MARINE GEOLOGICAL SURVEYING AND SAMPLING

Marine Geological Surveying

and Sampling

Edited by

E. A. HAlLWOOD
Oceanography Department, University of Southampton, Southampton, U.K.
and
R. B. KIDD
Geology Department, University College of Wales, Cardiff, U.K.

Reprinted from Marine Geophysical Researches

Vol. 12, Nos. 1-2 (1990)

KLUWER ACADEMIC PUBLISHERS

DORDRECHT / BOSTON / LONDON
Library of Congress Cataloging-in-Publication Data

Marine geologIcal surveying and samplIng I edited by Ernest A.

Hai lwood and Robert B. Kldd.
p. cm.
"Papers ... from a meetIng organized in May 1988 at the Geological
SocIety, London. under the auspices of its Marine Studies Group"-
-Introd.
"Reprinted from MarIne geophysical researches, v. 12, no. 1-2
(1990). "

1. SubmarIne geology--Congresses. 2. Geological surveys-

-Congresses. 3. Sampllng--Congresses. 1. Hallwood, E. A.
II. Kjdd. P~bert 8. III. Marlne geophy~ical researches.
QE39.M294 1990
551.46' 08--dc20 90-4537

ISBN-13: 978-94-010-6763-8 e-ISBN-13: 978-94-009-0615-0

DO I: 10.1007/978-94-009-0615-0

Published by Kluwer Academic Publishers,

P.O. Box 17, 3300 AA Dordrecht, The Netherlands.

Kluwer Academic Publishers incorporates the publishing programmes of

Martinus Nijhoff, Dr W. Junk, D. Reidel, and MTP Press.

Sold and distributed in the U.S.A. and Canada

by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.

In all other countries, sold and distributed

by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AH Dordrecht, The Netherlands.

printed on acid-free paper

All Rights Reserved

© 1990 Kluwer Academic Publishers
Softcover reprint of the hardcover 1st edition 1990
No part of the material protected by this copyright notice may be reproduced or utilized in
any form or by any means, electronic or mechanical, including photocopying, recording, or
by any information storage and retrieval system, without written permission from the
copyright owner.
Contents

E. A. HAlLWOOD and R. B. KIDD / Editorial Introduction vii

PART I: SURVEYING TECHNIQUES

K. G. ROBERTSON / Deep Sea Navigation Techniques 3

J. A. GRANT and R. SCHREIBER / Modern Swathe Sounding and Sub-Bottom Profiling Technology for
Research Applications: The Atlas Hydrosweep and Parasound Systems 9
R. C. SEARLE, T. P. LE BAS, N. C. MITCHELL, M. L. SOMERS, L. M. PARSON, and PH. PATRIAT / GLORIA
Image Processing: The State of the Art 21
J. M. REYNOLDS / High-Resolution Seismic Reflection Surveying of Shallow Marine and Estuarine Environ-
ments 41
C. M. R. ROBERTS and M. C. SINHA / A Fixed Receiver for Recording Multichannel Wide-Angle Seismic Data
on the Seabed 49
M. C. SINHA, P. D. PATEL, M. J. UNSWORTH, T. R. E. OWEN, and M. R. G. MacCORMACK / An Active
Source Electromagnetic Sounding System for Marine Use 59
Q. HUGGETT / Long-Range Underwater Photography in the Deep Ocean 69

PART II: SAMPLING TECHNIQUES

P. P. E. WEAVER and P. J. SCHULTHEISS / Current Methods for Obtaining, Logging and Splitting Marine
Sediment Cores 85
W. R. PARKER and G. C. SILLS / Observation of Corer Penetration and Sample Entry during Gravity Coring 101
M. A. STORMS / Ocean Drilling Program (ODP) Deep Sea Coring Techniques 109
R. B. KIDD, Q. J. HUGGETT, and A. T. S. RAMSAY / The Status of Geological Dredging Techniques 131
J. WHITE / The Use of Sediment Traps in High-Energy Environments 145
P. J. SCHULTHEISS / Pore Pressures in Marine Sediments: An Overview of Measurement Techniques and
Some Geological and Engineering Applications 153

List of Contributors 169

Editorial Introduction

This collection of papers originates from a meeting are in current use on board UK research vessels.
organized in May 1988 at the Geological Society, Marine geological exploration requires information
London, under the auspices of its Marine Studies under three further headings: (i) the "shape" of the
Group. The meeting was concerned with reviewing sea floor, (ii) the nature of the rocks and sediments
the present state-of-the-art of marine geological and which lie at its surface, and (iii) the nature of deeper
geophysical sampling and surveying techniques. structures. Studies of the shape of the sea floor
The pace of scientific exploration of the ocean (bathymetry) are based primarily on echo sounder
basins has increased dramatically over the past few and side-scan sonar surveying. Technology in this
decades in response to interest in the global tectonic field has seen major advances over the past two
processes which control their long-term evolution decades, with the development of new ceramic ma-
and the regional and local sedimentary and tectonic terials to provide more efficient and powerful trans-
processes which shape them, as well as more practi- ducers, the increasing use of digital data processing
cal questions such as the nature and extent of off- techniques to improve the quality of the signal from
shore mineral resources, problems of waste disposal the sea floor, and the introduction of new design
at sea and the response of sea level to global climatic concepts to provide higher resolution records.
change. Marine geological research commonly re- Notable advances in this field include the develop-
quires that investigations be carried out in a difficult ment of commercially available narrow-beam echo
and often hostile environment. Observations must be sounders, such as the Atlas Parasound system de-
made and instruments controlled at the sea floor, scribed in the paper by Grant and Schreiber, which
often many kilometres below the observer or provides greatly enhanced resolution compared with
recorder at the sea surface. In marine geology, more conventional broad-beam systems. Imaging of larger
than in most other branches of earth science, major areas of the sea floor is achieved through side-scan
advances have depended on corresponding develop- sonar surveying. The paper by Searle et al. discusses
ments in available technology. recent developments in the use of the UK Institute of
Although a number of general marine geological Oceanographic Sciences long-range side-scan sonar
texts have been published in recent years, few of system GLORIA, particularly in the fields of image
these address the specific problems of data acquisi- processing, accurate determination of backscattering
tion across the broad spectrum of marine geological levels (important in the use of these data for deter-
surveying. We felt that the pace of change in recent mining bottom type) and in the digital combination
years has been so great that there was a clear need of individual records into composite mosaics. The
for a text which reviews recent developments in this GLORIA system sweeps the sea floor to distances of
field in sufficient detail to be of value to practicing about 0.5 km either side of the research vessel. The
marine geologists as well as to others requiring a resulting record provides a more-or-Iess continuous
more general appreciation of marine geological sur- acoustic plan of the sea floor in which individual
veying techniques. This collection of papers attempts relief features are recognized from their geometry,
to meet this need. acoustic reflectivity and the acoustic shadows which
The first requirement of any study of the ocean they cast. However, precise bathymetric data over
floor is to know the position of the survey vessel the insonified area are not determined.
and the sampling/surveying instrument with the An alternative approach is that of "swathe sound-
maximum possible accuracy. The first paper, by ing" in which a number of sharply focused sound
Robertson, provides a summary of the position- beams making different angles with the vertical are
fixing techniques, for both ship and instrument, that transmitted simultaneously from a single transducer

Marine Geophysical Researches 12: vii-ix, 1990.

© 1990 Kluwer Academic Publishers.
viii EDITORIAL INTRODUCTION

array. Each beam provides a separate precise depth lific sampling and the particular limitations and ap-
determination at a different range from the vessel's plications of each type.
track and these are digitally combined to produce a Perhaps the greatest advances in scientific sam-
contoured bathymetric map of the survey area. The pling of ocean-floor rocks, particularly those at
US Sea Beam system utilizes this principle and has depths in excess of a few tens of metres beneath the
been widely used, particularly by North American sea bottom, have been made through the develop-
marine research institutes, over the past decade or ment and application of rotary drilling and sub-bot-
so. A more recently developed system, based on the tom coring techniques by the Deep Sea Drilling
same general principle is the Atlas Hydrosweep sys- Project and its successor, the Ocean Drilling Pro-
tem, described in the paper by Grant and Schreiber. gram, over the past two decades. The review by
On a smaller scale, information on the nature of Storms provides a detailed account of the particular
the sea floor can be obtained from direct observa- problems involved in the efficient recovery of core
tions using submersibles or underwater television samples by deep sea drilling and discusses recent
and photography. The paper by Huggett provides a advances in core recovery and low-disturbance
review of recent developments in deep ocean under- drilling techniques.
water photography, with a particular emphasis on In addition to obtaining samples of rocks and
the problems of light scattering and data retrieval. sediments from at and beneath the sea floor, marine
The second major requirement of marine geologi- geologists also require to obtain samples of the par-
cal surveying is t%btain samples of the rocks and ticles being deposited from the water column at the
sediments which form the upper layers of the sea present time, in order to explore processes of sedi-
bed, in order to investigate geologically recent pro- ment transport and deposition. This is achieved by
cesses in the ocean basins. The four techniques used the deployment of sediment traps. The paper by
to obtain such samples are grab sampling, dredging, White reviews the different types of sediment trap
coring and ocean drilling. The most widely used that are in current use and the extent to which
technique for sampling soft sediments is coring. The different designs provide a reliable measure of the
paper by Weaver and Schultheiss reviews the differ- true sediment particle flux to the sea floor.
ent kinds of sediment corer in current use and their Insight into geological processes occurring at or
various advantages and disadvantages. It also dis- beneath the sea floor can be obtained from measure-
cusses the techniques that are used for logging vari- ment of various physical and geotechnical properties
ous physical properties of the sediments, splitting the of sea-floor sediments. One such property which has
cores, describing them and taking sub-samples. Pene- received attention in recent years is pore-pressure.
tration of the corer into the sea floor commonly This property can provide important information,
compacts the sediment, so that the thicknesses of for example, on hydrothermal circulation processes
sedimentary units observed in the core are not neces- in the oceanic crust. Techniques for measuring this
sarily representative of the true thicknesses of the property and examples of the value of the data
geological units beneath the sea floor. Recent experi- obtained are reviewed in the paper by Schultheiss.
ments, in which penetration of the corer have been The third major requirement of marine geological
monitored acoustically are described in the paper by surveying is to obtain information on the structure of
Parker and Sills, which provides important new in- the sea floor to greater depths than can be explored
sights into the question of the fidelity of the geologi- by simple gravity coring and over wider areas than
cal record observed in sediment cores. can be represented by isolated deep drilling. Such
Harder rocks, particularly igneous and lithified information is obtained by geophysical surveying,
sedimentary rocks normally cannot be sampled by principally using seismic techniques. Multichannel
simple gravity-powered corers and the technique seismic reflection surveying is widely used in the oil
commonly employed for sampling such rocks at the exploration industry and current techniques are well-
sea floor is that of dredging. Despite the widespread described in the literature. The present publication
use of this technique, it is poorly documented in the focuses on two aspects of seismic techniques. The
literature. The contribution by Kidd et al. reviews first is the application of high-resolution single-
the different types of dredge in current use for scien- channel seismic reflection profiling to exploring the
 EDITORIAL INTRODUCfION ix

geology of shallow marine and estuarine environ- together with results of initial experiments which
ments, discussed in the contribution by Reynolds. demonstrate the very considerable value of this new
The second is the particular problem of recording marine geological surveying technique.
multichannel wide-angle seismic data on the conti- This collection of papers cannot be claimed to be
nental shelf. The latter topic is addressed in the comprehensive in the sense of a textbook. However,
paper by Roberts and Sinha, which describes the we believe that the reader will gain a broad perspec-
principles and operational procedures used in a new tive of the state-of-the-art in many areas of marine
system, the Pull-Up Multichannel Array (PUMA), geology through these papers and we express our
together with results of preliminary deployments of thanks to our contributors for their efforts to provide
the system on the UK continental shelf. this overview.
A relatively new technique of marine geological
exploration is electromagnetic sounding. This pro- E. A. Hailwood
vides information on the properties and distribution Oceanography Department,
of conductive fluids within the oceanic crust. Such University of Southampton
fluids include hydrothermal water and magma, so
that electromagnetic sounding can provide important
R. B. Kidd
new insights into geological processes such as hy-
Geology Department,
drothermal convection and magma emplacement at
University College of Wales
oceanic spreading centres. The paper by Sinha et al.
Cardiff
describes a new approach to the problem of electro-
magnetic sounding at sea, utilizing an active source, September 1989.
PART I

Surveying Techniques
Deep Sea Navigation Techniques
K. G. ROBERTSON
Natural Environment Research Council, Research Vessel Services, No.1 Dock, Barry, South Glamorgan CF66UZ, UK

(Received 27 April, 1989; accepted 1 September, 1989)

whilst maintaining an accurate knowledge of one's

Key words: Deep sea navigation, position fixing, scientific survey- position relative to known sea-bed features. For
ing.
example, underway geophysical studies using towed
air guns, hydrophone streamers, side scan sonar,
Abstract. Accurate navigation forms an essential part of all re-
search at sea and the deep ocean imposes it's own unique prob- magnetometers and shipbome gravity meters.
lems. This chapter discusses several of the techniques in current 3. To manoeuvre accurately a submerged package
use on the research vessels of the Natural Environment Research along a chosen track on the sea-bed, for example,
Council (NERC), concentrating on those systems which provide
global navigation facilities, as opposed to the more localised,
guiding deep towed instruments and Remote Oper-
coastal aids. Whilst most of the systems rely on surface propaga- ated Vehicles. This technique is also applicable to
tion of radio waves, the use of acoustics and sea-bed mapping refining the objectives described in 1. above.
instruments constitute accurate alternatives for some sub-sea
It will be obvious that success will often depend
applications.
upon using a combination of navigation techniques,
a knowledge of any previous work carried out in the
Introduction same region and access to the relevant, detailed
charts and geological maps, where they exist.
A wide variety of' navigation aids is available to
assist the scientist working in the marine environ-
ment. This chapter concentrates on those systems Navigation Techniques
which can be employed in deep sea applications and,
Techniques used by NERC have been refined over
in particular, those currently used by the Natural
many years as technological advances have given us
Environment Research Council on the research ves-
new tools with which to operate. With the aid of
sels RRS Charles Darwin, RRS Discovery and RRS
modem computing systems, installed on each of the
Challenger.
NERC ships, we are able to combine the outputs of
Although the main accent of this chapter is on
many devices and apply quality control parameters
marine geology, the techniques described are equally
that can reduce still further the individual instrument
well suited to other branches of science at sea.
errors and produce high quality, annotated track
charts in near real time.
What is Navigation?
The suite of navigation instruments one would
In the context of marine geology, navigation can normally expect to be present on board when work-
perhaps be defined as the means by which one ing in deep water would include the following.
achieves the following objectives.
TRANSIT Satellite Receiver.
1. To locate oneself accurately at a point on the
GPS Satellite Receiver.
surface of the ocean, above the sea-bed feature under
Gyro Compass.
investigation, for example when locating a site at
Electromagnetic and/or Doppler speed sensors.
which to take core or grab samples, or to make
underwater TV or photographic observations. For more specialised applications other instru-
2. To manoeuvre a ship along a chosen track, ments may also be included, such as:

Marine Geophysical Researches 12: 3-8, 1990.

© 1990 Kluwer Academic Publishers.
4 K. G. ROBERTSON

Radar Transponders. speed from the electromagnetic or doppler log, until

Microwave Ranging Systems. another satellite transmission is received. This proce-
Acoustic Transponders. dure is described as 'Dead Reckoning' and is open to
error from any inaccuracies in the calibration of the
Looking ahead, these may be supplemented in the
sensing devices. Cumulative errors can result in ap-
future by such instruments as:
parent jumps in one's assumed track when another
Swathe or Multi-Beam sounders. reliable spot fix is obtained (Fig. 2). With the use of
modern computing techniques the data can be repro-
Each of the techniques listed above is described in
cessed to minimise the uncertainties.
more detail in the following sections.
Although now beginning to be superceded by
TRANSIT GPS, the TRANSIT system remains a useful tool
and will continue to be operational for many years
The TRANSIT, satellite based system has been the
yet.
mainstay of deep sea navigation within NERC for
If one has the opportunity to accumulate a series
almost two decades. A number of polar orbiting
of spot fixes on a stationary target it is possible that
satellites circle the earth every 107 minutes, at an
an accuracy of 100 metres, or better, can be ex-
altitude of 600 nautical miles.
pected. For a moving ship this will be downgraded
The orbits trace a pattern like the sections of an
since the reliability of a fix will depend upon the
orange, wide at the equator and converging at the
quality of the data coming from the satellite, the
poles (Fig. 1). However, the orbits do not rotate with
characteristics of its orbit relative to the ship, the
the earth so they suggest a sort of cage, inside which
forward velocity of the ship and any ship motion
the earth turns on it's axis. Consequently, each point
induced accelerations.
on the earth's surface passes under each of the orbits
Typical accuracy in the deep sea is in the region of
approximately twice a day.
100 m to 1 nautical mile, dependent upon the
Each satellite transmits information, as a function
parameters described above. After prolonged periods
of time, about it's position relative to the centre of
of 'Dead Reckoning' the accumulated errors can
the earth. By measuring the change of Doppler
create an uncertainty of 2 km or so but this can vary
frequency of the received signals as the satellite
greatly.
approaches, passes and recedes, it is possible to fix
the position of the ship relative to the satellite and
GPS (NAVSTAR)
hence the precise position on the surface of the earth.
However, as there is a relatively small number of Undoubtedly, the navigation system upon which
satellites and they can exhibit a degree of "bunch- NERC will become increasingly reliant is the Global
ing", it is possible to encounter delays of up to a few
hours between position fixes.
In order to update one's position regularly it is
necessary to make a series of close estimates, based
on the ship's heading from the gyro compass and

• SATELLITE FIXES
- - ACTUAL COURSE
----- DEAD RECKONED COURSE

Fig. 1. Transit orbits. Fig. 2. The effect of 'Dead Reckoning' procedures.

 DEEP SEA NAVIGATION TECHNIQUES 5

Positioning System, or GPS. This consists of a con-

stellation of satellites, normally referred to as Space
Vehicles, or SV's, orbiting the earth every 12 hours,
associated ground stations and user equipment.
When completed, there will be 18 SV's in 6 differ-
ent orbital planes, inclined approximately 63° to the
equator and at an altitude of 20,000 km (Fig. 3). The
system will provide for three dimensional position,
horizontal and vertical velocity and extremely accu- Fig. 4. Radar transponder. Typical grid pattern survey.
rate time, on a continuous, worldwide basis. All SV's
were due to be launched by the end of the 1980's but RADAR TRANSPONDERS
the delay in the Space Shuttle programme has put
Although not used by NERC on a regular basis any
back full deployment by several years.
more this technique was employed for many years
The basic measurement made by the user's receiver
and much useful work was accomplished with it.
is that of the apparent propagation time of a timing
An active radar transponder is attached to the top
mark from the SV to the receiver antenna. By mak-
of a purpose made buoy which contains a sealed
ing this measurement from three SV's the user's
compartment to house the battery power supplies.
Latitude and Longitude can be determined. Further-
The buoy is tethered to an anchor weight by a length
more, by measuring the rate of change of range, the
of light steel wire and its surface position is deter-
user's speed and heading can be computed.
mined as accurately as possible by means of time
The satellite positions in space are measured to
series of satellite fixes. Surveys are then conducted
within a few metres and the satellite clocks are
whilst taking frequent readings of the range and
calibrated to a few nanoseconds so that computation
bearing of the transponder, using the ship's radar
of position on the surface of the earth is possible to
(Fig. 4). A series of buoys could give coverage of a
within a few metres.
larger area or the chance of greater precision, if
However, each satellite transmits two codes, the P,
needed.
or Precision code and the CA, or Coarse Acquisition
Unfortunately there are several practical limita-
code. To gain access to the full precision of the
tions which restrict the usefulness of the technique,
system a user would need a knowledge of the P code.
especially in deep water. The buoy remains subject to
Commercial users have access only to the CA code
movement from surface currents and winds, thus
and as a consequence the likely order of precision is
creating a degree of uncertainty with regard to its
in the region of 100 metres.
absolute position at any time (Fig. 5). In addition,
The advantages are that computation of position
the time taken to obtain enough satellite fixes to
will be continuous and without the cumulative errors
determine the buoy's position with sufficient accu-
associated with Dead Reckoning procedures. With
racy can become prohibitive.
the added facility of determining velocity and head-
ing it is obvious that GPS has a great potential as an
aid to marine science. -,
I
i
r-L,
I r-~
I -T:--------~------~--
L ..y t..-rJ
, I
\ I

-+-
\ I

--.J,-
\ /
\\ /
/
\ I
\ /

\ /
\ I

Fig. 5. Radar transponders. The effect of tidal and surface current

Fig. 3. GPS orbits (only 2 are shown for clarity). movement.
6 K. G. ROBERTSON

The practical limit to the range from a single buoy

is about 20 miles and is dependent on sea state x
conditions but nevertheless the technique allows al-
most continuous positioning and is still used for
some shallow water applications.
Fig. 7. Microwave ranging used to maintain separation of vessels.

MICROW AVE RANGING

With a normal operating range of some tens of commonly consist of a number of underwater acous-
kilometers this technique would not seem applicable tic transponders, a surface transducer to interrogate
to deep sea work, at first sight. However, geophysi- them and instruments to translate the data received
cists wi11 be aware that there are certain tasks which into positional information.
demand the use of two ships working together. Com- Having navigated the ship to the area of interest
monly one would be operating the acoustic energy by the use of conventional surface methods the first
sources while the other would be listening and step is to lay a pattern of transponders on, or close
recording with a multi-channel hydrophone system. to, the sea bed. When interrogated from the surface
When the two ships are traversing the same track, the transponders in range wi11 reply and the time
in the same direction, the Microwave Ranging sys- taken for this reply to be received will be a function
tem provides a constantly updated and accurate of the slant distance from the ship. As replies are
measure of their separation, thus allowing one to received from a number of transponders in the pat-
increase, or decrease its forward speed to maintain tern it is possible to compute the ship's position
the distance between them to within relatively fine relative to the sea bed instrument array.
limits (Fig. 6). This technique can be extended to cater for one or
When the ships are traversing the same track but more towed, or free swimming vehicles. For example,
in opposite directions, it is normal to maintain a if interrogation of sea bed transponders is conducted
steady speed and use Microwave Ranging to measure through a transponder mounted on the submerged
accurately their increasing, or decreasing separation vehicle then its position, relative to the array, can be
(Fig. 7). determined. A combination of surface and sub-
By this means NERC has frequently achieved a merged transducers allows one to track both the
precision of a few tens of metres over separation towed vehicle and the ship (Fig. 8).
ranges of several tens of kilometres. With the use of conventional surface navigation
aids the position of the sea bed transponder array
ACOUSTIC TRANSPONDERS
can be translated into geographic coordinates. All
such systems include calibration routines to achieve
All systems described so far can be grouped into the
this end if desired. Under some circumstances it is
category of radio navigation aids. That is, they oper-
possible to achieve relative positional accuracy of a
ate by transmitting to a receiver through the air, at
metre or better but as with all navigation systems,
high frequencies.
there are sources of possible error.
If, however, it is necessary to navigate a vehicle or
H is essential to establish a precise knowledge of
towed device beneath the surface of the ocean one
local variations in the velocity of propagation of
must look to a different technology, based on acous-
sound in water. Calculation of range from the mea-
tic communication at lower frequencies. The equip-
surement of transit time can be heavily dependent on
ment used for this type of application would
this parameter if inconsistencies are encountered.
Internal refraction in the water column can alter the
range over which reliable acoustic communication
x can be expected.

-- .J~ ~ Line of sight conditions must normally be estab-

lished for correct operation. In the presence of sea
Fig. 6. Microwave ranging used to maintain range of pair of bed projections it may be necessary to moor the
seismic survey vessels. transponders at distances above the ocean floor

MARl H381-Page 4
 DEEP SEA NAVIGATION TECHNIQUES 7

f
t
Fig. 8. Acoustic transponders. Acombination of acoustic transponders and satellite navigation allows tracking of the surface ship and sub
surface vehicle.

and hence they become more likely to be influenced and representations of the three-dimensional topo-
by currents. Thus distances relative to each other and graphic characteristics offer the scientist an opportu-
to surface coordinates have a degree of uncertainty. nity for positioning his sampling device with great
The accuracy to which one can translate positions precision.
relative to a transponder pattern into positions rela- The production of such detailed pictures will facil-
tive to geographic coordinates depends on the accu- itate very accurate selection of sites or survey tracks
racy of the radio navigation aids with which the ship
is equipped and how well they are operating at the
time of system calibration.
Acoustic transponders are not limited to use in the
manner described above. The techniques of Long I
Baseline, Short Baseline and Ultra Short Baseline I
I
I
surveying are well documented elsewhere. Applica-
tions are numerous but two areas which could be of I
I
interest to the marine geologist might be:-
The precise positioning of a corer within a sea
------ --.....--_-- ........
---
bed array (Fig. 9).
The tracking of arrays of seismic energy sources Fig. 9. The use of acoustic transponders to position a corer.
(Fig. 10).

SWATHE SOUNDERS

Swathe, or Multi-Beam sounders are discussed else-

where in this volume. Their capability to map, in
great detail, large areas of sea bed, coupled with Fig. 10. The use of acoustic transponders to monitor the towing
processing possibilities of constructing contour plots characteristics of near surface air gun or hydrophone arrays.
8 K. G. ROBERTSON

and if near real-time verification is obtainable this Conclusions

will establish a high level of confidence that one is
Many aids to navigation are available to the scien-
exactly where one wished to be.
tist at sea. This paper has concentrated on those
techniques which are most applicable to deep
The Role of Computers water work, where one is usually distant from the
localised hyperbolic systems and alternative Radar
As mentioned earlier, intelligent use of computing
targets.
can improve navigation by combining the outputs of
There is little doubt that the continued develop-
several devices and eradicating, or reducing, sources
ment of the GPS system will have the greatest
of errors. The ability to produce live track charts,
impact on the work of NERC for all types of
annotated with depth, position, gravity, magnetic
surveys. The combination of this high quality and
field, station number etc., reduces the need for post
continuous global positioning with the more local
cruise processing and allows for on-line decisions to
and specialised techniques of underwater acoustics ,
be made regarding the conduct of the surveyor
swathe bathymetry and microwave ranging offers a
sampling programme.
powerful package to the marine scientist.
By relating all measurements to a time reference
data can be correlated and assessed more readily. '
Modern Swathe Sounding and Sub-Bottom Profiling
Technology for Research Applications:
The Atlas Hydrosweep and Parasound Systems

1. A. GRANT and R. SCHREIBER

Krupp Atlas Elektronik Gmbh, Sebaldsbriicker HeerstrafJe 235, D2800, Bremen 44, FRG

(Received 27 April, 1989; accepted I September, 1989)

Key words: Swathe sounding, narrow-beam sub-bottom profiling, survey lines it is still not possible to guarantee full
acoustic surveying.
coverage of the seabed, and isolated features that lie
between the lines may be missed. Various ap-
Abstract. This chapter describes two separate but complementary proaches have been used in trying to solve this
research echosounder systems originally developed by Krupp At- problem including, with some success, echosounder
las Elektronik GmbH for the new German oceanographic research
systems such as the Atlas BOMA system using multi-
vessel Meteor.
The Hydrosweep is a Hydrographic wide-swathe sweep survey ple transducers mounted on booms, and the deploy-
echosounder for both shallow and deep water applications provid- ment of survey launches sailing on tracks parallel to
ing accurate bathymetric surveys and terrain-following navigation that of a mother ship (Stenborg, 1987).
capabilities.
The Parasound system is a hull-mounted dual channel paramet-
The 1970s saw the development of the first com-
ric narrow-beam deep sea survey and sub-bottom profiling echo- mercial swathe bathymetry systems (de Moustier,
sounder enabling particularly high vertical and horizontal 1988; Wentzell and Ziese, 1988). These were either of
resolution of seabed features.
the multibeam echosounder type, such as Sea Beam
(Farr, 1980) or the less accurate, towed interferomet-
ric sidescan type with swathe bathymetry capabili-
Swathe Sounding Systems
ties, typified in the USA by SeaMARC (Blackinton
The experimental work of Colladon and Sturm et at., 1986), and in the UK by Bathyscan (Cloet and
(1827) to determine the speed of sound in water, laid Edwards, 1986).
the foundations for the eventual development of the It has been shown that Sea Beam, with its compar-
echo sounder, which, a century later, had replaced atively narrow swathe width and restricted capabili-
the leadline as the traditional method of sampling ties in shallow water, suffers from a number of errors
water depth. The next 50 years saw advances in both resulting in incorrect depth determination, which
acoustic and electronic engineering which led to the may cause artifacts to be observed in the data col-
modern precision echosounders and the ability to lected (de Moustier and Kleinrock, 1986). Modern
acquire and evaluate depth data to a high order of oceanographic research increasingly requires depth
accuracy. data to be collected over large areas of the seabed
The conventional precision echosounder, however, with great precision and efficiency.
can only measure the water depth immediately below The multi beam system Hydrosweep is designed to
the survey vessel, and in order to achieve satisfactory meet this need, providing the geologist and hydro-
coverage of the seabed it is usually necessary to graphic surveyor with the capability of carrying out
survey sets of parallel sounding lines (Grant, 1985). accurate, wide-swathe surveys in water that may be
The number and spacing of these lines will depend as shallow as 10m or as deep as II 000 m.
upon a number of factors including the scale of the At the end of December 1987, the Atlas Hy-
survey and its purpose. Even with closely spaced drosweep on board R.V. Meteor was used to survey

Marine Geophysical Researches 12: 9-19, 1990.

© 1990 Kluwer Academic Publishers.
10 J. A. GRANT AND R. SCHREIBER

....
,.. .. ...
....1
....
""", '...

..
..
~ , ,,~
_ _ •• ROMANCHI! FRACTURE ZONE
"
~, ~ A.1 .. HVDROSWEEP
'I .- ' ,
... ... "' Htnlne'liUn.en

Fig. I. Part of the Romanche Fracture Zone, shown as a perspective representation of a three dimensional terrain model, generated during
post-processing at the Alfred Wegener Institute, Bremerhaven.

the Romanche Fracture Zone (Siedler et ai., 1987). The system accuracy. In multibeam systems with large
central area of this trench zone is about 10 km wide swathe angles there is an additional effect on the
with water depths of up to 7900 m, while the slopes divergence of the slant beams which may result in
at the sides of the trench have inclinations which can unacceptable errors in measurement, especially by
be greater than 45% (Fig. 1). Hydrosweep reached the refraction of the outermost beams. For this reason
bottom of the trench with its entire fan of sound beams, other swathe systems have had to make use of sound
a swathe width of more than 14 km (Schenke, 1988). velocity probes to overcome the problem. This is
time-consuming and therefore not cost-effective. Hy-
drosweep makes use of a unique, patented calibration
The Atlas Hydrosweep System
process which takes place while the vessel is underway
The Hydrosweep system is characterized by the very (Fig. 2).
wide area which is covered by the 59 pre-formed During the survey all vertical depth values are
beams (PFBs) transmitted in a fan from the hull- measured by the centre beam and stored. These
mounted transducers. The outer-beams may be radi- vertical depth determinations are, of course, not
ated at angles of 45° relative to the surface, thus subject to slant error. At regular intervals the fan of
providing the capability of achieving a total swathe beams is rotated through 90° to transmit along the
width equal to twice the water depth. ship's longitudinal axis and the resulting profile is
As with conventional vertical echosounders, varia- measured and stored.
tions of sound velocity in the water column influence In an iterative process the two profiles are com-
the travel time of the acoustic pulse and therefore the pared and a correction factor is applied to produce a
SWATHE SOUNDING AND SUB-BOTTOM PROFILING 11

berthy","';c
p<ofil.

Fig. 2. Atlas Hydrosweep calibration procedure.

Direction 01 travel

Trnnsducer arrays
inshrp'. bOHO~

TOUl$mitter
(S ..... eymode)
Receiver
(Calibration mode)

CalIbration mode

Fig. 3. Atlas Hydrosweep Transducer beam patterns.

12 J. A. GRANT AND R. SCHREIBER

'best fit' result. This calibration factor includes, system switches over to the calibration mode, the
among other parameters, the average sound velocity functions of the transducers are exchanged, and the
in the water column, and may be applied as a fan beam rotates through 90°.
correction to both the slant and vertical beam travel In order to maintain shallow and deep water
time in the survey mode. This process is repeated capabilities the appropriate transmission mode is
automatically at regular intervals as the vessel automatically selected by the system. In depths be-
travels along its track (Schreiber and Schenke, tween 10m and 1000 m, short pulse lengths and low
1989). energy emissions are used.
Analysis of data from system trials have shown
that the residual depth error may be reduced to depth transmission
better than 0.4% of water depth. In addition, the m mode beam width
Hydrosweep calibration process makes it possible to
10- 100 omnidirectional 90° x 4S
sail from shallow to deep water without the need to
100-1000 omnidirectional 90° x 1.9°
deploy sound-velocity probes, and therefore without
losing valuable operation time.
Hydrosweep uses two identical, solid-state piezo- In water depths greater than 1000 m, longer pulse
electric transducer arrays, each array consisting of lengths and higher energy are emitted using a Rota-
replaceable modules (Fig. 3). One of the arrays is tional Directional Transmission mode (ROT). The
installed along the longitudinal fore and aft axis of transmission is made in three steps, one centred on
the vessel and during the survey mode is used as the the vertical and the other at the 36° points with time
transmitter, while the second array is set at right separation between the pulses being kept to a mini-
angles to the first and acts as the receiver. When the
,
mum to ensure that the optimum sounding rate is

'IS J2 18 30 :8 28 -18 21 -18 22

o -16

... 0 '18

omao,cr - 0. I9 ~ 0'20
~A_EI .~
"0111.0 Ci.OO£T1C SIS E 1972

HYDROSVEEP
RO':MICHE RINNE
02.01.86
HYDROMAP
STRr: FENAUSIIERTUt'G
o -22

Fig. 4. Contour map of the Romanche Fracture Zone derived from Hydrosweep surveys. Overlapping parts of two separate surveys show
excellent correspondence.
 SWATHE SOUNDING AND SUB-BOTTOM PROFILING 13

maintained. Using RDT in this way ensures that the specific test measurements were made in the transi-
system achieves uniform insonification across the tion area between the deep sea floor and the steep
swathe. The pulse length will vary according to sides of the trench (Fig. 4.) and no 'omega-effect'
depth, and is separately controlled on the vertical was observed (Tyce, 1987; Schenke, 1988).
beam and on the outer sectors. The system is operated at a frequency of 15.5 kHz,
The so-called 'omega-effect' was experienced with which was chosen to achieve two objectives. On the
earlier multi-beam systems when descending a steep one hand, this frequency is low enough to achieve
slope, (de Moustier and Kleinrock, 1986). In such a the desired depth range, giving Hydrosweep with its
situation the intensity of an echo generated by side highly efficient narrow beam transducers the capabil-
lobes from further back up the slope may be com- ity to reach the full ocean depths in excess of
parable to the vertically reflected return from directly 10 000 m, while the low noise signal processing tech-
beneath the ship. The tracking gate within the system niques provide a precise determination of the depths
may then fail to follow the actual bottom, but will from the 59 return signals.
lock on to the false shallow depth produced by the On the other hand, the frequency is high enough
side lobe echo and lag behind the changing slope. to secure good vertical resolution, even in shallow
The false echo can produce a contour which in water, and the signal is reflected from the true bot-
classic cases resembles the Greek letter omega (n). tom, rather than penetrating the sediment layers to
The transmission pattern of Hydrosweep has been the sub-bottom surfaces. The ch~n frequency is
modified to neutralize this effect and during the also high enough to achieve a good "itafeigh" figure,
Meteor survey of the Romanche Fracture Zone, which determines the frequency at which a bottom of

Fig. 5. Atlas Hydrosweep real-time colour graduated contour map. Vessel position is shown by the cross lines at the top of the display.
The ships planned track is superimposed on the display and all other parameters such as depth, range across the display as well as the
incrementation of the contours is shown.
14 J. A. GRANT AND R. SCHREIBER

a certain roughness is seen as a diffuse reflector energy, which, by definition is the true bottom
(Urick, 1983). This will provide a better signal return (Schreiber and Schenke, 1989).
at the outermost slant beams. It should be noted that the uncertainty in the
The return signal may be affected by a number of horizontal position of a survey vessel currently exceeds
disturbing influences which may make discrimination the uncertainty in depth determination. In order to fully
of the true bottom echo difficult, and side lobe echoes utilize the achievable levels of operational accuracy of
may fall into the receiver time gates, falsifying the a survey system a pre-requisite is that high-accuracy
energy centre created by the main echo. Within the position fixing methods are employed together with
Hydrosweep System microprocessors exclusively compensation for heave, roll and pitch motion on the
allocated to echo discrimination carry out a routine vessel.
which is repeated several times. An 'adaptive window' The Hydrosweep system is controlled, and data
technique accurately determines the precise centre of collection monitored, from an operator's console.

Fig. 6. Atlas Hydrosweep: profile and performance monitor display. The display is "steady", and its contents are overwritten anew with
each receiving cycle.
Left
Navigation parameters.
Middle
Upper part cross profile display (during calibration: longitudinal profile):
(I) For each beam, the measured depth of water is displayed with a vertical linear scale.
(2) The "search window" for the depth digitizer is re-positioned automatically.
(3) The lateral boundary of the beams of which the data are to be recorded can be selected manually.
(4) The number of the beams are marked on this line.
Lower part for each of eight amplifier channels of the side beams, a bar graph is shown for the "attenuation", i.e. the relative retlectivity
of the bottom.
Right
Annotation of dynamic parameters of the ship at the instant of transmission.
 SWATHE SOUNDING AND SUB-BOTTOM PROFILING 15

Dedicated function keys are used to call up or enter bottom profiler and represents a major new develop-
the required operation parameters on a monochrome ment in advanced hydroacoustic engineering. The
alphanumeric display. Rapid system familiarization techniques employed make it possible to achieve
is made possible by the use of menu dialogue tech- very high vertical and horizontal resolution, facilitat-
niques, facilitating ease of operation even when ing the detection of fine layers of sediment, while
changing from one function to the next. giving excellent bottom penetration from a hull-
Data are displayed in real-time on a second high- mounted transducer (Schreiber et aI., 1988).
resolution VDU, either in the form of a graduated The parametric principle has been established for
colour contour map (Fig. 5) or as a cross-section some years and makes use of the non-linear acoustic
profile and performance monitor display (Fig. 6). properties of the water column (Berktay, 1965).
Navigation during survey operations may be en- When two signals at adjacent frequencies are radi-
hanced by being able to follow seabed contours, any ated simultaneously into the water column at very
deviation from pre-planned tracks being clearly high energy levels, they generate a secondary fre-
shown. All data are recorded on digital tape, and quency which is equal to the difference between the
may also be reproduced on isoline charts for immedi- two primary frequencies. The secondary frequency is
ate on-board evaluation. only generated in the main part of the beam, where
Other sophisticated routines allow system testing the highest energy levels occur. By this means a low
in support of fault location, quality-control checks frequency transmission, virtually free of side lobes,
and total replay of survey data. This playback and with a very narrow beam is produced (Fig. 7).
simulation function also allows operator training to
be carried out on-board the vessel, even in port
(Tyce, 1987).

The Atlas Parasound System

The importance of being able to investigate the

sedimentary structure of the seabed, whether for pure
geological research or commercial offshore engineer-
ing purposes, has led to the development of various
types of low-frequency echosounders which have been
used to good effect as sub-bottom profilers.
These systems typically have used frequencies of
around 3.5 kHz, the actual frequency depending _ ...._ _ 5·
upon the depth and nature of the bottom. While they 2,5 - 5.5 kHz
and 18 or 33 kHz
can penetrate quite deeply, they do nevertheless
suffer from a number of inherent disadvantages. The
use of such low frequencies implies the employment
of large transducers and wide beam-widths. This in
turn leads to the generation of side echoes which
interfere with the desired sediment echo, often ren-
dering it virtually useless.
To overcome this problem, transducers have been
mounted in deep-towed vehicles. However, this re-
quires the use of special cables, winches, hardware
and, of course, vessel operation at low speeds. The
result is degraded data at high mobilization and
operation costs.
The Atlas Parasound System combines a narrow Fig. 7. Atlas Parasound-narrow beam echosounder and sub-
beam, deep sea survey echosounder with a sub- bottom pro/Her.
 16 1. A. GRANT AND R. SCHREIBER

1 • • • • • • • • 1 • • • • ' • • • • • • • , • • • • 11 • • •

...-·. o
18 kHz
- 20

- 40

.,..... - 60

: - 80
..- .....
A
• -··
. . . . . . . . . . . . . . . . . . . . . 4o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "
~

........................... 100m

it t,.i,.
o
3.5 kHz
(18 + 21.5 kHz) - 20

B
-100m
Fig. 8. The acoustic parametric effect. Comparison of seabed records made by R. V. 'Meteor' in the Kattegat. Echogram A at 18 kHz and
B 3.5 kHz (18 kHz and 21.5 kHz transmitted simultaneously).

"£ •••••••• , ••••••••• ,.,', •• , •• t • • • • • , • • • , •• I • • • • • ' . t.t. i ••••••• t •••• I •••••••••••••••••• "'" t I
Z=Om

~ 110m .

........................................ Z = 50 m
o appr. 1.8 km

Fig. 9. Atlas Parasound record of giant active sandwaves with an elevation of up to 10 m at a water depth of approximately 30 m. (English
Channel).
 SWATHE SOUNDING AND SUB-BOTTOM PROFILING 17
Z = 50 m
••••••••••••••• ••••••• I ' •• i • • • • • • • • • • • • • • • • • """ I I t •••••••••••••• ··, •

Z = 100 m
I I
o 3.0 6_0 km

Fig. 10. Atlas Parasound record of alternating layers of slightly distorted hard and soft Mesozoic layers, with sand covering at water depth
of 60- 70 m (English Channel).

SBP ________~----~W~A~T~E~R SURFACE

NBS_
20Log 2r+ 2a.r

IINBS

a 19 . 5k Hz ~ 0 .00 3dBIm
a' 3.5kHz "'- 0 .7dBlm

Fig. II. Atlas Parasound-TVC behaviour in water and sediment. NBS - Narrow beam signal at the primary frequency acts as a pilot
pulse and detects the true bottom. SBP - Sub-bottom profiler signal at the secondary frequency. r - Range in meters. a - Decibels km-I.
18 J. A. GRANT AND R. SCHREIBER

Fig. 12. Atlas Parasound-high resolution colour display. The main display on the left shows the reflected sediment layers at an
appropriate scale (50 m in this case). The right-hand display is an expansion of the instantaneous signal.

