerBriefs in Oceanography, DOI 10.1007/978-3-319-02579-7
108 Index
S SVD (filtering), 48
Sampling, 2, 4, 9, 19, 24, 60, 94, 97, 104 Synthetic Aperture, 41, 42, 57, 58, 62, 68
Semblance analysis, 50, 51
Sonar T
multibeam, 6, 12, 15, 19, 48 Texture/textural character, 11, 22, 75, 76, 90
sidescan, 6, 10, 15, 19, 26, 93 Tidal turbines, 7
synthetic aperture (SAS), 41, 57, 61 Till (including glacial till), 3, 72, 77, 97
Specular, 11, 14, 20, 69, 94
Spudscan, 15 V
Stacking, 24, 50, 52, 99, 100 Volume/Volumetric, 9, 14, 19, 21, 36, 38, 42,
Stratigraphy/stratigraphic, 3, 4, 14, 22, 25, 43, 56, 60, 62, 69, 71, 81, 88, 94, 95, 99,
52, 55, 56, 78, 82, 86, 88, 96 103
Summation, 25, 41, 42, 60, 61
Surface, 2, 6, 9, 14, 19–21, 52, 62, 77, 93, W
95, 103 Wind farm, 7, 10, 85, 88, 104

Acoustic Investigation of Complex Seabeds

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

Acoustic Investigation of Complex Seabeds

Uploaded by

Copyright:

Available Formats

SPRINGER BRIEFS IN OCEANOGRAPHY

Acoustic Zoom Ltd.

ISSN 2196-1212 ISSN 2196-1220 (electronic)

© The Author(s) 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

For my parents: role models,

Marine environments are increasingly important to human activities, with

Paradise, NL, Canada Jacques Yves Guigné

1 Acoustic Coring—The Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

4.3 Sizing and Identifying Individual Targets . . . . . . . . . . . . . . . . . . . . 70

JYG-Cross Patented conﬁguration of transducers, designed by Jacques Yves

Abstract The scope and ambition of marine geotechnical investigations have

1.1 The Need for Accurate Assessments

© The Author(s) 2017 1

1.2 Importance to Present Maritime Operations

Mobile Offshore Units are the traditional beneﬁciaries of detailed geotechnical

1.3 The Technology Gap

knowledge on the composition of the seabed to address cost-effectively the issues of

Existing Spudcan Impressions in the Seafloor

offsets, render the data via migration/inversion protocols so as to discriminate

1.4 Filling the Gaps—The Case for “Acoustic Coring”

The concept of acoustic coring, also called Acoustic Sub-Seabed Interrogator

Keywords High-frequency ! Low-frequency ! Broadband ! Parametric arrays !

2.1 Capturing the Relevant Characteristics

© The Author(s) 2017 19

geotechnical characteristics of the surface (e.g. Wienberg and Bartholomä 2005).

2.2 Buried Objects and Benthic Habitats

at very distinct locations, for example by positioning receivers on moving AUVs.

A related approach by Guigné and co-workers (Schwinghamer et al. 1996) had

Fig. 2.8 Fractal dimensions

2.3 Signiﬁcance and Value—Moving Toward

multiple receivers on a vertical line array; (2) multi-aspect imaging of subtle

Blondel Ph (2009) Handbook of sidescan sonar. Springer, Heidelberg

3.1 Creating a Prototype

The recommendations concluding Chap. 2 were ﬁrst embodied with a prototype

© The Author(s) 2017 33

in an octagonal polyethylene framework held by a circular aluminium structure

3.2 Formulating the First Product

Fig. 3.4 Bespoke signal

3.3 The Next Generation: The Acoustic Corer

3.4 Data Processing—Coherent Summation

voxel (small computational volume) and with appropriate weighting

successively interrogating individual voxels. The answer product derived from

3.5 Processing Acoustic Corer Data

3.5.1 Data and Acquisition

Accentuating Specular and Non-Specular Seismic Events from within Shallow

3.5.2 JYG-Cross Processing

Fig. 3.18 The JYG-Cross

employed. Pre-stack migration is generally employed upfront along the regularly

DsNMO ð xÞ ¼ sð xÞ % sð0Þ ð3:1Þ

where sðxÞ is the two-way-travel-time to a speciﬁed depth point on a reflector, with

where vðsð0ÞÞ is the two-way travel-time, dependent on the stacking velocity

CMP-gathers or to the stacked data. The reflection mechanisms particular to

3.5.3 Synthetic Aperture—Rendering and Processing

Synthetic Aperture is a technique commonly used in radar remote sensing, and

previously indicated, the focusing methodology is based on straight-ray travel-time

reconstruction of the desired (approximately) 25-ms chirps is obtained via

Fig. 3.27 The respective

xanalytical ðk Þ ¼ xðkÞ þ j H fxðkÞg ð3:6Þ

3.5.4 Combining Low-Frequency and High-Frequency

Successful acoustic interrogation of seabed and sub-seabed structures requires a

Ainslie M (2010) Principles of sonar performance modeling. Springer-Praxis, Chichester

4.1 Using the Low-Frequency (LF) JYG-Cross Data