In principle, any of the ultrasonic frequencies may Horizontal resolution is optimized at depths
be used as primary frequencies. Parasound is oper- greater than 2000 m by operating Parasound in an
ated with one fixed primary frequency of 18 kHz, automatic multi-pulse mode. A pilot pulse is first-
which allows the system to reach full ocean depths, transmitted to determine the actual depth of the
with a second primary frequency which is operator seabed. The system processor then determines how
selectable to generate a secondary frequency between many pulses will fit into the travel time at that depth
2.5 kHz and 5.5 kHz. Figure 8 shows a comparison and emits a suitable train of pulses. This sequence is
between seabed records made at frequencies of then repeated and corrected with the measured depth
18 kHz and parametric derived 3.5 kHz. automatically. This approach greatly enhances the
The vertical resolution of a sub-bottom profiling resolution of small-scale features such as sedimentary
system is determined by the pulse length of the bed forms (Figs 9 and 10).
radiated signal, which in turn is a function of the In order to achieve optimum penetration and
transmitted primary frequency. The comparatively resolution of the sub-bottom layer, account must be
high primary frequency used by Parasound makes taken of differences between the behaviour of sound
it possible to radiate a considerably shorter pulse in water and in sediment. To set and maintain the
length than conventional sub-bottom profilers. This bottom TVC (Time Volume Control) independently
together with the ability to generate the low sec- of the water column TVC with changing water
ondary frequency with precise wave periods, results depth has always been difficult with earlier equip-
in greatly improved vertical resolution of up to ment. Parasound makes use of its relatively high
30 cm, even in deep sea sediments (Fig. 8). primary frequency to accurately define the true
 SWATHE SOUNDING AND SUB-BOTTOM PROFILING 19

seabed and therefore the start depth for the bottom References
TVC The correct setting is then automatically main- Berktay, H. 0., 1965, Various Papers on Non-linear Acoustics, J.
tained as the depth varies. Sound Vib.
A high-definition colour display unit is incorpo- Blackinton, J. G., Williams, J. F., Hills, D., and Kossalos, 1. G.,
1986, First Results from a Combination Sidescan and Seafloor
rated within the system to permit monitoring of the Mapping System, Proc. Offshore Tech. Conf., OTC, 4478.
operation. The display screen can be split to allow a Cloet, R. L. and Edwards, C. R., 1986, The Bathymetric Swathe
general overview of the bottom and a high-resolution Sounding System. The Hydrographic Journal 40.
Colladon, J. D., and Sturm, F. K., 1827, The Compression of
display of the sediment layers. The display may be
Liquids (in French). Ann. Chim. Phys. Series 2, 36, Part IV.
frozen and photographed if required (Fig. 12). Speed of Sound in Liquids, 236-257.
A high-precision echogram with depth digitization de Moustier, c., 1988, State of the Art in Swathe Bathymetry
provides a further record. It is possible to make Survey Systems Int. Hydro. Rev. 65, 29-38.
de Moustier, c., and Kleinrock, M. c., 1986, Bathymetric Arti-
simultaneous recordings of the bottom profile and facts in Seabeam Data, J. Geophys. Res. 91, 3407-3434.
sediment layers with annotation of the selected oper- Farr, H. K., 1980, Multibeam Bathymetric Sonar Seabeam and
ating parameters. Hydrochart, Marine Geodesy 4,77-93.
Grant, J. A., 1985, Filtering, Digitisation, Logging and Presenta-
tion of Depth. Proc. 3rd Hydro. Symp. Min of Defence.
Schenke, H. W., 1988, Unpublished Correspondence on Meteor
Conclusions Expedition M6 4, Alfred Wegener Inst. Bremerhaven, West
Germany.
The Atlas Hydrosweep has been proven to offer Schenke, H. W., and Ulrich, I., 1986, Flachenhafte Kartierung des
significantly improved performance levels over other Meeresbodens, Geowissenschaffen in unsere Zeit 4.
Schreiber, R. and Schenke, H. W., 1989, Atlas Hydrosweep,
swathe-mapping systems, not only with respect to Efficient Hydrographic Surveying of EEZ with New Multibeam
time saved, but also in the quality of data acquired. Echosounder Technology for Shallow and Deep Water, Pro-
The system accuracy is improved despite the ceedings EEZ Resources: Technology Assessment Conference,
JOTC Hawaii, U.S.A., Jan. 1989, pp. 3-16 to 3-30.
wider swathe coverage due to the unique self-calibra-
Schreiber, R., Wentzell, H. F., and Ziese, R., 1988, Efficient
tion process, which also allows depth data to be Seafloor Mapping and Sub-Bottom Profiling with Atlas Hy-
collected without interrupting other ship routines. drosweep and Atlas Parasound, in Proceedings of Techno Ocean
Indeed, terrain following navigation capabilities are '88 Conference, Kobe, Japan, Nov. 1988, pp. 103-110.
Siedler, G. A., Kuhl, A., and Zenk, W., 1987, The Madeira Mode
enhanced. Water, J. Phys. Oceanogr. (in press).
The Atlas Parasound has proved its capability for Stenborg, E., 1987, Swedish Parallel Sounding Method. Int. Hyd.
carrying out high-resolution sediment surveys from Rev. 64 (1), Monaco.
Tyee, R. c., 1987, Workshop Report on Hydrosweep Data Post-
hull-mounted tra~ducers with horizontal resolution processing, Univ. Rhode Island.
comparable tQ~that of towed systems, and much Urick, R. J., 1983, Principle of Underwater Sound, McGraw Hill,
better vertical resolution. The narrow beam function Inc.
Wentzell, H. F. and Ziese, R., 1988, New Echosounding Methods
inherent in the parametric principle has provided the
for Shallow Water and Deep Sea Surveying, Advances in UW
additional capability of high-precision bathymetric Tech. Vol. 16 Oceanology, pp. 25-41. Graham & Trotman,
surveys in both deep and shallow water. London.
GLORIA Image Processing: The State of the Art

R. C. SEARLP
Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Godalming, Surrey, GUS 5UB, UK

T. P. LE BAS
Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Godalming, Surrey, GUS 5UB, UK

N. C. MITCHELL
Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Godalming, Surrey, GUS 5UB, UK
Department of Earth Sciences, University of Oxford, Parks Road, Oxford OX 1 3PR, UK

M. L. SOMERS
Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Godalming, Surrey, GUS 5UB, UK

L. M. PARSON
Institute of Oceanographic Sciences, Deacon Laboratory, Wormley, Godalming, Surrey, GUS 5UB, UK

and

PH. PATRIAT
Institute de Physique du Globe de Paris, Tour 14-15, 4 place Jussieu, 75230, Paris, France

(Received 27 April, 1989; accepted I September, 1989)

Key words: GLORIA, sidescan sonar, image processing. years, during which time it has evolved from a
"test-bed" instrument to a highly sophisticated and
Abstract. This chapter presents a summary of the image-
reliable survey tool (Rusby, 1970; Somers, 1970,
processing techniques being used at present in the Institute of 1973; Rusby and Somers, 1977; Somers et ai., 1978;
Oceanographic Sciences Deacon Laboratory's GLORIA long- Laughton, 1981; EEZ-scan 84 Scientific Party, 1988).
range sidescan sonar system. It begins with a brief review of the
development of GLORIA, and then describes in outline the
A major part of this evolution has been concerned
present shipboard data acquisition, recording and replay system, with improving the image quality and presentation,
including simple image-processing techniques that can be used and in this chapter we shall briefly outline those
on-board ship. Next, a detailed form of the sonar equation is
developments, and then describe the state of the art
developed, and this is evaluated factor-by-factor, to demonstrate
the effects of beam directivity, refraction and water depth on the at IOSDL in 1988.
form of intensity variation to be expected in the final image. The first images had low signal-to-noise ratios,
Finally, we discuss recent developments in shore-based image-pro- were output on wet-paper recorders, and suffered
cessing. These include the development of improved radiometric
corrections to normalize range-dependent intensity variations, re-
from severe distortions of both geometry and signal
covery of true backscattering levels and estimation of backscatter- intensity. Slant-range, not horizontal range, was dis-
ing coefficients, and combination of GLORIA with other data sets played (so the horizontal range scale was non-
into single, colour digital images. As an example of the last
process we show a digital mosaic of sonar data from the South-
linear), no corrections were made for varying
west Indian Ridge, coloured as a function of depth derived from tow-speed, and even the average along-track and
Sea Beam data in the same area. cross-track scales were different. The recorders used
were unable to cope with the > 35 dB dynamic range
Introduction of the signal, so strong signals were saturated and
weak ones were lost.
The GLORIA long-range sidescan sonar system
Gradually these defects were remedied, with all
(Figs I and 2) has now been used for about twenty
improvements to the GLORIA system subsequent to
• Present address: Department of Geological Sciences, University 1971 being ultimately enabled by the introduction
of Durham, Durham, DH3 ILE, UK then of a linear correlation processor. This involves

Marine Geophysical Researches 12: 21-39, 1990.

© 1990 Kluwer Academic Publishers.
22 R. C. SEARLE ET AL.

~.~ _________________________ 45km--------------------~~~

Fig. I. Schematic diagram of the GLORIA tow-vehicle and its sound swath, with principle dimensions.

the use of a long transmission pulse (up to 4 s) mitter, the receiver signal-to-noise ratio depends
coded with a linear FM sweep. The correlator com- upon the pulse energy, while the correlator restores
presses the energy of the long pulse into a short the range resolution. The effect in GLORIA was
interval r that is inversely proportional to the FM equivalent to raising the transmitter power of a short
bandwidth. In a system with a power-limited trans- pulse system from 10 kW to 4 MW. This enabled

I TRANSMIT/REC£Ivt:
I REPLAY I
-------------------------

NAVlCATlON
+
8AlliYWE1llY

II-TRACK
TAPE
(AIICK\I[)

------- - --------- ----- ---

----- - ---------- --- ------

p~
AWPS

--- ---- --- --------------~

Fig. 2. Simplified block diagram of the present GLORIA shipboard data acquisition and processing system.
 GLORIA IMAGE PROCESSING 23

adequate SIN ratios to be obtained over long ranges, and recording system was replaced by a digital sys-
gave the reduction in vertical array aperture needed tem, opening up a wide range of further image
to obtain more even illumination of the swath, and correction, enhancement and processing possibilities.
thus allowed the introduction of a smaller, more The rest of this chapter will describe our progress in
robust vehicle configuration in GLORIA Mark II. these areas to date.
Signals were then replayed from tape through an
analogue facsimile recorder to produce good-quality
Current Data Acquisition and Processing Systems
photographic prints. An 'anamorphic camera' was
developed to correct these for varying tow-speeds. Figure 2 shows a schematic diagram of the present
This applied a variable along-track stretch by repho- shipboard signal-processing system. The towed sonar
tographing a moving copy of the original print vehicle carries two rectangular transducer arrays
through a slit on to a moving film, with the ratio of (one port, one starboard), with their axes directed
print speed to film speed set by the ship's speed over 20° below the horizontal. Each array consists of
the ground. At the same time a factor was applied to 2 x 30 transducers 0.17 m in diameter and spaced
set the average horizontal range scale equal to the 0.17 m apart, separately wired into six horizontal
along-track distance scale. To cope with the large sections of 5 x 2 transducers. Each section has inde-
dynamic range, data were recorded (on analogue pendent transmit and receive signals.
FM tape recorders) in two ways: with automatic gain Figure 3 shows the computed beam pattern for
control (AGC), and so-called 'fixed gain' (FG). In this arrangement, which produces a main beam
fact time-varied gain (TVG) was applied to both to 2.7° x 35° at half power points. The horizontal beam
provide a rough compensation for spreading and pattern has the sidelobe arrangement of any uni-
attenuation losses. Thereafter, the FG signal was formly illuminated line. The effect of these sidelobes
recorded directly, and preserved the true relative is to broaden the beam somewhat, and this can be
intensity of strong signals, but lost weak ones, seen especially on distant discrete targets. The verti-
whereas the AGe channel, in which the recorder cal beam pattern is the non-ideal compromise be-
gain was automatically reduced for stronger signals, tween several conflicting demands:
preserved signals of all intensities but on a non- - for mechanical engineering reasons a minimum
linear, and unrecoverable, intensity scale. vehicle diameter was needed;
This processing system was introduced in the early - the peak of the main vertical lobe should ideally
1970s, and within its limitations produced good- be about 5° -1 0° below the horizontal;
quality records throughout that decade. The major - the sensitivity should reduce uniformly to a low
limitations were the inability to correct for slant- value at the nadir;
range distortion and the continued need for AGe - and the sensitivity should reduce sharply above
recording. There was a further undesirable effect of the horizontal.
the AGe, associated with the correlator system. The low sensitivity in the nadir is required to
AGe was applied before correlation, and the non- reduce the occurrence of second and even third
linearity produced in this way caused image distor- bottom echoes over flat terrain, which were so un-
tions, of which the most noticeable was the tendency welcome a feature of the GLORIA I records. In fact
for a strong echo to suppress the reverberation just the pattern has a small null 20° or so off the nadir,
before and just after itself, producing erroneous and second bottom echoes are still occasionally a
shadows. Even so, in the early 1980s we were begin- problem with GLORIA II.
ning to apply a slant range correction to digitized The transmitted signal is a 100 Hz linear FM
tracings of interpreted features (e.g. Searle and Hey, pulse, usually 2-s long repeated every 30 s, although
1983), and improved tape recording technology, to- other pulse repetition periods (PRPs) and pulse
gether with the replacement of the earlier analogue lengths are available. Separate carrier frequencies of
signal correlators by digital ones in 1982 and intro- 6.2875 kHz (starboard) and 6.7625 kHz (port) are
duction of an improved TVG law on EPROM in used to reduce cross-talk to a negligible level. Nor-
1987, improved the dynamic range of the FG records mally only three sections per side are used for trans-
considerably. In 1980 the analogue signal processing mission, so the transmit beam is 5° wide, allowing
24 R. C. SEARLE ET AL.

, ' ,,
,,,
-, ,

,
I

1.0
1.0
PORT
STBD

t \
20'

\ --- --- , 35 '

20'
t

--- ---
FIRST
SIDE LOBE
FIRST
SIDE LOBE
6.3kHz
I
6.8kHz
MAIN LOBE
6 .3kHz
MAIN LOBE
6.8kHz
-----

FORE

27'
53 t
•
4 2 0
SINGLE FULL
SECnON ARRAY

AFT

Fig. 3. Power directivity diagrams for the GLORIA array. Top: vertical section normal to axis of vehicle. Bottom: horizontal section; right
half shows pattern of whole array (six sections), left half that of a single section to the same scale. Main lobe and first five side lobes are
numbered 0 to 5; each side lobe is one of a symmetric pair, of which only one (alternating fore and aft) is shown for clarity.

good insonification even in the presence of vehicle for a 30-s PRP. At 8 knots (4 ms -I) the near-ship
yaw. All sections are used for reception, and after along-track resolution is 120 m for the same PRP,
preamplification have TVG applied using the law although this resolution is degraded by the finite
beam width to about 900 m at maximum range. The
T= -3010gR -8.2 x 1O - 4 R (1)
data are then recorded on digital tape cartridges with
to match spreading losses and attenuation in the a sampling rate set to ensure 992 pixels per scan (496
water column (see equations (4) and (7), next sec- per side) and simultaneously displayed on a paper
tion), where T is gain in dB and R is slant-range in line scan recorder monitor. Thus for a 30 s PRP the
m. This assumes the attenuation coefficient IY. is pixel size is 60 ms or 45 m. The remaining 32 bytes of
0.82 dB km -1. The signals from the separate sections the 1024-byte record are used as headers to record
are then phased to steer the beam along the normal other pertinent information (see Table I). The output
to the long-period mean heading of the vehicle. This from the correlators is a 12-bit signal, but prior to
electronic beam steering is effective for yaw up to recording this is compressed to 8 bits using the "Bell
about ± 5°. Higher yaw may be experienced in sea (MU) 255 law"
states of roughly 6 and over, and can produce muted
y = C In( 1 + 255x /4096), (2)
lines ('dropouts') in the records.
After beam formation, the port and starboard where x and yare the input and output signals,
signals are correlated and digitized into 496 samples respectively, and C is a constant. Each cartridge
per side. (In fact the digitization rate is 500 samples normally holds two data files (or "passes" in our
per side, but for technical reasons the last 4 samples jargon), and for a 30-s PRP, each pass will contain
are discarded, causing no real loss since their infor- six hours' worth of data in 720 scans.
mation content is heavily eclipsed by the next trans- Individual scans to the LSR monitor are replicated
mission pulse). This yields a range resolution of 45 m so that the record is approximately isometric for a
 GLORIA IMAGE PROCESSING 25

TABLE I

Recording format used for GLORIA data

Byte Word in Contents

number IBM-AT

I I Low Pass number (binary, high byte)

2 I High Pass number (binary, low byte)
3 2L Scan number (binary, high byte)
4 2H Scan number (binary, low byte)
5 3L Flag for hour mark
6 3H Slant-range correction code
7 4L Binary zero
8 4H Pulse repetition period, seconds (ASCII)
9-11 5L-6L Vehicle heading, degrees (3 ASCII characters)
12-13 6H-7L Year (2 ASCII characters)
14-15 7H-8L Hexadecimal FF (edge mark)
16 8H Sonar sample I, far range port (binary)

512 256L Sonar sample 496, near range port

513 256H Sonar sample 497, near range starboard

1009 505L Sonar sample 992, far range starboard

1010-1011 505H-506L FF (edge mark)
1012-1014 506H-507H Julian day (3 ASCII characters)
1015-1016 508L-508H Hours (2 ASCII characters)
1017-1018 509L-509H Minutes (2 ASCII characters)
1019-1020 51OL-510H Seconds (2 ASCII characters)
1021-1022 511L-511H Checksum
1023-1024 512L-512H Unused

ship speed of about 8 knots, and a rough slant-range system. The anamorphic correction is achieved by
correction is applied via a look-up table using depths computing the number of times an individual scan
entered manually by the watchkeeper. must be replicated to achieve the desired "stretch"
At intervals throughout a cruise (usually every (since this is in general not an integer, slight over-
24 h), the digital data are processed further. The and under-corrections alternate). Slant-range correc-
reason for using cartridges will now be clear, since tion is achieved by computing the horizontal range
with a data rate of 3 Mbyte per 24 hours, a 2400-foot for a given slant-range, assuming a plane horizontal
9-track tape would hold over four days' worth of sea floor whose depth is that below the ship, and
data, whereas each track of a cartridge holds a moving the contents of the recorded pixel to the
convenient six hours' worth, leading to a much more appropriate new pixel. (Note that this implies some
convenient shipboard replay schedule. The cartridges replication of pixels, because the real time data and
are downloaded to disc on an IBM AT personal the slant-range corrected data have the same number
computer, and also transferred to standard 9-track of pixels, but the direct water path has been removed
magnetic tape as an archive medium. The cartridges from the latter).
are wiped clean and reused once the 9-track tapes are We can also carry out further processing at this
safely duplicated and at separate locations ashore, stage: for example, filtering to remove line drop-outs
but enough are available on the ship to hold 100 using the high-passjlow-pass method of Chavez
days' worth of data. (1986), application of radiometric corrections (along
The AT is then used to apply the anamorphic and the lines to be outlined in a later section of this
slant-range corrections in a precise way, using navi- paper) to produce more uniform 'illumination'
gation and depth data supplied (at present by tape across the image, and reversal of image polarity
transfer) from the ship's normal computer logging (white on black or black on white presentation).
26 R. C. SEARLE ET AL.

The processed files are then written to 35 mm film (5)

on a purpose-built laser film-writer. The film is devel-
where Pb is the backscattered power and A is the
oped on board, and photographic prints are made to
effective insonified area; the latter can be approxi-
the desired survey scale (usually between 1:500000
mated by
and 1:250000; 1:375000 has proved to be a particu-
larly popular and appropriate scale). If a 24-hr data A = (R(}b . cr . sec fJ) /2, (6)
batch ends in the morning, the complete processing where R is the range of the backscattering point, (}b
cycle, including production of photographic prints, is the - 3 dB horizontal beam width, r is the width
can be completed by the same evening, so that no of the function produced by passing the transmitted
data are delayed longer than 36 hr from recording to FM pulse through the receiver correlator, and
arrival on the plotting table. fJ(R, ¢) is the grazing angle of the ray incident on the
Further processing can be carried out ashore. At bottom (in general not equal to ¢ because of refrac-
present this is done at IOSDL, Wormley, using the tion in the water column).
United States Geological Survey's MIPS (Mini Im- Combining (4), (5), and (6), the backscattered
age Processing System) software (Chavez, 1986) run power at the seabed is
on a PDP 11/34 computer with Grinnell monitor and
interface (In fact the computer was upgraded to a Pb = PoD· b . exp( - r:xR) . Sb(}b . cr sec fJ /8nR.
Micro Vax II while this paper was in press). How- Allowing for further attenuation and spherical
ever, processing is not hardware-dependent, and one spreading on the return path, the intensity incident
of us (NCM) has done a considerable amount on a on the receiver array is
Sun workstation at Oxford University.
Ib = PoD· b . exp( -2rxR) . Sb(}b· cr sec fJ/32n 2 R 3,
and taking account of the receiver directivity, the
Acoustic Considerations
received acoustic power is
Before describing in detail our progress in shore-
Pr = PoD 2b 2 • exp( -2r:xR) . Sb(}b· cr sec fJ/32nR3.
based processing, it is instructive to consider quanti-
tatively the effects on the GLORIA images of both This produces a voltage V", Pr!, and applying TVG
the sonar system itself and the propagation of sound (T(R), Equation 1) and other constant receiver gains
through the sea. (G), we have
We use spherical coordinates where r is the range 1
V = (POSb(}b . cr sec fJ)2DbTG
from the sonar array, ¢ is the angle in the vertical
plane normal to the axis of the array, and () is the . exp( -r:xR)/4 . (2n)!R3, (7)
angle in the horizontal plane through the array. The In evaluating expression (7) we must remember
acoustic intensity I is the incident power P per unit that for a given horizontal range both the grazing
area A and is related to r.m.s. pressure variations p angle fJ and the initial ray angle at source (and hence
by D) will depend on R, the slant range measured along
(3) the acoustic ray. Moreover, the rays in general will
follow curved lines because sound velocity varies
where p is the density and c the sound velocity of with depth, giving rise to refraction. Moreover, this
seawater. Since the GLORIA transducers are linear refraction may cause rays to converge or diverge
in pressure, their output voltage should be propor- more than expected for spherical spreading in a
tional to the square root of the acoustic intensity. constant velocity ocean, thus changing the incident
At some distance r from a source of sound of sound intensity. We define
power Po, directivity factor D and directivity pattern
b«(}, ¢), the incident sound intensity will be F = Maspherical) /Ii (spherical) (8)

Ii(r, (), ¢) = PoD· b . exp( -r:xr)/4r 2 • (4) as the "refraction factor". So finally we have
1
V= (POSb(}b . cr . sec fJ)2DbTGF
The dimensionless backscattering coefficient Sb is
defined as . exp( -r:xR)/4 . (2n)!R3
 GLORIA IMAGE PROCESSING 27

or, explicitly including T (Equation 1): beam and first sidelobe) occurs closer to the ship and
is narrower in shallow water than in deep.
V = (POSb(}b . cr . sec p)!DbGF/4 . (2n)!. (9)
The loss of signal in this null is significant, especially
In the following sections we illustrate some of the
in deep water and, as we shall describe later, we have
features of this equation.
put considerable effort into ameliorating this effect in
our image processing.
RAY TRACING

In order to examine the effects represented in Equa- REFRACTION FACTOR

tion (9), we need to evaluate them for appropriate
To evaluate the refraction factor F (8), we note that
values of the slant range R. We did this by using a
the incident intensity at the sea floor is inversely
standard ray-tracing program to compute where
proportional to the spacing of rays there, and com-
rays emitted by GLORIA, at a typical tow depth,
pare this with the spacing in a uniform velocity
would strike the sea floor. Figure 4a, b shows two
ocean (i.e. straight rays from the source). The result-
such plots for typical operating conditions and two
ing factor is illustrated in Fig. 4e, f. In the shallow
different water depths. On these plots, which have
water case (Fig. 4e) this factor increases gradually
equal vertical and horizontal scales, the ray bending
with increasing range, and beyond about 10 km
due to refraction can clearly be seen. Note too that
climbs steeply to a value of about 2 at maximum
the rays are tending to converge or bunch up near
range. The cause is the tight bunching of rays near
the upper edge of the beam. The shallowest ray
the upper edge of the beam (Fig. 4b). The effect is
shown here is emitted at 10° above the horizontal;
critically dependent on the slope of the thermocline,
rays taking off at higher angles are reflected down-
and Fig. 5 shows the effect of increasing this slope in
ward by the sea surface even more steeply (or scat-
the region of the sound source, while keeping the
tered by it in rough weather), so the effective top
other parameters the same. This produces an en-
surface of the beam is roughly limited by the ray
hanced response at far range, which shows up on the
shown. Thus two important features are clear from
records as a line or band of increased intensity at the
the figure: first, the maximum range is limited by
edge of the sonar coverage (Fig. 6). It is usually most
refraction (i.e. there is the well-known-and mili-
severe in shallow tropical or subtropical waters,
tarily important-shadow zone between the top of
where the restricted range and strong refraction due
the beam and the sea surface), and this maximum
to a well-developed thermocline combine to facilitate
range increases with increasing water depth; sec-
it. Since the causative refraction mostly takes place
ondly, rays can become focussed near the top edge
near the sea surface, local variations in acoustic
of the beam, giving rise to rapidly varying sound
velocity there can modify the effect; in particular, the
intensity, as we shall see below.
presence of internal waves at density interfaces in the
near-surface waters can strongly modulate the fo-
BEAM DIRECTIVITY
cussing effect, leading to the appearance of speckling
Figure 4c, d shows the directivity pattern b(c/J) in the or even coherent fringes in the high-intensity band.
vertical plane mapped on to the horizontal plane of By contrast, in deep water the refraction factor
the sea floor, by evaluating it at the appropriate remains close to unity throughout the range, since in
value of c/J for each ray and plotting it against the this case the strongly focussed upper part of the
horizontal range where that ray strikes the sea floor. beam does not intersect the sea floor within the range
This is done for the same two water depths and the
of the sonar (Fig. 4a).
same velocity profile as illustrated in Fig. 4a, b. The
important features to note are that
TOTAL EFFECT
( 1) the maximum of the function (centre of main The spherical spreading (1/R2), incidence area
beam) occurs closer to the ship, and the response (sec () /R), and attenuation terms all vary monotoni-
falls off more rapidly with increasing range, in shal- cally and show no particular points of interest; ex-
lower water; and cept for the sec () term they should be offset by the
(2) the minimum (the null between the main system TVG. In Fig. 4g, h, we show the combined
 28 R. C. SEARLE ET AL.

CENTR AL INDIAN OCEAN MAy-TUNE

SOURCE-DEPTH'MI. 46
~AVErRONT INrERv ALCMS I. 1000
VELDCII Y . MIS INIT RAY ANGLES 'rIRST . LAST. INCRI. 89 .0-10.0 -1.0
I I
o~~~~~~~~~~

1000 1000
z:
2000
:I:
>-
lOOO Q.
W
o
4000

a RANGE. M

BEAM DIRECTIVITy

I.
~H++' , I ,
0.8 +++++ ++
0.6 + + + +
0.4

0.2

O.
O. 5000. 10000. 15000. 20000 . 25000.
C RANGE. M

REFRACTION FACTOR

5.

4.

l.

2.

I. ++++ + + + + + +
O.
O. 5000. 10000 . 15000. 20000. 25000.

e RANGE. M

TOTAL SIGNAL. PRESSURE LINEAR

121.

96 .8 + +
72.5

48 .4

24.3

O.
O. 5000 . 10000. 15000. 20000 . 25000.
RANGE. M
9

Fig. 4.
 GLORIA IMAGE PROCESSING 29

CENTRAL lN01AN OCEAN MAY - JUNE

SOURCE - DEPTH< MI - '0
uAvErRONT INTERVALIMS I. 1000
INIT RAY ANGLES IFIRST. LAST. INeRt- 89.0-10.0 -1.0
I
0
c 000
1200 1200
;: 1800
0. 1800
w
0 2 400
3000 ~~~~~r-~~
... __~~-+__~~~__~~2'00
3000
20000 25000
b RANGE. M

BEAM DIRECTIVITY

I.

0.8

{l.0

0.'

0.2

O.
O. 5000. 10000 . 15000. 20000 . 25000.
d RANGE. M

RErRACTION r ACTOR

5. I I I

4. - i-
3. - -
2. - -

I. ""
"" -
O.
I I 1
o. SOJa. 10000. 15000. 20000. 25000.
f RANGE. M

TOT AL SIGNAL. PRESSURE LINEAR

95 .2

70.2

57.

38.

19.

O.
O. 5000. 10000. 15000 . 20000. 25000.
h RANGE . M

Fig. 4. Computed ray diagrams and effects of various instrument and propagation effects on GLORIA image intensity for typical operating
conditions. Left-hand column computed for sea floor at 5000 m, right-hand column for 3000 m. Horizontal scale is the same throughout.
(a, b) ray paths. The assumed velocity/depth function is shown on the left (based on data from the Central Indian Ocean in May- June).
Right-hand part of plot shows rays with take-off angles 1° apart between _10° (above horizontal) to 89° (near nadir). Vertical and
horizontal scales are equal. Arcs show wavefront at I s intervals. Sound source at 46 m. (c, d) the beam directivity pattern (b). (e, f) the
refraction factor F = Ii (aspherical) /1;( spherical). (g, h) the product of all effects (text Equation 9) assuming a constant backscattering
coefficient Sb = 1.
 30 R. C. SEARLE ET AL.

CEN,RAL INOIAN OCEAN MAY-JUNE - MOOlr lEO HlERMOCLl t

SOURCE -O(P,HI M 1- 46
WAVEFRONT INTERVALI MS ) _ 1000
INH RAY ANGLES IfIRST. LAS,. INCRI. 89 .0- 10.0 -1.0

E 600
a
600 E
1200
1200
1800 ~
1800 0-
2400
~
3000 +~"--'-t>-'-"""""'-'-1~'-"-'-

20000 25000
RANGE. M

BEAM DIRECtIVITY
I.

0.8
=t-+++-
0.6
1-++++
O ••
+++,
0.2

o.
O. 5000. 10000. 15000. 20000. 25000.
RANGE. M

REFRACTION FACTOR
5. I I I I

'. - -
3. -

-L..w..++~
2. -
l-
I. "'"1111
""" r-
O.
o. 5060. 100~0. 150~0. 20obo. 250-JO.
RA GE. M

TO, AL SIGNAL. PRESSlRE L I EAR

95 .3

76. 2

57.2

38.1

19.

O.
O. 5000. 10000 . 15000. 20000. 25000.
RANGE. M

Fig. 5. As Fig. 4, but using a slightly modified velocity profile in which there is a steepeI thermocline in the vicinity of the sound source.
Note that this leads to increased response at far range due to enhanced refraction.

effect of all the factors. This is the same as equation to-far range, and the refraction caustic, are also
(9) for a constant backscattering coefficient Sb = I. preserved.
The variation of intensity with range is dominated by However, although these features are qualitatively
the effect of the beam directivity, especially the null similar to those observed in actual signals, we find
between the main beam and the first sidelobe. Other the quantitative match is variable. In particular, the
features, such as the gradual fall-off in gain at mid- response at far range often falls off somewhat faster
 GLORIA IMAGE PROCESSING 31

SHADING CORRECTION

Our first attempt at a shading correction was carried

out on the NERC FS image-processor then installed
in Swindon in 1984. The method we adopted was to
find an empirical function to approximate the instru-
ment/propagation response (Equation 9). We call
this the standard profile. It would seem best to calcu-
late the standard profile from Equation (9) itself;
however, even though all the appropriate parameters
are supposedly known, experience with the similar
SeaMARC II instrument (Reed, 1987) has shown
that in practice such computed profiles do not accu-
rately reflect actual instrument response. We intend
to try this approach in the near future, but for the
time being we present the results of the more empir-
ical approach.
The empirical function we used first was a
smoothed average of several data scans over a fea-
o, KM ,
10
tureless abyssal plain (Searle and Kidd, 1984; Searle
and Hunter, 1986). By dividing each pixel of an
Fig. 6. Examples of the caustic produced by focussing of the outer image by the value of the corresponding (same
part of the beam as a result of near-surface refraction (narrow range) pixel in our standard profile, we effectively
white bands at the edges of the data swath). Black areas beyond
removed the major, cross-track instrument response.
the caustic lay above the sonar beam so were not imaged. Sono-
graph from the north flank of Saya de Malha Bank, Indian Ocean, This was very effective in enhancing the image, giv-
near 9° S, 62° E, water depth about 2000- 2500 m. The broad ing it more uniform contrast, revealing previously
white band crossing the swath arises from a debris flow down the unseen information in the weak near- and far-range
margin of the bank. Parallel, broken white bands along right edge
are artefacts. areas, and enabling better comparisons between fea-
tures at different ranges. However, this method is not
than Fig. 4 predicts, and the first null is somewhat ideal, since the standard profile also incorporates the
narrower. The former may be largely a result of effects of possible variations in back scattering as a
using an oversimplified backscattering law in the function of range or incidence angle. Since the exact
calculations (we know for example that the fall-off in nature of these variations will probably be a function
apparent system response is different for different sea of the seabed lithology, different standard profiles
floor lithologies), while the latter may largely be due might be needed for different areas. Moreover, since
to poor modelling of the near-nadir beamform. the standard profile now contains (and this shading
method removes) this information, potentially im-
portant geological information is being lost.
Image Processing
Meanwhile the USGS followed a slightly different
In this section we describe three areas in which we approach (Chavez, 1986). Instead of taking their
have made recent progress. These are (I) develop- standard profile from a known featureless area, they
ment of a radiometric or 'shading' correction to use the along-track average of a single 6-hour data
further flatten the response curve shown in Fig. pass (implemented by subroutine SHAD2 in the
4g, h; (2) implementation of that correction to enable MIPS software). This works well in areas of rela-
true levels of backscattered energy and values of the tively low relief, and is convenient in that it allows
backscattering coefficient to be estimated; and (3) for automatic shading corrections, taking into ac-
digital mosaicking and combination of GLORIA count the variations in average backscattering be-
and Sea Beam (multi beam echosounder) data into tween different terrains. It still suffers from the
single images for enhanced presentation of the data. disadvantage that knowledge of these variations is
32 R. C. SEARLE ET AL.

A B

c o

o, KM 10
,

Fig. 7. The results of different forms of 'shading' correction on an image crossing the Rodriguez Triple Junction near 25S S, 70° E. Closely
spaced lineaments are fault scarps and volcanic ridges of the tectonic spreading fabric; cross-cutting feature is a minor fracture zone. (a)
'raw' image after slant-range and anamorphic corrections; (b), shaded using the along-track average profile (SHAD2): (c), shaded using
the along-track maximum profile (SHADX); (d), shaded using the function shown in Fig. 8. Note that d particularly produces a more
uniform 'illumination', and enhances features at very near and at far range. The possible improvements are usually even greater over
well-sedimented areas (see, e.g. Searle and Hunter, 1986).
 GLORIA IMAGE PROCESSING 33

lost, but it also has a more serious defect in rugged entirely on the response computed from Equation
areas. Where there is strong sea floor relief, the (9).
maximum range of the sonar beam will vary (Fig.
4a, b). If the pass is now averaged along track, the BACKSCATIERING DETERMINAnONS

averages for the outermost pixels will include both The backscattering coefficient Sb (Equation 5) is
real data where the swath is wide, and zeros from usually quoted in its logarithmic form, 10 log Sb
areas of shadow where it is narrow. The standard (dB). A large number of acoustic experiments have
profile will thus be too low at its margins, and the been carried out on the sea floor to determine the
corresponding parts of the corrected image will be dependence of Sb on angle of incidence, bottom type
too bright (Fig. 7a, b). (roughness and acoustic impedance) and acoustic
To overcome this defect, we construct the standard frequency (for examples see Urick, 1975; McKinney
profile not from the along-track average of all pixels and Anderson, 1964). These typically show a de-
at a given range, but from the maximum occurring crease of 20-40 dB with decreasing grazing angle,
along the whole pass for that range (subroutine and a variation of 20 to 25 dB between different
SHADX). The rationale is that in areas of strong bottom types.
backscattering (which also tend to be the ones with In view of the obvious benefits to be expected from
high relief) there is a strong chance that the maxi- direct comparison with experimental results such as
mum signal received from a relatively small area will these, we have derived a simple model for estimating
be close to the maximum value that could be re- the backscattering strength from GLORIA data and
ceived from that area (Le. the area will contain at can display the value of Sb as an image. This allows
least one optimum scatterer), and this will accurately for quantitative estimates of the likely causes of
convey the instrument response. Fig. 7c shows the backscatter observed on sidescan images.
result of applying shading in this way, which is We compute Sb from the logged signal voltages V,
clearly better in this case than the averaging version. using a simple model (equation 9) that assumes only
However, even this version of the correction spherical spreading, and approximate values for
suffers from a dependence on the local geology, and GLORIA's voltage gains G. Our estimates of the
will in general vary from place to place. Therefore, gains and other parameters are probably accurate to
we now employ a user-defined version of the stan- within only about 10 dB at present, so our discussion
dard profile, which is constructed using the SHAD2 concerns the variation in Sb, not its absolute value.
and SHADX profiles as a guide (Figs 7d, 8), This The beam pattern calculation is poor near the nadir
allows the user to tailor the correction to his require- (ranges less than 3 km), so the computed values of Sb
ments, including the use of a fixed standard profile in that range are unreliable. So may be those at far
over a whole survey area or beyond. In the near range, where the effects of refraction have been
future we intend to try a standard profile based ignored.

Standard profile
2.0 ,- - .... 200
I '

.,
-
~

o
o o
01 ~

0.

~ 1.0 100 ~
Correct i on factor
01
.,o
~
'0

o (/)
<..l

0~-~2~0~---------~1~0~---------0~----------1~
0-----------2~0~ 0

km

Fig. 8. 'Standard profile', and its inverse the correction factor, used to produce the shaded image shown in Fig. 7d.
34 R. C. SEARLE ET AL.

'.

. '.
- ;;:
.: ·c~/;·· .:-, .

} "1f}~(~~~~~~~i~#j!~~~
,' .r '~:_

.'

·ro+--------.,--------.--------.--------.~------,_--------._------_.------~
·2001X) -lSfXI() ·IOOIX) ·5fX1() 0 SfXI() 100IX) ISfXI() 2001X)
Horizontal range (m) .

Fig. 9. Profile of inferred backscattering coefficient Sb, against horizontal range, from a sonograph over very thinly sedimented basaltic sea
floor (similar to that of Fig. 7) on the crest of the Southeast Indian Ridge near 26° S, 70° E. The scan direction was parallel to the principal
topographic fabric. Points from six adjacent scans have been stacked. Absolute values are only reliable to about ± 10 dB because of poorly
known instrument gains and propagation parameters, but relative levels should be approximately correct except within 3 km of the nadir,
where the beam directivity is poorly modelled.

Figure 9 shows inferred values of Sb plotted Figure 10 shows a similar display of Sb from the
against horizontal range. The figure shows the same area as Fig. 9, but for a pass in which the scans
stacked determinations from six adjacent scans par- were normal to the lineated topographic fabric. Here
allel to the topographic grain over the crest of the Sb displays much greater scatter (> 20 dB) for a
medium spreading rate Southeast Indian Ridge near given range because of the greater variation in graz-
26° S, 70° E. This is an area of wdl-lineated seafloor ing angle for this scan direction. Very low values due
spreading (abyssal hill) fabric, with only a few me- to shadowing are also more common, and become
tres or less of sediment covering young basaltic lavas. more frequent with increasing range.
There is a small decrease in Sb with range, consistent Despite the shortcomings of this simple model, we
with the decrease in average grazing angle. The feel that reasonable values of Sb can be calculated
'" 10 dB scatter in Sb at a given range is to be from a fair proportion of the data. For more accu-
expected from changes in bottom roughness and rate estimates, we need to take account of refraction
slope. Large decreases (e.g. near - 10 km) are proba- and the effects of real bottom topography. We are
bly due to acoustic shadowing by intervening objects. now progressing in that direction, estimating the

-IOOIX) ·SfXI() 0 5fX1() 100IX) ISfXI() 2001X)

Horizontal range (m).

Fig. 10. Profile of inferred backscattering coefficient Sb, against horizontal range, from a sonograph over very thinly sedimented basaltic
sea floor on the crest of the Southeast Indian Ridge. As for Fig. 9, except that here the look direction was normal to the principal
topographic fabric. Alternating peaks and troughs in Sb are due to reflections and shadows from the sea floor ridges.
 GLORIA IMAGE PROCESSING 35

refraction effect as outlined previously, and using Sea to use parts of one or the other or an average of the
Beam data to make precise estimates of grazing two. Our state of knowledge is too primitive at this
angle. The results of these investigations will be stage to consider a computed "cut and paste" al-
reported elsewhere. gorithm, and the difficulties of developing one would
be compounded by the fact that overlap areas are in
MOSAICKING AND COMBINATION WITH SEA BEAM
general insonified from widely different directions.
Assuming that the system and propagation parame- However, we have in mind in the longer term an
ters are known, the back scattered acoustic intensity algorithm based on contouring the cross-correlation
recorded in each pixel by GLORIA can be regarded between small-windowed areas of the overlaid im-
as a function of two variables: the backscattering ages and looking for a ridge in the contours to
coefficient Sb (itself a function of the seabed rough- determine the best match.
ness and acoustic impedance), which is linked to The Sea Beam data have to be handled slightly
surface geology, and the angle of insonification, differently. The input consists of numerical integer
which depends on range and the slope of the sea depths and the corresponding ranges from ship. Thus
bottom. Thus two pixels could contain the same for each depth value a location has to be calculated
intensity (and be interpreted as similar geological from the position of the ship and the range value
features) even though they actually represent differ- (which is perpendicular to the ship's track). The
ent lithological and topographic entities. It is there- individual point soundings are then placed on the
fore important for a proper interpretation to consult nearest points of a 50-m grid. The results are stored
other sources of data, particularly bathymetry, to aid in a large two-dimensional array held in dynamic
geological identification and correlation. This can be memory before writing the array to a disk file when
a slow operation, and the initial absence of the all the input data have been read. This 'image' is
supplementary information reduces the ease of un- then transferred to the image-processor for combina-
derstanding of a mosaic of raw GLORIA images. tion with the GLORIA image.
The Sea Beam multi-narrow-beam echosounder It is more convenient for us to grid the Sea Beam
system (Renard and Allenou, 1979) produces high data on the IOSDL mainframe IBM computer rather
precision bathymetric maps of the sea floor. Mixing than the MIPS one, but memory restrictions on the
of Sea Beam and GLORIA data in a single digital former mean that the Sea Beam data have initially to
image can therefore complement the two data sets be assigned to a 50-m grid-a slightly larger spacing
and aid interpretation. We have used this technique than the 45-m optimum for GLORIA gridding. The
in an area having 100% coverage of GLORIA data next operation therefore is to interpolate the Sea
and over 80% coverage of Sea Beam data (the lack Beam data on to a 45-m grid, which can be done by
of full coverage is due to the much smaller simple linear interpolation without loss of real data.
( '" 2.5 km) swath width of the Sea Beam system, The same interpolation can also make up for the lack
combined with real time navigational uncertainties of of full Sea Beam coverage by filling in gaps where
a similar size, and the finite time available for sur- necessary. We now have two images of the same area
veys). The area chosen is a section of the Southwest with the same pixel size for combination, both with
Indian Ridge near 27.so S, 66° E. 100% coverage.
The first step is to produce digital mosaics of each GLORIA images are displayed as black and white
data set. Creating the GLORIA mosaic requires images usually with 256 density levels (or 'grey lev-
every track which crosses the area to be digitally els') requiring 8 bits. The Sea Beam images, however,
processed (slant-range, anamorphic and shading cor- contain integer values between 0 and 15000 m, re-
rections, and filtering to remove line drop-outs), quiring at least 14 bits. Most MIPS programs are
producing separate images of each track segment. designed to work with 8-bit data and therefore the
These are then mosaicked by digitally adding them Sea Beam data need compression. A new scale is
into a master image in their correct geographical given to the data where 0 represents the shallowest
locations, making a final mosaic covering the whole depth (in this case 2000 m) and 255 the deepest data
area (Fig. 11). Where segments overlap, the operator (5500 m). Thus vertical resolution of the data was
can view both images in different colours, and decide reduced in this case from 1 m precision to about
36 R. C. SEARLE ET AL.

Fig. II. Digitally mosaicked GLORIA images. Mosaic is approximately 100 km square, straddling the axis of the slow-spreading Southwest
Indian Ridge near 27.5" S, 66° E.

14 m which is unlikely to affect the geological inter- TABLE II

pretation, and is in fact close to the real resolution of
Colour table used in producing Figs lla and b
the Sea Beam system (Renard and Allenou, 1979). A
colour scale is then attached to these 256 values, each Level Red Green Blue Approximate colour
colour being represented by predefined ratios of red
255 0 255 Deep purple
to green to blue. We use a table of colour values that 42 204 0 10 Deep red
produce a 'rainbow' colour succession, from red for 104 255 255 0 Yellow
shallowest depths to violet for deepest. 126 0 204 0 Green
168 0 255 255 Cyan
The GLORIA and Sea Beam information are 210 0 0 204 Deep blue
combined by noting that the colour and brightness of 255 204 0 204 Magenta
any image pixel can be defined in terms of the values
Intermediate levels were produced by linear interpolation between
of three independent primary colours: red (r), green
these values.
(g), and blue (b). The hue of the pixel is determined
by the three ratios r :g :b, and its brightness by their
absolute values. We use the hue to represent the sensltlVlty of the eye to different colours, and the
depth (coded from the Sea Beam pixel value), and strengths of different CRT phosphors, photographic
the intensity to represent the GLORIA value. In dyes, and plotting inks, are very variable, different
principle one would expect the brightness to be a values of the sum r + g + b must be used to give the
simple sum r + g + b; however, because both the same perceived brightness for different colours. To
 37

(a)

(b)

Fig. 12. (a) (above) coloured image of same area as Fig. 11, made by combining sidescan (GLORIA) and depth (Sea Beam) information;
the hue is a function of depth, while the intensity is a function of sonar backscatter. Blue (deep) areas near the centre of the image are en
echelon basins marking the spreading axis. Note that the addition of depth information aids greatly in "reading" the image. (b) (below)
shaded relief image of the area shown in (a), for comparison, made from bathymetric (Sea Beam) information alone. The image intensity
represents the amount of light that would be reflected from the given topography for an illumination source 20° above the horizon to the
south of the area, assuming reflection according to Lambert's Law; image colour is related to depth. Note that (a) shows finer detail than
(b) (and also some textural information, though that is less important in this region where overall sediment cover is low).
38 R. C. SEARLE ET AL.

some extent the choice of levels must therefore be used in near real time on board ship. Other programs
made subjectively. Table II gives the r g b table used provide enhanced display of images, either alone or
for the examples shown in this chapter. combined with other data sets. Finally, a start has
As an example, assume a Sea Beam depth of been made in recovering true values of the acoustic
3000 m and a GLORIA intensity of 47. The position back scattering coefficient. These developments are
in the colour table is (3000-2000)*256/(5500- the latest in a series of continuous improvements to
2000) = 73. Colour value 73 is predefined as the ratio the system over a period of twenty years. Together
r:g:b = 230:128:5, a bright (at maximum intensity) with efficient and reliable digital data acquisition,
yellowy-red (see Table II). Therefore the required they now enable GLORIA to be used as a truly
output values are: quantitative research and survey tool.

r = 230 x 47/256 = 42;

g = 128 x 47/256 = 24; Acknowledgments
b = 5 x 47/256 = 1.
We gratefully acknowledge the help and advice of all
Total output is therefore 42:24:1, a dark yellowy-red. those who have been involved in developing and
The resultant Sea Beam/GLORIA image is shown operating the GLORIA system over the years. We
in Fig. 12a. The positions of structural highs and are especially thankful to Pat Chavez and his col-
lows in relation to sonar information can now be leagues in USGS, not only for making the MIPS
seen at a glance. The combined image also helps software available to us, but for many hours spent
explain the presence of certain bright or dull spots, with us explaining it and image processing in general.
as being due either to the angle of insonification or We also thank Angela Morrison, John Baker, and
to real lithological variation. It gives a somewhat other staff at the NERC Thematic Information
similar appearance to a shaded relief map con- Servies for helping us in our early struggles with the
structed just from Sea Beam data (Figure 12b), but I 2S system.
the latter can give no information on lithology or
surface roughness, and also lacks the resolution to
reveal some of the very fine detail seen in the GLO- References
RIA data. This is clearly a powerful technique for Chavez, P. S., 1986, Processing Techniques for Digital Sonar
combining different data sets, with all geographical Images from GLORIA, Photogrammetric Engineering and Re-
cross-referencing having been incorporated into the mote Sensing 52, 1133-1145.
EEZ-Scan 84 Scientific Staff, 1988, Physiography of the Western
image. United States Exclusive Economic Zone, Geology 16, 131-134.
Although we have not yet tried them, many other Laughton, A. S., 1981, The First Decade of Gloria, J. Geophys.
possibilities for combining data by this method exist. Res. 86, 11511-11534.
McKinney, C. M. and Anderson, C. D., 1964, Measurements of
One could reverse the combination, and relate image Backscattering of Sound from the Ocean Bottom, J. Acoust.
hue to sidescan intensity (or, better, to the derived Soc. America 36, 158-163.
back scattering coefficient), and image intensity to Reed, T. 8., 1987, Digital Image-Processing and Analysis Tech-
niquesfor Sea MARC II Side-Scan Sonar, Ph.D. thesis, Univer-
topography in a shaded relief representation. The
sity of Hawaii.
method would be equally suitable for combining Renard, V. and Allenou, J. P., 1979, Seabeam, Multi-beam Echo-
sidescan or topography with, for example, gravity Sounding in "Jean Charcot", Int. Hydr. Rev. 56, 35-67.
and magnetic fields, or sediment thickness, to great Rusby, 1. S. M., 1970, A Long-Range Side-Scan Sonar for Use in
the Deep Sea (GLORIA Project). Int. Hydrogr. Rev. 47, 25-39.
advantage. It would even be possible to combine Rusby, J. S. M. and Somers, M. L., 1977, The Development of the
more than two types of data, by using different "Gloria" Sonar System from 1970 to 1975, in M. Angel (ed.), A
intensities of a single colour for each. Voyage of Discovery, 611-625, Oxford, Pergamon (Suppl. to
Deep-Sea Research).
Searle, R. C. and Hey, R. N., 1983, GLORIA Observations of the
Propagating Rift at 95.5" W on the Cocos-Nazca Spreading
Conclusions
Center. J. Geophys. Res. 88, 6433-6447.
A suite of programs is now available for applying Searle, R. C. and Hunter, P. M., 1986, The Use of GLORIA
Long-Range Sidescan Sonar for Deep-Ocean Mapping, in M.
geometric and radiometric corrections to GLORIA Blackmore (ed.), Autocarto London, 2, Digital Mapping and
and other sidescan images, most of which can be Spatial Information Systems, 339-388.
 GLORIA IMAGE PROCESSING 39

Searle, R. C. and Kidd, R. B., 1984, GLORIA Digital Image NATO Advanced Study Institute, Loughborough, 1972), Aca-
Processing and Interpretation of the Saharan Sediment Slide, demic Press, London, 757-767.
Eastern North Atlantic, EOS, Trans. Amer. Geophys. Union 65, Somers, M. L., Carson, R. M., Revie, J. A., Edge, R. H., Barrow,
1082-1083 (Abstract). B. J., and Andrews, A. G., 1978, GLORIA II-An Improved
Somers, M. L., 1970, Signal Processing in Project GLORIA, A Long-Range Side-Scan Sonar, in Proc. IEEE/IERE Sub-
Long Range Side-Scan Sonar, in D. G. Tucker (ed.), Electronic conference on Offshore Instrumentation, Oceanology Interna-
Engineering in Ocean Technology, Proc. Inst. Electronic and tional '78, Technical Session J, pp. 16-24, BPS Publications,
Radio Engineers Conference, Swansea, 1970, Institution of London.
Electronic and Radio Engineers, London, 109-120. Urick, R. J. 1975, Principles of Underwater Sound for Engineers,
Somers, M. L., 1973, Some Recent Results with a Long-Range 2nd edition. McGraw-Hill, New York.
Side-Scan Sonar, in Signal Processing (Proceedings of the
High-Resolution Seismic Reflection Surveying of
Shallow Marine and Estuarine Environments
J. M. REYNOLDS
Department of Geological Sciences, Polytechnic South West, Drake Circus, Plymouth, PL4 BAA, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: Seismic reflection profiling, geophysical interpretation cations (Table I) by way of three types of seismic
methods. source systems, namely, Pingers, Boomers and
Sparkers. More details of the types of sources are
Abstract. High·frequency seismic reflection profiling is a well- given by Sieck and Self (1977), Lugg (1979) and
established and often used technique in marine investigations. McQuillin et al. (1984), and will be discussed in more
Traditionally seismic data are viewed as two-dimensional time
sections. Given closely spaced profile lines, it is possible to produce
detail later. The main functions of seismic profiling
posted two-way travel time maps of sub-surface reflectors which, in these applications are two-fold: to define basic
when plotted as isometric displays, clearly show the three-dimen- engineering and geological parameters, and to dis-
sional spatial morphology of the sub-surface topography. With
cover any potentially hazardous sub-surface geologi-
borehole control, such information can be used to provide a series
of images which indicate temporal as well as spatial relationships cal conditions. Examples of the former are
of sub-surface reflectors. With the high-resolution afforded by rippability parameters, overburden thickness, and
high-frequency methods, detailed information on palaeo-environ- the recognition of simple rock types. Examples of the
ments can be reconstructed. Using the geophysical database as a
basic framework, other aspects of the same environment can be latter include seabed pockmarks (which may indicate
examined in considerable detail. To demonstrate the effectiveness degassing of sediments), buried rock valleys (which
of these procedures, examples will be given from Plymouth Sound may cause problems with foundation stability and
where a series of nested buried rock valleys has been mapped in
detail from Sparker and Boomer surveys which have been inter- drilling), gasified sediments (low-pressure gas) which
preted in the light of newly-acquired borehole information. Isomet- would be problematical if large structures were con-
ric plots of the various sub-surface interfaces show how the structed over them, and gas accumulation (high-
channels have developed as sea level has risen over the last ca.
pressure gas) which can cause blowouts during
10,000 years. Different sedimentological facies can be resolved
within the channel system thus providing information about the drilling (as for example has happened in the Gulf of
processes involved in their formation. Methods of improving the Mexico). This chapter will concentrate more specifi-
resolution and subsequent geological interpretation of high- cally on applications to the near-shore and estuarine
resolution seismic reflection surveys are being developed for shal-
low marine and estuarine environments such as those found in environments with water depths in the general range
Plymouth Sound. 0-25 m.
One of the reasons why high-resolution seismic
profiling is such an attractive option in commercial
Introduction
offshore site investigation is that it provides a cheap
Although high-frequency seismic reflection profiling and effective way of obtaining images of the sub-
is a well-established and often-used technique in surface. A typical single-channel analogue acquisi-
marine site-investigations, the guidelines (British tion system produces a graphical output for which
Standard 5930; 1981) for its application are notori- subsequent data processing is rarely possible, but
ously inadequate. The question "What constitutes an which can yield much valuable geotechnical informa-
adequate interpretation of high-resolution seismic tion. The seismic sections produced are two dimen-
data?" has still to be addressed. Nevertheless, the sional images of the geological environment at the
method is commonly employed in a variety of appli- time of the survey. For many applications the level

Marine Geophysical Researches 12: 41-48, 1990.

© 1990 Kluwer Academic Publishers.
42 J. M. REYNOLDS

TABLE I rig drilled two boreholes within Plymouth Sound in

February 1988. Examples from the work in Ply-
Principal applications of commercial offshore site-investigation
using high-resolution seismic methods. (Modified from Carter mouth Sound will be used in this chapter to illustrate
et al., 1986) particular aspects of seismic interpretation.
Near-shore marine/estuarine environments
Bridges, tunnels, viaducts Survey Constraints
Harbours, jetties, quay walls, marinas
Pipelines and tunnels for sewage outfalls POSITION FIXING
Dredging for access channels to ports/harbours
Marine As with any survey, navigation is of vital impor-
Hydrocarbon pipelines tance. The accuracy of the position fixing methods
Hydrocarbon production platforms/well heads
Siting of drilling rigs employed will constrain the spatial interpretation of
the seismic data. If the ground conditions vary
markedly over short distances (e.g. 30 m), it would
prove an almost impossible task to locate a borehole
of interpretation which this affords may well be over a specific spot if the positions of the survey
acceptable. For others it is not, and one of the stations were known to no better than, say, 20 m.
objectives of this chapter is to demonstrate that Within Plymouth Sound, for example, the depth to
single-channel analogue seismic records can provide bedrock beneath the mudline within the estuary can
a four-dimensional spatial and temporal model of vary from about 3 m to 20 m in a horizontal distance
the sub-surface geological environment. The infor- of only 50 m. For positioning the seismic lines and
mation so obtained can be of immense value in such the two boreholes of this survey, Trisponder systems
areas as environmental impact assessment investiga- were used. These resulted in a standard deviation on
tions and hydrogeological studies, as well as in estab- position of between 0.5 and 1 m. In the future, it is
.lishing the nature of sedimentary sequences, all of likely that positional accuracies to within ± 1 m will
which can aid engineering design. It is important to become more commonplace once the Global Posi-
recognize that there are sophisticated digital multi- tioning System (GPS) becomes fully established
channel high-resolution seismic systems available, (Bullock, 1988). This should help to solve the prob-
the data from which can be enhanced using commer- lems associated with uncertainties in position, permit
cial processing packages, and which, consequently, closer line spacings (e.g. 5 to 10 m separations) and
are very expensive to use. The more usual site-inves- allow greater opportunities for spatial interline inter-
tigation practice in near-shore and estuarine environ- pretation of seismic sections.
ments is to employ the low-cost single-channel
analogue graphical systems. It would be an advan- SOURCE PARAMETERS
tage to be able to improve the level of geotechnical One of the factors which controls the resolution of a
interpretation of data acquired by these analogue seismic survey is the shape of the source pulse.
systems without incurring a prohibitive increase in Ideally this should be a spike (Dirac) pulse where all
costs. The purpose of this chapter is to demonstrate the energy is applied at a single instant of time. In
that judicious survey design, coupled with a limited reality, best resolution occurs with those sources
amount of computer power, can enhance greatly the which produce a wave shape as close to the ideal as
information obtainable from these basic surveys and possible, typically a minimum-phase wavelet (Wa-
provide a very useful foundation for further investi- ters, 1978, p. 164). A seismogram, denoted by s,
gations. comprises the convolution of the source wave w with
The Department of Geological Sciences and the e, a time series which contains the acoustic
Institute of Marine Studies at Polytechnic South impedance function of the media through which the
West have undertaken high-resolution seismic reflec- signal is propagating such that:
tion surveys within Plymouth Sound and the River
Tamar since 1982 (McCallum and Reynolds, 1987;
s = w*e( +n),
Reynolds, 1987). Following financial support from where * denotes convolution and n, additive noise
the National Advisory Body, a commercial jack-up (Hatton et ai., 1986). Consequently, the resolution of
 HIGH-RESOLUTION SEISMIC REFLECTION 43

TABLE II nals are attenuated more rapidly than those at lower

frequencies, so that the higher the source frequency,
Theoretical resolution and depth penetration of three common
high-frequency seismic sources the shallower is the depth penetration (Table II).
However, such specification summaries (e.g. Conway
Source Frequency Resolution Depth of
et al., 1986) are misleading. With Sparkers, for exam-
bandwidth penetration
ple, the pulse shape depends critically on the output
Pingers 3.5-7 kHz 0.1-1.0 m s; tens of metres power of the system and the number and configura-
Boomers 400 Hz-S kHz :::olm tens to 100 + m tion of electrodes in the sparker array. Examples of
Sparkers 200 Hz-I.S kHz 2-3m ~ 1000 m
this source variability are given by Lugg (1979) and
are illustrated in Fig. I. Better pulse shapes are
achieved by increasing the number of electrode tips.
a seismic section is dependent upon the quality of the Within Plymouth Sound, a 3-tip Sparker array at
source wavelet w. 500 J produced significantly broader pulses with a
Parameters of the source wavelet which particu- corresponding decrease in resolution than did an
larly affect the quality of the seismic sections are the 18-tip array at the same output energy. In addition,
pulse shape and the pulse length. the source depth (below the sea surface) also affects
the pulse shape by virtue of the interference of the
PULSE SHAPE surface ghost and bubble pulse. The effect of these
The higher the source frequency, the shorter the parameters is to cause the source wavelet to become
wavelength of the signal and hence (in theory) the a complex non-ideal shape and, furthermore, to
better the resolution. However, high-frequency sig- differ from the manufacturer's specification for the

3 e'tctrod,s
b
a

500J

I I ,
o 5 10 15 20m,

300 electrexln

0 .4 bar [+- 500 J

30 kJ

o
t
5
I I
10 15 m,
I o 60 ms

c
CSP ? ? ?

~
--_.
... ..,
.
.-
- ...... - - +
+

~~ -.
_ L _" _ +
• _ _ __ #T". ' +

Fig. 1. Sparker source pulse shapes with (a) 3 and 300 electrode tips with 500 J output (after Lugg, 1979), and (b) a far-field signature for
a system with 30 kJ output (after Hatton et al., 1986). (c) Analogue signal form from Plymouth Sound (left) with pulse polarity but with
no amplitude information, and right, a variety of waveforms with the same pulse polarity and phase but different amplitudes and which
could represent the observed signal.
44 J. M. REYNOLDS

device used, because of the dependence of perfor- a multi-tip Sparker at the same energy output
mance on the local ambient conditions of operation. (0.5 kJ). The quality of the Sparker data was suffi-
ciently good that the main geological interfaces pre-
PULSE LENGTH
dicted from the seismic interpretation were all
The ability to resolve between reflectors in an inter- confirmed to within 0.5 m by subsequent drilling. It
pretation depends upon the pulse length (Meckel was also found that the Sparker results degraded
and Nath, 1979), the amount of reverberation which slightly throughout the survey as some of the elec-
occurs and on the distribution and nature of the trode tips burned back to the insulation and eventu-
sub-surface media. If a pulse length of, say, 5 ms is ally ceased to work, thereby reducing the number of
used and the two-way travel time difference between active tips and the effective output both of which
two events is greater than the pulse period, then it served to degrade the pulse shape. It is important to
•
should be possible to resolve each reflection. How- check on the state of electrode tips during each day's
ever, if reverberation of the pulse occurs and the tail survey and to monitor signal quality.
of the first reverberation obscures the onset of the
second event and interferes with it, then the second
Detailed Seismic Interpretation
reflection may remain unresolved and effective de-
convolution of the acoustic impedance function of Traditionally, analogue seismic data are viewed as
the sub-surface media is unsuccessful. It is therefore two-dimensional time sections which can be inter-
to be recommended that the source pulse shape and preted simplistically in terms of line drawings of the
duration should be recorded at the time of each principal reflections. This level of interpretation is
survey. The observed pulse shape can then be used, if commonly used in non-critical site investigations
necessary, in the generation of any synthetic seis- where only the depth to bedrock is of importance
mograms to aid the interpretation. Another factor is and detailed interpretation is not required. It is
the nature of the media through which the acoustic possible to go a stage further and interpret the
pulses travel. Penetration through and the seismic sections in terms of seismic stratigraphic sequences.
character of specific types of material also affect the Despite the apparent relative coarseness of the reso-
choice of source. Some indication of this is given by lution of Sparker data, it has been possible to define
Carter et al. (1986). Within Plymouth Sound, for detail within a sequence of sediment progradational
example, and EG & G Boomer produced more surfaces with a time resolution ofless than 1 ms (Fig.
smeared seismic sections and poorer resolution than 2). The sigmoidal form of the reflections is very

PROFILE A SEABED PROFILE B

20-

-----=-== - -:+::
-
- ;::;::::::;=:;:

100m --
100m

Fig. 2. Two extracts from sparker records in the range 20-50 ms Two-Way Travel Time (TWTT) of the same ground within Plymouth
Sound showing sigmoidal reflections associated with a known sand body (after Eddies and Reynolds, 1988). The two records were acquired
with different filter frequency bandwidths.
 HIGH-RESOLUTION SEISMIC REFLECTION 45

similar to that of reflection events described by trates stacked isometric projections of Two-Way
Mitchum et al. (1977). Where low-frequency multi- Travel Times, which have been normalized to a time
channel data have been recorded digitally the sophis- datum of 60 ms below Ordnance Datum (Newlyn)
tication of interpretation is well developed (e.g. Todd for graphical display. Four surfaces have been se-
and Mitchum, 1977; Brown et al., 1984), but with lected: seabed, reflector 2, reflector 1 and bedrock.
analogue records the options for further interpreta- Reflector 1 correlates with the shelly gravel horizon
tion are much more restricted. Nevertheless, given a (Fig. 3) which separates organic muds from overly-
sufficient number of closely spaced seismic profile ing sands. Reflector 2 is a prominent event associated
lines, such as the one illustrated in Fig. 3, it is with the sigmoidal reflections seen in Figs 2 and 3
possible to produce posted Two-Way Travel Time and has been interpreted as being a point bar foreset
(TWTT) maps (e.g. Fig. 4) of sub-surface reflectors (Eddies and Reynolds, 1988). Two Way Travel Time
which, when plotted as isometric displays (Fig. 5), maps and isometric projections were produced by
clearly show the three-dimensional spatial morphol- digitizing line interpretations of the seismic sections
ogy of the sub-surface topography. Figure 5 illus- and processing the data using the SURFER graphi-

FIX NUMBERS
0-

20-

TWT(ms)

30-

40-

Fig. 3. Analogue Sparker record over a buried rock valley within Plymouth Sound with the main geological sequences as found by drilling.
Two preliminary average interval velocities have been determined for the main sequences corresponding to the depth ranges 0-11.4 m and
11.4- 26.6 m using the observed Two-Way Travel (TWT) Times and the measured depths to the corresponding interfaces from the borehole
lithological log.
46 J. M. REYNOLDS

Plymouth Sound Bedrock TWTT (ms)

247432 247582 247732 247882 248032 248182 248332 248482
52483 I=JT1~TI71"T"mrrm-nTTTTT"ITTT1TTTTTTTT""I'l~"TTTlTT"rnn"l'l"TTTTTT>lVT1lTlTT1rmnr17f7TYT"T7T1rTT17TTTTTrTT1TTT1VTTTT1ITTiTTTTTTT"TT+l

52333

5218.3

247582 247732 247882 248032 248182 248332 248482

Fig. 4. An example of a Two-Way Travel Time map to bedrock from a part of Plymouth Sound. Values along the X and Y axes are
Universal Transverse Mercator (UTM) co-ordinates. The location of the borehole (details in Fig. 3) is also shown.

cal package (Golden Software, Inc., Colorado, process rates within the sedimentological regime as
USA) on an IBM PC-XT compatible microcom- well as to how the dynamics of a particular environ-
puter. Hardcopy output was produced via a 24-pin ment relate to larger scale climatic events such as
printer. The stacked isometric projections illustrate sea level changes (Eddies and Reynolds, 1988). Al-
the upward and lateral migration of the main though only provisional dates are available at
drainage channel of the River Tamar through Ply- present, the core material is currently being sub-
mouth Sound as a consequence of rising sea level sampled, not only for further detailed micro- and
and subsequent sedimentation. With borehole con- macro-palaeontological, palaeo-botanical and pa-
trol, such information can be used to provide a lynological studies, but also for heavy metal concen-
series of images which indicate temporal as well as trations, clay mineralogy, and organic geochemistry.
spatial relationships of sub-surface reflectors. A With the high-resolution afforded by single-channel
borehole with good recovery of material sited on a seismic profiling methods (Table II), it should be
well-interpreted seismic line provides a means for possible to develop more detailed palaeo-environ-
calibrating the seismic sections. More than that, mental reconstructions.
interdisciplinary investigations of the retrieved ma- Survey techniques in deeper water and on land
terial can provide information about when certain have received a great deal of attention in terms of
events may have occurred. For example, within the both data acquisition and in interpretation. How-
cores from Plymouth Sound, extremely well-pre- ever, seismic surveys in very shallow marine and
served micro-fossils have been found which have estuarine environments, especially those associated
been used to date major horizons. These, in turn, with the intertidal zone, are assuming increasing
have been correlated with major reflections. Conse- importance. Such an environment is a very difficult
quently, some reflections then serve as isochrons. one in which to work logistically and which fits
For example, one horizon (top of the organic muds neither land nor marine surveying techniques com-
shown in Fig. 3) has been dated provisionally at fortably. However, an example of a three-dimen-
8300 BP (EddIes and Hart, in press) on the basis of sional seismic survey over tidal mudflats in The
microfossils; similar mid-Flandrian ages have been Netherlands has been described recently by Corsmit
obtained from samples of the same material from et al. (1988). The success of the survey was depen-
preliminary palynological investigations (J. Scourse, dent upon (a) the acquisition of multifold coverage
pers. comm.). This can give some indication as to and (b) data processing on a minicomputer.
 HIGH-RESOLUTION SEISMIC REFLECTION 47

ment of Geological Sciences at Polytechnic South

West, Plymouth, for shallow marine and estuarine
environments such as those found in Plymouth
Sound. Such developments will be of benefit to hy-
60 drographers, geologists and civil engineers as well as
to mariners and environmental conservationists.
50

Conclusions
<le
Factors which are known to constrain the quality of
6e data acquired during an analogue seismic profil-
ing survey include: navigational accuracy, the shape
5e of the seismic source wavelet, and the nature and
geometry of the sub-surface reflectors.
Analogue seismic records are usually interpreted
(J) <l0
E two-dimensionally and with only the main reflections
'-- 60
(!)
being picked out on line-drawing representations.
E However, the use of seismic stratigraphic sequence
f-
50 interpretation allows inferences to be drawn concern-
(!) ing the dynamics of the depositional regime and the
;:..
processes involved in producing the resultant sub-
.::
'lJ
<l0
surface geology. This in itself does not go far
6e
::... enough. The use of a sufficient number of closely-
=t
I
spaced seismic survey lines permit the definition of
0 50 isometric surfaces of the principal reflections and
t2 thus the production of a three-dimensional model of
40 the spatial distribution of particular seismic events.
In addition, by modelling events from the rockhead
through to the sea/river floor it is possible to obtain
3e
an impression of both the spatial and temporal
relationship between features. Allied with borehole
20 control which can give probable dates for specific
events, a dynamic model of the sub-surface regime
10 can be developed. This approach can be of great
value in aiding the geotechnical, hydrogeological and
geological interpretation of analogue seismic data,
0
and it offers the potential of aiding civil engineer-
ing design in the evaluation of possible post-
development environmental problems.

Fig. 5. Examples for part of Plymouth Sound of stacked isometric

Acknowledgements
projections of normalised Two-Way Travel Times to seabed, re-
flector 2, reflector I and bedrock respectively. Values along the X
I am grateful to Capt. K. McCallum, Mr. C. Sawyer,
and Y axes are Universal Transverse Mercator (UTM) co-ordinates. and Mr. T. Parrott, all of the Institute of Marine
The location of the borehole (details in Fig. 3) is also shown. Studies at Polytechnic South West, Plymouth, for
their collaboration and assistance, without which the
Methods for improving the resolution and subse- field work would not have been possible. The Ply-
quent geological interpretation of shallow seismic mouth Sound Project has received valuable financial
reflection surveys are being developed at the Depart- support from the National Advisory Body. Thanks
48 J. M. REYNOLDS

are also due to Mr. R. D. Eddies for fruitful discus- Hatton, L., Worthington, M. H., and Makin, J., 1986, Seismic
sions concerning the seismic work and for his help in Data Processing, Oxford, Blackwell Scientific Publications, 177
pp.
drafting some of the diagrams, and to Dr E. Hail- Lugg, R., 1979, Marine Seismic Sources, in Fitch, A. A. (ed.),
wood for his many editorial improvements. Developments in Geophysical Exploration Methods -1, London,
Applied Science Publishers Ltd, 143-203.
MacCallum, K. D. and Reynolds, 1. M., 1987, High-resolution
(Plymouth Sound Research Project Contribution Seismic Profiling in Plymouth Sound and the River Tamar
No.8). (Abstract), Proceedings of the Ussher Society 6, 562.
McQuillin, R., Bacon, M., and Barclay, W., 1984, An Introduction
to Seismic Interpretation, London, Graham and Trotman, 287
References pp.
Meckel, L. D. and Nath, A. D., 1977, Geologic Considerations for
Stratigraphic Modelling and Interpretation, in Payton, C. E.
BS 5930, 1981, Code of Practice for Site Investigations (Formerly (ed.), Seismic Stratigraphy-Applications to Hydrocarbon Ex-
CP 200 I), British Standards Institution. ploration, Memoir 26, Tulsa, Oklahoma, The American Associ-
Brown, A. R., Wright, R. M., Burkart, K. D., Abriel, W. L., and ation of Petroleum Geologists, 417-38.
McBeath, R. G., 1984, Tuning Effects, Lithological Effects and Mitchum, R. M., Vail, P. R., and Sangree, J. B., 1977, Seismic
Depositional Effects in the Seismic Response of Gas Reservoirs. Stratigraphy and Global Changes of Sea Level, Part 6:
Presented at the 46th Annual Meeting of the European Associ- Stratigraphic Interpretation of Seismic Reflection Patterns in
ation of Exploration Geophysicists, London, June 21, 1984. Depositional Sequences, in Payton, C. E. (ed.), Seismic Stratig-
Bullock, S. 1., 1988, Future and Present Trends of Navigation and raphy-Applications to Hydrocarbon Exploration, Memoir 26,
Positioning Techniques in Exploration Geophysics, Geophysical Tulsa, Oklahoma, The American Association of Petroleum
Journal 92(3), 521. Geologists, 117-33.
Carter, P. G., Pirie, R. M., and Sneddon, M., 1986, Marine Site Reynolds, 1. M., 1987, Geophysical Detection of Buried Channels
Investigations and BS 5930, in Hawkins, A. B. (ed.), Site in Plymouth Sound, Devon (Abstract), Geophys. J. Roy. Astron.
Investigation Practice: Assessing BS 5930, Geological Society Soc. 89, 457.
Engineering Geology Special Publication No.2, 163-66. Sieck, H. C. and Self, G. W., 1977, Analysis of High Resolution
Conway, B. W., McCann, D. M., Sarginson, M., and Floyde, Seismic Data, in Payton, C. E. (ed.), Seismic Stratigraphy-
R. A., 1984, A Geophysical Survey of the Crouch/Roach River Applications to Hydrocarbon Exploration, Memoir 26, Tulsa,
System in South Essex with Special Reference to Buried Chan- Oklahoma, The American Association of Petroleum Geologists,
nels, Quarterly Journal of Engineering Geology 17(3), 269-82. 353-85.
Corsmit, J., Verteeg, W. H., Brouwer, J. H., and Helbig, K., 1988, Todd, R. G. and Mitchum, R. M., 1977, Seismic Stratigraphy and
High-Resolution 3D Reflections on a Tidal Flat: Acquisition, Global Changes of Sea Level, Part 8: Identification of Upper
Processing and Interpretation. First Break 6(1), 9-23. Triassic, Jurassic, and Lower Cretaceous Seismic Sequences in
Eddies, R. D. and Reynolds, 1. M., 1988, Seismic Characteristics the Gulf of Mexico and Offshore West Africa, in Payton, C. E.
of Buried Rock Valleys in Plymouth Sound and the River (ed.), Seismic Stratigraphy-Applications to Hydrocarbon Ex-
Tamar, Proceedings of the Ussher Society 7, 36-40. ploration, Memoir 26, Tulsa, Oklahoma, The American Associ-
Eddies, A. P. and Hart, M. B., 1989, Late Quaternary Foraminife- ation of Petroleum Geologists, 145-63.
rida from Plymouth Sound: Preliminary Investigation, Proceed- Waters, K. H., 1978, Reflection Seismology, New York, John
ings of the Ussher Society 8 (in press). Wiley and Sons, 377 pp.
A Fixed Receiver for Recording Multichannel Wide-Angle
Seismic Data on the Seabed
C. M. R. ROBERTS
Department of Geological Sciences, University of Durham, UK

and

M. C. SINHA
Bullard Laboratories, Department of Earth Sciences, University of Cambridge, UK

(Received 27 April, 1989; accepted 1 September, 1989)

Key words: seismic refraction, lithosphere, continental shelf. channel data from a fixed sea-bottom receiver. Com-
pared with earlier techniques, this method provides a
closer determination of both the velocity structure
Abstract. Previous experiments to record seismic data at wide and the composition of the continental lithosphere,
angle on the continental shelf have generally been unsuccessful in in particular the lower continental crust and Moho.
determining velocity structure in the lower crust; either the lines
were too short or shot-receiver density too sparse to identify lower Currently used seismic techniques at both normal
crustal arrivals. In contrast, deep normal incidence profiles show incidence and wide-angle are first reviewed. Existing
good structural resolution in the crust and uppermost mantle. A seismic experimental methods are shown to have
sea-bottom multichannel instrument has been developed to record
datasets containing closely spaced traces, in order to improve the
certain limitations when used on the continental
resolution of reversed wide-angle experiments on the continental shelf. A new instrument is then described which has
shelf. been used to overcome these disadvantages by
The Pull-up Multichannel Array (PUMA) is a 1200 m, 12-
recording more closely-spaced data at wide angle on
channel hydrophone array for remotely recording seismic data on
the seabed. It consists of 12 short hydrophone sections linked by the seabed.
100 m-long passive sections. A pressure case is attached to the
array at one end, in which recording electronics, cassette tape
recorders and a battery power supply are housed. The PUMA is
designed for deployment in water depths less than 200 m from a Review of Existing Normal Incidence and
research ship and is moored to buoys for recovery. Wide-Angle Seismic Techniques
The instrument, which was successfully used in an experiment
west of Lewis, Outer Hebrides, UK (Powell and Sinha, 1987) was The resolution of structures within the continental
specifically designed to provide a reliable determination of the
lithosphere has greatly increased with the develop-
velocity structure of the crust and uppermost mantle over part of
the BIRPS WINCH deep normal incidence profile. Because the ment of deep normal incidence seismic profiling
traces are closely spaced it is easy to correlate phases across the (Meissner et al., 1983). Phinney (1978) estimates that
record section and to monitor changes in amplitude. A velocity the current resolution of deep reflection profiling is
structure for the continental crust and uppermost mantle has been
devised from these data, using amplitude modelling. approximately 150 m vertically and 500 m horizon-
tally, probably an order of magnitude smaller than
the uncertainties in crustal structure determined from
seismic experiments at wide angle.
Introduction
Deep reflection data are good for imaging deep
The purpose of this chapter is to show that it is structure, but give no quantitative insight into the
possible to achieve significant improvements in the seismic velocities of the deep crust because of the
resolution of conventional wide-angle seismic experi- small amount of moveout across the receiver array.
ments on the continental shelf by recording multi- Wide-angle experiments, although suffering from

Marine Geophysical Researches 12: 49-57, 1990.

© 1990 Kluwer Academic Publishers.
50 C. M. R. POWELL AND M. C. SINHA

poorer structural resolution, do provide detailed able parameter when modelling the data. A further
velocity-depth information which can be related to advantage of collecting this type of wide-angle seis-
experiments on material taken from exposures mic data is that multichannel techniques can be used
of crystalline basement for determining crustal in data processing. These include tau-p methods
composition. (Diebold and Stoffa, 1981), slant stacking and the
Information derived from seismic studies at wide- application of normal moveout corrections to study
angle and normal incidence is complementary, and reflections. All of these factors help to improve the
so there has been an increasing number of combined resolution of wide-angle seismic experiments.
normal incidence and wide-angle experiments (e.g.
Mooney and Brocher, 1986). Many of these show
Wide-Angle Data Acquisition Techniques on the
good correlations between the two datasets (Mueller,
Continental Shelf
1977; Barton et al., 1984). When both normal inci-
dence and wide-angle techniques are used at the On land it has been relatively easy to improve the
same locality, the depths of reflectors observed on resolution of wide-angle experiments by increasing
normal incidence profiles can be calculated using the number of receivers. This approach presents
velocity-depth structures determined at wide angle. problems at sea because it is not practical to deploy
Only then can the position of the Moho or other a large number of fixed recording instruments at
structures of interest be identified on normal inci- close quarters. Similarly, the cost of detonating ex-
dence records. Accurate determinations of velocity plosive charges at 100-m intervals would be pro-
are also required for migrating deep reflection data. hibitive and the operation would be beyond the
During the last fifteen years major advances have capabilities of many navigation systems. As a result,
been made in wide-angle seismic deep sounding to data from wide-angle experiments at sea still have a
improve the resolution of this technique. Interpreta- typical trace spacing of 2-4 km (e.g. Barton and
tion of wide-angle data sets has evolved from one-di- Wood, 1984).
mensional models representing two or three layers, There are, however, existing marine multichannel
to complex, laterally-varying multi-layer structures. wide-angle seismic techniques. One of these, which
This has become possible as a result of two principal recently has been used extensively, is two ship, ex-
developments. First, a new generation of inversion panding spread profiling (ESP) (Stoffa and Buhl,
techniques has been developed and become widely 1979). Here two ships steam apart on opposite head-
used (Meissner, 1986). These permit modelling not ings at constant speed. The seismic source is fired
only of travel times, but also of wave forms in the from one ship and seismic arrivals are recorded by
observed P- and S-wave data. Secondly, multi- the other, which tows a multichannel streamer (Fig.
channel data acquisition methods and experimental I). The resulting dataset is densely sampled, and
configurations are now being used which give trace from it an accurate, one-dimensional velocity-depth
spacings of 100 m or less in crustal wide-angle curve can be determined below the mid-point of the
seismic experiments. line. However, if the crust beneath the line is not
There are several advantages to obtaining such laterally homogeneous, the interpretation becomes
closely-spaced traces. First, it is much easier to pick ambiguous, since the line is not reversed. Two-
first breaks, simply because the character and posi- dimensional structure cannot be resolved because for
tion of an arrival vary only slightly from trace to an ESP, unlike conventional reversed wide-angle ex-
trace. For the same reason later arrivals can be periments, there are no ray paths crossing at low
identified reliably and correlated across a record angles. The ESP technique also presents particular
section from sub-critical reflections to wide-angle problems with static corrections if there are variable
reflections and refractions. All these phases then can thicknesses of sediment below both the source and
be used to constrain the model when the data are receiver, neither of which is fixed (Calvert, 1985).
inverted. In addition to improving travel-time infor- ESP has proved most successful in studies of the
mation, closely spaced traces also allow the variation oceanic crust, where shot-receiver offsets of 50 km or
in signal amplitude to be monitored across a record less are sufficient for observing upper mantle refrac-
section without spatial aliasing. This too is a valu- tions as first arrivals. On the continental shelf, where
 MULTICHANNEL WIDE-ANGLE SEISMIC DATA 51

a)
without the limitations of ESP techniques. The sys-
Receiving Ship
tem is an array, and therefore offers all the process-
0
ing advantages of data recorded by a multichannel
E streamer, but avoids some of the statics problems
-"

5 associated with ESPs because the PUMA is a fixed

:J:
l-
ll. receiver. Large explosive shots can also be recorded
UJ
0
10
at long range because the firing programme is not
restricted by logistical problems as it is for an ESP;
for example, there is no need to fire all the explosive
shots during a single pass of the shooting line. The
RANGE(km) paths taken by rays travelling from a given shot to
TWO SHIP EXPANDING SPREAD PROFILE each hydrophone will be different. When the data are
plotted in record section, it is assumed that lateral
variability at depth does not occur on a scale smaller
b)
0 than the length of the recording array.
E Receiver The PUMA is a 1.2 km-Iong, 12-channel hydro-
:'!
5
phone array and recording system which can be
:J:
l- deployed on the seabed in water depths of up to
ll.
UJ
0 200 m (Fig. 2). It is suitable for use in reversed,
10
wide-angle experiments, in which airgun or explosive
shots are fired at spacings of not more than one
15~--------------~~--~~~--~--~
array length, to provide a final record section with
40 30 20 10 0 10 20 30 40
trace spacings of 100 m or less from normal inci-
RANGE(km)
dence out to refracted mantle first arrival ranges.
REVERSED REFRACTION PROFILE

Instrument Description
Fig. 1. Comparison of the geometries of (a) expanding spread
profiles (ESPs) and (b) reversed wide-angle experiments using The PUMA's seismic sensors consist of 12 short
fixed receivers. Two-dimensional structure can be resolved with hydrophone sections containing piezo-electric trans-
the latter configuration, where there are many rays crossing at low ducers in an oil-filled PVC sleeve (Fig. 3). Hydro-
angles, but not with ESPs, which contain no crossing ray-paths.
phone sections are separated by 100 m-long passive
array sections. Each is an oil-filled PVC tube con-
the crust is much thicker, offsets of 120 km or more taining a steel strength member and conductors
are needed to see mantle first arrivals. At these which carry power supply to the array and seismic
ranges, repetitive airgun array sources are not suffi- signals back to the recording package. When the
ciently powerful to provide a detectable signal. It is PUMA is assembled, adjacent hydrophone and array
also difficult to prepare and fire large explosive shots section terminations are pushed together and sealed
(ideally 100 kg or more) spaced half a receiver array by '0 rings', leaving air-filled chambers in between
length ( "" 1200 m) apart while steaming at a steady for multi way connectors and hydrophone pre-am-
speed of around 6 knots. plifiers. An alloy sleeve is bolted across hydrophone
The problem of recording closely spaced wide- sections, which are weaker than the interconnecting
angle datasets on the continental shelf has been cables, to maintain the strength of the array along its
overcome by developing a fixed, multichannel sea- length (Fig. 3). The sleeve contains machined slots to
bottom receiver (Powell et al., 1986). The Pull-Up ensure good acoustic coupling between the hydro-
Multichannel Array (PUMA) was designed with the phone and the surrounding sea water.
specific objective of measuring seismic velocities in Eleven of the passive sections are identical and
the lower crust and upper mantle with a greater therefore interchangeable. In case of damage at sea,
resolution than had previously been possible in re- a faulty section could be removed on deck and the
versed refraction experiments recorded at sea, but PUMA redeployed as an 1L-channe1 system_ The
52 C. M. R. POWELL AND M. C. SINHA

PULL UP MULTI-CHANNEL ARRAY (P.U.M.A.)

Buoy

Hydrophones

/ ,,I \~
Recording
Package

12 2
100m

Fig. 2. A sketch showing the PUMA system deployed on the seabed. Twelve long passive sections separate the twelve hydrophones and
carry seismic signals to the recording package. The instrument is moored to buoys for recovery.

twelfth passive section connects the hydrophone ar- been modified for four-track recording. Each is capa-
ray to the recording package. ble of recording seven channels of frequency modu-
The recording package is contained within a cylin- lated (FM) analogue data plus clock and flutter
drical pressure vessel with flat end caps (Fig. 4). One correction channels. One cassette recorder can there-
end cap carries the connections to the array, while fore record from half the array (6 channels) while the
the other is fitted with a watertight bulkhead connec- seventh data channel records one hydrophone at
tor. This can be used for external monitoring and lower gain. Cassette tape recorders are run in pairs
resetting of the PUMA's internal clock. The record- with odd and even numbered channels being
ing package includes a battery power supply and a recorded on odd and even numbered recorders
set of eight cassette tape recorders as well as record- respectively.
ing and control electronics. The tape recorders have Recording electronics filter, amplify and frequency
modulate incoming signals and record them on to
tape. This method of recording mixed FM signals on
cassette tape was implemented on earlier seismic
instruments built at Cambridge (Duschenes et ai.,
1985; Smith and Christie, 1977; Owen and/Mason,
1982). The recorded signals are separated and de-
modulated by a separate replay system, which also
produces a signal for digitizing from the flutter
channel.
Control electronics monitor, control and activate
the recording system, including the tape recorders.
They comprise an internal clock, microprocessor
controller and relay drivers for switching the tape
recorders and recording electronics. The clock is
based on a crystal oscillator from which all timing
Fig. 3. The assembled hydrophone unit connected between two
passive array sections. The external slotted sleeve maintains me-
signals are derived. Its output is recorded on to each
chanical strength across the hydrophone module. cassette tape for timing, and.> its drift calculated by
 MULTICHANNEL WIDE·ANGLE SEISMIC DATA 53

Fig. 4. The recording package (foreground) consists of a chassis holding (from left to right) a battery power supply, 8 cassette tape
recorders and the instrument control panel and electronic circuitry. This is housed in a pressure case, which is attached at one end to the
PUMA array (left). The pressure vessel is sealed by removable end·caps bolted on at either end.

comparing it with a shipboard clock before deploy-

ment and after recovery. The microprocessor unit
(MPU), which derives its timing from second pulses
from the crystal controlled clock, remotely controls
the state of the PUMA. It is responsible for switch-
ing the instrument on and off and for activating the
correct pair of tape recorders for each pre-
determined recording window. A table of recording
windows may either be downloaded to the MPU
directly, or from a disk file on a separate micro-
computer. The system allows for up to 128 recording
windows to be programmed, each of an integral
number of 5-second periods, starting on any second.
The MPU will also perform repeated instructions of
up to 60 identical on and off cycles. A full descrip-
tion of the PUMA with system specifications is given
in Powell et al., 1986; and Powell, 1986 (a) and (b).

PUMA Deployment
Fig. 5. The 1200-m PUMA on an array winch mounted on the
The PUMA is deployed from a standard hydrophone stern of RRS Challenger. The ends of each hydrophone section
array winch mounted on the stern of a research ship are supported on the curved drum by wooden blocks. The end of
the array (here protected by a dummy end) is attached to the
(Fig. 5). Its mooring system consists of two sets of recording package before deployment. One set of mooring wires is
buoys, mooring lines, swivels and anchor weights wound on to the winch drum underneath the PUMA.
(Fig. 6), so that the instrument may be recovered
from either end, depending on tide and weather weight reaches the sea bottom before deployment of
conditions. The length of cable between the buoy the PUMA itself begins. The anchors and ground
and anchor weight is twice the water depth. The lines decouple the instrument from any noise gener-
length of wire linking the PUMA to the anchor is ated by the buoys and their moorings. Swivels are
also greater than the water depth, so that the anchor used on all the mooring wires, so that any torque
54 C. M. R. POWELL AND M. C. SINHA

PUMA MOORING SYSTEM Cr.n.e

__ Deck Winch
(al Deck Winch
3m Strop
\
C Unulzed Shackle
/ I Un •• l ud

\
" Shackle
I A \ B
Una.lud (Buoyl Mooring Line Ground line Working Wlr.
(Anchor)
Shackle

(bl Array Winch PUMA

PUMA. Pressure Vessel
H)ldrophone , 12 Ground nn_ Mooring Line Working Wire
PUMA \ Chain
\ \ 3m .Stro p \

a ........ Shaokl.
0 ········ Ring
D
( Anchor)
leU~YI I-
Un.elzed Shackle
A,,'Y Winch

D········ Swivel

Fig. 6. The PUMA Mooring System. Deployment. (a) Firstly, the toroidal buoy attached at ring A is deployed, followed by a (350-m)
mooring wire from the deck winch, the first anchor weight (Ring B), a (250-m) ground line, also from the deck winch, and the PUMA
pressure case (Ring C). The end of the PUMA is led out via a protecting chute and attached to the recording package. (b) Once the PUMA
has been laid on the seabed a second set of mooring lines, wound on to the array winch underneath the PUMA, are deployed. These begin
with a short length of chain shackled to hydrophone 12, and continue with the second (250-m) ground line, anchor weight (Ring D),
(350-m) mooring wire and buoy (Ring E). Figures in brackets indicate cable lengths if water depth is 200 m. Recovery. (c) One of the two
buoys is first retrieved. The shackle at the end of the 3-m strop is then attached to the main warp, the buoy unfastened from ring A (or
E) and the rest of the mooring system and PUMA recovered in reverse order to deployment.

produced in them under tension is relieved at the The operation is completed by re-attaching the pres-
. swivels rather than the less robust PUMA. sure case to the array and lowering the PUMA back
The weighted pressure vessel is attached to one into position.
ground line via a two-legged strop (Figs 4 and 6).
The array and recording package are connected to-
Experiment and Results
gether on deck before being deployed.
Deployment and recovery each begin at either The PUMA was first deployed in a reversed wide-
buoy in the mooring system, and then proceed se- angle experiment west of the Outer Hebrides (Fig. 7;
quentially, finishing at the second set of moorings. Powell and Sinha, 1987) in summer 1984. The aim of
However, it is usual to deploy the recording package this experiment was to determine the crustal velocity
end first (Fig. 6), and recover it last. The PUMA is structure with high resolution along the Winch line,
laid in a line by advancing the ship at low speed in a BIRPS deep reflection profile (Brewer et al., 1983).
the desired direction while unwinding the PUMA The PUMA, which was deployed at the northern
from the winch drum and through a protecting chute end of the line, successfully recorded all 99 explosive
on the after-deck into the water. The PUMA is shots. These were detonated approximately one ar-
negatively buoyant and therefore sinks to the sea ray length (1100 m) apart at ranges of 18-138 km.
floor. During deployment the array is kept slightly At closer ranges an airgun source was used. Other
taut and directly astern to avoid laying it in a recording instruments included shallow water seis-
non-linear configuration. Launching the PUMA is mometers (PUSSs; Smith and Christie, 1977) located
attempted only in reasonable weather conditions at Al and A3, and land seismometers (SCRAPs;
(less than Force 4) and at slack water, to avoid Owen and Mason, 1982) on the Isle of Lewis
damaging the array. (Fig. 7).
It is also possible to supply the recording package A 5-channel PUMA record section for explosives
with new cassettes and batteries and to reprogramme line A is shown in Fig. 8. Trace amplitudes are scaled
the MPU while leaving the array on the seabed. The for shot size and range, and traces are plotted with a
pressure case is recovered and detached on deck, and reduction velocity of 7.0 km s-'. The signal to noise
the array temporarily redeployed with a dummy end. ratio is good, and it is easy to follow phases across
 MULTICHANNEL WIDE-ANGLE SEISMIC DATA 55

S' 30'W S' W 7 3' 0' W 7' W 6 30'

' W

-
5S ' 40' N ~~======:J========~~~~------~~------~lf~~--~

5S' 20' N

5S'N ~--------~--------~~------~~~----~~~~~~--~

Explosives
-line

57'40' N ~---------l'---------~,......,.~~-'#:'';;''--+- •. _ ..... Airgun line

- - WINCH
• P.U.S . S_

'" P.U.M .A.

/ • S.C.R.A.P.
57 ' 20' N ~~ ______~__________~__________~__________________~

Fig. 7. Schematic representation of the PUMA Experiment, west of Lewis. Explosive line A is coincident with a section of the BIRPS
WINCH profile. Shots were recorded by sea-bottom instruments at Al and A3 and by land-seismometers on Lewis.

the section and to monitor changes in amplitude with a first arrival is altered by noise, by comparison with
offset. The amplitude and travel times of arrivals can nearby traces.
be determined with greater accuracy from the multi- The first arrival can be identified along the
channel dataset than from the other single channel whole record section. It reaches a maximum amplitude
data because of the similarity and close proximity of at 45-65 km range. At around 85 km its energy
adjacent traces. It is also obvious where the shape of decreases sharply, and the first arrival would be

Fig. 8. A 5-channel PUMA record section from explosive line A reduced at 7 km s - I. The traces are plotted with amplitudes scaled for
range and shot size, and have been band-pass filtered between 2 and 40 Hz. Gaps in the record section occur because the PUMA was not
aligned parallel to the explosives line during the experiment, and because four shots misfired. Note the good signal-to-noise ratio, and the
ease with which arrivals can be traced across the record section. The two most prominent phases are the first arrival from the upper and
mid-crustal layers, and the wide-angle reflection from the Moho.
 56 C. M. R. POWELL AND M. C. SINHA

65 km, and between here and 110 km both its dura-

Compressiona l Wav e Velocit y (km/s ) tion and its frequency content increase. Its maximum
5 6 7 8
amplitude occurs at 95 km. The Moho refracted
O r-------~~----------~~~ phase, Pn, is small in amplitude. It does not occur as
a first arrival on the PUMA record section, but does
S
on the data from the land seismometers on Lewis.
Shear wave phases can also be clearly identified on
10
PUMA record sections. Particularly distinct on the
c records are high amplitude shear wave reflections
from the Moho and diving waves from within the
1S crystalline crust.
E
~
r.
Travel times and amplitudes of P-wave data have
...
1i been modelled by ray-tracing (Cerveny et al., 1977)
o 20
d and using the reflectivity method (Fuchs and Muller
1981). The model which best fits the data is displayed
25 as a 1-D velocity depth curve in Fig. 9, and con-
MOho verted to two-way vertical travel time and super-
imposed on a line drawing of the BIRPS reflection
30
UM profile in Fig. 10.
The results of both experiments show much
35L---------------________ ~ __ ~
similarity; however, the Moho, which on deep
reflection profiles is often considered to lie at the
Fig. 9. Velocity-depth curve derived from 2-dimensional ampli-
base of the reflective lower crust, appears to coin-
tude modelling of explosives line data using asymptotic ray theory.
LVZ = low velocity zone; UM = upper mantle. cide with the top of a distinctive reflective zone at
8.3 s two-way travel time. Amplitude modelling of
difficult to identify beyond here without such closely PUMA data does suggest that this boundary is
spaced traces. Amplitudes increase again at around structured (Powell, 1986b).
125 km offset. The wide-angle Moho reflection can It is also significant that the PUMA recorded the
be seen beyond 50 km range reduced at a two-way first wide-angle dataset on the Hebridean shelf from
travel time of about 4 s. Its amplitude increases at which lower crustal arrivals (layers 'c' and 'd', Fig.

o
1
Scale: km 50
1

s N
A3
0T__________~~--~~~~~--~-~-~,~-~-~~~~~~~.n~T~'~~gh~~~~~~~~~~~~~~~~~~A1 ~ "'~ ... -:..---~ .ftI'"l:. .

TWTT
(s) ~======~====~==================~====~~z -- -
•••••••••••• , •••••••••••• ~ •••••••••••••••••••••••••••••••••••••••••••••••••••••• P••••••p,..."... •••• ,.... ••••••••••••••........................................
-
. . --- ~-=- - "
c , . . / "~--- --.--
•••~ ............ - ~ .u~........ ~ .....:..... ...........,;-;:....
_ _ _ _ ... ....... _ _ ~...::-- _ ... .........
5 . :......- . - •••••••• ~--=-. . ~ !!!! ••••••••••~ •••••••.:...... ....... ... .
-~--
~~
- --~...... --.,.:oo.;:::, ---
-- . . . -
- -- - ---- ""'--
-
- d '""::.....~
~~-=-
~~ " ---=---
~

- -.::::.--
-- -
-:::-
-~- .......- " '

--<"" -
.... _- - - - --....
-:::.- _-- -- MOho
..:;...... -~
5 ............. .
---- --
- ~ _ ~ -_ -- - -
~~~~=~.!!...~~ ~ ~
-
a.A..*.. _ """--::.• • • • • •

10 ./' / ...~-...;;:".". .............. ~ .

/'
" / '>- -
/'
UM --

154-____________________________________________________________________________-1
-.....---
Fig. 10. Comparison of the BIRPS WINCH and PUMA experiment data recorded at wide angle. A two-dimensional model composed of
the velocity-depth curve in Fig. 9 for the Flannan Trough has been converted to two-way travel time and superimposed on a line drawing
of the WINCH data. Dotted lines denote interfaces. The low velocity zone (LVZ) is indicated by a continuous line. The Flannan Trough
is shaded. Annotation as in Fig. 9. The Moho coincides with the top of the banded reflections at 8.3 s two-way travel time.
 MULTICHANNEL WIDE-ANGLE SEISMIC DATA 57

10) have been successfully identified and modelled Bott, M. H. P., Armour, A. R., Himsworth, E. M., Murphy, T.,
(e_g_ Jones et at-, 1984; Bott et at-, 1979)_ This can be and Wylie, G., 1979, An Explosion Seismology Investigation of
the Continental Margin West of the Hebrides, Scotland, at
attributed to the improved resolution of low ampli- 58° N, Tectonophysics 59, 217-231.
tude arrivals from the multichannel dataset Brewer, J. A., Matthews, D. H., Warner, M. R., Hall, J., Smythe,
D. K., and Whittington, R. J., 1983, BIRPS Deep Seismic
Reflection Studies of the British Caledonides, Nature (London)
Conclusions 305, 206-210.
Calvert, A. J., 1985, Seismic Studies of the Atlantic Fracture
It has been shown that a new approach was needed Zones: Charlie-Gibbs and Tydeman, Unpublished PhD Thesis,
to recording seismic data at wide angle on the conti- University of Cambridge.
Cerveny, V., Molotkov, I. A., and Psencik, I., 1977, Ray Method
nental shelf, in order to study the velocity structure in Seismology, Univ. Karlova, Praha, Czechoslovakia, 214 pp.
of the continental lithosphere in greater detail. A sea Diebold, J. B. and Stoffa, P. L., 1981, The Traveltime Equation,
bottom multichannel hydrophone array was de- tau-p Mapping and Inversion of Common Midpoint Data,
Geophysics 46, 238-254.
signed and built to meet this need_ The success of the Duschenes, J., Potts, C. G., and Rayner, M., 1985, Cambridge
PUMA instrumentation is illustrated by its perfor- Deep Ocean Geophone, Mar. Geophys. Res. 7, 455-466.
mance at sea, where it recorded 100% of the data Fuchs, K. and Muller, G., 1971, Computation of Synthetic Seis-
mograms with the Reflectivity Method and Comparison with
from explosive shots detonated on the sea floor. The
Observations, Geophys. J. R. Astron. Soc. 23,417-433.
quality of this dataset is extremely good. Noise levels Jones, E. J. W., White, R. S., Hughes, V. J., Matthews, D. H., and
are low and, for much of the data, almost every Clayton, B. R., 1984, Crustal Structure of the Continental Shelf
'wiggle' on the seismograms can be traced across a off Northwest Britain from Two-Ship Seismic Experiments,
Geophys. 49, 1605-1621.
single shot gather, and many phases are easily corre- Meissner, R., 1986, The Continental Crust: A Geophysical Ap-
lated between adjacent shots. proach. International Geophys. Series, 34, (W. L. Donn, ed.),
Indications are that, providing the instrument is Academic Press, London 423 pp.
Meissner, R., Luschen, E., and Fluh, E. R., 1983, Studies of the
continually modified to keep step with advances in Continental Crust by Near-Vertical Reflection Methods: A
technology, the PUMA, with its unique survey geo- Review, Phys. of the Earth and Planetary Interiors 31, 363-376.
metry will remain a vital tool for recording multi- Mooney, W. D. and Brocher, T. M., 1986, Coincident Seismic
Reflection IRefraction Studies of the Continental Lithosphere: A
channel data on the continental shelf.
Global Review, Presented at the Second International Sympo-
sium on Deep Seismic Reflection Profiles of the Continental
Lithosphere 15-17 July 1986, Cambridge, England, 42 pp.
Acknowledgements Mueller, S., 1977, A New Model of the Continental Crust, in The
Earth's Crust, Geophysical Monograph 20, American Geophys-
We wish to thank the master, officers and crew of ical Union, 289-317 (1. G. Heacock, ed.), Washington, DC.
RRS Discovery and the staff of NERC Research Owen, T. R. E. and Mason, M., 1982, A Cassette Recorder for
Vessel Services for their assistance at sea. A great Explosion Seismology, Internal Report, Bullard Laboratories,
University of Cambridge.
many people at Bullard Laboratories were involved
Phinney, R. A., 1978, Interpretation of Reflection Seismic Images
in various stages of this project, and their contribu- of the Lower Continental Crust, EOS., Trans. AGU 59, 389.
tions are gratefully acknowledged. The work was Powell, C. M. R., 1986a, The PUMA Manual, Internal Report,
financially supported by the Natural Environment Bullard Laboratories, University of Cambridge.
Powell, C. M. R., 1986b, A Wide-angle Multichannel Seismic
Research Council, Grant GR3/5200. University of Study of the Continental Lithosphere, Unpublished PhD Thesis,
Cambridge, Department of Earth Sciences contribu- University of Cambridge.
tion number ES 1249. Powell, C. M. R. and Sinha, M. c., 1987, The PUMA Experi-
ment West of Lewis, UK, Geophys. J. R. Astr. Soc., 89,
259-264.
References Powell, C. M. R., Sinha, M. c., Carter, P. W., and Leonard, J. R.,
1986, A Sea-Bottom Multichannel Hydrophone Array, Mar.
Barton, P. J., Matthews, D., Hall, J., and Warner, M., 1984, Geophys. Res. 8, 277-292.
Moho beneath the North Sea Compared on Normal Incidence Smith, W. A. and Christie, P. A. F., 1977, A Pull-Up Shallow
and Wide-Angle Seismic Records, Nature (London) 308, 55-57. Water Seismometer, Mar. Geophys. Res. 3, 235-250.
Barton, P. J., and Wood, R., 1984, Tectonic Evolution of the Stoffa, P. L., and Buhl, P., 1979, Two-Ship Multichannel Seismic
North Sea Basin: Crustal Stretching and Subsidence, Geophys. Experiments for Deep Crustal Studies: Expanded Spread and
J. R. Astron Soc. 79, 987-1022. Constant Offset Profiles, J. Geophys. Res. 84, 7645-7660.
An Active Source Electromagnetic Sounding System for
Marine Use

M. C. SINHA, P. D. PATEL, M. 1. UNSWORTH, T. R. E. OWEN and M. R. G. MACCORMACK

University of Cambridge, Department of Earth Sciences Laboratories, Madingley Road, Cambridge CB3 OEZ, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: electrical conductivity, instrumentation. situ a range of geological processes that occur in the
crust and uppermost mantle at oceanic spreading
centres. A property of many such processes (for
Abstract. Instrumentation has been developed for carrying out example, the injection of magma from the mantle to
active source electromagnetic sounding experiments in the deep
oceans. Experiments of this type are directly and uniquely sensi- crustal levels; the accumulation of molten or par-
tive to the presence of molten or partially molten material, to tially crystallized material in crustal magma cham-
temperature structure and to the porosity of upper crustal rocks bers; the production of new oceanic crust by
such as those that accommodate hydrothermal circulation sys-
emplacement and crystallization of this melt; and the
tems. Electromagnetic sounding experiments therefore represent
an extremely desirable addition to the existing range of activity of hydrothermal circulation systems) is that
geophysical techniques for studying geological processes in ther- they are all concerned with the presence of fluid
mally, hydrothermally or magmatically active regions-for ex-
phases within the crust and upper mantle, and the
ample, at oceanic spreading centres.
The instruments can be operated in regions of rugged, unsedi- interactions between the fluids and the surrounding
men ted sea bottom terrain, and are designed for investigating rocks.
the distribution of electrical conductivity within the oceanic Geological processes in which the activity of fluids
crust and uppermost mantle. The instrumentation consists of a
deep towed, horizontal electric dipole transmitter and a set of plays a major role are by no means limited to mid
free-fall, sea bottom, horizontal electric field recording devices. ocean ridge environments; they occur commonly in a
The transmitter is a deep-towed instrument, which is provided wide range of geological settings. However, it has
with power from the towing ship through a conducting cable.
The transmitter package is fitted with an integral echo sounder,
become evident in recent years that ridges represent
which allows it to be towed safely a short distance above the the sites of a unique and complex set of interactions:
seabed. Electromagnetic signals are transmitted from a neu- between tectonic, magmatic and hydrothermal pro-
trally-buoyant antenna array, which is streamed behind the deep
cesses; between mantle, crust and hydrosphere; and
tow.
The sea bottom receiving instruments each consist of a recov- between geothermal, geochemical, hydrological and
erable package which contains the instrumentation and digital biological systems. Many of the processes and inter-
recording system, an acoustic release unit, four low-noise, actions involved are understood poorly, if at all, and
porous electrodes arranged in two orthogonal, horizontal
dipoles, and a disposable bottom weight. a number may prove to be of substantial economic
The instruments have been used at sea on three occasions. significance (for example, the role of ridge hy-
On their most recent use, active source signals were successfully drothermal systems in the genesis of polymetallic
recorded during an experiment to investigate crustal magmatism
sulphide ore bodies).
and hydrothermal circulation beneath the axis of the East
Pacific Rise. The distribution of electrical conductivity within
the oceanic crust and upper mantle is very largely
determined by the presence, even at low concentra-
Introduction
tions, of magmatic and hydrothermal fluids, conduc-
In this chapter we describe a set of instrumentation tivities of which can be several orders of magnitude
developed to carry out active source electromagnetic greater than those of the surrounding host rock; and
sounding experiments in the deep oceans. The pri- by temperature. Other physical parameters accessible
mary incentive for doing this is a desire to study in to geophysical measurements, such as seismic veloc-

Marine Geophysical Researches 12: 59-68, 1990.

© 1990 Kluwer Academic Publishers.
60 M. C. SINHA ET AL.

ities or density, exhibit much weaker dependence on A second and entirely different technique, known
these factors. Experiments which allow reliable de- as active-source electromagnetic sounding, has been
terminations of electrical conductivity structure are described by Chave and Cox (1982), and Young
therefore directly and uniquely sensitive to the pres- and Cox (1981). In this method, which provides
ence of fluid phases and temperature anomalies, and probably the best approach to the problem of how
are therefore potentially a more fruitful source of to obtain detailed conductivity information from
structural information about fluid or thermally re- the deeper parts of the oceanic crust and from
lated processes than other marine geophysical meth- the sub-oceanic upper mantle, horizontal electric
ods. The development of practical, electrical or dipoles are used for both the source and the receiv-
electromagnetic sounding techniques that can be ap- ers. The technique is a frequency-domain sounding
plied in the deep oceans is therefore a highly desir- system, which relies on the variations in the re-
able goal for the study of these processes, either at sponse of the earth to electromagnetic signals of
mid ocean ridges or elsewhere. different frequencies for determinations of conduc-
The most commonly used technique for deep elec- tivity structure.
trical sounding on land is the magneto-telluric (MT) This chapter describes a newly developed instru-
method. However, this technique suffers from a seri- mentation system for active source electromagnetic
ous limitation in the deep oceans. This is that the sounding of the oceanic lithosphere which is based
presence of several kilometres of highly conductive on the principles described by Chave and Cox, and
sea water effectively screens the seabed from the which has recently been developed at Cambridge
ionospheric sources of MT signals at all frequencies University.
higher than a few cycles per hour. The very long
period signals that do penetrate to the seabed can be
used for large-scale sounding of the sub-oceanic
Active-Source Electromagnetic Sounding
mantle (e.g. Filloux, 1981), but are so severely lim-
ited in terms of depth resolution that they are quite The instrumentation system developed at Cam-
unsuitable for intra-crustal investigations. bridge for electromagnetic sounding consists of
An alternative approach to MT sounding is to two main elements (Fig. 1). The first of these is a
generate artificial source currents in the water Deep-towed Active Source Instrument (DASI).
column, and to measure the resulting electric or DASI is towed from a surface ship on a conducting
magnetic fields at the seabed. This approach makes it cable, and transmits electromagnetic signals at ex-
possible to exploit the high conductivity of sea water, tremely low frequencies (0.0625-256 Hz) into the
both by using it as the return path for the artificial seabed. The second element is one or more Low-
source currents and by taking advantage of the frequency, Electro-Magnetic Underwater Recorders
electrically quiet environment at the seabed which its (LEMURs). LEMURs are free-fall instruments
screening properties provide. which are deployed on the seabed at an appropriate
One successful example of an artificial source tech- range (typically a few kilometres) from the trans-
nique for electrical sounding of the upper oceanic mitter, and used to record the resultant oscillating
crust has been described by Edwards et al. (1981). electric fields.
Their technique, known as MOSES (Magnetometric In the. active source electromagnetic sounding
Off-Shore Electrical Sounding), involves the use of a method, the earth's conductivity structure is deter-
vertical transmitter array and of magnetometers for mined by measuring its response to radiation over a
measuring the resultant fields at the seabed. It is wide range of frequencies. The deep towed source
essentially a galvanic resistivity method, in which a instrument generates oscillating electric and magnetic
long vertical transmitter with its lower electrode fields close to the seabed, by passing a large, alter-
close to the seabed is used to inject current into nating electric current between a pair of electrodes
sub-sea bottom formations. Changes in the magnetic mounted in an antenna streamer which is towed
field at the seabed due to reversing the source current horizontally behind the instrument package. The
are measured by vertical component, sea bottom return path between the transmitter electrodes is
magnetometers. provided by sea water. Towing the transmitter
 ACTIVE SOURCE ELECTROMAGNETIC SOUNDING 61

Research Vessel

Water dept h
up to 6 km
Low-Frequency Electro Magnetic
Underwater Receiver

Floatat ion
Deep Towed AcoustiC Release
Aetive Source Transmitter Array Instrument
Instrument -..-~ ~_ Package
Cable Cutter

Electrode Array

-4-- - Source - Receiver Range up to 10 km

Fig. I. Schematic arrangement of an active source electromagnetic sounding experiment, showing the deep-towed active source instrument,
the surface research vessel and a low-frequency electromagnetic underwater recorder (not to scale).

antenna sufficiently close to the seabed ensures that Deep-Towed Active Source Instrument (DASI)
the electric and magnetic fields are well coupled into
The deep towed active source instrument, or DASI,
the sub-bottom rocks.
constitutes the transmitter end of the Cambridge
Because of the high conductivity of sea water,
radiation from the transmitter is very rapidly atten- electromagnetic sounding system (Fig. 2). It consists
uated in the water column. However, in the rela- of three major sub-systems: the ship-board power
tively resistive seabed, the signal is able to supply unit, control unit and echo-sounder display
propagate for much greater distances. By placing (Fig. 3); the deep-towed instrument package (Fig.
suitable recording instruments on the seabed at an 4); and the transmitter antenna array. The deep-
appropriate distance (typically a few kilometres for towed package is connected electrically and me-
oceanic crustal studies), the attenuation and phase chanically to the ship-board systems by means of an
shift of the transmitted signals can be measured, armoured, co-axial cable.
and related to sub-bottom conductivity. Since the The ship-board power supply unit converts the
depth of penetration of the transmitted signal into ship's 3-phase mains supply (440 V, 50 Hz, unstabi-
the seabed depends on its frequency, signals at de- lized) to a variable, regulated, frequency stabilized
creasing frequencies contain information about the supply at up to 2000 V RMS and 5 A RMS. The
conductivity of the earth to increasing depths below frequency of the output supply is 256 Hz, and this
the sea floor. Thus by making measurements at is tied to a frequency standard derived from the
suitable ranges and over an appropriate set of fre- ship's master clock system. This stabilized, high-
quencies, it is possible to derive information about voltage output is fed to the deep tow cable to
the variation of sub-bottom conductivity with power the deep-towed transmitter.
depth. Use of several receivers and numerous source At the deep tow, the power received from the
positions can obviously provide information about ship is first transformed down to a lower voltage,
lateral, as well as vertical, variations in sub-bottom and then switched by a semiconductor bridge to
conductivity. provide a pattern of half-cycles of either positive or
62 M. C. SINHA ET AL.

Deep Tow Cable

Conducting Swivel - _...I.

Stabilising Fins

Electronics Package
Antenna Array
Towing Bridie Floatation Hose

Antenna Cable

Transformer & Switching Package

Connector Near Electrode Cable to Far Electrode

(
; I
/
I Far Electrode Tall Rope
I

Fig. 2. The deep-towed active source instrument, DASI. The instrument package is suspended from the surface ship by an armoured
conducting cable. Two pressure vessels house the active source signal generator, the echo sounder transceiver and the associated control and
monitoring equipment. The l80-m long, neutrally buoyant antenna array is streamed horizontally behind the deep-tow package.

negative polarity to the transmitter antenna array. is simultaneously carrying the high voltage, 256 Hz
The maximum output current to the antenna is power supply. Data from the echo sounder are de-
200 A RMS. By controlling the switching to provide modulated at the towing ship and displayed on a
blocks of appropriate numbers of half-cycles of each graphic recorder. A winch operator on the ship uses
polarity to the antenna, pseudo-square waves of any the display to maintain the deep tow as closely as
desired frequency that is an integer factor of 256 Hz possible at a fixed height above the seabed, by
can be synthesized. Obviously, since the transmitted veering or heaving in cable. Since the echo sounder
frequency is derived directly by dividing the power transducer is effectively omni-directional, a reflection
supply frequency in this way, the transmitted fre- from the sea surface is usually seen, as well as that
quency has the same stability characteristics as the from the seabed. Consequently, the echo sounder
power supply unit\and the ship-board clock system. indicates the depth of the deep tow beneath the
DASI's height above the seabed is monitored by surface, as well as its height above the seabed.
means of an integral echo-sounder. The frequency of Local control of the deep-tow instrument is per-
this is selectable at either 3.5 or 7 kHz. The lower formed by a microprocessor unit. The microproces-
frequency has the advantage that it provides signifi- sor controls the switching of the semiconductor
cant penetration in thinly sedimented areas. It is bridge in the electromagnetic transmitter, and also
therefore possible to determine the instrument's the pulse length, frequency and receiver gain of the
height above igneous basement, as well as its height echo sounder. The microprocessor also performs
above the seabed, in such areas. The signals from the monitoring and diagnostic tasks via a multiplexed,
echo sounder are frequency modulated on a 50-kHz analogue to digital interface. This allows it to mea-
carrier, and telemetered to the towing ship along the sure the time-averaged voltage and current at the
deep-tow cable. Tuned coupling circuits at the ship antenna; the instantaneous current and voltage, in a
and deep-tow end enable the carrier signal to be high-frequency sampling mode which logs the wave
injected into the cable and retrieved from it while it form of the transmitter and stores it in memory,
 ACfIVE SOURCE ELECTROMAGNETIC SOUNDING 63

VDU FSK 50 KHz Carrier

Terminal
rr-_-VI Encoder/K~============:::;l
Decoder

EPC
ecorder AM
Echo
IA--~ Sounder
Demod.
K;::============:::::50 KHz CarrIer

PSU EMU 2000 V SA 256 Hz AC

Control Pow er ~1imi1imi1imi1imiffil1imi1imiffilffil~ffil~ I FII t e r
Unit Circuit.
Supply
Unit

Winch
Control

Fig. 3. EMU Transmitter-shipboard systems for the deep-towed transmitter. These systems provide a stabilized, high-voltage power
supply to the deep tow; serial communications with its microprocessor controller; and display of data from its 3.5/7.0 kHz echo sounder.

for subsequent transmission to the towing ship; the with a VDU terminal at the surface ship through the
temperature of the cooling oil in the pressure vessel deep tow cable by using a serial communications link
housing the transmitter bridge and transformer; and running via a frequency-shift keyed (FSK) carrier
the state of the instrument's various battery packs. In centred on 50 kHz. The FSK link shares this carrier
addition, a moisture-detection circuit continually frequency with the echo sounder data. An additional
tests for high impedance between a pair of electrodes task of the microprocessor is to organize time sharing
located at the bottom of the oil-filled pressure vessel, of the carrier between outgoing, analogue echo
in case of seepage of sea water into the system. sounder data, outgoing FSK data and incoming FSK
The microprocessor in the deep tow communicates commands from the operator's terminal on the ship.

Anlenna
power Coupllngl Step- EMU

,
Down Wave -
Filter Form

.
Trans-
Ira m ship CIrcuIts former jGene ..... o
",. r-
h
'"
Cable • FSK IA. J.
Mlcro-
ADC
Encoderl .' Processor
Driver --y Unit
Oecoder !contro".rl""
"'\ >- '"
Echo- Echo-
lA. Sounder
Sounder Trans.- Clock
Modulator cefver

8,,"·,,"'"'
Fig. 4. Block diagram of the deep-towed active source instrument. The step down transformer and wave form generator convert power
supplied from the ship and feed it to the antenna array. The remaining electronic systems provide an acoustic indication of the instrument's
position in the water column, and control and monitoring functions.
64 M. C. SINHA ET AL.

Software running on the microprocessor allows for DASI's antenna array consists of a pair of heavy-
simple control of the deep tow by selection of com- duty, underwater cables connecting the deep-tow
mand options from menus displayed on the VDU package to two electrodes. The electrodes are made
screen. Once a particular mode of operation has been from 6-m lengths of 12-mm diameter steel wire rope.
selected and initiated, no further operator interven- The near electrode is towed 30 m behind the deep tow,
tion is required until one or more operational parame- the far electrode another 100 m behind that. The
ters need changing. In the meantime, regular status array is supported by a number of 30-m lengths of
reports (induding data from the various analogue to 63-mm diameter, flexible PVC hose filled with low
digital interface channels) are sent to the console by density, low compressibility oil. Neutral buoyancy of
the microprocessor at 2-minute intervals. In practice, the streamer is achieved by adjusting the volume of oil
a personal computer is used as the shipboard console, in each buoyancy section, and by distributing small
so that data and commands sent between the deep lead weights along the array. A 50-m long, polypropy-
tow and the console can be logged to magnetic disc, lene tail rope terminating in a small drogue provides
providing a permanent record of deep tow operations. the streamer with additional towing stability.
The antenna is deployed from a large-drummed
winch, identical to those used for handling seismic
streamers. Once the array is in the water, its inboard
end can be connected electrically and mechanically to
the deep-tow package. The entire deep-tow system
can then be deployed relatively easily over the stern
of the towing ship (Fig. 5). The deep tow cable is
terminated by a deep water conducting swivel system,
fitted immediately above the deep tow package, to
minimize torsional stresses in the cable.

Low-Frequency, Electromagnetic, Underwater

Recorder (LEMUR)
The receiving instruments, or LEMURs, are free fall,
ocean bottom instrument packages which digitally
record changes in horizontal electric field across an
orthogonal pair of electric dipole antennas. Each
LEMUR consists of: an electronics, instrumentation
and recording package; an acoustic release unit; a
mechanical release and cable cutter assembly; a dis-
posable bottom weight; a set of disposable, porous
electrodes and electrode deploying arms; and a buoy-
ancy package.
When deployed from the ship, the LEMUR is
configured as a vertical string, with the buoyancy
package at the top, the instrumentation and acoustic
release units suspended beneath it, the mechanical
release and cable cutter below that, and the bottom
Fig. 5. Photograph of the deep-tow system while it is being
weight and electrode array assembly at the bottom
deployed. The conducting swivel which terminates the deep-tow (Fig. 6). The instrument sinks to the seabed, where
cable is positioned immediately above the instrument. The elec- after the bottom weight has landed the electrode
trode array is visible in the background, suspended from the
deploying arms fold down into a horizontal position,
instrument package and streaming astern of the ship just below
the sea surface. The array storage and handling winch can be seen leaving the electrodes in a cross configuration of two
in the right foreground. orthogonal, 13.5-m dipoles.
 ACTIVE SOURCE ELECTROMAGNETIC SOUNDING 65

tion under a partial vacuum to remove air bubbles,

and then electrolysed so that a uniform layer of silver
Flag
chloride forms on the outer surface of the silver foil.
This process makes low noise, low impedance elec-
trodes which are sufficiently robust to withstand
operations in a marine environment. However, once
wet and electrolysed, the electrodes have a relatively
short storage life of 6 to 8 weeks, so the final stages
of preparation are carried out at sea immediately
before use.
Acoustic
Release
Unit __
Each electrode is fitted into the end of a 6-m long
- Instrument P.cklge
deploying arm, made of polypropylene tube. The end
of the tube, where the electrode is housed, is pierced
by a pattern of holes to allow good electrical contact

Mechanical
R.I ....
II between the electrode and sea water. The inboard
end of each arm is attached by a hinge to the bottom
weight. The hinge has a restricted travel, allowing the
arm to move in a vertical plane from 10° below
horizontal to 70° above horizontal. This movement

, allows the arms to fold upwards during the instru-

ment's descent through the water column. Solid-glass
==#~~~~~~~=~6~m~e~
tr.~ r{ Electrode yo,
Arm
weights placed in the ends of the arms close to the
electrodes ensure that once the instrument reaches
BoUom Weight Ag/ AgCI
Electrode the bottom, the arms lie flat on the seabed.
The bottom weight itself consists of a rectangular
Fig. 6. Schematic arrangement of a low-frequency, electromag- steel frame with sides of 1.4 and 1.56 m weighing
netic underwater recorder (LEMUR). The instrument package
65 kg. A set of rope strops attach it to the bottom of
and acoustic release unit are mounted on a stainless steel frame
and suspended beneath the floatation package. The bottom weight the cable cutter assembly in an arrangement that
and electrode array assembly are connected to the instrument maintains the instrument package above it in a fixed
frame by the mechanical release and cable cutter units. orientation relative to the electrode array.
The cable cutter is a guillotine, actuated by the
mechanical release unit of the acoustic command
At the end of an experiment, the instrument is system. The mechanical release breaks the load-
recovered by acoustically activating the release unit. bearing connection between the instrument package
The acoustic release system operates the mechanical and the body of the cutter unit. The cutter then falls
release and cable cutter, severing all connections rapidly towards the sea floor, driven by a 4-kg
between the instrument package and the bottom weight attached to it. The guillotine blade remains
weight. The instrument package, still attached to the attached to the instrument package by a short length
buoyancy, then floats back to the surface. of stainless steel wire. When the cutter reaches the
The LEMUR's electrical sensors are a set of four end of the wire's travel, the blade is pulled out of the
low noise, porous, silver-silver chloride electrodes cutter body, severing the four electrical cables which
which are deployed in a horizontal array around the connect the electrodes to the recording package. The
bottom weight. The electrodes are based on a design cutter blade returns to the surface with the instru-
developed at Scripps (Webb et aI., 1985). A silver foil ment package, but the cutter body remains on the
is mounted on a cylindrical PVC former, and sur- seabed.
rounded by a mixture of Kieselguhr (diatomaceous The acoustic release system is the standard unit
silica) and silver chloride. The mixture is held in made by the Institute of Oceanographic Sciences,
place by an external sleeve of porous polyethylene. Deacon Laboratory (Phillips, 1981). It operates at a
The electrode is immersed in sodium chloride solu- centre frequency of 10 kHz, and makes use of the
66 M. C. SINHA ET AL.

Pad
RAM

Acoustic
R.I •••• Time
Unit Clock

4 Olglta I Tape Recorders

Fig. 7. Low-frequency Underwater Electromagnetic Recorder (LEMUR). Block diagram of the LEMUR electronics and digital recording
system (see text for a description).

ship's precision echo sounder system for graphical four cassette drives, totalling approximately
display of the range and status of the release unit. 20 MBytes of data storage. The data can be syn-
The instrument's buoyancy is provided by three chronously stacked as they are recorded. Since in
0.43-m diameter glass spheres, mounted on an alu- general, the frequency content of the active source
minium frame. A mast and flag and a pressure signal is known in advance, selection of suitable
deactivated flashing light are fitted to the buoyancy stacking parameters allows the collection of a great
unit to aid recovery of the instrument by day or deal more data than would otherwise be possible,
night. without significantly degrading the data quality. Au-
The LEMUR instrument package is contained in a tomatic gain control software dynamically adjusts
cylindrical pressure vessel, capable of operating at the gains of the amplifiers, to maintain maximum
water depths of up to 6 km. In the instrument pack- dynamic range of recording without saturating the
age, voltage signals from two data channels, corre- analogue to digital converters.
sponding to opposite pairs of electrodes, are All aspects of the recording system, including the
amplified, frequency filtered, digitized, stacked and sampling rate and resolution of the analogue to
digitally recorded on magnetic tape (Fig. 7). digital converters, the fold of stack, the automatic
Each electrode is coupled via an electrolytic capac- gain control and the cut-off frequencies of the high
itor to one side of a differential preamplifier. At the and low pass filters, are under the control of a
preamplifier, the voltage between electrodes is microprocessor unit. Data are recorded on tape in
chopped by a field effect transistor bridge, and then discrete recording windows, each representing 16
coupled through a step-up transformer to a high-gain kilobytes of data from each channel. Data tables
amplifier which is tuned to the chopping frequency. containing the start times and recording parameters
After amplification, the chopped signal is demodu- for up to 620 windows are stored in random access
lated, frequency filtered, amplified further and then memory. The microprocessor continually scans these
digitized. The analogue to digital converter is based tables, and initiates recording with the appropriate
on a voltage-controlled oscillator and counter-timer. settings at the beginning of each window. All of the
This forms a 16 bit, time-integrating digitizing system, parameters controlled by the microprocessor and
which has better signal-to-noise characteristics than mentioned above can vary from window to window,
sample-and-hold types of converter. allowing great versatility in the instrument's overall
The digitized signals from both channels are stored recording programme. The data tables can either be
in random access memory, and subsequently trans- prepared directly on the microprocessor, or more
ferred to magnetic tape. The digital tape system uses conveniently they can be prepared in advance on a
 ACTIVE SOURCE ELECTROMAGNETIC SOUNDING 67

personal computer and downloaded to the instru- Valu Fa Ridge, a back-arc spreading centre located
ment through its serial communications port. in the southeastern part of the basin. Again, a num-
Calibration circuits are built into the preamplifier ber of instrumental problems affected the electro-
stage, and can be activated by the microprocessor. magnetic sounding equipment, and we were not able
The calibration circuits allow step functions of to collect an interpretable electromagnetic data set.
voltage or current to be applied individually to either The cruise did, however, act as a useful proving
sensor channel. In this way, the calibration circuits ground for the modifications made to the instru-
provide measurements of the sensor impedance and ments since the first cruise, and provided experience
of cross-talk between channels, as well as of the of operating them over the difficult terrain of a
system gain and sensitivity. spreading centre ...
Orientation of the instrument is achieved by a During the 1988 cruise, we were able to operate
photographically recording compass system which is the transmitter close to the seabed at antenna cur-
located in the instrument package. A photographic rents of up to 120 A RMS for a total of 16 hours.
image of a compass needle is recorded on film during Electric field data collected by LEMURs deployed
data recording. Since the orientation of the recording on the Valu Fa Ridge indicated RMS noise levels as
package is fixed relative to the bottom weight, this low as 1O-7-1O- 8 Vm- 1 in the frequency band
allows the orientation of the receiver electrode array 0.0625-8 Hz. The noise is dominated by low fre-
relative to magnetic north to be determined. quencies, and represents a combination of environ-
The design of the LEMURs draws to a large mental background noise, which predominates at the
extent on that of the Cambridge Digital Ocean Bot- lowest frequencies, and electrode noise, which pre-
tom Seismometer (Owen and Barton, 1989). The dominates at higher frequencies. The levels are con-
recording package pressure case, the cable cutter, the sistent with measurements of sea bottom electric
microprocessor unit, the digital tape recording sys- fields reported by Webb et al. (1985).
tem and the buoyancy package make use of the The combination of sustainable transmitter power
same, or identical, hardware. Use of this existing and low receiver noise levels achieved during the
technology has made possible the development of a cruise gave us confidence that, provided the remain-
sophisticated sea bottom instrument within the re- ing instrumental problems could be overcome,
sources of a relatively small university research the equipment would be capable of generating data
group in a short period of time. that would provide useful information about crustal
conductivity structure beneath an active ridge
system.
Experience of Use
After further changes, the equipment was used at
To date, the equipment described here has been used sea for a third time on the East Pacific Rise in May
at sea three times. The first occasion was in Septem- and June 1989. Improvements to the instrumenta-
ber 1977, during an instrument test cruise in the tion for this cruise included modifications of the
north Atlantic. Experience on that cruise brought to LEMUR bottom weight and electrode arm assem-
light a number of problems, among the more serious blies, to allow them to be deployed more reliably on
of which was the difficulty encountered in telemeter- areas of unsedimented, ridge axis sea floor.
ing echo sounder and serial digital data through the During the cruise, an electromagnetic sounding
deep-tow cable while simultaneously using it to sup- experiment was successfully carried out on the axis
ply power to the electromagnetic transmitter. How- of the East Pacific Rise at 13°15' N. The experiment
ever, the cruise did confirm that the equipment was was performed in collaboration with the active
being developed along broadly correct lines. source electromagnetic sounding group at Scripps
The instruments were used for a second time, after Institution of Oceanography. Four Cambridge
modifications and improvements, in July and August LEMURs and four similar receiving instruments
1988. This was during a cruise to the Lau Basin, a from SIO were deployed in two lO-km long lines,
back-arc, marginal basin in the western Pacific. The one along the ridge axis and one parallel to it but
objective was to carry out a combined programme of offset by 5 km to the east. The Cambridge deep
seismic and electromagnetic sounding studies of the towed transmitter was towed repeatedly along each
68 M. C. SINHA ET AL.

line in tum for a total of two and a half days, ous types will play an increasingly important role in
transmitting signals at between 0.125 and 8 Hz. future geophysical investigations of the sub-sea-
Having at last solved the problem of telemetering bottom lithosphere.
echo sounder signals up the deep-tow cable while
sending power down it, we were able to maintain
Acknowledgements
DASI's height above the seabed at 30 ± 10 m. With
the aid of acoustic navigation, it proved possible to We are indebted to the members of the active-source
manoeuvre DASI along the desired tracks to within sounding group at Scripps Institution of Oceanogra-
50 m laterally for much of the time, and to within phy, and especially to Chip Cox, Steve Constable and
200 m for virtually all of the time. Tom Deaton, for their encouragement of our endeav-
At the time of writing, the data from the 1989 cruise ours, for much useful advice, and for making freely
have yet to be processed and interpreted. However, available to us many details of their instrumenta-
the experiment represents the first successful electro- tion-much of which we have simply copied. We also
magnetic sounding of an oceanic spreading centre, acknowledge the contributions of John Leonard, Pe-
and as such its significance lies not only in the results ter Carter, Melvyn Mason and Roger Theobald, who
that will accrue from it but also in the demonstration carried out much of the work on instrument develop-
that such experiments are possible. ment and construction; the officers and crews of the
research vessels Discovery and Charles Darwin, and
the staff of NERC Research Vessel Services, Barry,
Conclusions
who made possible the work at sea; and Phyl Fisher,
The electromagnetic sounding instrumentation de- who prepared the figures. The work was financially
scribed in this chapter represents one of a very limited supported by the Natural Environment Research
number of systems capable of exploring the conduc- Council, under grants GR3/5851, GR3/6671 and
tivity structure of the sub-oceanic crust and upper- GR3/6673.
most mantle. The development of a deep-towed University of Cambridge, Department of Earth
transmitter that can be towed reliably a short distance Sciences contribution number ES 1399.
above the seabed, and which uses a neutrally buoyant
streamer for its antenna, means that for the first time
it is possible to generate high power, low frequency, References
electromagnetic signals over unsedimented crust in Chave, A. D. and Cox, C. S., 1982, Controlled Electromagnetic
areas of rugged sea bottom topography, such as exist Sources for Measuring Electrical Conductivity beneath the
Oceans-I, Forward Problem and Model Study, J. Geophys.
at mid ocean ridges. Our experience and that of other
Res. 87,5327-5338.
groups indicates that it is possible to exploit such Edwards, R. N., Law, L. K., and DeLaurier, J. M., 1981, On
signals successfully for investigating fluid-controlled Measuring the Electrical Conductivity of the Oceanic Crust by
a Modified Magnetometric Resistivity Method, J. Geophys. Res.
processes in the seabed beneath the deep oceans.
86,11,609-11,615.
This chapter cannot be a final or definitive descrip- Filloux, J. H., 1981, Magnetotelluric Exploration of the North
tion of the Cambridge electromagnetic sounding sys- Pacific: Progress Report and Preliminary Soundings near a
tem since the instruments are under continual Spreading Ridge, Phys. Earth Planet Inter. 25, 187-195.
Owen, T. R. E. and Barton, P. J., 1989, The Cambridge Digital
development. However, our aim has been to demon- Seismic Recorder for Land and Marine Use, Tectonophys. (in
strate that it has proved possible to construct instru- press).
ments suitable for electromagnetic sounding of Phillips, G. R. J., 1981, The I. O. S. Acoustic Command and
Monitoring System, Report No. 96, Institute of Oceanographic
tectonically active, unsedimentented areas-where Sciences (Deacon Laboratory), Wormley, Surrey.
technical limitations have previously prevented such Webb, S. c., Constable, S. C., Cox, C. S., and Deaton, T. K., 1985,
work from being carried out. The potential scientific A Sea Floor Electric Field Instrument, J. Geomag. Geoelectr. 37,
1115-1129.
and commercial benefits from such investigations are
Young, P. D. and Cox, C. S., 1981, Electromagnetic Active Source
enormous, and it seems likely that the development Sounding near the East Pacific Rise, Geophys. Res. Lett. 8,
and use of electromagnetic sounding systems of vari 1043-1046.
Long-Range Underwater Photography in the Deep Ocean

Q. HUGGETI
Institute of Oceanographic Sciences, Deacon Laboratory, Brook Road, Wormley, Godalming, Surrey, GUS SUB, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: Photography, cameras, television, film, surveying, Until the mid 1970s underwater photography in
visibility, light sources. the deep ocean was carried out mostly at close range
to the sea floor (Heezen and Hollister, 1971). This
Abstract. Although the optical properties of seawater at extreme
was done either in a random "pogo-stick" fashion
depths are more suitable for underwater photography than those
at the surface or on continental shelves, they still impose severe (Laughton, 1957), in stereo (e.g. Ohta, 1983), using a
limitations on 'long-range' wide area bottom photography. Addi- "pop-up" time-lapse system (Lampitt and Burnham,
tionally, deep ocean operations impose technical limitations on 1983), from a submersible (Ballard and Moore,
control, power and bandwidth. This chapter reviews the ap-
proaches contemplated or made towards improving the camera-
1977) or using a sledge which remains in contact with
to-target range in underwater photography in the deep ocean. the sea floor (Aldred et al., 1976; Chardy et al., 1980)
Further significant improvements await advances in control, (Fig. 1). Though even at such close ranges good
power/light sources and bandwidth reduction. With the develop-
ments now contemplated, TV and video systems will eventually
sharp photographs can be obtained, they cannot
present a strong challenge to emulsion film techniques. cover sufficient ground to give a truly representative
picture of the sea floor. These systems are particu-
larly useful for detailed studies of the benthos (Rice
Introduction et al., 1979; Rice and Collins, 1985) and in the
identification of small objects seen in long-range
Since William Thompson's production of the first
photographs (Fig. 2). However, if we wish to map
underwater photograph in 1856 (Brown, 1985) sev-
some of the larger-scale benthic processes, wide-
eral technological advances have made underwater
angle or long-range photography is essential.
photography an important tool for use in ocean
The desire to view larger areas of the deep ocean
surveying. Rebikoff (1967) presents a succinct his-
floor has led us into two principal areas of difficulty
tory of underwater photography and traces the de-
in underwater photography, those of light scattering
velopment of underwater photography from Louis
and data retrieval. These two topics will be discussed
Boutan's early work with glass plate cameras en-
with reference to recent developments in deep ocean
closed in brass and iron pressure cases through to the
underwater photography.
wide-angle lenses of the 1960s. A useful bibliography
on advances in underwater photography up to 1968
was published by the Kodak company (Anon. 1968).
Light Scattering
As early as 1899, when Louis Boutan lowered his
camera system to a depth of 50 m, he had identified Since the early days of underwater photography the
the four main hurdles to underwater photography: problems of light absorption and scattering in sea-
water have been paramount. In the absence of ambi-
- Scattering of light by sea water;
ent (natural) light, and at short ranges from the
- Data retrieval;
target, colour balancing can be achieved by using
- Need for pressure-resistant housings;
broad band light sources and suitable filters. This
- Provision of adequate power sources.
was amply illustrated by Mertens (1970), who com-
The first two of these problems have yet to be pared underwater photographs, illuminated by day-
overcome and continue to plague us today. light, taken with and without colour filters and at

Marine Geophysical Researches 12: 69-81, 1990.

© 1990 Kluwer Academic Publishers.
70 Q. HUGGETT

CAM ERA

Fig. I. Illustrations of (A) a 'Pogo-Stick' camera, and (8) the I.O.S. Epibenthic Sledge (after Aldred e/ al., 1976).

different depths. At longer camera-to-subject ranges are well tried and tested. They have also produced a
(i.e. > 5 m) light absorption by the seawater renders wealth of good-quality photographs from the world's
much of the visible waveband unuseable for under- oceans (e.g. Ballard and Moore, 1977; Phillips et al.,
water photography. Figure 3 illustrates the effective 1979; Huggett, 1987). In order to further enhance
absorption distances for the various colour bands. image definition or camera-to-target ranges, increas-
Not only is much of the visible waveband strongly ingly complex (and expensive!) techniques can be
attenuated and redundant as a source of illumina- used to reduce backscatter; these are:
tion, but it can also scatter close to the camera. Thus
3. Light polarization;
light which does not reach the sea floor can still
4. Monochromatic light sources;
contribute backscattered noise to the image.
5. Range gating;
Over the years, two main approaches have been
6. Synchronous scanning.
adopted to ameliorate the effects of backscattering.
Both involve the geometric separation of source and These six techniques will be described in order of
receiver; they are: relative cost and complexity and their merits dis-
cussed.
1. Horizontal camera-to-light separation;
2. Light Behind Camera (LIBEC). HORIZONTAL SEPARATION

Both these approaches have the advantage of cost This approach takes advantage of the attenuation
effectiveness as all the components in the system are properties of the sea by increasing the distance be-
readily available (see Anon. 1986; Anon. 1987) and tween camera and light source to at least 2 m. This
 LONGE-RANGE UNDERWATER PHOTOGRAPHY 71

Fig. 2. Two underwater photographs of the same field of manganese nodules taken from (A) the Epibenthic Sledge and (B) the Wide Area
Survey Photography (WASP) system to illustrate their differing resolution and scales.
72 Q. HUGGETI

Ol
C "-.
~
C 0::1-
WI
0 ::::E<.!>
E
Q)
<l: ::J
U
Om
~

-+-'
..r::
Ol
-.J

Q)
Ol
c
5m 0
~
-+-'
Q)
Ol
L
0
t-
o
-+-'

0
10m L
Q)

E
0
u

15m
GREEN

Fig. 3. Diagram illustrating the absorption of light in water. The absorption has been calculated for two-way travel. This illustrates the
'transmission window' in the blue-green waveband.

moves the volume of water common to both light camera from the sea floor every 2 s. This informa-
and camera away from the camera so discriminating tion is sent both to the camera, where it is printed on
against the highly attenuated wavelengths which the appropriate negative, and to the ship (via a
produce negligible image. Although this technique 10 kHz pinger), for the winch operator to control the
does not eliminate all of the back scattered light, at camera height. The system is programmed to take
least some of the wavebands which only produce photographs when it is within 8 and 18 m of the sea
noise are eliminated by the increased light-to-camera floor.
separation. As an indication of the popularity of this ap-
Figure 4 illustrates the Wide Area Survey Photog- proach to backscatter reduction, several camera sys-
raphy (WASP) system developed at the Institute of tems of this type exist, for example:
Oceanographic Sciences (UK) which is typical of
systems using horizontal separation. It was designed - The ANGUS system, operated by Woods
along the lines of the ANGUS system (Phillips et al., Hole Oceanographic Institution (Phillips et
1979) and uses a still camera with a 37 mm lens and al., 1979).
has a 1600 frame capacity. Light comes from 6 - The RAIE system, operated by CNEXO
Vivitar flashguns powered by NiCad battery stacks (Mauviel, 1982).
capable of at least 2000 discharges per run. A high- - The USGS Camera system, operated by
frequency echo sounder measures the height of the USGS, Menlo Park. (Chezar and Lee, 1985).
 LONGE-RANGE UNDERWATER PHOTOGRAPHY 73

10kHz Pinger
(to ship)

Flash Units

35 kHz Altimeter

Fig. 4. Wide Area Survey Photography (WASP) system.

In addition to still camera, the USGS Cam-

LIGHTS
era system has a video camera and recorder.
ODD
- The ARGO system, operated by Woods Hole
Oceanographic Institution (Harris and Bal-
i Zone

lard, 1987). This has a TV system in addition 10m

to the still camera.

LIBEC
This was developed at the US Naval Research Labo-
ratory (Patterson 1972). Figure 5 illustrates the
camera-to-light geometry of the system. This set-up
was developed because the introduction of wide-an- 25m
gle lenses increased the volume of water common to
both light and camera, so increasing image deteriora-
tion. Again, this system separates the light and cam-
era in order to reduce the common volume, so
reducing backscatter. Of the geometric separation
techniques, this one produces the greatest camera-to-
target ranges (up to 21 m; Patterson, 1972). The only
(minor) disadvantages of this system are its greater Fig. 5. LIBEe geometry (after Patterson, 1975).
vulnerability to damage and cost of light sources.
LIGHT POLARIZATION
source and a correspondingly crossed polarizer in
This idea came from experiments using polarized front of the camera. It works on the principle that
light to see through fog (Nathan, 1957). It involves a the sea-floor can depolarize light, whereas scatterers
plane polarizing filter placed in front of the light in the intervening water do not. Thus the camera,
74 Q. HUGGETT

with its crossed polarizer, can only receive light from The disadvantage of using polarizers is that
the sea floor. Experiments carried out by Briggs and light levels are significantly reduced. With a reduc-
Hatchett (1965) produced a 20% increase in th~ tion of 50% for each filter crossing, only 25% of the
visible range in muddy water. The problem with this ·light remains for film exposure. Thus extra expense
technique, however, is that it depends on the depo- in film ·type and/or light source is incurred when
larizing characteristics of both the water and the sea using this type of system. For the WASP system,
floor, neither of which are absolute. It also depends this ~echnique was deemed too expensive as it
upon the accurate alignment of the two polarizing would have required replacing all the lights. The film
filters. type could not be changed since in the UK, only one
Gilbert and Pernicka (1966) improved on this type of thin-based film is available in long lengths
technique by using circular instead of plane polariz- (Ilford X650; 400ASA; HP5 process) and a compro-
ers. Circular polarization is obtained by passing light mise on the number of exposures per run was not
through a quarter wave retardation plate after it has acceptable.
been linearly polarized. If the light passing through
MONOCHROMATIC LIGHT SOURCES
the retardation plate is polarized at 45° to its axes, it
emerges with the two optic axes having equal ampli- Figure 3 shows how much of the visible spectrum is
tude but with a phase lag of 90° between them. In unsuitable for illumination in long-range underwater
this condition the light is said to be "circularly" photography. If only the optimum wavelengths are
polarized and can propagate with a right- or left- used then backscatter will be significantly reduced.
handedness, depending upon which way round the There are three ways in which "monochromatic"
optic axes were orientated in the retardation plate. (narrow-band) light may be produced for underwa-
The principle behind this approach is similar to ter photography. These are by filtering, and the use
that of the plane polarizing technique, as it depends of arc lamps and lasers.
upon the different polarizing characteristics of the Filtering is best achieved by using a filter which
sea floor and suspended particles. This is because mimics the attenuation characteristics of the antici-
whenever circularly polarized light is reflected it pated camera-to-target range. The Wratten Filter
changes its polarity ( or handedness). Suspended par- Book (Anon, 1953) gives technical data for Kodak
ticles are small and only produce single reflections filters and may be used to choose the filter most
and therefore opposite polarity light (to the source). suitable for the water and range conditions antici-
On the other hand, the sea floor produces multiple pated. Mertens (1970) compared pure water at four
reflections which result in equal amounts of both target ranges (5, 10, 50 and 100 m) with the four
opposite and normally (or left- and right-handed) most suitable filters. In turbid waters allowances for
polarized light. a shift of the "transmission window" towards the
The advantages of this system are that the polariz- yellow part of the spectrum must be made. The
ing characteristics of the sea floor and suspended disadvantage of filtering is that it reduces the overall
particles match the circular more than the plane light level, so that more, or stronger, light sources
polarizing technique. Circular polarizers are also bet- are needed. The advantage of this system is that it
ter because the orientation of the filters is not criti- can be adapted to suit different water and range
cal. Cocking (1976) pointed out that the effectiveness conditions.
of circular polarization depends upon the size of Arc-lamps have been built to produce dominant
particles causing the backscattering. It appears that outputs in the blue-green light band. They are based
circular polarization is most effective in reducing upon the ionization of a mixture of mercury and
scattering from particles smaller than 1 micron. With thallium under high pressure in a quartz tube. Har-
particles greater than 6 microns circular polarization ford (1968) tested this light source and found that
can actually make matters worse. Generally, 90% of the dominant emissions occur at a wavelength of
particles suspended in oceanic waters within 50 m of 535 nm (blue-green) and that it has an output
the sea floor are smaller than 0.8 lim (microns) and efficiency 801m W- 1 (lumens/watt) (i.e. much higher
99% of particles are smaller than 6 lim (W. R. than incandescent tubes which produce up to 40 1m
Simpson, pers. comm., 1988). W- 1). The thallium arc-lamp is the most suitable
 LONGE-RANGE UNDERWATER PHOTOGRAPHY 75

arc-light source for long-range photography (Hittle- rather than space. This is done by generating a short
man and Strickland, 1968)_ It combines efficiency pulse of light and opening the camera shutter only
with monochromatic output in the "transmission when the pulse is expected to return from the target
window" of seawater. The disadvantage of arc-lamps and for a time equal to the duration of the original
is that they must be operated continuously (for all pulse (Heckman, 1966). Figure 6 illustrates this tech-
practical purposes). They therefore need to be cable- nique and shows how most of the back scattered light
rather than battery-powered. is prevented from entering the camera.
A wide range of lasers of differing spectral output Two range-gated systems are under development
have been produced for underwater application. The at present; Dixon et al. (1984) describe a range-gated
most suitable as a monochromatic light source is the system using a pulsed copper chloride laser with
copper chloride laser, used in conjunction with a a photomultiplier tube (PMT) as the receiver.
range-gated imaging system (Dixon et al., 1984). A laser is required in order to produce a short
Although this laser has not been used as a light enough pulse to eliminate unwanted scattering (i.e. a
source for non-gated cameras its spectral distribution pulse length 10 ns long = 2 m water path length). In
matches the 'transmission window' of seawater and this system the laser is pulsed and scans the sea floor.
in this respect would be ideal for long-range photog- The signal is received in a PMT and converted into
raphy. The main disadvantage is that the cost and raster format. From tests with their prototype sys-
complexity of a laser may not be justified for conven- tem, Dixon et at. (1984) envisage a system capable of
tional photography alone. Also, the relatively low being operated at 50 m altitude covering an area
power output of this source requires an image inten- 100 m wide with a 20-cm resolution. Wilson (1986)
sifier to detect the sea floor images. described a range-gated system that is likely to be-
come commercially available in the near future.
RANGE-GATING This system uses an image-intensifier camera and
In the LIBEe system described earlier (Fig. 5), the a yttrium-aluminium-garnet (YAG) laser. The
signal-to-noise ratio (i.e. the ratio of image to YAG laser (which operates at a wavelength of
backscattered light) was improved by moving the 1.06 jlm) is fitted with a frequency doubler to pro-
camera closer to the target. In range-gating the effect duce light at 532 nm, and a beam-spreader in order
is similar except that the camera is moved in time to match it to the camera viewing angle. So far this

(i) gate clos ed (ii) gate open

OOfE
backscatter

- -7\) 1::ririIO outgoing

incom ing
signal

pulse

zon e

Fig.,6. Diagrams illustrating (A) Range-Gating and (8) Angle-Gating.

76 Q. HUGGETT

system has been tested successfully over a maximum illustrated by Funk (1973). It must be only a matter
viewing range of 40 m. In principle, range gating of time before image intensifiers allow range-gated
should be capable of much greater ranges than this. systems to exceed the 50-m maximum range dis-
However, the beam is spread to illuminate the whole cussed here.
viewing area at once. The power per unit area of the
laser is reduced as a result and this is the limiting SYNCHRONOUS SCANNING
factor at present. For comparison, were it possible Synchronous scanning (or angle-gating) removes
for the lamp to be pulsed, a thallium arc would scattered light by using a narrow beam ( < 0.1 degree
produce 6 J per pulse at 50 Hz compared to 25 mJ divergence) to scan the target, reflections from which
for the frequency-doubled, YAG laser. The narrow are analysed by a synchronously scanning receiver
beam width of the Dixon et al. (1984) system puts (Fig. 6) (Wall, 1968). The receiver cannot see light
more power per unit area on to the target (by up to which is being scattered outside the receiver lightpath
5 X 104) and is less power-limited. so large-angle forward scattered light is also elimi-
Range-gating gives underwater photography a nated in this system. To optimize scatter reduction
leap ahead in survey coverage from the previous the beamwidth must be small, and there should be
record of 21 m altitude set by LIBEC to the 50 m some separation between source and receiver to re-
proposed above (corresponding to a sixfold increase duce the common volume.
in the area covered). There are limitations to this Wall (1968) proposed that synchronous scanning
technique however, the main ones being target reso- should be carried out in a TV format. Therefore the
lution and receiver sensitivity. All optical images are target would be scanned at a speed that requires time
degraded by forward scattering and refraction from as well as angle synchronization between source and
the target. Duntley (1963) found that in macro- receiver. This is because at target ranges in excess of
scopically uniform water, the apparent contrast of about 5 m, the travel time for light to and from the
fine details decreased inversely as the cube of the target is longer than its 'dwell' time at any point.
subject distance. The "flickering" of a far away Thus for any target range, a delay between the
object is an example of this on land. In water, output and receive scans must be inserted to achieve
forward scattering is more severe and the resulting synchronization. This delay introduces an element of
loss in contrast can limit underwater photography range-gating to the system so further reducing
according to the size and type of target to be iden- backscatter.
tified. In designing an underwater viewing system The obvious disadvantage of this system is its
the type of target to be identified must be considered complexity, and likely cost. No synchronous scan-
first. This is because target shape and colour will ning systems are available commercially at present
define the contrast conditions that will enable its and most effort seems to be concentrated on range-
identification. For this reason a simple rule of gated systems. The main advantages of this system is
thumb for system design is difficult to establish. A that it improves contrast through the removal of
good example of this is given by Ryan and forward scattering.
Rabushka (1985), who used a camera system (in- It is clear, therefore, that light scattering will al-
corporating horizontal separation) from altitudes up ways be a big problem in underwater photography.
to 30 m to search for the Titanic. High-contrast Until range, or angle-gated systems are further devel-
targets such as this are exceptional in underwater oped and become widely available, we are left with a
photography so altitudes quoted in this paper are short menu of techniques for long-range photogra-
based upon natural (low contrast) sea floor targets. phy. These will allow us to photograph the sea floor
Duntley (1963) has written a comprehensive account from a maximum altitude of 21 m (using conven-
of the problems of image contrast and describes tional film and LIBEC geometry) or 40 m (with
some field experiments to investigate them. Receiver range-gated video). If man-made objects (e.g.
sensitivity can limit the maximum operating range wrecks, etc.) are the targets for underwater photog-
because pulsed laser sources give a relatively low raphy, LIBEC altitudes can be increased by approxi-
power output. However, image intensifiers are read- mately 25% since image contrasts will be much
ily available (Anon, 1986) ap.d their usefulness is greater. Range-gated video is power limited at
 LONGE· RANGE UNDERWATER PHOTOGRAPHY 77

present but may increase in range as this problem is (Fig. 3). It is a convenient coincidence that most
overcome. black and white films have a narrow response in the
An interesting but extremely low-resolution ap- blue/green wavebands. Red-sensitive monochromatic
proach to eliminating optical scattering is to adopt a or colour films are not normally used in long-range
system that uses sound rather than light to illuminate underwater photography as broad band-emulsion
the target Jones and Gilmour (1976) describe two sensitivity is wasted when the image comprises
slightly different sonic cameras operating at 2 MHz narrow-band light A possible exception to this rule
and 3 MHz for use in turbid water. A 2 MHz sonic is proposed by Fritz et al. (1972), who describes a
camera was also tested in a variety of conditions by two layer emulsion "colour" film tailored to the
Andrews (1980). He comments, " ... with optical transmission window. Underwater trials of this film
visibility of less than 1 cm ... there was no deteriora- do not appear to have been made however, and its
tion of the (sonic) image quality ... A comparison usefulness is unknown. Unfortunately, the very
with a low-light TV camera was revealing in that narrow-band blue-sensitive films have very low
even with good natural lighting in the shallow tank sensitivities (around 8 ASA) which limits their use.
none of the objects could be detected". Andrews Typically, cameras are designed with 90 mm (di-
(1980) envisaged a theoretical resolution of 86 mm at ameter) spools capable of holding up to 30 m of film.
10 m altitude. In turbid and extreme circumstances Some films are manufactured with an "Estar" base
the sonic camera could prove to be a useful tool. which can be made thinner than normal films. About
60 m of thin-based film can be loaded on to a 90-mm
spool, thereby doubling its capacity. If the increased
Data Retrieval capacity of thin-based film is desired, then film
choice is restricted. In the UK only Ilford X650
In 1893 Louis Boutan was only too aware of the
(HP5) meets the requirements outlined above.
problems of data retrieval as he struggled with the
Mechanical cameras cannot be range- or angle-
reloading of his 5 x 7 in plate camera. Since those
gated, so the maximum altitude from which photo-
days a variety of imaging systems, with increasing
graphs can be taken using cellulose film is 21 m.
sensitivity, have been developed. These range from
Assuming an average camera altitude of 15 m and
the first cellulose films of the 1890s (with emulsions
camera viewing angles of 55 x 35° (i.e. not wide
capable of up to 30° BS) to the I.S.I.T. cameras of
angle), the average picture area is 15 x 10 m. At a
today, capable of up to 200000 ASA.
towing speed of 1 m/sec (2 kn), and allowing 10%
In long range photography the type of receiver
overlap between frames, one survey can cover a strip
used controls the sensitivity, resolution, endurance
of sea floor 21.5 km long by 10 m wide in about 6.2
and playback time of the system. These aspects of
hours. About the same as a 14-mile strip of two-lane
image quality will be discussed for the two principal
highway!
types of receiver:
TELEVISION
1. Emulsion film cameras;
2. Television. The commercial development of television for do-
mestic users has provided us with a wide variety of
EMULSION FILM CAMERAS
highly complex, but relatively cheap cameras and
First invented at the turn of the century, cellulose- recorders. A potential drawback is that we are forced
based emulsions were first used in underwater pho- to follow the domestic market down whatever devel-
tography in about 1915 to produce underwater, opment route it is taking. This means using 625 (UK
motion pictures for the classic film 20000 Leagues standard) or 525 (US standard) lines picture format
Under the Sea. and a corresponding loss in resolution compared to
Most underwater cameras use 35 mm cellulose film that of emulsion films. On a standard 625 line set,
on to which a great variety of emulsions can be the image has a resolution less than 1/3 that of most
coated. For long-range photography the emulsion emulsion films. This loss of resolution is compen-
must be blue sensitive in order to record images that sated for by a high scanning rate which enables the
have passed through the "transmission window" detection of movement The human eye is deceived
78 Q. HUGGETT

by TV scanning owing to our ability to correlate yet is the combination of circular polarization and an
from scan to scan subconsciously, identifying objects S.I.T. camera. This combination may increase view-
and rejecting much of the noise. Freeze-frame images ing ranges by a factor of two (Gilbert and Pernicka,
(which would be used for detailed and quantitative 1966), bringing average operating altitudes up to
analysis) reveal the flaws of TV and its relatively 25 m for normal or in excess of 40 m for LIBEC
poor resolution, a feature that should not be over- geometry.
looked. On the other hand, TV signals can be moni- As components in a range-gated system, image
tored in near real time; also, TV tubes can operate intensifiers have the potential of significantly extend-
at extremely low light levels, and provide motion ing viewing ranges over conventional (emulsion) film
pictures. techniques. The most important feature of an image
Television (or photo-electric scanning) can be intensifier in this respect is that it can be electronically
achieved using three principal types of receiver: Stan- gated with reasonably high precision. In the previous
dard tube, Silicon Intensified Target (S.I.T.), and section the two range-gated systems under develop-
Charged Couple Device (C.C.O.). The standard TV ment were introduced. Only the Wilson (1986) system
consists of a target which is scanned by an electron seems likely to become available, and despite its range
beam. The target may be either photoconductive limitation (40 m), in more turbid waters (on the
(e.g. the Vidicon) or photosensitive (e.g. the Image continental shelf) it will enable photography in
orthicon tube) and may detect images of 0.5 Lux (at conditions that were previously unviewable.
the camera). The chief advantage of a TV system is that the
The S.I.T. tube consists of a silicon target vidicon- data can be sent in near real time to the operator on
type tube incorporating an image-intensifier (Vigil, board ship. In ocean floor-searching projects this is a
1972). S.I.T. tubes can detect images down to 10- 3 prime consideration (Ryan and Rabushka, 1985),
Lux. For even greater sensitivities an extra image- however it does bring its own problems. If data are
intensifier can be added to form an Intensified S.I.T. to be sent direct to the ship, then a standard (or
(I.S.I.T.) which can detect images down to 10- 4 Lux domestic) TV image requires more signal bandwidth
(Harris, 1980). than is available in a single element coaxial cable.
The c.C.O. operates on a different principle, Standard TV requires 4.5 MHz for transmission at a
which does not use a scanning electron beam. In- scan rate of 30 frames per second. A coaxial cable is
stead, the optical image is projected on to a target capable of transmitting a bandwidth of 2.0 MHz
consisting of an array of photoelectric sensors. Scan- over 8000 m (Mosley, 1988). Fibre optical towing
ning can be performed via a simple shift operation cables have almost been developed to a point where
reading sequentially through the array, since the they are available for deep ocean operations. For
sensor elements are coupled electrically as a shift underwater TV photography the consequences are
register. c.C.O.s have the advantage of wide dy- far reaching as these cables will eliminate all prob-
namic range combined with rugged construction. lems of data retrieval and storage. In the absence of
The latest commercially available C.C.O. cameras fibre optic cables there are only two realistic options
can detect images of 2 Lux (Martin et ai., 1987). to solving the bandwidth problem. The first is to use
Direct comparisons between the effectiveness of a multicore cable. The second is to cut or compress
TV tubes and films are not simple (see Patterson, the signal to fit the bandwidth available.
1981) as sensitivity, resolution and spectral response Signal cutting would involve slowing down the
are interlinked and affected by factors such as expo- scanning rate or cutting down the number of lines.
sure time and operating temperature. Figure 7 at- Neither option is attractive since reduced scanning
tempts to compare the sensitivity of various viewing would introduce flickering, and line reduction would
systems in a semi-quantitative way. cut resolution. In a system that is already at one
When used with conventional (thallium iodide) third of the resolution of emulsion film and at long
light sources, and without filtering, TV cameras have range from the sea floor, such measures could prove
similar range limitations to those of emulsion sys- intolerable. In the case of range-gated systems, cut-
tems in the absence of LIBEC geometry (Hittleman ting may have already occurred (Wilson, 1986) as the
et al., 1975). A possibility that. has not been explored laser cannot cope with a 50 Hz repetition rate. Most
E::: -~:x r __-s7 • • • • • • •
·.<\ii •..•.• .• • . .•. . .
I

I m - :-""):>.- : ':"
HU<tAN '. ::'. ':.. .~ v.~IO~ ·· . . . Colou e! Vision 5z
Cl

-+. ~~. ~">'

\"

. . ... ".'. '.... .

~
z
Cl

~-...
tTl
c:
Z
NEWVICON ... ' ......... . .. .. ..... ....... '.
.•..... ........ .........
.... .... .... ... .. .. ... . . .
. . . "

"
tTl

N.D. IFilter Required

J ~
~
S.I .T. 'tI
:z:

LS. U. N.D. Fil ter ReqUirt d §

I I Ful l i ~
:z:
Scene Ili um. Star light Moon light Twilight Int1rior Over ost Sun I -<
I 3
(LUX)
-4
10
-3
10 10
-2
10
-1
10
.
110
2
1 10 L10 4 I
Fig. 7. Chart comparing viewing system sensitivities (after Harris 1980), see text for explanation.

-..I
\0
80 Q. HUGGED

underwater (emulsion) cameras have a finite recy- tion of emulsion film. The problems of signal storage
cling time of approximately 2 sec. If cutting the scan and transmission combined with the relatively poor
rate is the chosen option, then 2 sec could be adopted resolution of TV systems make emulsion films an
as the minimum scan time for TV images to compete attractive option if real-time monitoring and extreme
with. A thorough discussion of signal cutting (i.e. (> 21 m) ranges are not a prime requirement.
narrowband television) is given by Deutsch (1968). Range-gating may ultimately become the best
Signal compression is more complex but prefer- technique for long-range photography, as higher res-
able. It involves the analysis of the picture before olution and higher sensitivity TV systems are devel-
transmission so that only those elements of the pic- oped. However, at present photography in relatively
ture that change between scans are transmitted, thus clear, deep oceanic waters is probably best carried
cutting the bandwidth required. With this approach, out using emulsion films with polarizing filters and
the more complex the target and the faster the arc lamps. If video cameras are to be used for image
system is towed over the sea floor, the poorer the retrieval then the problems of poorer resolution and
pictures will become because there is a limit to the limited bandwidth will have to be accepted. At long
"compressibility" of a moving picture. At present, ranges, images are low-contrast and any reduction in
signal compression is not a commercially available resolution or transmitted bandwidth will affect the
option; however, it should be considered as a possi- ability to analyse them in any detail. For operations
ble way forward. involving searching for man-made objects this is not
Video recording using commercial systems placed so important. However, in the study of subtle and
in pressure housings has been successfully accom- low-contrast sedimentary features on the sea floor,
plished (Chezar and Lee, 1985). Even though the images must be of high quality. Transmitting a
bandwidth of domestic VCRs is limited to between 2 blurred, low-contrast image at high resolution is a
and 3 MHz and the image resolution degraded on waste of bandwidth. In practice';05tha:ransmission
recording (by approximately 50%), video systems system need only be better than the source image,
are cheap and reliable. A video recorder can record not perfect.
up to 4 hours of signal or can be switched to a
half-cut signal to double its recording time (with a
corresponding loss of information). Until signal Acknowledgements
compression techniques become available, one solu- The author would like to thank Mike Conquer and
tion to TV viewing problems would be to store the Arnold Madgwick for their advice and help in the
full signal on one or more VCRs mounted on the production of underwater photographs. Invaluable
towed vehicle and to cut the monitor signal for assistance was also provided by the IOSDL applied
transmission up to the ship. This way the operation physics department in the construction and opera-
of the vehicle can be monitored and the eventual tion of underwater cameras.
replay (after retrieval of the vehicle) can be of rela-
tively high quality.
References
Summary Aldred, R. G., Thurston, M. H., Rice, A. L., and Morley, D. R.,
1976, An Acoustically Monitored Opening and Closing Epiben-
Since the late 1960s the problems of optical scattering thic Sledge, Deep-Sea Research 23, 167-174.
Andrews, W. B., 1980, Inspection, Maintenance and Repair:
and signal transmission in seawater have been under- Ultrasonic Imaging for Offshore Underwater Operations, in
stood but until recently no original advances have Oceanology International '80, Technical Session J; B.P.S. Exhi-
been made in either field. Compared to the develop- bitions Limited, 19-28.
Anon., 1953, Wratten Light Filters, Kodak, London, 92.
ment of aerial photography which has risen in range
Anon., 1968, Bibliography on Underwater Photography and Pho-
from 1000 to 100 000 mover 100 years, underwater togrammetry, Eastman Kodak, New York, 23.
photography has struggled through its physical limi- Anon., 1986, Underwater Cameras-Product Review, International
Underwater Systems Design 8(5), 6-8.
tations to an increase in range from 4 to 40 m.
Anon., 1987, Current Trends in Underwater Photography; A
Cocking (1976) pointed out that images viewed on Company Perspective, International Underwater Systems Design
t
standard TV screens are betw(;!en and ~ the resolu- 9( 1), 30 pp.
 LONGE-RANGE UNDERWATER PHOTOGRAPHY 81

Ballard, R_ D. and Moore, J. G., 1977, Photographic Atlas of the Muds by Feeding Traces Observed in Deep-Sea Photographs, in
Mid-Atlantic Ridge Rift Valley, Springer-Verlag, New York, P. P. E. Weaver and J. Thomson, (eds.), 1987, Geology and
114 pp. Geochemistry of Abyssal Plains, Geo!. Soc. Spec. Pub. No. 31,
Briggs, R. O. and Hatchett, G. L., 1965, Techniques for Improv- pp. 105-112.
ing Underwater Visibility with Video Equipment, Ocean Science Jones, C. H. and Gilmour, G. A., 1976, Sonic Cameras, J. Acoust.
and Ocean Engineering 2, 1284-1308. Soc. Am. 59( I), 74-85.
Brown, J. F., 1985, The First Underwater Photograph, The British Lampitt, R. S. and Burnham, M. P., 1983, A Free Fall Time-
Journal of Photography 32, 894-895. Lapse Camera and Current Meter System 'BATHYSNAP' with
Chardy, P., Guennegan, Y., and Brannelec, J., 1980, Photographie Notes on the Foraging Behavior of a Bathyal Decapod Shrimp,
sous-marine et analyse des peuplements benthiques. Deep Sea Research 30, 1009-1017.
C.N.E.x.O., Rapports Scientifiques et Techniques 41, 32 pp. Laughton, A. S., 1957, A New Deep-Sea Camera, Deep-Sea
Chezar, H. and Lee, J., 1985, A New Look at Deep Sea Video, Research 4, 120-125.
Deep-Sea Research 32(l1A), 1429-1436. Martin, G. J., Womack, K. H., and Fischer, J. H., 1987, A High
Cocking, S. J., 1976, Improving Underwater Viewing, (139-190) Resolution CCD Camera for Scientific and Industrial Imaging
in E. A. Drew, J. N. Lythgoe and J. D. Woods, (eds.), Under- Applications, Proc. Society of Photo-optical Instrumentation
water Research, Academic Press, London. 430 pp. Engineers 818, 301-319.
Deutsch, S., 1986, Narrowband Television Pictures for Transmis- Mauviel, A., 1982, La bioturbation actuelle dans Ie milieu abyssal
sion via Oceanographic Sound Waves, Ocean Engineering 1, de 1'000an Atlantique Nord, These, Diplome de Docteur de
9-16. troisieme cycle, Universite de Bretagne Occidentale, 103 pp.
Dixon, T. H., Pivirotto, T. J., Chapman, R. F., and Tyce, R. c., Mertens, L. M., 1970, In-Water Photography: Theory and Prac-
1984, A Range-Gated Laser System for Ocean Floor Imaging, tice, John Wiley & Sons, New York, 391 pp.
Marine Tech. Soc. Journal17( 4), 5-12. Mosley, C., 1988, Video Systems in Deep Water Mining and
Duntley, S. Q., 1963, Light in the Sea, Journal of the Optical Soc. Oceanographic Applications, Underwater Systems Design
of America 53(2) 214-233. 10( 4), 20-22.
Fritz, N. L., Specht, M. R., and Needler, D. G., 1972, A New Nathan, A. M., 1957, A Polarizing Technique for Seeing through
Colour Film for Water-Penetration Photography, Preprints of Fogs with Active Optical Systems, New York University, Tech.
the 8th Annual Conference of the Marine Technology Society Rep. 362.01.
1972, pp. 737-738. Ohta, S., 1983, Photographic Census of Large Sized Benthic
Funk, C. J., 1973, Predicted System Performance of Improved Organisms in the Bathyal Zone of Suruga Bay, Central Japan,
Underwater Light Sources. (pp. 7-17) in Anon (ed.), OCEAN Bull. Ocean Res. Inst., Univ. Tokyo, No. 15, 244 pp.
'73 Engineering in the Ocean Environment, IEEE, New York, Patterson, E., 1981, Underwater Television-An Art in a State,
623 pp. International Underwater Systems Design 3( 5), 11-12.
Gilbert, G. D. and Pernicka, J. C., 1966, Improvement of Under- Patterson, R. B., 1972, Increased Ranges for Conventional Under-
water Visibility by Reduction of Backscatter with a Circular water Cameras, Proc. S.P.I.E. 24, 153-161.
Polarization Technique, in Underwater Photo-Optics: Seminar Phillips, J. D., Driscoll, A. H., Peal, K. R., Marquet, W. M., and
Proceedings, Society for Photo-Optical Instrumentation Engi- Owen, D. M., 1979, A New Undersea Geological Survey Tool:
neers, Redondo Beach, Ca. pp. A-I1I-1 to II. ANGUS, Deep-Sea Research 26( A), 211-226.
Harford, J., 1968, Underwater Lighting-A Status Report, in F. Rebikoff, D., 1967, History of Underwater Photography, Pho-
Alt, (ed.), Marine Sciences Instrumentation, Vol. 4. Plenum, togrammetric Engineering 33(8),897-904.
New York, pp. 373-380. Rice, A. L. and Collins, E. P., 1985, The Use of Photography in
Harris, R. J. 1980, Improving the Design of Underwater TV Deep-Sea Benthic Biology at the Institute of Oceanographic
Cameras, Underwater Systems Design 2(1),7-11. Sciences, (pp. 153-164) in J. D. George, G. I. Lythgoe and
Harris, S. E. and Ballard, R. D., 1987, ARGO: Capabilities for J. N. Lythgoe (eds.), Underwater Photography and Televisionfor
Deep Ocean Exploration, OCEANS '86 Conference Record, Scientists, Clarendon, Oxford, 184 pp.
Vo!. 1 MTS, Washington, pp. 6-8. Rice, A. L., Aldred, R. G., Billett, D. S. M., and Thurston, M. H.,
Heckman, P. J., Jr., 1966, Underwater Range-gated Photography, 1979, The Combined Use of an Epibenthic Sledge and a Deep-Sea
pp. B-IX-I to 9 in Underwater Photo-Optics: Seminar Proceed- Camera to give Quantitative Relevance to Macro-Benthos Sam-
ings; Society of Photo-optical Instrumentation Engineers. Re- ples. AMBIO Special Report No.6, pp. 59-72.
dondo Beach, Ca. Ryan, P. R. and Rabushka, A., 1985, The Discovery of the
Heezen, B. C. and Hollister, C. D., 1971, The Face of the Deep, Titanic by the U.S. and French Expedition, Oceanus 28(4),
Oxford University Press, 659 pp. 16-33.
Hittleman, R. L. and Strickland, C. L., 1968, The Application of Vigil, A. E., 1972, A New Low Light Level Television Camera for
Incandescent, Mercury Vapour, and Thallium Oxide Lighting Underwater Applications, Preprints of the 8th Annual Confer-
to Underwater Tasks, Society of Photo-optical Instrumentation ence of the Marine Technology Society, 1972, pp. 581-600.
Engineers. Seminar Proceedings Vol. 12, pp. 153-165. Wal, M. R., 1968, Underwater Viewing, Proceedings Society of
Hittleman, R. L., Vigil, A. E., Coffman, J. E., and Hatchett, G. Photo-Optical Instrumentation Engineers 12, pp. 21-32.
L., 1975, The Performance of Low-Light Level Television Cam- Wilson, T. B., 1986, An Experimental Laser Gated-Television
eras on Underwater Remote Controlled Vehicles and Towed Underwater Viewing System, Proceedings of Electro-Optics and
Sensor Platforms, S.P.l.E. Vol. 64, pp. 128-134. Laser, U.K. '86, Published by Communications in Print, Essex
Huggett, Q. J., 1987, Mapping of Hemipelagic versus Turbiditic UK.
PART II

Sampling Techniques
Current Methods for Obtaining, Logging and Splitting
Marine Sediment Cores
P. P. E. WEAVER and P. J. SCHULTHEISS *
Insititue of Oceanographic Sciences (DL) Brook Road, Wormley, Godalming, Surrey GU8 5UB, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: coring devices, box corer, gravity corer, piston corer, using a variety of devices to sample the upper 20-
giant piston corer, 'P' wave log, wholecore logging, kevlar, warp. 25 m of the sediment column. The largest of these
ships are also capable of handling the relatively new
Abstract. The main types of deep-sea sediment coring devices are giant piston corers which currently can sample to
described and their relative merits and drawbacks are discussed.
These devices include box corers, gravity corers, piston corers, depths of 35 m below the sea floor, and in the future
giant piston corers and vibrocorers. Recent utilisations of kevlar may be able to sample as deep as 50 m.
and polyester coring warps are also discussed, since these are the There are two important considerations to be
only warps capable of handling the large weights associated with
the larger devices. Recent developments in wholecore logging,
taken into account before taking a deep-sea core: (a)
including 'P' wave, density and magnetic susceptibility, are de- how hard is the substrate?; and (b) what is the core
scribed as are methods of subcoring and core splitting to obtain to be used for? In some circumstances it may be
the maximum amount of detail on the split surfaces. The
prudent or necessary to use more than one type of
wholecore logs together with a good colour photograph of the
recently split sediment surface provide a lasting unambiguous corer to ensure that a complete record of a site has
record of the core. been obtained to the depth of interest or to a depth
limited by the available technology. This is often the
Introduction case if a good-quality core of the sediment surface is
required together with a sediment record several
Our knowledge of the deep-sea floor is built up from
metres long. Scientists wishing to make physical
interpretations of a combination of imaging tech-
property measurements, and those interested in an
niques (geophysical mapping and profiling) and by
accurate record of the thicknesses of all lithological
direct sampling. A vast range of devices has been
units in a core, should be particularly careful and
developed for sampling the seabed, and these fall
aware of problems arising from core disturbance and
into various categories, with each category being
core mis-sampling. In some cases these are not easily
aimed at either particular depth penetrations, partic- detected and may lead to incorrect interpretations.
ular sediment types or a specific use for the sample,
Coring techniques have developed over the years
The deepest and most complete record of deep-sea with a lot of ingenuity but without any significant
sediments and basement rocks has been provided by "high technology". The success of a coring operation
the Ocean Drilling Program (ODP), using a combi-
has all too often been judged by the criteria of "how
nation of downhole hydraulic piston coring and full or how long is it?". Attention has, more recently,
rotary drilling. The drilling capabilities of the ODP
been focused on the problems associated with core
and its' predecessor the Deep Sea Drilling Project
disturbance and ensuring the general quality and
(DSDP) have been well documented elsewhere (see,
representativeness of the core rather than solely that
Storms, this volume) and are not considered further
of maximizing the depth of penetration. There has
in this chapter. Smaller and less expensive research
been a move towards the design of more sophisti-
vessels than those used for drilling are capable of
cated corers in recent years, often involving the
* Present address: Schultheiss Geotek, Fern Cottage, Marley Lane, mounting of monitoring equipment on the corer,
Haslemere Surrey GU27 3RF, UK both to improve penetration and to produce a de-

Marine Geophysical Researches 12: 85-100, 1990.

© 1990 Kluwer Academic Publishers.
86 P. P. E. WEAVER AND P. J. SCHULTHEISS

tailed record of that penetration. Although some penetrated the sediment. Although this system po-
work has been completed on the design and testing tentially allows the corer to bounce on the sea floor,
of this type of equipment (see Parker and Sills, this no problems have been experienced to date.
volume), there is no operational corer which uses it
routinely.
(iravity Corers
Several other major reviews of coring equipment
(e.g. those of Bouma, 1969; Moore and Heath, These corers are very simple, but variable in design,
1978; Lee and Clausner, 1979) and of core handling consisting of a large headweight, to which is at-
(Bouma, 1969) have been published in the past two tached a barrel of variable length with a core cutter
decades. Details of several corers not mentioned and catcher at the lower end. The ship's warp is
in this chapter will be found in those papers. attached directly to the core head, except in the
This chapter summarizes the main types of coring device described by Hvorslev and Stetson (1946), in
equipment currently in use, including recently which the corer is triggered by a trigger weight.
developed giant piston corers, as well as modern Most round barrel corers of this type (Fig. lc)
methods for handling and describing cores. There is utilize core liners, but those with long box-shaped
now a greater emphasis on whole core logging and barrels, such as the Kastenlot corer (Fig. Id), do
high-quality core photography, which can replace not. Most incorporate a flap valve at the top to
and improve much of the laborious core description allow water to escape during coring but which is
work frequently carried out by inexperienced closed during pullout, ascent and recovery to pre-
personnel. vent sediment being washed out. Problems with
washout occur particularly at the sea surface during
retrieval as a result of the loss of buoyancy when
Box Corers
the core is lifted from the water.
The main advantage of box corers is that they ob- Gravity corers are inexpensive and easy to use
tain large-volume cores of surface sediment with and can take high-quality cores of the upper few
minimum disturbance. They consist of a square box metres of the sea floor. They are, however, subject
(occasionally a large diameter cylinder), a head- to two major sources of error. These are (1) mis-
weight and a spade-type lever arm (Fig. la). Some sampling due to sediment plugging in the core bar-
designs also include a tripod support frame to en- rel, and (2) repenetration caused by vertical
sure vertical coring (Fig. Ib). The cross-sectional oscillations of the ship's warp. Emery and Dietz
area of the box may be up to 0.25 m2 , but the (1941) showed that open barrel gravity corers were
length is generally less than 1.2 m. The spade lever capable of taking very shortened sections of core,
arm lies horizontally during deployment and de- and Hvorslev and Stetson (1946) showed how the
scent, but on recovery it is pulled into its vertical sediment layers would be shortened ahead of the
position, thus closing the bottom of the core box corer. Weaver and Schultheiss (1983b) were able to
prior to pull-out. In many of the designs the spade show evidence of gravity corers "bouncing" on the
lever arm also closes the top of the box, thus ensur- sea floor, enabling multiple sampling of the upper
ing an undisturbed sediment-water interface. Some sediment section by repeated penetrations. They
box corers utilize a scissor arrangement of two also showed how each successive repenetration was
spade lever arms (Fig. la) which ensures that the shortened relative to the previous one, and how the
box is maintained in a vertical position, thus elimi- shortening was more pronounced in the softer more
nating the need for a tripod frame. It has become plastic, sediment layers. This suggests that gravity
apparent, however, that single-spade corers are of- corers can be unreliable in sediments of mixed com-
ten much better at coring sandy sediments. Most position, and produce increasingly unreliable results
box corers activate the spade closure by means of a with increasing penetration. The length of core that
"no-load" release. Although this system is simple, it can be taken with an open barrel gravity corer is
can lead to pre-tripping, especially in rough seas. limited by the increasing friction between the cored
The lOS box corer (Fig. la) has an acoustically sediment and the inside of the barrel. At some stage
activated release which is fired after the corer has the force required to move the core up the barrel.
 MARINE SEDIMENT CORE METHODS 87

a) b)

c) d) e)

100 100em
em
100

50
o

Fig. I. Corers used for near-surface sediment sampling: (a) lOS box corer (double spade) (after Peters et al., 1980); (b) box core with
single-spade and tripod frame (see Bouma, 1969); (c) hydrostatic gravity corer (Richards and Keller, 1961); (d) square barrel Kastenlot
core (Kogler, 1963) ; (e) free-fall, pop-up boomerang corer (Moore, 1961).

exceeds the force required for the barrel and core to detailed study of this effect has been made by
act as a solid rod for further penetration. This is Parker and Sills ( this volume) in one sediment type
known as "plugging" and commonly occurs be- using a technique to continuously monitor the sedi-
tween 3 and 6 m penetration. Intermittent or partial ment surface both inside and outside the core barrel
plugging is the cause of core shortening. A more during penetration.
88 P. P. E. WEAVER AND P. J. SCHULTHEISS

Gravity cores are frequently used as trigger since the round-trip time is about 15 min/lOOO m
weights for piston corers and often produce a more water depth and hence there is little time for the
representative record of the surface sediment than equipment to drift during the operation. The disad-
that obtained from the piston core. Care must be vantage of the system is that only short cores
taken, however, to ensure that no repenetration of ( < 1.2 m) can be obtained and it can sometimes be
the trigger corer has occurred since this is common difficult to find the floating core in poor visibility or
during piston coring (McCoy, 1980). rough weather.
The Kastenlot corer (Fig. Id) bridges the gap
between gravity and box corers since it has a rela-
Piston Corers
tively large cross-sectional area (225 cm 2), and can
take cores up to 6 m long (Kogler, 1963). The box The piston corer (Fig. 2) was first developed by
splits longitudinally into two halves and so it must Kullenberg in 1947 to overcome the depth penetra-
be laid horizontally to open the core, which often tion limitation, caused by plugging, of gravity cor-
causes some slumping and disturbance of the sur- ers. It consists of a series of connected barrels, a
face sediment. This corer does not use a liner and is large, heavy headweight, a trigger arm and a trigger
usually sub-sampled immediately after opening. Sec- corer. A plastic core liner is usually employed, but
tions of the cored material can be stored in I m- the core may be extruded on deck after recovery.
long boxes, which have the same cross-section as The head may be designed with removeable weights,
the corer but with closed ends for later examina- with fins, or with housings for instruments such as
tion. It is also possible to remove one corner of the cameras, flash-guns or pingers. Most trigger arms
core into a V-shaped trough which can be stored in are fitted with safety pins to prevent accidents dur-
a "D" tube as an archive section. The large cross- ing deployment. Hydrostatic pins usually retract in
sectional area allows this corer to provide high- the upper or mid water column, but acoustic re-
quality cores which often suffer less disturbance and leases can be used, which can prime the corer by
core shortening than those taken by other gravity command from the ship at any depth. These offer
corers. However, the Kastenlot corer is susceptible increased safety in rough weather or on less stable
to repenetration (Weaver and Schultheiss, 1983b). ships, where temporary reductions in load on the
The Boomerang corer (Fig. Ie) is a modification trigger arm due to ships' heave can cause premature
of the open-barrel gravity corer which operates as a release of the corer.
freefall/pop-up corer. It consists of a ballast section The advantages of the piston corer over gravity
comprising a steel barrel, weight, steel float protec- corers is the increased depth of penetration as a
tion and lead pilot weight. The float section consists result of the prevention of plugging. This should
of a core liner, valve release and two glass spheres. ensure higher quality cores. The depth of penetration
The whole system is released from the ship, after depends on sediment type, but in relatively soft
which it quickly rights itself and attains a terminal muds, over 20 m can be obtained (Kuijpers et al.,
velocity of approximately 7 m/sec. A hollow rubber 1984). The action of the piston, which reduces the
ball prevents release of the float portion during internal friction, enables these corers to generally
deployment, but this is compressed and released recover more complete, and less disturbed sediment
during descent. When the corer impacts with the sequences than open-barrel gravity corers. The at-
seabed the pilot weight slides up the barrel, releas- tachment of the main warp to the piston, however,
ing the float section which then ascends, closing the causes ships' heave motions and elastic rebound of
valve-release at the top of the liner and pulling free the warp after corer release to be transmitted to the
the liner tube with the core inside. The float section piston during coring, which can result in core short-
ascends through the water column and is identified ening and/or other sediment disturbances (McCoy,
at the surface by an electronic flash in one of the 1980, 1985). Deformation is often particularly severe
spheres. The advantage of this system is that several in the vicinity of sandy layers which may be tilted,
cores can be obtained rapidly by deploying them as bent downwards at the sides of the liner or totally
a series prior to recovery. Furthermore, the core site mixed by the coring action. Igarashi et al. (1970)
is likely to be directly beneath the ships position, showed that coarse particles could be dragged from
 MARINE SEDIMENT CORE METHODS 89

coble to shii! free fall coble

/

"pi3ton stop

< barrel t hoin

cobl e run n in g
through barrel

piston>
< catcher
cutter>

trigger corer>

Fig. 2. Typical piston corer shown (a) as deployed and (b) after triggering at the sea-floor.

the sediment surface and embedded in the deeper liner diameters of 83 mm and 145 mm, and have
parts of the core without there being obvious visual found that 83 mm liners give good-quality cores of
disturbance to the core. Burns (1962) showed that comparable length to the small-diameter liners, but
the upper sediment layers could be mis-sampled by obviously with much larger volumes of material re-
improper adjustments of the trip chain length trieved, (Kuijpers et al., 1984). This can be particu-
or incomplete immobilization of the piston during larly useful when sediments are required for
coring. geotechnical and geochemical studies.
The sediment surface is often lost in piston cores,
and if the relative lengths of freefall and trip-chain Giant Piston Corers
are not calculated correctly, the upper few metres of
THE ADVANCED PISTON CORER (APC)
the sediment column can be omitted. Trigger cores
usually provide a more accurate record of the surface The concept of a long piston corer, capable of
sediment, but care must be taken in their interpreta- penetrating up to 50 m of sediment was first pro-
tion because of the repenetration problems men- posed by the Wood's Hole Oceanographic Institu-
tioned previously. The piston may be of a solid or tion in 1967, and reports of the first trials were
breakaway type; the breakaway version splits into published by Hollister et al. (1973) and Driscoll
two at the beginning of the pull-out and thus pre- and Hollister (1974). The first model was called the
vents material flowing into the core barrel if it is not "Giant Piston Corer" (GPC) and took sediment
completely full. samples 110 mm in diameter. An account of the
The majority of piston corers have internal liner design and early modifications of this corer is given
diameters of 65-70 mm with many being 66 mm, the by Hollister et al. (1973). The most serious problem
same as used by the ODP. The Geological Survey of was encountered with the core barrels, which were
the Netherlands, however, have experimented with standard well casing pipes that frequently bent or
90 P. P. E. WEAVER AND P. 1. SCHULTHEISS

broke at the barrel couplings on 30- 35 m deploy- acoustic release, a tilt alarm, a rotation recorder, a
ments. Further problems were encountered with sta- piston control system, an acoustic telemetry package
bilizing the piston and tripping the core. Driscoll and and a power and control package. These instruments
Hollister (1974) discussed several modifications to were designed to enable close control and monitoring
the system, including step tapered barrels, a re- of the system during the coring operation. A simpler
designed head with provision for instrumentation version of this system (Fig. 3) was used on the Giant
and an acoustic release. In 1977 a major redesign was Piston Corer (Silva et al., 1977) and on the Marion
proposed (Driscoll and Silva, 1977) to incorporate Dufresne ESOPE expedition (Schuttenhelm, in
these features into a technically sophisticated system press). In the simple version, a parachute is attached
for taking 50 m-Iong cores. Initially called the Long to the coring warp just above the core head, the warp
Coring Facility, it was later renamed the Advanced continues down the barrels to a standard-type piston
Piston Corer (APC). (This should not be confused and is held in position at the core head by the
with the Ocean Drilling Program's hydraulic release. The parachute takes some of the core weight
"APC".) during lowering, but on approach to the seabed the
The instrumentation of the APC head includes an winch payout is speeded up so that the parachute

5
parachute
100m ~

~
10
ship

corer head
with electronics
SOm

Om

Fig. 3. The giant piston corer with parachute (a) during descent; (b) during penetration (redrawn after Driscoll, 1981).
 MARINE SEDIMENT CORE METHODS 91

takes the whole weight of the corer, thus producing the barrel, and since the baseplate is 1.5 m across, it
a quantity of loose warp and eliminating any re- remains at the sediment surface and holds the piston
bound effects of the warp when the load is released. at the same position while the corer slides past. This
The core is tripped with minimal freefall, and slides corer takes cores with a 110 mm internal diameter
past the piston which is held at the same position by and, because of the stationary piston, generally
the slowly descending parachute. achieves recovery approaching 100% with excellent
A very complex system has been considered for core quality.
stabilizing the piston by decoupling it from the main The STACOR system does not depend on sophis-
ship's warp (Driscoll, 1981). Although this system ticated technology and has a high success rate. It has
has never been built, it illustrates the extent to which so far been launched in three different ways: over the
technology can be utilized in coring equipment. It is stern of a ship, over the side rail and via a moon
known as the Hydrostatic Accumulator Piston pool (Montarges et al., 1987). The long barrel-length
(HAP), and utilizes a parachute located just above and the large loads, particularly those associated
the core head. The HAP piston concept consists of with pull-out (about 200 kN) restrict its use to the
two long chambers, the upper one air-filled and larger oceanographic vessels, or to vessels such as the
connected to the lower oil-filled one by a valve. Nadir, which have a large free deck for deployment
There is also an accelerometer to assess the piston over the stern. The loads can be drastically reduced
motion and a sonar transducer at the lower end to by using Kevlar cables which are almost weightless
monitor the position of the sediment surface during in water (Schilling et al., 1988). One other drawback
coring. The piston is connected to the parachute as with the current STACOR system is the length of
before, but after triggering the valve between the oil time required to complete a coring operation, which
and air filled chambers allows oil to pass, thus can be as much as 16 hours. A large proportion of
compressing the length of the piston. The rate of this time is spent removing the core from the barrel,
contraction of the piston is designed to keep the base which detracts from the corer's usefulness for geo-
of the piston stable during penetration (which takes chemical work where cores often need to be placed
only a few seconds). The contraction rate needed can rapidly under nitrogen (Schuttenhelm et al., in
be calculated from the parachute sink rate, or from press). However, some redesign features of the sys-
the accelerometer which monitors the vertical motion tem could probably overcome these drawbacks.
of the bottom part of the piston.
The drawbacks of the APC system are its great
Vibrocoring and Rockdrilling
complexity and the very limited numbers of ships
capable of dealing with the large loads experienced The vibrocorer extends the range of sediments which
during pull-out. can be sampled to include stiff and stony clays, soft
rock and sands, all of which are difficult to penetrate
STACOR
using conventional gravity or piston corers. The
The only other corer which has taken cores in excess British Geological Survey's vibrocorer (Fig. 5) con-
of 25 m is the STACOR (stationary piston corer), sists of a tripod frame with a base-mounted rotary
developed by the Institut Fran~ais du Petrole, Elf drive table which drives a 6 m-long hexagonal drill
Aquitaine and Total in the mid-1970s (Montarges et barrel. The twin vibrator motor gives a force of 6
al., 1983, 1987). This corer utilizes a mechanical tonnes at 50 Hz and the core can be withdrawn from
system for maintaining a stationary piston, an idea the sediment by a base-mounted winch capable of
first suggested by Kullenberg (1955) and later devel- exerting 12 tonnes force. This corer takes 83 mm
oped by Kermabon and Cortis (1968). The principle diameter cores in plastic liners. The system is electri-
of the stationary or recoilless piston is that it is cally powered from the ship, from where the opera-
coupled to a baseplate on the outside of the corer via tion can be monitored and controlled (Ardus et al.,
a series of pulleys and wires (Fig. 4). The main ship's 1982). The present system has been tested to 1800 m
warp connects directly with the corer head and is water depth, with the main limiting factor being the
thus decoupled from the piston. The baseplate and power supply cable.
piston begin the coring operation at the cutter end of The British Geological Survey's rockdrill is
92 P. P. E. WEAVER AND P. J. SCHULTHEISS

corer rei eaSe

wire connecting
piston to baseplat e

pis ton

pulley

boseplote
ossembly

trigger we igh t

Fig. 4. The stationary piston corer, Stacor, showing the operating principle of how the base-plate and piston remain at the sea-floor during
penetration (redrawn after Montarges et al., 1983).

mounted in the same tripod structure as the vibro- capability of most oceanographic winches. The desire
corer (Pheasant, 1984). Up to 6 m penetration is to produce piston corers of larger diameter, and
possible with this system and, as before, withdrawal giant piston corers has, however, increased the po-
of the core is achieved by the 12-tonne winch. Brook tential maximum loads to such an extent that con-
and Pelletier (1970) report a deep-sea rock drill ventional steel warps are no longer adequate. The
which works by hydraulic power, but which is also problem with steel warps is their large weight in
limited to 1800 m water depth. This system utilizes water, so that in deep water, beyond a certain limit
three heavy-duty gas cylinders which allow water to the increased strength of a warp is largely used to
enter by hydrostatic pressure which turns a hydro- offset the increased weight of the warp itself.
static motor for a 7.S-minute cycle. The first half of Two new materials have been used in recent years
the cycle drills the core in and flushes the drill bit and to overcome this problem; Kevlar and braided
the second half of the cycle withdraws the core. polyester. Both of these materials have negligible
weight in water. The Geological Survey of the
Netherlands has experimented with Kevlar and
Winches and Warps
found that it gives excellent results when used with
The maximum loads experienced during coring oper- piston corers of large and small diameter. The
ations are those associated with pulling the corer out Kevlar is encased in a polyethylene sheath to prevent
of the sediment. For most gravity, box and piston chaffing and is stored on a drum like steel warp
coring work the loads will b!! generally within the (Schilling et al., 1988). The polyester cable was
 MARINE SEDIMENT CORE METHODS 93

ElectrIC Vibrator Hca_

Retraction Wire tension lin~s

Winch
Wire

Vertical \:uldc
Ilelllbers

300cm

200

100

r-;
ELEVATION

Hydraulic rotary
drill table

Hydraulic Electro-hydraulic
rctractlon power pack ~ reservoir
,

Acoustic penetration
measurement transducpr

PLAN

Fig. 5. The British Geological Survey's vibrocorer/rotary drill (from Pheasant, 1984).

chosen for the APe corer (Dzwilewski and Driscoll, than steel; they do not corrode; they can be cut and
1980), even though this cable cannot be stored on a spliced, and yet Kevlar is stronger than a steel cable
drum because it requires a long time for strain of equivalent diameter and polyester has about half
recovery after it has been loaded. This strain recov- the strength of steel. Polyester has about the same
ery would exert tremendous force on the drum and cost as steel for equivalent strength cable, but it has
so the cable is stored loose in a container on deck the disadvantage of needing a long recovery time
(Driscoll, 1981). Both types of cable stretch signifi- after it has been stressed, and it cannot, therefore be
cantly under load, which must be allowed for when wound on a drum. Kevlar, by contrast, costs about
calculating the length of wire required for the free- three times as much as steel, but it can be stored on
fall distances of piston corers. Both of these cables a drum.
have considerable advantages over steel cables: they Many oceanographic cruises today use charter
are light in water; they have smaller bending radiuses ships which may not have the necessary combination
94 P. P. E. WEAVER AND P. J. SCHULTHEISS

of winch, warp and "A" frame required for piston Continuous whole-core logging, with a range of
coring. Houbolt ( 1971) described a system different sensors would be of considerable scientific
combining all these features which could be trans- value for many reasons:
ferred from ship to ship. The Houbolt winch has (1) Complete continuous records of P-wave velocity
been considerably improved and is now marketed by and density (from gamma ray attenuation) are
the Dutch company Seabed BV. The winch is com- necessary to construct synthetic seismograms for
pletely independent of the ship, having its own diesel comparison with seismic reflection profiling
hydraulic power pack, 10 000 nm of Kevlar cable, a records and borehole velocity logs.
telescopic boom, a traction unit with 20-ton lifting (2) Continuous density logs are needed to evaluate
capacity, and two handling winches. The whole sys- the state of compaction of sedimentary se-
tem has the dimensions and fittings of a standard quences.
container and can thus be freighted around the world (3) Magnetic susceptibility logs provide rapid iden-
and bolted on to any ship with suitable container tification of terrestrially derived sediments and
anchorage points. can provide excellent, high resolution, interhole
correlations.
(4) Any logged parameter that is sensitive to subtle
Whole Core Logging
changes in sediment type or sediment structure
Having obtained a core in a plastic liner, there are a can be invaluable as a guide for later detailed
number of useful measurements that can be made sampling. They may, for example, reveal cycli-
prior to sectioning or splitting. These come under the city which is caused by climatic changes and
general heading of whole-core logging. Whole-core could be used directly for climate spectral analy-
logging is defined as non-destructive measurements SlS.
that can be automatically made on cores within plas- An example of the type of automated whole-core
tic core liners at frequent intervals or on a continuous logging system that is required is the P-Wave-Logger
basis. The most frequently used is gamma ray atten- (PWL) developed at the Institute of Oceangraphic
uation which essentially provides a log that is a func- Sciences Deacon Laboratory (Schultheiss and
tion of the sediment density. This technique has been McPhail, in press) for measuring the compressional
used on cores obtained from the Deep Sea Drilling wave velocity of sediments. The PWL (Fig. 6) acc-
Project for many years. Other non-destructive whole urately measures and automatically records the
core logging techniques which are being more fre- P-wave velocity of soft sediments within a cylindrical
quently used are P-wave velocity logs and magnetic plastic core liner and produces a very detailed velo-
susceptibility. It can be argued that much data is lost city log of the whole core.
from many sediment cores (especially those from the Velocity measurements are automatically taken at
ocean drilling legs) because the time constraints on- regular intervals along each core section as it travels
board ship preclude many measurements being made between a pair of ultrasonic transducers. The PWL
at appropriately small sampling intervals. Perhaps provides fine-scale velocity profiles that (a) enable
any useful parameter that can be measured continu- accurate correlations of adjacent cores or holes to be
ously and rapidly and does not require extra person- made, (b) provide high-quality data for synthetic
nel should be measured on a routine basis. seismograms, (c) help in the interpretation of seismic
The potential value of some whole-core logs is records and (d) indicate the nature of sedimentary
already clear (Schultheiss and Mienert, 1988); other features not easily detected by conventional means.
non-destructive logging techniques may produce The mechanical arrangement of the transducers used
data of only limited intrinsic value, or of a value that on the PWL is shown schematically as part of the
is currently unclear. However, while any single overall system diagram in Fig. 6a. An example of the
logged parameter may at worst be of marginal value, type of data obtained from very high resolution
the combination of logs is almost certain to provide P-Wave logging is shown in Figure 6b which is taken
a valuable diagnostic data set in the same way that from the ODP Leg 108 Initial Report (Schultheiss
down-hole logs are most useful when used as combi- and McPhail, in press), clearly illustrating the cyclic
nation logs. nature of the sedimentary sequence.
 VELOCITY <../sl
(a) P-Wr1te logger elect nI1ICS
& ~uter inlerloce
COlllJUler malib pIott~ lrove( time, (b) 0
1500 1550 1600
..,
16'0

cc.'e dilJlltler ond signal ~hllMle

i!

J 3:
;.-
5 :>:>
...~ Z
m
I!/
'"0m
4

...~ ~

..
i m
I Z
::>
-!
(')
0
6 :>:>
m
3:
m
-!
:t
0
0
8
I ~ '"

')
I
s..uc DtTDv........
108-6591-19
fref Qlps 10

Fig. 6. (a) Schematic diagram of the P-Wave-Logger (PWL); (b) an example of a detailed P-Wave velocity record (after Schultheiss and
McPhail, in press).

\0
V.
96 P. P. E. WEAVER AND P. J. SCHULTHEISS

Currently there is scope for the development of it in place during cutting as it does with a blade
a multi-sensor whole-core logging apparatus that, system.
in addition to providing a tool for the currently Core cutters which use blades operate by pulling
established techniques (gamma ray attenuation, two blades (one on each side of the core) along the
P-Wave velocity and magnetic susceptibility), length of the core. It has been found that sharp
would include other non-destructive measurements blades do not work because they cannot be made to
such as electrical resistivity, spectral natural gamma, cut absolutely straight, and once the cut has gone
radar, ultrasonic reflection and possibly neutron off line it is impossible to prevent the blade from
activation. breaking. The blades are therefore filed flat or bro-
ken, and this blunt edge is used. The force required
to pull two blades through PVC or polycarbonate
Subcoring
liner is considerable, and readily bends the liner if it
Box cores retrieve large volumes of sediment which is not held very rigidly. This requires very strong
are often subcored as soon as the core arrives on core holding equipment and it has been found that
deck. The lOS box corer is fitted with a vertical row I m-Iong cores are much easier to handle than
of sample ports which can be easily opened, giving 1.5 m-Iong ones.
access to a series of sediment layers which can be
sampled with syringes before the core is opened
Core Splitting
(Peters et al., 1980). If the upper closing spade of
the box corer is removed the sediment surface will The opening of stored cores is extensively discussed
be revealed which can be sampled by pushing in by Bouma (1969). The techniques are all simple,
lengths of core liner. Better cores will be taken if a using blades or wires to cut the sediment, although
piston is used which can be held at the sediment Bouma suggests applying a DC electric current to
surface with a small jig which is attached to the achieve an electro-osmotic effect if possible. The
corer frame. The box corer described by Papucci et present authors use a constant current DC power-
al. (1986) has a cylindrical barrel with a segmented supply capable of providing 0.5 amp at 30 volts,
liner thus allowing the core to be serially sectioned. and attach the cathode to the cutting blade and the
For geological investigations it is often important to anode to a platinum wire which is kept in contact
transport the whole core to the laboratory, and this with the sediment and in close proximity to the
may be done by using a plastic liner as described by blade during cutting (Weaver and Schultheiss,
Karl (1976), or by using interchangeable boxes as 1983b). Sediment surfaces cut in such a way are free
described by Bouma (1969). of smearing, show burrows and lithological
boundaries in minute detail (Fig. 8) and can reveal
small-scale sediment disturbances due to coring. A
Core Cutting
good quality colour photograph of an osmotic
Core liners can be cut with either a saw or a blade. knife-cut surface combined with a description pro-
Rotating saws are noisy and potentially dangerous duces a much better record of a core than a descrip-
and produce large amounts of plastic swarf which tion alone. We also use the osmotic knife to cut off
frequently becomes embedded in the sediment. the archive corner from the Kastenlot cores, and to
Other methods have therefore been sought which trim samples for geotechnical testing.
overcome these problems. Kawohl and Kudrass One other method of core splitting is to fracture
(1987) report the use of a vibrating saw which is the core in a controlled way. Fractured surfaces can
less dangerous than a rotating saw and produces a reveal the maximum detail within the sediments be-
curled thread of plastic instead of swarf. In their cause the surfaces have not been distorted due to
system two pairs of saws are used to cut a slice of local shearing as with a cut surface. This method is
sediment (Fig. 7) which can then be X-rayed and unpopular because of the semi-random split which is
sampled, leaving two intact strips of core for stor- achieved and the poor visual appearance of the
age. This system has the advantage that the core surface. It may be necessary to photograph the core,
does not require such heavy equipment to hold thus producing a two-dimensional image, before all
 MARINE SEDIMENT CORE METHODS 97

worll gur

..e

V-Ih,ptd
bid

Io'or" gur

Fig. 7. Vibrating saw for splitting sediment cores into three sections. The centre section is used for X-radiography (from Kawohl and
Kudrass, 1987).

the features are readily apparent. An accidental lon- Core Description and Photography
gitudinal core fracture was produced in a Kastenlot
core by Weaver and Schultheiss (1983a) which re- Cores are usually described immediately after split-
vealed numerous minute open burrows each about ting before any deterioration of the sediment
0.5 mm across. Such features would have been virtu- surface occurs. Traditionally, the sedimentologist
ally impossible to see in a cut surface even with only records all visible features of the core, such as
small amounts of smearing. colour, texture, lithological boundaries, burrows etc.
98 P. P. E. WEAVER AND P. J. SCHULTHEISS

Fig. 8. Examples of heavily burrowed sediment showing the clarity of detail produced by an electro-osmotic knife-cut surface.

The combination of a good quality colour pho- directly from the core, even if the core is not avail-
tograph from an unsmeared osmotic knife-cut sur- able, without having to also interpret the potentially
face with a multiple whole core log of the core can poor core description with its mis-interpretations or
provide much of the information in a more objective omissions.
way than from a simple description. Ideally, the
photographs and whole core logs should be backed-
Conclusions
up by comments from an experienced sedimentolo-
gist, and the whole must be synthesized into The present range of available corers includes box
diagramatic form for publication. One advantage of corers, gravity corers, piston corers, giant piston
this method is that information can be interpreted corers and vibrocorers. If -care is taken in selecting
 MARINE SEDIMENT CORE METHODS 99

the corer most appropriate for the sediment type and Burns, R. E., 1962, A Note on Some Possible Misinformation
most appropriate for the analyses to be conducted, a from Cores Obtained by Piston-Type Coring Devices, J. Sedim.
Petrol. 33, 950-952.
maximum of 20-30 m of core should be obtainable. Driscoll, A. H. 1981, The Long Coring Facility, New Techniques
Box corers provide good quality cores with large in Deep Ocean Coring, Oceans 81 1, New York, IEEE. Inc.,
volumes from the upper 0.5 m and often retrieve the 404-410.
Driscoll, A. H. and Hollister, C. D., 1974, The W.H.O.l. Giant
sediment-water interface. Single-spade box corers
Piston Corer; State of the Art, in Marine Technology Society,
are more successful in obtaining cores in sandy sub- 10th Annual Conference, 663-675.
strates. Gravity corers should be avoided, or cores Driscoll, A. H. and Silva, A. J., 1977, Report of the Engineering
from them treated with caution for physical property Workshop on Deep Sea Coring, Vols 1 and 2. Wood's Hole
Oceanographic Inst., November, 1977.
studies or any work which requires precise data on Dzwilewzki, P. T. and Driscoll, A. H., 1980, Long Core Facility
the depths of boundaries within the sediment. This is Winch and Cable System, American Society of Mechanical
because they are subject to errors of mis-sampling Engineers, Winter Annual Meeting, 7 pp.
Emery, K. O. and Dietz, R. S., 1941, Gravity Coring Instrument
which cannot always be assessed. The one exception
and Mechanics of Sediment Coring, Bull. Geol. Soc. Amer. 52,
may be the Kastenlot corer, which, because of its 1685-1714.
large cross-sectional area gives relatively undisturbed Hollister, C. D., Silva, A. J., and Driscoll, A. H., 1973, A Giant
cores in soft sediments. Piston corers produce more Piston Corer, Ocean Engineering 2, 159-168.
Houbolt, J. J. H. c., 1971, Transferable Deep-Sea Coring Gear,
representative results than gravity corers, but even Marine Geol. 10, 121-131.
these are subject to mis-sampling errors due to move- Hvorslev, M. J. and Stetson, H. c., 1946, Free-fall Coring Tube:
ment of the piston during the coring operation. A New Type of Gravity Bottom Sampler, Bull. Geol. Soc. Amer.
57, 935-950.
These corers can penetrate up to 20 m in suitable
Igarashi, Y., Ridlon, J. B., Campbell, J. R., and Allman, R. L.,
sediments. Giant piston corers are capable of coring 1970, Note on a Mode of Piston Core Disturbance, J. Sedim.
to 30-50 m. At present, the STACOR is operational Petrol. 40, 1351-1355.
and functions well, retrieving relatively undisturbed Karl, H. A., 1976, Box Core Liner System Developed at the
Sedimentology Research Laboratory, University of Southern
cores. The GPC has not completed its design phase California, Mar. Geol. 20, M1-M6.
and so its ability to recover long undisturbed cores is Kawohl, H. and Kudras, H. R., 1987, The Use of a Multiple-disc
currently unknown. Vibrating Saw for Cutting the Liners of Sediment Cores, J. Sed.
Pet. 57, 789-790.
New developments in corer handling include the Kermabon, A. and Cortis, U., 1968, A Recoilless Piston for the
use of Kevlar cables which allow more precise moni- SACLANTCEN Sphincter Corer, Sac/ant ASW Research Cen-
toring of operations and the production of a con- tre, Technical Report No. 112, 22 pp.
Kogler, F. c., 1963, Das Kastenlot, Meyniana 13, 1-7.
tainerized winch which allows access to a much
Kuijpers, A., Rispens, F. B., and Burger, A. W., 1984, Late
larger number of ships for running coring opera- Quaternary Sedimentation and Sedimentary Processes of the
tions. Madeira Abyssal Plain, Eastern North Atlantic, Meded. Rijks
Whole-core logging techniques combined with Geol. Dienst 38-2, 91-118.
Kullenberg, B., 1947, The Piston Core Sampler, Svenska Hydrogr.
colour photographs of osmotic knife-cut surfaces of Bioi. Kommn. Skr 1, 46 pp.
split cores provide excellent and permanent records Kullenberg, B., 1955, A New Core-Sampler, K. Vet. O. Vitterh.
of the core which are of more value than descriptions Samh. Handl. 6, 17 pp.
Lee, H. J. and C1ausner, J. E., 1979, Seafloor Soil Sampling and
on their own. The authors strongly commend the
Geotechnical Parameter Determination - Handbook, Technical
adoption of these techniques by other workers in the Report, Civil Engineering Laboratory, Port Hueneme, California
field. TR-873, 128 pp.
McCoy, F. W., 1980, Photographic Analysis of Coring, Mar. Geol
38, 263 - 282.
References McCoy, F. W., 1985, Mid-Core Flow-In: Implications for
Stretched Stratigraphic Sections in Piston Cores, J. Sedim. Pet.
Ardus, D. A., Skinner, A., Owens, R., and Pheasant, J., 1982, 55, 608-610.
Improved Coring Techniques and Offshore Laboratory Pro- Montarges, R., Fay, J-B., and Le Tirant, P., 1987, Soil Reconnais-
cedures in Sampling and Shallow Drilling, Oceanology Interna- sance at Great Water Depth, 4th International Conference on
tional2, 18 pp. Deep Offshore Technology, 2/18-2/29.
Bouma, A. H., 1969, Methods for the Study of Sedimentary Montarges, R., Le Tirant, P., Wannesson, J., Valery, P., and
Structures. J. Wiley & Sons, New York, 458. Berthon, J-L., 1983, Large-size Stationary Piston Corer, 2nd
Brooke, J. and Pelletier, B. R., 1970, Sea Drilling Techniques of International Conference on Deep Offshore Technology, 63-74.
the Bedford Institute, J. Underwater Sci. and Technol. 2, 165- Moore, D. G., 1961, The Free Corer: Sediment Sampling without
167. Wire and Winch, J. Sed. Petrol. 31, 627-630.
100 P. P. E. WEAVER AND P. J. SCHULTHEISS

Moore, T. C. and Heath, G. R., 1978, Sea-floor Sampling Tech- (Sites 657-668), in W. Ruddiman, M. Samthein, J. Saldauf et
niques, in J. P. Riley and R. Chester (eds.), ChemicalOceanog- al., Proc., Init. Repts. (Pt. A), ODP 108, 1015-1046.
raphy, 7, Academic Press (London), 75-126. Schultheiss, P. 1. and McPhail, S. D., 1989, An Automated
Papucci, C., Jennings, C. D., and Lavarello, 0., 1986, A Modified P-Wave Logger for Recording Fine Scale Compressional Wave
Box Corer and Extruder for Marine Pollution Studies, Conti- Velocities in Sediments, in W. Ruddiman, M. Samthein, J.
nental Shelf Research 6, 671-675. Baldauf et al., Proc., Init. Repts. (Pt. B), ODP 108 (in press).
Peters, R. D., Timmins, N. T., Calvert, S. E., and Morris, R. 1., Schuttenhelm, R. T. E., Auffret, G. A., Buckley, D. E., Cranston,
1980, The lOS Box Corer: Its Design, Development, Operation R. E., Murray, C. N., Sheppard, L. E., and Spijkstra, A. E.
and Sampling, I.O.S. Report No. 106, 16 pp (unpublished (eds). 1990, Geoscience Investigations of Two North Atlantic
manuscript). Abyssal Plains- the ESOPE International Expedition, Vols 1
Pheasant, J., 1984, A Microprocessor Controlled Seabed Rock- and 2, CEC-Joint Research Centre, JRC Report (in press).
drilljVibrocorer, Underwater Technology 10, 10-14. Silva, A. J., Hollister, C. D., Laine, E. P., and Beverly, B. E.,
Richards, A. F. and Keller, G. H., 1961, A Plastic-Barrel Sedi- 1977, Geotechnical Properties of Deep-Sea Sediments: Bermuda
ment Corer, Deep-Sea Research 8, 306-312. Rise, Mar. Geotechnol. 1, 195-232.
Schilling, J., Van Weering, T. C. E., and Eisma, D., 1988, Weaver, P. P. E. and Schultheiss, P. J., 1983a, Vertical Open
Advantages of Lightweight Kevlar Rope for Ocean Bottom Burrows in Deep Sea Sediments 2 m in Length, Nature 301,
Sampling with Piston Corer and Box Corer, Mar. Geol. 79, 329-331.
149-152. Weaver, P. P. E. and Schultheiss, P. J., 1983b, Detection of
Schultheiss, P. 1. and Mienert, J., 1988, Whole Core P-Wave Repenetration and Sediment Disturbance in Open Barrel Grav-
Velocity and Gamma Ray Attenuation Logs from ODP Leg 108 ity Cores, J. Sedim. Petrol 53, 649-654.
Observation of Corer Penetration and Sample Entry during
Gravity Coring
W. R. PARKER and G. C. SILLS
Blackdown Consultants, Taunton and Department of Engineering Science, Oxford University, UK

(Received 27 April, 1989; I September, 1989)

Key words: gravity coring, core-shortening, corer penetration, the corer as it is pulled from the seabed. This chapter
sample entry, observations. described measurements of penetration of the corer
into the bed and of entry of the sample into the corer,
made during coring. Any losses that occur at this
Abstract. Gravity core samples provide the basic data source for a stage will be crucial in subsequent analyses which rely
wide range of geological, geotechnical and geochemical studies. on a knowledge of depth of the sample in the seabed.
However, the length of the core recovered is often less than the
penetration achieved by the corer, such cores being described as
"shortened". If the penetration of the corer has been measured, Equipment and Methods
and it is assumed that no dropout of the core occurs as the barrel
is withdrawn from the seabed, it is present practice to reconstruct A system to measure corer penetration and sample
in situ dimensions using an overall correction factor based on this
penetration and the length of core recovered. However, measure-
entry was developed using two acoustic transducers
ments, reported here, have been made of corer penetration and mounted on an adaptor fitted between the core
sample entry and these show that the entry deficit (penetration barrel and corer weight chassis. One transducer was
minus sample entry) develops in some instances continuously, and
directed parallel to the barrel outside the corer to the
in others intermittently. These results indicate that an overall
correction factor is unlikely to be appropriate to any given section seabed and the other inside the core barrel to the
of the core. surface of the sample. The transit time from sending
a signal to receiving its reflection from the sediment
was measured with a dual channel echo-ranging
Background
system. The time interval was converted to distance
A number of authors (e.g. Emery and Deitz, 1941; by measurement of the speed of sound in water,
Hvorslev, 1949; Emery and Hulsemann, 1964; Ross using an external target at a known distance on the
and Riedel, 1967; McCoy, 1971; Lebel et al., 1982; outside of the corer barrel and the core catcher as a
Weaver and Schultheiss, 1983) have reported obser- internal target. An analogue DC output proportional
vations made during the process of gravity coring to distance was recorded onboard ship. The opera-
that the recovered core length is frequently less than ting frequency of the transducers was chosen to be
the depth of penetration of the corer into the sedi- centred around 500 kHz, and the pulse repetition
ment. McCoy (1971), Lebel et al., (1982) and rate was approximately 400 Hz. The system allows
Weaver and Schultheiss (1983) have discussed the continuous monitoring of the corer penetration and
causes of this phenomenon and have investigated its withdrawal, the sample entry and, if it occurs, subse-
nature by a variety of comparative observations. quent dropout. Under field conditions the overall
Although penetration of the corer has sometimes resolution of the system is 5 mm.
been measured during the coring process, the corre- Two types of corers were studied: first a cylindrical
sponding measurement of the sample entering the barrel gravity corer with liner LD. = 83 mm, barrel
corer has not, in general, been made. Sample lengths O.D. = 101 mm and barrel lengths of 3.17 and
can be measured after recovery, but these values may 3.30 m; and secondly a 150 mm x 150 mm square
be affected by the further process of "dropout" from section Kasten corer with a 2.3 m barrel. The area

Marine Geophysical Researches 12: 101-107, 1990.

1990 Kluwer Academic Publishers.
lO2 W. R. PARKER AND G. C. SILLS

ratio (Hvorslev 1949, Emery and Hulsemann 1964) of with peak values of the order of 10 kN/m 2 and
the cylindrical barrel corer was 0.33 and that of the residual values of 2 kN/m 2 reached by a depth of
Kasten corer was 0.05, values which are typical for 2m.
such corers.

SITES Data Analysis

Cores were collected at five sites in the northeast Irish Changes in distance to the sediment surface outside
Sea. Detailed observations are presented for two the corer give approach speed and penetration, P,
sites, IS05 at 54° 24' N, 3° 41' Wand ISO 1 at and inside the core barrel yield sample entry, E. In
54° N 3 °39'W. The sediment of the area is a silty sand this analysis core entry was determined from the
mud (Pantin, 1978, Williams et aI., 1981), which has chart records at each 100 mm increment in penetra-
been heavily bioturbated to considerable depth (Ker- tion. Core entry deficit (P-E) and core entry ratio
shaw et al., 1983, Kershaw et al., 1984). X-radio- (E / P) were thus determined throughout corer pene-
graphs of cores show that generally it is of tration. The incremental entry deficit, which repre-
homogeneous texture (Williams et al. 1981; Kirby et sents the entry deficit arising in each 100 mm of
al. 1983). At each site, seabed density profiles were penetration, and the incremental entry ratio (the
measured in situ using a nuclear backscatter density ratio of entry to penetration in each penetration
probe (Parker et al. 1975). Since the sediment is increment) were also calculated. Increases in range
contaminated by discharges of low-level radioactive from the external echo sounder allow withdrawal to
waste from Sellafield, the results were corrected for be monitored and increases in range from the inter-
background radiation. Density in core subsamples nal sounder indicated sample dropout. These values
was measured onboard using a vibrating tube den- can be correlated with recovered core lengths.
simeter (that converts from natural frequency of
vibration to density) and, in the laboratory ashore,
Results
by moisture content determination. These data
taken together suggest a similar pattern at both Core entry ratios have been calculated for 34 deploy-
sites. Densities increase downwards in the top 0.4 m ments of a 3 m-long cylindrical barrel corer and 14
of the bed to approximately 1.7 Mg/m 3 , then increase deployments of a 2 m-long Kasten corer. The results
more slowly to about 1.9 Mg/m 3 by 1.4 m, thereafter are plotted against approach speed in Fig. 1. The
remaining more or less constant to 2 m. Shear groupings of the data reflect available winch speeds,
strength was measured on core samples to a depth of but the actual approach speed was derived from the
2 m using a standard vane and sensitive torque trans- chart record to an accuracy of 0.05 m/sec. The entry
ducer. Results show strength increasing with depth, ratios vary non-systematically from one core to

~ o·g
1·0 , •• • •
l1J
•
•
0
;; o·e
00
~
~ 0·7
<f!>0
Ilf
o 0
'boo
S
8
• 0

0
>- 0·6 0 0
t:t:
ffi 0·5 0

tl! 0·4
o o Cylindrical barrel
u 0.3
w • Kasten square barrel
c5 0·2
::c
~ 0·1
O L-~-L~__L-~-L~__L-~-L~~L-~-L~~L-~-L~~
a 0·1 0·2 0·3 0·4 0·5 (}6 0·7 o·e o·g 1·0 1-1 1·2 1·3 1·4 1·5 1·6 1·7 I·e 1·9 2 ·0
APPROACH SPEED (metres/sec )

Fig. I. Calculated care entry ratios for corer types.

 CORER PENETRATION AND SAMPLE ENTRY 103

another, even when speeds of entry are similar. The tion. The differences between whole-core entry ratio
cylindrical barrel cores show no significant trend in and local entry ratio are also clearly visible.
entry ratio for approach speeds up to 1.9 m/sec,
although the range of values increases at the higher
Discussion
speeds. If the core entry ratios are calculated for the
barrel cores for the first 2 m of penetration (to Previous authors ( e.g. Emery and Deitz, 1941;
make them directly comparable with the Kasten Hvorslev, 1949; Emery and Hulsemann, 1964; Mc-
core results), the individual values change, some Coy, 1971; Label et al., 1982 and Karnes et al.,
increasing and some decreasing, by amounts gener- 1980) have suggested that approach speed influences
ally up to about 0.04, but the overall pattern of entry ratio in barrel corer operation. For the barrel
variation remains very similar. At approach speeds corer tested and at the approach speeds used these
below 0.6 m/sec, higher entry ratios are achieved data show little systematic variation in entry ratio.
with the Kasten corer than with the cylindrical bar- The Kasten corer tested provided cores with gener-
rel corer. At speeds above 1.0 m/sec, the entry ratio ally good entry ratio except at high approach-speed.
in the Kasten cores declines to that for the cylindri- Several authors have suggested that entry deficit
cal barrel cores. develops linearly (Emery and Deitz, 1941; Emery
and Hulsemann, 1964; and Lebel et al., 1982) or
ENTRY DEFICIT AND ENTRY RATIO PROFILES
increases with penetration (Ross and Riedel, 1967;
Profiles of cumulative and incremental entry deficit Weaver and Schultheiss, 1983) and that the entry
vs total penetration for the barrel corer are shown ratio for the whole core is representative of that
in Figs 2 and 3. The cores illustrated were taken in for a particular level (Emery and Hulsemann,
a period of 2 hours with the ship at anchor at site 1964; Lebel et aI., 1982). Data illustrated in Figs
IS05. Only the approach speed varied and the profi- 2-4 show clearly that whole-core values mask the
les are arranged in order of increasing speed. Al- irregularity in the distributions of entry deficit or
though the cumulative entry deficit increases entry ratio which do not necessarily develop either
steadily in cores 44 and 45 (Fig. 2), the incremental continuously, linearly or increasingly with depth.
deficit is irregular. This is more clearly shown in For example, in Fig. 2a, the incremental entry
core 47 (Fig. 3a). An extreme example is shown in deficit is distributed around the whole-core entry
core 46 (Fig. 3b), where adjacent sections of core ratio over the whole depth of penetration, while
show incremental deficits of zero (complete entry Fig. 2b shows the incremental entry deficit to be
locally) and 60 mm (in 100 mm penetration only generally lower than the whole-core entry ratio in
40 mm entered). Also indicated in Figs 2 and 3 are the upper 2 m of the core, and higher in the metre
the whole core entry ratio and the incremental entry below.
ratio (scale from right to left) clearly illustrating the It has been suggested that sample losses may be
local variability in entry ratio compared to the due to compression. However, on the timescale of
whole core value shown. corer penetration (10-60 sec) compression of the
Cumulative and incremental entry deficit profiles sediment framework requiring expulsion of pore
for four Kasten cores from site ISO} are plotted in water is not feasible for the type of sediment sam-
Fig. 4. Also marked are the whole-core entry ratio pled. The difference between penetration and sam-
and incremental entry ratio scale. Only the approach ple length must therefore arise as a consequence of
speed was varied and the profiles are arranged in sediment not entering the core barrel but being
order of increasing speed. The profiles for cores 68 pushed aside to a greater or lesser degree with, at
and 78 are typical of the high entry ratio achieved. times, no sediment entering. It is possible that mate-
Cores 76 and 73 illustrated the values of direct rial may be carried down ahead of the corer to
continuous observation by revealing the distribution enter later when the balance of forces influencing
of entry deficit. Core 73 shows significant deficit sample entry changes. Thus, the entry deficit at the
growth in the upper two thirds of the core with none particular level of penetration does not necessarily
below 1.6 m, whereas core 76 shows most of the indicate the absence of the stratigraphic level in the
entry deficit arising during the latter I m of penetra- core. Cores which are shorter than corer penetra-
104 W. R. PARKER AND G. C. SILLS

CUMULATIVE ENTRY DEFICIT (met res) INCREMENTAL ENTRY D:FICIT

o0 0,1 0·2 0·3 0·4 0·5 006 0·7 0·8 0·9 1·0 1-1 o 20 40 60 80 (mm)

0·2
+

+
0·6 +
+
0·8 +
+
1·0 +
+
+
+
+
z +
o 1·6 +
.... +
....~ 1·8 +
+
~ 2·0 +
w +
n.
2-2 +
+
2-4 +
+
2·6 +
+
2·8 +
+
3·0 +

INCREMENTAL ENTRY RATIO

CORE NO SPEED m/sec E/P,0'65
1544 BCA 0·15

Fig. 2(a) Cumulative and incremental entry deficit profiles: barrel core 44.

CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT

(mml
0·1 0·2 0-3 0·4 0·5 0·6 0·7 0·8 O·g 1-0 H 1·2 0 20 40 60 SO 100
00

0·20 +
+
0·40 +
+
0·60 +
+
0-80 +
+
_ 1·00 +
+
=1·20
~
+
'"
~ 1'40
+
+
z +
01·60 +
;:: +
~ 1·80 +
....
w +
~2-00 +
n. +
NO +
+
HO +
+
2-60 +
+
2-80 +
+
3{)0 +
1·0 o-S 0·6 0·4 0·2 0
INCREMENTAL ENTRY RATIO

CORE NO SPEED m/sec E/P,O·SI

IS 45BCA 55

Fig. 2(b) Cumulative and incremental entry deficit profiles: barrel core 45.
 CORER PENETRATION AND SAMPLE ENTRY 105

CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT

o0 0·1 (>2 0·3 0·4 0·5 0·6 0·7 0-8 (}9 1·0 )·1 o 20 40 60 80 100
(mml
0·2 +
+
0·4 +
+
0·6 +
+
0·8 +
_ 1·00
+
+
~ +
0; 1·20 +
E +
1·40 +
z +
o 1-60 +
~ +
g:w 1·80 +
&2·00 +
+
+
2·20 +
+
NO +
+
2-60 +
2·80
+
+
3·00
+
+
1·0 0·8 0·6 0·4 0·2 0
INCREMENTAL ENTRY RATIO
CORE NO SPEED mIse:: E IP=0·63
IS 47 SCA 1·21

Fig. 3( a) Cumulative and incremental entry deficit profiles: barrel core 47.

CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT

o 0·10 0·20 O·Xl 0·/,(l. (}5() (>60 0·70 0·00 (}9() HXl o 20 40 60 80 100 (mm)
Or-'--r--r-'--.-'--'--.-,~

0·20
0·40 +
+
(}60 +
+
o-ao +
+
1·00 +
+
-;;; 1·20 +
~ + i
~ 1-40 + , WHOLE CORE

z 1·60
+
+ ~ ENTRY RATIO
0 +
~ 1·80 +
Q: +
Iii 2·00 +
z +
~ 2·20 +
+
NO +
+
2·60 +
+
2-80 +
+
3·00 +
)·0 (>8 0·6 0·' 0·2 0
INCREMENTAL ENTRY RATIO

CORE NO SPEED m/sec E/P =O·77

IS 46 SCA 1·64

Fig. 3(b) Cumulative and incremental entry deficit profiles: barrel core 46.
106 W. R. PARKER AND G. C. SILLS

CUMULATIVE ENTRY
DEFICIT (metres) INCREMENTAL ENTRY DEFICIT I mm) CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT
o 0·1 0·2 0·3
o ;........;:..,-..c:..,::.--, a 20 1.0 SO 80 100 a O·XJ 0·20 0·30 O ·~ 0·50 0 20 1.0 60 eo 100 (mm)

0·20
+
0·1.0
WHO LE CORE
ENTRY RATI O +
060 +
+
OSO +
iii
I!! +
!
1·00 +
+
z 1-20+
..... + 0
;= 1-1.0
+
UJ 1·1.0 + <I: +
z
V
w + Q: +
+ WHOLE CORE
a.. HiO + UJ 1·50
I-

+ z + ENTRY RATIO
w +
l·eo a.. 1·80
+ +
2'00 + 2·00 +
+ +
NO + 2·20 +
1·0 0·8 o·s 0'1. 0'2 0 +
INCREMENTAL ENTRY RATIO 2·1.0
1·0 oe os 01. 0·2 0
INCREMENTAL ENTRY RATIO

CORE NO SPEED m/sec E/P.0·97

IS se KCA 0-11. CORE NO SPEED m/sec E/P,0·97
IS7SKCA 0·1.7

CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT (mm)

CUMULATIVE ENTRY DEFICIT (metres) INCREMENTAL ENTRY DEFICIT (mm) a 0.10 0·20 030 O·~ 050 060 0-70 0 20 1.0 60 80 KlO
Or-'--.--r-.--r--'~

00
0·1 02 0·3 0·1. 0·5 0·5 o 20 ~ 60 80 100 +
0'20 +
+
0·2 0,'0 +
+
0·' 0-60 +
+
~ ~0-80 +
l; + +
E D-8 + ~ 1-00 +
+ +
z 1·00 + z 1·20 +
0 + 0
+
~ 1·20 + ~ 1-1.0 +
I- + II::
I- +
w 1-1.0
z + w 1·60 + • WHOLE ~E
z
~ 1-60
+
+ ~ 1'80
+
+
~ ENTRY RATIO
+ +
I-SO + 2·00 +
+ + I
2-00
1·0 0·8 OS 0·4 0-2 0 2-20 +
+ i
INCREMENTAL ENTRY RATIO 2-40
~o 0·8 0·6 0-4 0·2 0
INCREMENTAL ENTRY RATIO
CORE ~ SPEED m/sec E/P: Q.8S CORE NO SPEED m/sec E/P. 0·72
IS 751KA 0·20 IS 73 KCA 144

Fig. 4. Cumulative and incremental entry deficit profiles for four Kaster cores: site ISO I.
 CORER PENETRATION AND SAMPLE ENTRY 107

tion can occur as a result of entry decifit of are due to Dr. D. MeG. Elder, currently with Soils
"dropout". The description "core-shortening" sus- and Foundations Ltd, New Zealand, and previously
tains the concept of overall volume reduction during at Oxford, who provided the shear strength and
coring which is a misconception of the process in- density data.
volved. Ideally, data derived from gravity cores
should be accompanied by an entry deficit or entry
ratio profile. References

Emery, K. O. and Hulsemann, J., 1964, Shortening of Sediment

Conclusions Cores Collected in Open Barrel Gravity Corers, Sedimentology
3,144-154.
The Kasten corer produces substantially less entry Emery, K. O. and Deitz, R. S., 1941, Gravity Coring Instrument
deficient samples than the barrel corer. This is most and Mechanics of Sediment Coring, Bull. Geol. Soc. Amer. 52,
probably due to the difference in area ratio between 1685-1714.
Hvorslev, M. J., 1949, Subsurface Exploration and Sampling of
the two corers. For the cylindrical barrel equipment Soils for Civil Engineering Purposes, US Corps Engs. Water-
and sediment type reported here whole-core entry ways Expt. Sta. Tech. Repts.
ratios vary non-systematically from 0.50 to 0.82 for a Karnes, C. H., Burchett, S. N., and Dzwilewski, P. T., 1980,
Optimised Design and Predicted Performance of a Deep Ocean
range of approach speeds up to 1.9 m/sec. The incre- 50m Piston Coring System, I.E.E.E. Publication No.
mental entry ratio varies with penetration and may 80CHI572-7, pp. 231-239.
be significantly larger or smaller than the whole-core Kershaw, P. J., Swift, D. J., Pentreath, R. J., and Lovett, M. B.,
1983, Plutonium Redistribution by Biological Activity in Irish
entry ratio. Thus, whole-core ratios are not a guide
Sea Sediments, Nature 306, 774-775.
to sample deficiency at particular levels within a core Kershaw, P. J., Swift, D. J., Pentreath, R. J., and Lovett, M. B.,
and their use in "linear stretching" of core profiles 1984, Science of the Total Environment 40, 61-81.
may compound existing deficiencies. Entry deficit Kirby, R. Parker, W. R., Pentreath, R. J., and Lovett, M. B.
1983, Sedimentation Studies Relevant to Low-level Radioactive
profiles do not allow complete reconstruction of the Effluent Dispersal in the Irish Sea, Rep. No. 178, Institute of
position of individual sections of core. However, Oceanographic Sciences, Taunton.
they do identify those parts of the penetration event Lebel, J., Silverberg, N., and Sundby, B., 1982, Gravity Core
Shortening and Pore Water Chemical Gradients, Deep Sea
during which entry of the sample is deficient or Research 29, 1365-1372.
absent and thereby indicate the quality of the core McCoy, P. W., 1971, An Analysis of Piston Coring through
for studies which rely on its dimensional integrity. Corehead Camera Photography, Underwater Soil Sampling,
Testing and Construction Control, A.S.T.M., STP SOl, 90-105.
These results emphasize the need to choose a corer
Pantin, H. M., 1978, Quaternary Sediments from the North-east
which generates minimum entry deficit and demon- Irish Sea: Isle of Man to Cumbria, Bulletin of the Geological
strate the desirability of measuring the pattern of Survey of Great Britain, No. 64.
entry deficit for each core. Parker, W. R., Sills, G. c., and Paske, R. A. E., 1975, In situ
Bulk Density Measurement in Dredging Practice and Control,
First IntI. Symp. Dredging Technology, BHRA, Cranfield.
Ross, D. A. and Riedel, W. R., 1967, Comparison of Upper
Acknowledgements
Parts of some Piston Cores with Simultaneously Collected
Open-Barrel Cores. Deep Sea Research 14, 285-294.
This work was supported by the Department of the Weaver, P. P. E. and Schultheiss, P. J., 1983, Detection of
Environment under contract No. PECD/7/9/172- Repenetration and Sediment Disturbance in Open-Barrel
150-83 and by the Ministry of Agriculture, Fisheries Gravity Cores, Jour. Sed. Pet. 53(2), 649-654.
Williams, S. J., Kirby, R., Smith, T. J., and Parker, W. R., 1981,
and Food. The contribution of colleagues in Oxford,
Sedimentation Studies Relevant to Low-Level Radioactive Efflu-
and in Lowestoft, and the officers and crew of R.Y. ent Dispersal in the Irish Sea. (Part II), Rep. No. 120, Institute
Cirolana is gratefully acknowledged. Special thanks of Oceanographic Sciences, Taunton.
Ocean Drilling Program (ODP) Deep Sea Coring Techniques*
M. A. STORMS
Ocean Drilling Program, Texas A & M University, U.S.A.

(Received 27 April, 1989; accepted I September, 1989)

Key words: coring, wireline, scientific, ocean, downhole, deepwater mission to recover marine geology core samples by
Resolution, Challenger, ODP, DSDP. drilling a series of holes (transect) across the Mid
Atlantic Ridge. The intent of this research program,
Abstract. The coring techniques and systems of the Ocean Drilling managed by the University of California, San
Program (ODP) were developed to satisfy a scientific need for Diego, Scripp's Institution of Oceanography, was to
better quality and improved recovery of oceanic core samples. prove or disprove once and for all the theory of sea
Some of the ODP systems in use today evolved from refinements
to earlier systems developed by the Deep Sea Drilling Project floor spreading. The effort utilized a specially de-
(DSDP) or were adaptions of available industry technology. Other signed, technically advanced, dynamically posi-
systems were conceived and designed by ODP engineers. The tioned drill ship, called the Glomar Challenger (Fig.
evolution of these progressive scientific coring systems began with
the Rotary Core Barrel (RCB) used by the DSDP and proceeded
1). This drill ship was designed and operated by
to the highly advanced Diamond Coring System (DCS) currently Global Marine Incorporated (GMI). Using the
under development by ODP engineers. During the evolution sev- "Challenger" the DSDP research program success-
eral other key systems were developed. These included hydraulic
fully recovered the deepest cores ever recovered
piston coring, extended coring, pressure coring, bare rock spud-
ding and several systems using high speed diamond coring tech- from an offshore drill ship and successfully proved
nology. These include the positive displacement coring motor the theory of sea floor spreading. The predominant
(PDCM), the "Navi-drill" core barrel, and a top driven diamond coring system used for this drift research effort was
coring system (DCS). This article describes the evolution and
conceptual design of these systems including the required bottom a wireline retrievable rotary coring system devel-
hole assemblies, and the sinker bar/sandline configurations. oped by Hycalog, a prominent "oilfield" service and
supply company. Due to the dramatic success of
this "pilot" program and the immense scientific po-
Historical Background tential of continued scientific coring, the project was
extended and went on to recover literally miles of
The coring systems in use today by the Ocean
deep-sea core samples and over a period of 15 years
Drilling Program (ODP) evolved both as refinements
proved and disproved many marine geological theo-
to earlier systems developed by the Deep Sea Drilling ries and ideas.
Project (DSDP) and because of a continual need by
The follow-on project to DSDP began in 1983
the marine geology community as a whole. In gen-
and was christened "the Ocean Drilling Program"
eral, the coring systems were developed from unique, (ODP). The ODP is managed by Texas A & M
internally generated, concepts or were adapted from University and operates the state-of-the-art scientific
existing industry systems to satisfy a scientific need
drill ship Joides Resolution registered as the
for better quality and improved recovery of oceanic
SEDCO/BP 471 (Fig. 2). The principal funding
core samples.
agency for the ODP remains the National Science
In 1968, the National Science Foundation-funded
Foundation; however the program receives signifi-
(DSDP) embarked on an eighteen-month scientific
cant monetary, scientific and technical support from
its international partners which include: the Federal
* All data, text, figures, photo's etc. contained within this chapter Republic of Germany, France, Japan, the United
are public domain and free from copywrite. All rights of ODP,
NSF, JOI, or other private individuals to use any or all of this Kingdom, Canada and Australia and the European
material in future published documents is reserved. Science Foundation Consortium.

Marine Geophysical Researches 12: 109-130, 1990.

110 M. A. STORMS

check valve installed in the swivel allowed water to

be displaced out of the top as the core entered the
barrel at the bottom. This feature was designed to
minimize core erosion and lost recovery. The check
valve also prevented the core from being washed out
of the core barrel during retrieval back to the drill
ship.
Although the coring system was effective initially
for recovering hard-rock core, it was not optimum
for other types of lithologies, particularly the softer
sediments encountered at or near the mudline. The
upper 100 to 300 m of the sedimentary column were
Fig. I. DjV Glomar Challenger.
often highly disturbed by the rotary drilling process.
Attempts at detailed sampling for disciplines such as
paleoceangraphy, paleoclimatology, magnetostrati-
graphy, and high-resolution stratigraphy were all but
impossible. Initially, improvements were made to the
basic rotary coring system such as new spring-loaded
dog-type core catchers and more efficient venting at
the top of the core barrel to minimize the back
pressure that tends to inhibit core entry. These
changes helped to improve overall system perfor-
mance, particularly core recovery, but still fell short
of the scientific requirement for good quality, soft
formation cores.
It was apparent that some means to overcome the
limitations of rotary drilling in unlithified sediments
was required. Oceanographic piston cores studies
Fig. 2. DjV Joides resolution. had provided means of distinguishing events re-
corded in sediments to a precision of thousands of
The initial coring system used by DSDP was wire- years; such events were being homogenized by
line retrievable and represented a reasonably efficient DSDP's rotary coring. Although oceanographic ves-
system for recovering continuous core samples from sels were routinely taking piston cores of mudline
the deep ocean. The 2.44 in. (6.20 cm) diameter by sediments, these "conventional" piston coring sys-
32 ft (9.76 m) long core was cut by the rotation of a tems were limited to just a few tens of meters of the
tungsten carbide insert (TCI) roller cone core bit (see surface material and lacked the capability to achieve
ODP operation in Fig. 5). The core was received in a any significant depth of penetration.
clear plastic, acetate butyrate core liner installed The original DSDP hydraulic piston coror (HPC)
inside the wireline retrievable core barrel. A collet- was born of rotary coring shortcomings. The inten-
type core catcher retained the hard, crystalline rock tion was to develop a high-quality coring system,
core and kept it from falling out of the bottom of the compatible with development through the drill string
barrel during retrieval. The core barrel was equipped and capable of penetrations up to 200 m below the
with a latch at the top to counteract the torque sea floor. The HPC system was developed to meet
generated by the coring process. A swivel assembly, this mandate. It has since been refined through sev-
installed immediately below the core barrel latch, in eral iterations to the enormously successful Ad-
conjunction with a lower support bearing installed as vanced Piston Corer (APC) system in use today in
a landing shoulder in the outer core barrel assembly, the ODP (see ODP operation in Fig. 10). Major
prevented the inner core barrel from rotating during hydraulic piston coring refinements that have taken
the coring process and thus disturbing the core. A place over the years include expanding the core
 ODP DEEP SEA CORING TECHNIQUES III

bottom hole assembly required for piston coring

APClXC B/NCB ReB
operations was not compatible with the rotary coring
system. The core bit for HPC coring has the cones
spread apart to allow passage of the piston core
barrel out into the formation ahead of the bit face.
(2) 5- 1/2" HEAV Y "'All
DRILL PIPE This bit would cut a 3.88 in (9.86 cm) OD core,
much too large to fit inside the standard 2.88 in
(7.32 cm) ID rotary core barrel. This meant that
7-1/ 4" DRILL COLL AR
upon completion of each piston cored hole, the drill
string would have to be tripped out and the BHA
changed to allow deeper objectives and/or basement
8- 1/4" DRILL COLLAR S
to be reached. Obviously this was a very time-con-
suming operation and was not an efficient way to
HEAD SUB DR
NCB L ATCH SUB continuously core in the deep oceans.
The extended core barrel (XCB) was developed to
rectify this situation by allowing a "driven" rotary
L ONG TOP SUB coring system to be deployed in the APC BHA (see
L ANDING SAVER SUB
ODP operation in Fig. 13). The XCB design required
' SMOOTH BORE'
OUTER CORE BARREL
an independent cutting shoe to trim the core from
'CONTROL LENGTH' the 3.88 in (9.86 cm) OD size provides from the
OUTER CORE BARREL
BIT SUB "'/lOCKABLE
roller cone bit to a nominal 2.31 in (5.87 cm) OD size
flOAT VALVE
(OR MBR ASSEMBL Y) BIT SUB that could be recovered inside the core barrel. The
(DR MBR/HBR
ASSEMBL Y> development of the XCB system allowed the deepen-
XCB BIT
11 - 7 / 16 X 3 8
.7 ing of single bit HPC holes without requiring the
tripping of the drill string and subsequent re-spud-
ding of the hole. As with the HPC, the XCB develop-
ment has continued through many iterations
Fig': 3. Typical ODP bottom hole assembly configurations. throughout the years. Enhancements to the system
included better circulation to the cutting shoe,
length from the original 14.8 ft (4.5 m) to the present strengthened connections for deeper drilling into
31.2 ft (9.5 m), increasing thrust to allow coring into more indurated formations, the incorporation of a
stiffer formations, adding a core orientation capabil- "venturi" venting assembly resulting in further re-
ity, adding temperature measurement capability, duction of core barrel back pressure, and many
keyed piston rods to eliminate core barrel spiraling, hardware improvements geared toward simpler,
modified shear subs for more consistent shearing more efficient shipboard operations. The current ver-
pressures, and finally refining all components for sion of the XCB has drilled in excess of 1000 m
more simple and efficient operations. below the sea floor and has significantly improved
The current APC system has cored in excess of recovery and quality in most formations.
300 m below the sea floor (mbsf) and routinely pro- Lithologies which still create problems for the
vides 90 to 100% core recovery. It has the capability XCB are interbedded soft/hard formations (e.g.
to penetrate into semi-indurated formations which chalk/chert), unconsolidated material such as turbid-
can then be rotary cored without inducing significant ities or loose-flowing sands, and basement-type crys-
core disturbance. As such, the APC is now the more talline rock. The ODP is pursuing the evaluation of
widely used coring system in the ODP arsenal. hydraulic percussion hammer (vibra-coring) tech-
The tremendous success and acceptance of hy- niques as a possible way to improve recovery in
draulic piston coring within DSDP led to the devel- unconsolidated formations. Improved basement and
opment of the extended core barrel (XCB) coring crystalline rock coring systems are being developed
system. With the almost exclusive use of the HPC using small diameter, narrow kerf, high speed, dia-
system for spudding holes a dilemma arose. The mond coring techniques.
112 M. A. STORMS

The development of ODP diamond coring began tems (APCjXCB/NCB and PCS) are completely in-
with the navi-drill core barrel (NCB). The NCB is terchangeable at any point in the coring cycle. Any
the third of several coring systems either designed or of the coring systems may be substituted for another
under development within DSDP/ODP to enable when necessary to meet the scientific objectives for
single-bit holes to be drilled with the optimum coring the particular formation being cored.
system from the mudline down to and possibly into Independent of the single bit hole "optimized"
basement. As with the APC and XCB coring systems coring systems described above, the ODP was also
the NCB is fully compatible with the APCjXCB directed to develop the capability to spud (initiate
BHA. This system is being developed to allow coring drilling and coring) holes on bare rock, where the
to continue once basement or crystalline rock has basement rock is exposed to the sea floor. These
been reached with the XCB. When fully operational "zero-age" crustal formations are typically highly
the NCB should allow limited coring (i.e. 10-50 m) fractured, poorly cemented, and generally very un-
into basement. stable. The ability to successfully spud holes, make
The most recent ODP coring system currently significant penetration, and maintain any reasonable
under development to be compatible with the APCj core recovery percentage in this environment had
XCB BHA is the pressure core sampler (PCS). The been a major problem for the DSDP. Early in the
predecessor to the ODP pes was the DSDP pressure ODP (November 1985) the capability was estab-
core barrel (PCB). The DSDP version, although lished to allow successful spud-in on bare rock. The
successful at recovering cores under pressure, had sea floor structure developed was referred to as a
several major shortcomings that eventually led to the "hard rock guide base" (HRB) and consisted of a
scientific mandate for an improved pressure coring massive steel base plate mated with a gimbled re-
capability for the ODP. The PCB was developed entry cone above the cement filled bags below. The
before the APC revolution. It therefore was only technique was time-consuming and the hardware
compatible with the rotary coring system (RCB). relatively expensive, but the capability to deploy the
Today the RCB system is only rarely used for coring HRB and spud-in on bare rock was demonstrated on
softer lithologies and so the PCB was incompatible ODP legs 106 and 118.
for most desired applications. The pressure chamber Although the ability to spud-in on fractured rock
was very long and cumbersome causing special hand- was proven, achieving significant penetration and
ling difficulties on deck. The PCB operating pressure core recovery in this environment continued to be a
was limited to 5000 psi (340 bar), which negated it's problem for the ODP. This led to a dramatic depar-
use in many deep-water situations. Finally, the PCB ture from conventional offshore coring techniques.
was extremely complex to fabricate (hence expen- The development of a coring system designed to
sive) and was very difficult to assemble and operate. allow the deployment of high speed, narrow kerf,
The new ODP PCS system is designed to recover diamond core bits from a floating vessel was ini-
61 in. 3 (1000cm 3) samples, 1.65 in. (4.19cm) in di- tiated. This system, dubbed the diamond coring sys-
ameter and recover them under in situ (hydrostatic) tem (DCS), was based on proven diamond coring
pressure. This system has a maximum operating systems successfully in use by the mining industry for
pressure of 10000 lb/in? and, as was stated earlier, is ore body exploration. The DCS system required
fully compatible with the APCjXCB BHA. The PCS mounting a smaller, high speed, top drive on to a
is significantly cheaper to fabricate and is much platform hung from the main 400 ton (363000 kg)
easier to assemble, handle and deploy at sea. When heave compensator in the derrick and then drilling
completed, the PCS system will have a special core through standard ODP drill pipe with a smaller,
processing chamber on-deck to receive the pressur- nominal 3.5 in (8.89 cm) diameter, DCS work string
ized core sample. The chamber will be pressure/tem- (tubing). The cores are then wireline retrieved
perature compensated, and will have provisions for through the work string. Although using proven
adapting to other scientific equipment provided by diamond coring technology (for land drilling), the
independently funded principal investigators. DCS was highly developmental as an offshore coring
Besides sharing total compatibility within the system because it required very precise weight on bit
APCjXCB BHA the four afor\!mentioned coring sys- control. Whereas conventional offshore coring oper-
 ODP DEEP SEA CORING TECHNIQUES 113

ations with TCI roller cone bits might require may include providing heave compensation, stabi-
25 000-35 000 Ib (11 350-15890 kg) Weight On Bit lization for drilling straight (vertical) hole, jarring
WOB and fluctuate 10 000-15 000 Ib (4540- capability for feeding the BRA when stuck in the
6810 kg) due to vessel motion, the DCS small, 4.0 in hole, and landing shoulders, latches, etc. required for
(10.16 cm) diameter narrow kerf, bits required coring operations.
4000-120001b (1810-5440 kg) WOB with fluctua- The Ocean Drilling Program utilizes two basic
tions of plus or minus 5001b (227 kg). To maintain BRA configurations with many special variations
that level of control for the DCS coring required the depending on the task at hand. The BRA configura-
development of a secondary "active" heave compen- tion for a particular hole is determined by such
sation system. This second compensator utilized load considerations as the drilling objective (i.e. desired
cells, accelerometers and hydraulic servo valves to depth of penetration), coring system to be used, core
reduce the residual heave resulting from the large orientation requirements, type and size of core bit
compensator inefficiencies and differential stretch of (which dictates the required weight on bit), antici-
the two concentric pipe strings. Initially conceived as pated hole conditions, etc. Specialty operations such
a coring system for fractured crystalline rock for the as deploying re-entry cones, casing strings, and/or
DCS has demonstrated great potential for coring in drilling ahead also affect the BRA selection.
interbedded formations, atoll/guyot coring, shallow The principal BRA used in spudding single bit
water carbonate coring such as coral and reefal holes for the ODP is commonly referred to as the
limestones, and as a possible high-temperature cor- "APCjXCB BRA". This BRA is compatible with
ing system for future hydrothermal drilling environ- most of the specially developed ODP coring systems.
ments. It allows spudding (initiating at the sea floor) a
Future coring systems that may well be in opera- sedimentary hole with the Advanced Piston Corer
tion some day include a 24600 ft (7500 m), 5 in (APC), switching to the Extended Core Barrel
(12.7 cm) diameter DCS with the platform require- (XCB) as the formation becomes more indurated,
ment eliminated and the main drill string hung off on and finally allows deployment of the Navi-Drill
riser tensioners, a "smart" core barrel that yields real Core Barrel (NCB) coring system for coring in very
time feedback of core blockage and/or loss of recov- indurated or crystalline rock types. This BRA also
ery to the driller, and rock coring systems using high allows deployment of the ODP Pressure Core Sam-
pressure water jet technology. pler (PCS) when it is scientifically desireable to re-
Raving described the basic evolution and coring cover core samples under in situ "hydrostatic"
philosophy of the DSDP and ODP scientific coring pressure.
projects the remainder of this chapter will deal with All of these coring systems are interchangeable at
the design and operation details of each coring sys- any time within the BRA. The APCjXCB BRA,
tem in use or under development including a section with all required cross-over subs, typically consists of
on the bottom hole assemblies into which these an XCB core bit, one bit sub with lockable float
coring systems are deployed and the wireline sinker valve (LFV) or one mechanical bit release sub as-
bar assemblies with which the coring systems are sembly with LFV, one "smooth bore" outer core
deployed and/or recovered. barrel, one landing/saver sub, one latch sub, one top
sub, one head sub, five 8.25-in. drill collars, one
7.35-in. drill collar, and two stands, at a nominal
Bottom Hole Assembly
93.5 ft (28.5 m) per stand, of 5.50 in. "heavy wall"
The bottom hole assembly or BRA as it is com- drill pipe. A typical APCjXCB BRA is approxi-
monly referred to in industry represents that hard- mately 423 ft (129 m) long. For oriented coring oper-
ware that is attached below the lower most joint of ations a "non-magnetic" (austenitic stainless steel
the primary drill string (Fig. 3). This commonly material with low magnetic permeability) drill collar
includes such items as the core bit, outer core barrel is substituted for one of the basic alloy steel (AISI
assembly, drill collars, stabilizers, drilling jars, 4345) 8.25 in drill collars.
bumper subs, etc. The primary purpose of the BRA The BRA used most often for coring deep (i.e.
is to provide weight for the co,re bit. Other functions greater than 1000 m sub bottom penetration) sedi-
114 M. A. STORMS

mentary holes, crystalline rock, or for coring base- Sandline and Sinker Bar Assembly
ment, is called the "RCB BRA". The RCB or Rotary
Core Barrel is the only coring system compatible with All coring systems used in the Ocean Drilling Pro-
the RCB BRA. Although there are many variations, gram are "wireline retrievable". A wireline system
this BRA typically consists of a 9.88 in. (25.1 em) OD means that the core barrel may be lowered or
by 2.31 in. (5.87 cm) ID tungsten carbide, roller cone pumped down to the bottom of the drill string and
core bit (one of several available models), one bit sub recovered through the drill string utilizing a wireline
or one bit release assembly (either hydraulic or or "sandline" as it is often referred to in industry.
mechanical) with support bearing and standard float This is a much faster method of coring than the
valve, one "control length" outer core barrel, one top "conventional" system most often used in industry.
sub, one head sub, seven 8.25 in. drill collars, one With a conventional coring system the core barrel is
7.25 in. drill collar, and two stands of 5.50 in. "heavy attached to the end of the drill string. Cores are
wall" drill pipe. Occasionally additional drill collars recovered by tripping (recovering) the drill string
are added if more drilling weight on bit (WOB) is itself rather than by tripping a wireline. Use of this
desired. A typical RCB BRA is approximately 479 ft system for continuous coring in 15000-20000 ft
(146m) long. (4573-6098 m) water depths is therefore very slow
A set of drilling jars is sometimes included in the and would be totally impractical for typical ODP
BRA to assist in extracting the BRA from the continuous coring operations.
formation should it become stuck during the coring The sandline used aboard the Joides Resolution is
operation. The ODP drilling jars are mechanically a 0.5 in. (1.27 cm) diameter, extra improved plow
actuated and impart an instantaneous upward/down- steel, 3 x 18 swaged wire rope. This rope has a
ward force or "jarring" action to the bottom hole minimum tensile breaking strength of thirty thou-
assembly. Their location in the BRA varies accord- sand pounds and is terminated on the end with a
ing to the specific situation. "rope socket". The rope socket adapts the end of the
Stabilizers are occasionally used to help maintain wire rope to a threaded connection for attaching
vertical hole, aid in drilling in-gauge (accurate and other components of the sandwire assembly.
consistent diameter) hole, and also help to prevent Several pieces of equipment are made-up (at-
drill string or BRA sticking. tached) below the rope socket. They make up what is
Bumper subs, used routinely by the predecessor called the "sinker bar assembly" (Fig. 4). This con-
Deep Sea Drilling Project (DSDP) for heave com- sists of a wireline swivel assembly (to prevent cable
pensation (adjustment of varying drill string length twist), one 3.50-in. (8.89 cm) diameter by 5.0 ft
due to vertical vessel motion), are not used by the (1.52 m) longer sinker bar, one set of mechanical
ODP unless there is a specific "space-out" require- link jars (for jarring loose potentially stuck core
ment in the BRA. Reave compensation, aboard barrels or for shearing the retrieving tool shear pin),
ODP's drill ship Joides Resolution (SEDCO/BP 471), a quick release assembly (to allow the retrieving tool
is left to the massive 400 ton (363000 kg) Western to be changed rapidly if necessary), and the core
Gear hydraulic heave compensator located below the barrel retrieving tool itself.
travelling block and above the electric top drive Sinker bars are used to provide weight to the end
hung in the derrick. This hardware is commonly of the wire rope to prevent "floating" while running
referred to in industry as the "traveling equipment". in the hole. If the sinker bar assembly starts to float
Several special BRAs or variations of standard (fall slower than the sandline speed) the wireline may
BRAs can be deployed depending on the operational be overrun causing the wire rope to become tangled
requirement. These include such variations as adding inside the drill pipe. The sinker bars also provide
a specially modified (bored out rotor) 9.50 in. posi- weight for jarring.
tive displacement mud rotor (for bare rock spud- The core barrel retrieving tool is specially designed
ding), adding a double-J running tool (for re-entry so it will automatically attach to the pulling neck
cone deployments), using special 9.50 in. (24.1 cm) (top) of the core barrel assembly when it is time to
drill collars (for more concentrated BRA weight), recover the core barrel from downhole. The ODP
etc. uses two types of core barrel retrieving tools, a "GS"
 ODP DEEP SEA CORING TECHNIQUES 115

deviation and core orientation. The retrieving tool is

modified with a set screw to prevent rotation of the
multishot pressure case relative to the crew barrel.

Rotary Coring
MALE
QUICK The rotary core barrel (RCB) coring system is used for
RE LEASE routine coring in medium to hard formations including
basement and other crystalline rock formations. The
BARS
RCB recovers a nominal 31 ft (9.5 m) long core,
NON-MAG
SINKER BARS 2.31 in. (5.87 cm) in diameter, and is the most rugged
of all the ODP coring systems (Fig. 5). An inner barrel
MUL TISHOT
PRESSURE swivel assembly, located at the top of the tool, and a
CASE support bearing, located at the bottom of the outer
core barrel assembly allow the RCB inner core barrel
to remain stationary (non-rotating) during the coring
operation. The support bearing also provides a 3 6. 2 in.
NON-MAG
SINKER BARS (9.21 cm) ID landing shoulder where the RCB lands
in the BHA and helps to center the inner core barrel
LUG
FEMALE with the core bit.
QUICK
RELEASE Two inner barrel latch assemblies are available for
use with the RCB: a standard (single finger) or a
CORE BARREL
\ ~66~IEVING J .......
'"....
~ ~
- Pull i ng Neck

CONVENTIONAL Adjustable Latch Sleeve

Latch Finger
Fig. 4. Typical ODP sinker bar assembly configurations. Swivel

Spacer / Adapter

internal "spear" type or an "RS" external "overshot"

Quick Release
type. Each is equipped with a shear pin release system
.1:*1-.1-1- 1-- Check Valve
which allows the latching mechanism to be disabled
when required. This gives the sandline operator the Inner Barrel

ability to deactivate the retrieving latch mechanism Core Liner

should the core barrel become stuck in the drill pipe.
Without this capability, excessive tension would have
to be applied to the wireline to cause it to break. The
break would be uncontrolled and therefore could
occur at any point in the line, thus possibly causing
the loss of a large portion of the sandline and creating
significant operational problems for the rig crew while tt+-~---,f-- Landing Shoulder
~1Io-l~~-f-- Supporl Bearing
tripping the drill string with the extraneous wire rope
I/J-'hIf--f--- Float Valve
inside. BII Seals -+*f-~ ~-t+\f-t- Core Clltchers
When orienting cores with the APC system two non- Core Bit ---1~H
(9~'OO x 2~'6 " 10)
magnetic sinker bars with a non-magnetic "multishot"
pressure case between them is added into the sinker
bar string directly above the core barrel retrieving tool.
The pressure case contains a standard industry multi-
shot camera used for determining hole inclination, Fig. 5. ODP Rotary Core Barrel (RCB).
116 M. A. STORMS

special double finger type. Both serve to latch the assemblies. Several different types of core catchers
core barrel down under a latch sleeve located in the can be run depending on the type and characteristics
outer core barrel assembly, restraining the ReB dur- of the formation to be cored (Fig.6).
ing the coring operation. The core barrel latch re- Washing ahead, or drilling without attempting to
leases mechanically when sandline pull is applied via recover a core, can be accomplished with a normal
the core barrel retrieving tool. The bolted latch ReB barrel in place, called a "wash barrel", or by
sleeve, located in the outer core barrel assembly, can using a center bit. To make the center bit rotate with
be easily adjusted on deck to provide for proper the outer barrel the inner barrel swivel is replaced
latch down spacing even though assembled tool with a drilling sub which locks the upper and lower
lengths and outer core barrel lengths may vary. sections of the tool and allows the drive shoulder on
The ReB is retrieved by the sandline using a the latch sleeve to transmit torque to the inner
standard sinker bar assembly and a core barrel re- barrel. Holes in the wall of the drilling sub allow
trieving tool. Two types of retrieving tools are uti- circulation to the jet holes in the center bit. The
lized by the ODP: an Otis 3-in. RS-type "external" check ball in the end of the male quick release is
overshot or an Otis 3-in. GS-type "internal" spear. removed when the center bit and drilling sub are
Both retrieving tools are designed with a shear pin used. The double finger latch is considered superior
which can be jarred loose to disengage the pulling for transmitting the torque to the center bit when
tool from a stuck core barrel. Mechanical link jars rugged drilling conditions are anticipated.
are normally included in the sinker bar assembly to A bit deplugger is often used in place of the center
provide additional jarring force for this purpose. The bit when attempting to remove coring debris which
ReB is deployed by "free falling" or pumping it occasionally becomes jammed in the throat of the
down the drill pipe without the sandline attached. core bit.
The inner core barrel contains a clear butyrate core Typical core bits used with the ReB are 9.88 in.
liner. This liner protects the core during the coring (25.1 cm) OD by 2.31 in. (5.87 cm) ID tungsten car-
operation and also serves as a convenient means for bide insert (Tel), journal bearing, four cone roller
transporting and storing cores once removed from the type. The tungsten carbide inserts may be buttons,
tool. A core catcher sub, located at the bottom of the short chisels, long chisels or a combination of each
inner core barrel, houses one or more core catcher (Fig. 7).

Fig. 6. Selected Core Catchers-dog, collet and spring-type. Fig. 7. RCB 9-7/8 TCI roller cone core bit.
 ODP DEEP SEA CORING TECHNIQUES 117

Hydraulic Piston Coring injected into the formation. The stored energy in the
drill string provides the dominant source of energy for
The technique of hydraulic piston coring is based on
the coring system. The HPC system generates a large
conventional oceanographic piston coring. A cylin-
amount of thrust over a very short period of time.
der or tube is pushed into the formation, without
With the core barrel accelerating at 6-12 m/sec, the
rotation, at a high rate of speed. At the same time a
entire coring cycle typically lasts less than two sec-
piston inside the tube displaces the water in the core
onds. The benefits of this type of coring system are
barrel out of the top to prevent core erosion and
many. The high rate of speed generated acts to
disturbance during core entry. The piston is designed
decouple the coring system from the heave induced
to minimize the back pressure applied to the core
vessel motion hence there is minimal core disturbance
during the coring process.
(Fig. 9). The high coring force results in the ability to
To allow the piston core barrel access to the
penetrate into semi-indurated formations at which
formation the cones on a standard RCB core bit are
point rotary coring can be initiated without inducing
spread apart. The resultant bit, with an OD of
significant core disturbance. Continued penetration
11.44 in. (29.06 cm) and an ID of 3.88 in. (9.86 cm),
into the formation is facilitated by "washing down"
is referred to as "an APC;XCB bit" (Fig.8).
with the primary APC;XCB bit to the point where the
The primary difference between hydraulic and con-
previous core run ended its stroke.
ventional piston coring is the type of energy source
The advanced piston corer or "APC" (Fig. 10)
and the magnitude of the applied force. With con-
utilizes the technology of past DSDP hydraulic
ventional piston coring a weight is tripped at or
(HPC) and variable length hydraulic (VLHPC) pis-
slightly above the mudline which falls through the
ton corers. It incorporates a much more advanced
water column, pushing the core barrel into the for-
and simpler seal system which results in a 76%
mation ahead of it. The falling weight provides the
greater coring force of 28 OOOlb (125000 N (new-
energy or coring force. This system is limited to
tons». The APC uses a dynamic seal acting between
relatively soft formations and shallow penetrations
the scoping piston corer and a special honed-bore
into the seabed (commonly 10-50 m). With hy-
outer core barrel, commonly referred to as a "seal
draulic piston coring the core barrel is landed and
bore drill collar". The inside diameter of this outer
sealed off inside the BHA. Rig circulating pumps are
core barrel is 3.80 in. (9.65 cm) minimum, which
used to pressurize the drill string until the preset
constitutes the tightest restriction in the BHA.
shear pins fail and allow the core barrel to be
The non-scoping section of the APC incorporates
an adjustable flow-by. landing shoulder sub where
speed control set-screws can be added or removed to
control velocity. An anti-spiral system prevents rota-
tion of the scoping section of the corer relative to the
piston rod. An anti-spiral key located in the male
quick release tracks down a special groove machined
in each of the piston rod sections. All rod sections
are interchangeable, even if fabricated separately.
When assembled and locked a complete set of piston
rod sections, upper, center and lower, will automati-
cally have an aligned anti-spiral groove running the
entire length of the piston rod assembly.
The APC recovers core inside a plastic acetate
butyrate core liner just as the other ODP coring
systems. The core is retained in the barrel primarily
with a flapper-type, full-closure core catcher (Fig.
11), although the other core catchers used in the
rotary coring system are compatible and can be used
Fig. 8. APC/XCB 11-7/16 TCI roller cone bit. if the formation warrants.
1I8 M. A. STORMS

STANDARD ROTARY CORE

~___________________ ~.,~~~~~S~
TONCORE __________~______~

Fig. 9. Core comparison-rotary vs. hydraulic.

A magnetic orientation system is used to deter- also be taken as part of the routine coring cycle (Fig.
mine the orientation of cores taken by the APC with 12). A small microprocessor and thermistor can be
respect to magnetic north. The system consists of an installed into a special APC cutting shoe and used to
Eastman magnetic multishot survey instrument, a record temperature data after the piston corer has
non-magnetic sinker bar system, and a non-magnetic been injected into the formation. Thermistor equili-
drill collar. Adjustment for baseline alignment be- bration typically requires waiting 5-8 min before
tween the Eastman multishot and the double refer- retrieving the core barrel from the formation.
ence line on the core liner is accomplished by
following two special alignment sequences during the
Extended Coring
initial assembly of the coring system. Ten to fifteen
minutes of additional handling time in the coring The Extended Core Barrel (XCB) coring system
cycle are required when orientation data is required. (Fig. 13) has been developed from a conventional
If desired, an in situ temperature measurement can oilfield concept. The XCB is a rotary core barrel
 ODP DEEP SEA CORING TECHNIQUES 119

ORIENTATION
ALiGNABLE
PULLING NECK

SHEAR PINS

INNER SEALS
OUTER
SEALS
OUICK
RELEASE
VENTS

ROD

HONED
1.0.
DRILL
COLLAR

Fig. II. APC flapper-style core catcher.

LOWER
used for APC/XCB coring has a 3.80 in. (9.65 cm)
SUPPORT
BEARING
throat which is large enough to allow the piston
SNUBBER corer to be extended beyond the bit face and into
PISTON virgin formation. The XCB core barrel has a cutting
HEAD a. SEAL
CUTTING shoe structure which rotates in conjunction with the
SHOE
roller cone bit and trims the core down to 2.31 in.
- 3.80"BHA (5.87 cm). The abrasive drag style cutting action
INNER DIA.
produces better core quality than a roller cone bit in
most types of formations.
With the XCB system the cutting shoe can extend
about 7 in. (9.83 cm) beyond the face of the core bit.
The lower portion of the core barrel is spring loaded
over a hex spline to retract into the core bit under a
1500-20001b (6700-8900 N) load. In the extended
Fig. 10. ODP Advanced Piston Corer (APC).
position the cutting shoe shields the incoming core
from disturbance caused by the primary bit jet hy-
system similar in many ways to the standard RCB draulics. A variable amount of circulation is diverted
system. It uses the same butyrate core liner and the to the cutting shoe to lubricate, cool and flush the
core barrel is deployed by free-falling and/or pump- core trimmer. The spring loaded extension/retraction
ing down the drill string. The barrel latches in place capability provides overload protection of the core
in the outer core barrel and accepts a 2.31 in. barrel assembly and is also helpful when attempting
(5.87 cm) diameter core up to 32 ft (9.8 m) in length. to recover cores from hard and soft interbedded
The latch mechanism causes the XCB inner core lithologies which cannot be successfully cored with
barrel to rotate with the outer barrel (BHA). either the RCB or APC systems.
The XCB offers several important advantages over Different types of cutting shoes are available, in-
the RCB system. The most significant is its ability to cluding a hard-faced "soft formation" shoe and sev-
be deployed in the same bottom hole assembly as the eral diamond-type shoes using impregnated, geoset
Advanced Piston Corer (APC) after piston coring or natural diamond-cutting structures. Combination
operations have terminated. Rotary coring can there- diamond/saw-tooth bit profiles are also available.
fore be continued without a pipe trip to change over A special XCB latch has been designed for rugged-
to a different bit and BHA configuration. The bit ness, simplicity and positive transmission of torque.
120 M. A. STORMS

ENLARGED VIEW OF
MINIATURE ELECTRONIC
TEMPERATURE SENSING /
RECORDING PACKAGE
(BATIERY POWERED)

PRESSURE · TlGHT CAVITY

IN ENO OF CORER

THERMISTOR

CONTINUOUS TEMPE RATURE

MEASUREMENTS WITH CORER
EMBEDDED IN SEDI MENTS

Fig. 12. APe in situ temperature measurement system.

The XCB lands on the 3.0 in. (9.91 cm) ID landing/ widens out to a 5.50 in. (13.97 cm) ID below the
saver sub, just as the APe. The latch dogs engage in latch sleeve in the BHA. When the XCB lands the
a double-window latch sleeve which ensures that the latch is positioned in this section. The spring forces
latch will lock down in any orientation without the dogs to lock out over the high points on the
requiring a "muleshoe" feature . pulling neck. At this point the tool is locked under
Major components of the XCB include the latch, the latch sleeve and the dogs cannot shift if an
hex splined scoping section, spring shaft with helical upward force is applied to the tool (downward force
compression spring, quick release, vent sub, core on top of dogs). When it is desired to release the tool
barrel with non-rotating core liner, and cutting shoe. and recover the core barrel, a core barrel retrieving
The latch was designed and developed specifically tool is lowered by sandline and locks onto the head
for use with the XCB but may be used with any of the pulling neck. A pull on the sand line shifts the
other tool which requires a hold down or torsional pulling neck out from under the locking dogs and
transmission mechanism. The latch body is 3.75 in. allows the dogs to fall in and release from under the
(9.53 cm) in diameter. The two latch dogs lock out to latch sleeve.
a 5 .0in. (12.7 cm) diameter. When the XCB enters The hex shaped shaft of the male drive sub en-
the pipe the latch dogs a re forced up against the gages a similar profile in the hex landing sub to
spring until they fall into the pulling neck detents. provide for up to 8.0 in. (20.32 cm) of axial displace-
The dogs remain depressed in the 4.12 in. (10.48 cm) ment while continually transmitting torque.
bore as the tool travels down the pipe. The pipe bore A 34.00 in. (86.36 cm) long helical compression
 ODP DEEP SEA CORING TECHNIQUES 121

under evaluation, uses jetting geometry to reduce

CORING CORING back pressure at the top of the inner core barrel.
SOFT HARD
SEDIMENT SEDIMENT Up to 9.8 m (32.1 ft) of core can be recovered in the
standard butyrate core liner. The lower end of the
LATCH vent sub provides the inner race for a bearing device
OUTER BARREL called the "liner hanger", which is the upper support
COIL SPRING for the core liner. The liner is also supported at the
bottom where it rides on a low-friction bushing.
LANDING
SHOULDER Several types of cutting shoes are available for use
depending upon the nature of the sediment or rock
7" to be cored. The soft formation cutting shoe employs
STRDJa/ SPRING
turltACf'/ON
Ih+-+---.- SHAFT a serrated cutting profile hardfaced with tungsten
QUICK carbide grit. A portion of the circulation flow to the
RELEASE
core bit is diverted to directly lubricate the extended
VENTURI
VENT CIRCULATING cutting shoe. The flow enters through inlet holes at
FWID
I~j~-t-" SYSTEM the top of the shoe, and is directed through an
NON- ROTATING annulus created by the isolation sleeve and out small
CORE UNER
LINER
jet holes at the bottom of the shoe.
BEARINGS

FLO" TO
CUTTING SHOE

BIT SEAL ~-tr'--.;sl

CO~E~~~~I!~
F"-"""'1.E. QUCK.
BIT RE.LE.'-SE.
ROLLER
CIRCULATION I-lOz.
JETS
CORE CATCHERS
CUTTINC SHOE

Fig. 13. ODP Extended Core Barrel (XCB).

spring with a spring rate of 250lbjin (44000 Njm)

initially maintains the XCB at full extension. A
compression force of 2000 lb (8900 N) will cause the
spring to fully compress and allow the male drive sub
to shoulder on the hex landing sub.
A three-lug quick release is currently used on both
the XCB and APe. This mechanism reduces the turn
around time between successive cores by providing a
rapid means of connecting and disconnecting the
core barrel from the upper section of the tool. The
male and female sections engage and rotate 50° to
lock together.
The vent sub is fitted with a one way check valve
to allow fluid to exhaust from the core barrel into the
drill string annulus as the core enters. This valve also
prevents flow in the opposite direction, protecting
the core from being washed out during retrieval. A
second pair of slanted holes allows circulation to exit
at the vent sub after passing through the quick
release mechanism.
An optional "venturi" vent (Fig. 14), currently Fig. 14. XCB Venturi Vent Assembly.
122 M. A. STORMS

MJ..DC t-.rD<
\ / TIIRUSTER NOZZLE SUD

II EX MAL E SPL IN E
ItEAO SUD

THRUSTER UNIT TORQUE LATCH ASSEMBLY

SUBASSEKBLY
UNIVERSAL TOP SUB FEMALE SPLINE
THRUSTER SEALS I INNER SEALS)

OUTER SEAL ASSEMBLY

].7S NAVI -OfUl..L.

MuD MOTOR I"QSITrvr:DISP~11..O
SUBASSEMBLY

1l.8"1.0.1

SEALED QUICK RELEASE ASSEt:DL\'

n...oi OlVl011l'

ANTI-JAM ASSEMBLY
Fig. 15. Selected XCB cutting shoes (bits).
l.CNG BJT SUI)
1W>4 •.H.l1f 'CORE BARf.I.n (13 ' J
CORl: BARR£L
SUBASSEMBLY
The diamond cutting shoes (Fig. 15) may consist NON-ROTATING METAL INNER TUBE

of impregnated, surface set, geoset or natural dia- M'C/'N:B 00At BIT U.'''1.0.1
PLASTIC LINER

monds. Various matrix designs are used. Circulation DIAMOND REAMINC SHELL

flow to the cutting edge is similarly diverted down an IIIClI SPWl f\ARfO'I' to:"»'

internal annulus, and may exit directly onto the core OlJlt-OlV CORJ: I!FAD
13.750' x 2 . 2S ··)
or at the face of the diamond bit.
Specially designed combination cutting shoes em- Fig. 16. ODP Navi·Drill Core Barrel (NCB).
ploy various types of diamonds set in a saw-tooth
pattern similar to the soft formation shoe. Yet an- or subassemblies: a thruster unit for hydraulically
other experimental shoe uses geoset diamonds applying weight on bit (WOB), a small 3.75-in.
imbedded in an impregnated diamond matrix. (9.53 cm) OD positive-displacement mud motor for
generating downhole rotation (torque), a non-rotat-
ing core barrel assembly for receiving the core, and a
Navi-Drill Coring
3.75-in. (9.53 cm) OD narrow kerf diamond core bit
The Navi-Drill Core Barrel (NCB) is a prototype to cut the core.
coring system currently under development by the The thruster unit is comprised of several compo-
ODP (Fig. 16). The primary goal of the NCB devel- nents performing a wide variety of functions. The
opment is to allow single-bit APC/XCB holes to be primary task of the thruster unit is to translate
extended to greater depths and into more indurated hydraulic force into mechanical weight on bit
formations particularly fractured crystalline base- (WOB). The pressure drop by circulation through a
ment rocks. A secondary goal is to improve recovery nozzle sub at the top of a hexagonal spline assembly
in hard/soft interbedded formations such as soft results in a downward force applied to the diamond
chalks laced with chert stringers. core bit. Removable nozzles allow optimization of
This wireline-retrievable hard rock coring system the desired WOB at various flow rates. The reaction
can be deployed at any point in the coring operation. torque generated by mudmotor rotation is trans-
It is fully interchangeable with the APC/XCB BHA, ferred through the spline assembly and torque seg-
thus following the coring system to be optimized ments to the main outer barrel assembly. When the
from the mudline down to and into indurated forma- tool is freefall deployed, the thruster unit dampens
tions and/or basement rock. The NCB recovers a the landing impact thus preventing premature un-
nominal 2.25 in. (5.72 cm) diameter core, 13.25 ft latching and mechanical failure. The tool may also
( 4.0 m) long in a plastic polycarbonate liner. be deployed using a wireline delivery system; how-
The NCB is comprised of four main components ever, this requires an additional wireline trip and
 ODP DEEP SEA CORING TECHNIQUES 123

results in a less efficient coring operation. The

thruster unit also seals downhole causing all of the
circulating fluid to be channelled through the mud-
motor. In addition, it maintains the stroking portion
of the tool (hex male spline, core barrel and dia-
mond core bit) in a latched position until after the
tool has landed and rotation is initiated.
The NCB is powered by an Eastman Christensen
(EC) 3.75 in. (9.50 cm) OD, 7/8 lobe, Mach I posi-
tive displacement mud motor. This motor has been
recently developed by EC, primarily for industry oil
well drain hole drilling applications. At a pump rate
of 170 gal/min (635 l/min) and a pressure drop
across the motor of 1160 psi (80 bar), this "drain-
hole" motor is capable of generating 1250 foot-
Fig. 17. Selected NCB cutting shoes (bits).
pounds (1695 N/m) of torque and 410 rpm. The
motor operates at 90% plus efficiency and can de-
velop 96 hydraulic horsepower. The standard core nated diamond bits are available as well as surface
barrel assembly used with the NCB is a modified set and geoset diamond bits. Appropriate bit selec-
version of a standard Christensen Mining Products tion is determined by the type of formation to be
(CMP) HWD4 "HQ" type core barrel. It is attached cored. A surface set diamond reaming shell is in-
to the mud motor with a modified (sealed) three-lug stalled directly behind the core bit to enhance stabi-
quick disconnect to allow handling in the same lization and help maintain hole gage.
efficient manner as the other primary ODP coring Coring operations commence by freefall deploying
systems. The barrel contains a non-rotating inner the NCB assembly to land within the APCjXCB
tube which may be run with or without a polycar- BHA. As circulation with the rig mud pumps is
bonate liner. The core is retained by using either a established rotation of the coring assembly begins
standard core spring installed in an inner tube shoe inside the BHA (Fig. 18). Increased flow rate causes
(for hard formations) or a special spring loaded the NCB coring assembly to release and drop into
dog-type core catcher (for soft formations). Installed contact with the formation to be cored (Fig. 19).
between the inner-tube and the core catcher shoe is a Coring proceeds until the NCB reaches full stroke at
breakoff sub which allows for easy retrieval of the which time the driller sees a pressure spike due to a
core liner. A CMP anti-jam system can be installed choking of the flow area within the tool. Coring is
directly above the inner tube. In theory, when a core suspended and the NCB is retrieved. Additional
blockage occurs, the inner tube lifts up, energizing penetration is made by drilling out the NCB cored
the anti-jam system. The resultant "jarring" action is "pilot hole" with the main XCB bit and then deploy-
designed to free the blockage and allow unrestricted ing the NCB system again.
core entry to resume. Installed directly above the
core barrel assembly is a flow divider sub. This sub
Pressure Coring
allows the proper amount of circulating fluid to be
directed to the diamond core bit (typically lO- The Pressure Core Sampler (PCS) is an ODP devel-
IS gal/min or 38-571/min) while the remainder is opmental coring system capable of retrieving core
diverted to the annulus for hole cleaning and cut- samples at near in situ hydrostatic pressures (Fig.
tings removal. Removable nozzles allow the flow 20). The PCS utilizes both current conventional oil-
distribution to be optimized as required. field pressure coring technology and technology de-
The NCB is designed to operate with several types veloped by the Deep Sea Drilling Project (DSDP).
of narrow kerf diamond core bits (Fig. 17). All bits The PCS is also completely compatible with the
have an OD of 3.75 in. (9.50 em) and cut a 2.25 in. APCjXCB/NCB BHA. The PCS is being developed
(5.71 cm) core. Both hard and soft matrix impreg- in response to a scientific mandate for retrieving core
124 M. A. STORMS

DEPLOYMENT CORING AHEAD

• LAND NCB IN OUTER CORE

• USINC CIRCULATION PRESSURE
BARREL ASSEMBLY. DISENCACE THE LOCKINC BALLS
ON THE SPLINE ASSEMBLY.
• INCREASE THE FLO" RATE TO
• PLACE XCB BIT ON BOTTOM AND PRODUCE THE DESIRED "EICHT
CIRCULATE TO DETEIIJIINE ON BIT.
BEGINNING FLO" .AND
PRESSURE PARAMETERS. • MONITOR FLO" AND PRESSURE
PARAMETERS WHILE CORINC. .A
CONSTANT HICH PRESSURE
PROBABLY INDICATES MOTOR
• SPUNE ASSEMBLY IS LOCKED
STAL~ A CONSTANT LO"
IN THE UP POSITION .AND
NCB LATCH DOCS ARE ENGAGED PRESSURE PROBABLY INDICATES
IN THE LATCH SUB. A CORE BLOCK
• THE NCB SYSTEM PENETRATES
AHEAD OF THE XCB BIT AS
CORINC COMMENCES.

Fig. 18. NCB deployment-schematic operational sequence.

Fig. 19. NCB coring ahead-schematic operational sequence.

samples while maintaining near in situ hydrostatic
pressures of up to 10000 psi (690 bar). The system is
deployable in soft-to-moderately-indurated sedi- The pes comprises five main components of sub-
ments and will ultimately have the capability to assemblies: latch, actuator, valve-accumulator, ball
transfer a 61 in 3 (1000 cm 3) core sample from the valve and detachable sample chamber.
downhole tool to a pressure/temperature controlled The pes latch subassembly is a modified XeB
laboratory chamber while maintaining downhole latch which serves five functions. The latch subassem-
pressure. The core sample can then be accessed bly contains the landing point for the pes. The latch
directly for scientific evaluation under near in situ subassembly has a 4.0 in. ( 10.16 cm) outside diameter
pressure and temperature conditions. shoulder which cannot pass the 3.82 in. (9.70 cm) ID
The pes is a wi reline retrievable, free fall deploy- throat of the landing saver sub in the BHA, thus
able, hydraulically actuated pressure coring system. preventing the pes from passing completely through
When the pes is deployed, it lands and latches into the BHA. By latching into the BHA, the latch sub-
the BHA and is rotated with the BHA during coring assembly transmits torque from the BHA to the pes,
operations. It is fully interchangeable with the APe allowing it to train the core to proper size for entry
and XeB coring systems thus allowing a pressurized into the sample chamber. The latch subassembly
core sample to be taken at anytime from the mudline holds a check ball used in the actuation of the ball
down to indurated formations. The pes covers a valve subassembly. When the latch subassembly is
nominal 1.65-in (4.19 cm) diameter core sample, engaged by the wireline and an upward force is
34 in. (0.86 m) long at pressures up to 10000 psi applied, it automatically releases a check ball allowing
(690 atm). the ball to fall into the actuation subassembly.
 ODP DEEP SEA CORING TECHNIQUES 125

CORING SAMPLE CHAMBER

AHEAD CLOSED

BALL COLLET

LANDING
SHOULDER

VALVE

Fig. 20. ODP Pressure Core Sampler (PCS).

Finally, the latch subassembly diverts all flow ball has been released, the actuation subassembly
through the pes and provides a place for the wireline unlatches and strokes through itself pulling the core
to automatically attach itself during core barrel re- tube containing the core sample through the ball
trieving operations. The latch subassembly is attached valve into the sample chamber. As the core tube is
to the pes by a three lug quick-release allowing for pulled into the sample chamber the ball valve is closed
handling in the same efficient manner as the other and the upper end of the core tube is pulled into a seal
primary ODP coring systems. The pes actuator sub- receptacle thus closing the sample chamber at both
assembly serves two functions. It catches the check ends and trapping the core sample at hydrostatic
ball when released by the latch subassembly stopping pressure inside the pes. When the actuation sub-
all flow through the pes until stroking occurs. Also, assembly reaches the end of stroke it latches once
when pressure is applied to the pes and the check again and opens a circulation path through the pes.
126 M. A. STORMS

The pes valve-accumulator subassembly contains into the cutting shoe. When the actuation subassem-
a pressure-maintaining mechanism, safety pressure bly is activated and the core tube has been pulled
relief mechanisms, a sampling port, temperature and through the ball, the ball is rotated into the closed
pressure monitoring devices and the core tube. The position sealing the lower end of the sample cham-
pressure maintaining mechanism is a built-in accu- ber. The ball valve subassembly also provides a
mulator that maintains the pressure inside the sam- means for connecting the sample chamber to a pres-
ple chamber when a small volume change occurs surized laboratory chamber. This is done by remov-
during sealing. It also compensates for any minor ing the cutting shoe and using the threaded end to
seal leakage should it occur. The safety pressure connect to the test chamber. The ball valve sub-
relief mechanisms include an adjustable pressure re- assembly also contains the pressure containing body
lief valve set to automatically vent pressure above of the sample chamber and the seal receptacle used
10 000 psi (690 atm). Should the pressure relief valve to seal the upper end of the sample chamber.
fail to release pressure a burst disk will rupture at The detachable sample chamber is made up of the
12500 psi (862 atm) relieving all pressure from inside ball valve and valve-accumulator subassemblies. It is
the pes. An access port allows sampling of gasses or 3.75 in. (9.53 cm) in diameter, 5.0 ft (1.52 m) long
fluids directly from the pes sample chamber. A and is attached to the pes by quick release connec-
built-in thermistor and pressure transducer allows tions which allow the pressurized sample chamber to
for the connection of monitoring equipment to con- be removed from the rest of the pes for earlier
stantly monitor the temperature and pressure inside handling. Since the valve-accumulator subassembly
the pes sample chamber. The sample tube is a is an integral part of the detachable sample chamber,
non-rotating metal tube with integral core catchers the pressure and temperature can be continuously
used to contain the core sample. During coring monitored. Also, gas and fluid samples can be taken
operations the core tube is extended through the ball directly from the sample chamber (Fig. 21).
valve subassembly into the cutting shoe. When the The pes uses a specially designed pilot-type cut-
actuator is activated the core tube is pulled through ting shoe. The available cutting shoe cutting struc-
the ball valve into the sample chamber. tures for the pes are both hard- and soft-matrix
The pes ball valve subassembly is the sealing impregnated diamonds, surface-set diamonds, geoset
mechanism on the bottom of the pes sample cham- diamonds as well as standard hard facing.
ber. It also is the connection point for the pes The pes is free fall deployable and therefore is
cutting shoe used to trim the core sample to size. dropped down the drill pipe and landed in the BHA.
During deployment and coring operations the ball The pes is rotated by the top drive via the latch and
valve is open with the core tube extended through it drill stringJBHA.

Fig. 21. pes tool with sampling manifold.

 ODP DEEP SEA CORING TECHNIQUES 127

During coring operations the rig pumps maintain 6''''1IOl£O RE· ENTRY
flow down the drill string to keep the hole open and CONE

to cool/lubricate the pes cutting shoe. Once the core

has been cut the rig pumps are secured, the wireline SONAR

is attached to the pes and an up strain is applied to REFLE CTOAS

the pes latch to release the check ball. The wireline

is then slacked off and the rig pumps are restarted
slowly, letting the pressure build to activate the
actuator and stroke the sample chamber closed.
When circulation is once again established the sam-
ple chamber has been closed and the pes is retrieved
like any other wireline core barrel. Once on deck the
detachable sample chamber is removed from the
pes, placed in a portable temperature controlling HOLLOW ROTOR po",
CORINe. ",vo MOTOR
bath/safety shroud where temperature and pressure
monitoring equipment is attached. The sampler
chamber can then be safely moved off the rig floor
for scientific evaluation.
Fig. 22. ODP Hard Rock Guide Base (HRB) schamatic-drilling
in young fractured basalt.
Bare Rock Spudding

The development of new technology for spudding to 18000 ft (5488 m) of water and is designed to
holes and coring young, fresh rocks in areas with little land on up to a 20° sloping sea floor with 3 ft
or no sediment cover (i.e. mid-ocean ridges, spreading (0.92 m) diameter boulders present.
centers, bare rock basins, seamounts, etc.) is a major Upon drilling/coring 100-200 ft (30-60 m) of sur-
scientific and engineering objective of the ODP. face hole through the guide base, a special gimbaled
The Hard Rock Guide Base (HRB) is designed to reentry cone with a string of 16.0 in. (40.6 cm) or
sit on the sea floor and provide the necessary lateral 11.75 in. (29.8 cm) casing is run and cemented in
support for the bottom hole assembly and bit con-
finement to spud a hole on bare rock (Figs. 22 and
23). The HRB also serves as a temporary re-entry
cone. The guide base is 17 ft (5.18 m) square and 11 ft
(3.35 m) tall. The cone inside the guide base is 16 ft
(4.88 m) in diameter and 6 ft (1.83 m) deep. Due to
the size of the guide base, it is necessary to assemble
the HRB in two halves in the ship's moon pool. The
HRB is then run through the moon pool vertically on
two running cables and once below the keel of the
ship, it is rotated into the horizontal running position.
The HRB is then lowered to the sea floor suspended
from the drill string. Once on the bottom, the HRB
and attached cement bags are filled with 2000 ft3
(56.68 m 3) of cement to provide additional mass.
The HRB is divided by bulkheads into four com-
partments. The four compartments are isolated
from each other so that in the event of HRB damage
during deployment the individual compartments in-
sure that at least a portion of the guide base can be
filled with cement. The HRB can be deployed in up Fig. 23. HRB photograph.
128 M. A. STORMS

place. The purpose of the gimbal is to provide ade- body with a 2.50 in. (6.35 cm) 10. A plastic liner is
quate surface contact to support the reentry cone at run inside the core barrel to allow easy removal of
angles of up to 20°. the core from the barrel. The rotor is attached to the
drill string through the drive subassembly. Torque is
Mud Motor Coring generated as fluid mud passes down the drill string and
between the rotor/stator assembly. This torque ro-
The Positive Displacement Coring Motor (PDCM) tates the 40 ft (12.2 m) long outer motor housing,
is a modified version of a standard Eastman Chris- which in turn induces torque and rotation to a
tensen 9.5 in. (24.1 cm) OD-positive displacement 10.5 in. (26.7 cm) 00 by 2.25 in. (5.72cm) 10 roller
mud motor (PDM) designed for use with the hard cone core bit. After cutting a 30 ft (9 m) core, the
rock guide base. The primary modification required core barrel is retrieved by wireline, completing the
replacement of the standard "solid rotor" with an coring cycle.
"inverted hollow rotor" (Fig. 24). The PDCM provides up to 6000 foot-pounds
A nominal 30 ft (9 m) core barrel deployed via (8077 newton-meters) of torque at a flow rate of
wire1ine is landed in the hollow rotor section of the 600 gal/min (757-227ll/min). Bit rotational speeds
motor. The core barrel has 3.13 in. (33.4 cm) 00 of 90-120 r.p.m. can be generated by the motor with
upset connections and a 2.92 in. (7.42 cm) 00 mid- a maximum pressure drop across the motor of
640 lb/in? (44 bar).
With this system the entire drill string remains
stationary. Only the core bit and the motor housing
UPPER RADIAL rotate. The enhanced stability of the drill string and
BEARINC, BHA as a result of using the coring motor allows
exploratory coring operations to be performed on
THRUST BEAR INC,
bare rock efficiently and reliably.
BEARINC, ASSEMBLY
LOWER RADIAL
BEARING

LATCHING
ASSEMBLV

8EARINC, ASSEMBLY
OUTER CORE BARREL

FLEX SHAFT

NON - ROTAT INC,

INNER CORE BARREL
STATOR

ROTOR
MOTOR SECTION
NAV I- DRILL. MACH I

BI T SUB

IN NER BARREL
SUP POR T

COR E BIT

Fig. 24. Positive Displacement Coring Motor (PDCM). Fig. 25. Diamond Coring System (DCS) top drive concept.
 ODP DEEP SEA CORING TECHNIQUES 129

Diamond Coring pipe body and a 2.94 in. (7.47 em) ID with 3.87 in.
(9.83 cm) OD upset connections. A 4.0 in. (10.16 em)
The diamond coring system (DCS) is a developmental OD by 2.40 in. (6.10 cm) ID high-speed diamond
coring system designed to drill and core both sedi- coring bit is run with a core barrel assembly on the
mentary and crystalline rock formations. A scaled- drill rod string. The drill rod string is presently
down version of the DCS was tested during January rotated with a hydraulic top drive at 60-500 r.p.m.
1989 in the Luzon Strait, just north of the Philippines. An electric top drive is currently under evaluation as
The purpose of the test was to evaluate the potential a successor to the hydraulic unit. The small-diameter,
and validate the use of a top-driven high-speed dia- narrow kerf, diamond bits are typically operated with
mond coring system deployed from a floating vessel. 2000-12 000 lb (908-5448 kg) of drilling weight.
This unique coring system involves running a A secondary heave compensator system is used to
small-diameter drill rod string inside 5.50 in. maintain precise control of weight on bit. This sec-
(13.97 cm) OD drill pipe to the sea floor (Fig. 25). ondary "active compensator removes load fluctua-
This "working" string has a 3.50 in. (8.89 cm) OD tions resulting from the mechanical inefficiencies of

ELEVATOR BArul
ICCISTING
DElUIlCIC
GUIDI IWL
SECONDARY HEAVE COWPENSATOR
TRJ.CJC

HYDRAULIC
TOP DRIV!

10 n.
DRILLING PUP APPROX. 58'
JOINT DRILU:R

HE.I.VE COKl'ENlIATOR
IilICROPROCE9S0R

ELEVATOR BArul
DRILL ROD (HYDRIL
3-1/2" SERIES 500 TUBDlG)
TUBING
STRIPPBR
DRILL PIPE ELEVATORS
ELEVATOR STOOL

RIG FLOOR

Fig. 26. DeS platfonn configuration.

130 M. A. STORMS

the primary 400 ton (363000 kg) "passive" heave

compensator. The result is weight on bit control for
the small diamond core bits of ± 500 lb (227 kg). ODP
All of the diamond coring operations and drilling HEAVE COIlPENSATOR
(.00 roIlS)

functions are controlled and conducted from a DCS FEED CYUNDER

(SECONDARY HEAVE COI.W)
manned platform, suspended in the derrick (Fig. 26).
HYDRAUUC TOP DRIVE
From the driller's console on the coring platform,
the driller operates the top drive, secondary heave
compensator, wireline winch and controls the make-
up/break-out of the drill rod joints. Upon tripping
the drill rod joints to bottom, the driller activates the
secondary heave compensator and automatic feed
system. At that point, the diamond core bit is auto- 3- 1/ 2" HYDRIL TUBING
WORK STRING
matically fed to bottom and the desired bit weight is
established for the coring run. Upon completing the
coring run, the bit is retracted off bottom and the
core barrel is retrieved. A empty core barrel is REENTRY CONE (OPTIONAL)
1I1TH CASI G STRING
dropped (free-fall deployed) down the drill rod string M--"""-'

and coring is resumed. 1I1RE UNE CORE BARREL

The "manned" diamond coring system platform is DIA)dOND CORE BIT
(4" x 2 .)
45 ft (12.43 m) tall and weighs 40000 lb (18 160 kg).
The work area on the platform is 8.0 x 12.0 ft (26.2 x
39.4 m) square. Two-to-four people are stationed on Fig. 27. DeS platform, Varco top drive, and Hydril tubing.
the DeS platform while operating in the derrick.
When not in use the DeS platform is stored out of control, allowing undisturbed sedimentary cores to be
the way on the rig floor. It is rolled into position via cut with the drill ship heaving significantly.
a portable dolly/track system for deployment. During Phase II of the diamond coring system
A scaled-down diamond coring system was suc- development, the drilling systems (top drive, feed
cessfully deployed with a 1700-m drill rod string cylinder, and secondary heave compensator system)
during ODP leg 124E in January/February 1989 will be redesigned for deployment in up to 14760 ft
(Fig. 27). Several coring runs were made in heavy ( 4500 m) of water. This deep-water system is sched-
seas, with good results. The secondary active com- uled for testing on the ODP drill ship Joides Resolu-
pensator system maintained excellent weight on bit tion early in 1990.
The Status of Geological Dredging Techniques

R. B. KIDD', Q. J. HUGGETT 2 and A. T. S. RAMSAY'

'University of Wales, College of Cardiff, PO Box 68, Cardiff CF1 3XA, UK
2/nstitute of Oceanographic Sciences, Deacon Laboratory. Wormely, Godalming GU8 5UB, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: geological dredging, basement rock samples, pinger single lead weight towed ahead of the frame. A range
telemetry, transponder navigation, swathe mapping, sorting crite- of variations on this basic design are in current use.
ria, dredging and drilling comparisons.
The main purpose of this chapter is to review tech-
niques in dredge sampling for geological purposes,
Abstract. Scientific sea-floor dredging is currently used in marine including operations and navigation at sea, monitor-
geology primarily by the "hard-rock" community interested in ing of the dredge, and assessment of rock recovery.
the recovery of basement rock samples from the unsedimented
Sediment coring has become the marine geologist's
deep ocean floor. The technique has generally been eclipsed by
ocean drilling for recovery of sedimentary rocks, because of main sampling tool for detailed analysis of the (gen-
perceived uncertainties in the location of sampling and in the erally Late Quaternary) record held in the uppermost
representativeness of recovered material. This contribution re- tens of meters of unlithified deep sea sediments
views dredging equipment currently in use by marine geological
institutions and refers to pinger attachments that allow precise (Weaver and Schultheiss, this volume). Dredging
information on the behaviour of the dredge to be telemetered was, until the late 1960s, the only means available for
back to the ship. We argue that improvements in ship navigation sampling older lithified sedimentary rocks or the
and transponder navigation at the seafloor, when used in con-
igneous oceanic basement. Great care was taken
junction with surface and/or deeply towed sidescan and swathe-
mapping surveys, now allow for considerably less uncertainty on through accurate navigation, the use of various
the location of dredge sampling. Refined sorting criteria for methods of geophysical surveying and monitoring of
dredge hauls are now also available. Recent comparisons of
dredge behaviour, to ensure that dredge sampling
regional sample recovery by ocean drilling and by dredge sam-
pling indicate that the dredge hauls can usefully supplement the was as representative as possible of the geological
drilling data in the construction of sedimentary and tectonic terrains being investigated. However, through the
histories of seafloor areas. 1960s and 1970s, none of these techniques had ad-
vanced sufficiently to eliminate uncertainties associ-
ated with this sampling method.
Introduction
The advent of scientific ocean drilling, begun in
Marine scientists have, since the earliest days of 1968 by the Glomar Challenger, made it possible to
oceanography, used a variety of towed equipment to sample continuous sedimentary rock sections to
sample the deep ocean floor. Dredges were used in basement and ushered in a period when dredging was
the 1860s on the four-year voyage of HMS Chal- regarded as a relatively poor tool on which to base
lenger to obtain biological and geological samples studies of the sedimentary or tectonic history of an
from all the major ocean basins. The basic design of oceanic area. The Deep Sea Drilling Project (DSDP)
dredges for sampling seaward of the continental was never able to drill directly into igneous rock over
shelves have changed little over the intervening cen- the fifteen years of Glomar Challenger operations and
tury (Fig. 1). A solid metal frame "mouth" is towed it is only since 1985 that its successor, the Ocean
on a wire rope from a deep-sea winch on the ship. Drilling Program (ODP), could offer a supported
Behind the frame is a collecting bag, usually of hard-rock spud-in system to accomplish this task
interlinked metal grommets. The dredge is weighted (Storms, this volume). Because of this and because
to stay in contact with the sea floor either by trailing the drillship is only infrequently committed to such
weights on a heavy chain behind the bag or with a basement studies, the marine "hard rock" petrology

Marine Geophysical Researches 12: 131-143, 1990.

© 1990 Kluwer Academic Publishers.
132 R. B. KIDD ET AL.

(a)

(b)
strong link
t

Fig.!. Two commonly used dredges: (a) Pipe Dredge. This can either be open ended with a collecting bag or closed for use over recent
lavas (see text). The mouth diameter is usually around 50 em. (b) Rock dredge. Based on the Nalwalk design (Nalwalk et al., 1961), dredges
like this form the mainstay of geological dredging operations.

community has retained dredging as its primary sam- together with an assessement of the status of geolog-
pling tool throughout the past two decades, supple- ical dredging in the late 1980s in the light of the
menting it more recently with submersible operations developments in drilling and submersible technology.
(e.g. Fox and Stroup, 1981). Drilling operations,
also, cannot satisfy the need for large-volume aerially
Types of Dredges
distributed samples for use in investigations of re-
gional diversity. Early dredges were designed to collect for study
Over this same period, techniques of swath map- whatever material was present on the sea bottom, be
ping (Grant and Schreiber, this volume); accurate it biological or geological. Over the years dredge
navigation and near-bottom survey have advanced design has diversified and the more sophisticated
markedly. Furthermore, results are now available dredging equipment is presently used by the biologi-
from precisely-positioned continuously-drilled sec- cal community (Fig. 2). However, parallel develop-
tions in areas where knowledge of the geological ments in techniques for navigating the dredge and its
history previously had been based only on dredge behaviour have occurred for both applications of
surveys. This enables us to consider the representa- dredging. In some cases, biological benthic dredges
tiveness of dredge haul material in regional geologi- have been used by geologists, for example to study
cal studies. Examples are given later in this review, ice-rafted and other rock debris on abyssal plains

weak link

Fig. 2. Benthic sledge used for sampling sediment surfaces. Used mainly by biologists for sampling benthic communities, this dredge will
also collect manganese nodules and other rocks lying on sedimented seafloors.
 GEOLOGICAL DREDGING TECHNIQUES 133

o (a) (b) o
o
"'
o
a

\I~ pinger housing

I
meaS16ing ins1rument
box

Fig. 3. Prototype sampling sledge specifically designed for man-

ganese nodule collecting, dimensions are in centimeters (after
Kinoshita et al., 1975).

(Kidd and Huggett, 1981). Some of the largest

dredges aimed at rock collection, with frames over
1.5 m-Iong, were designed as part of prototype sys-
strong link-
tems (Fig. 3) for manganese nodule mining, again
for use on sedimented abyssal plains (Cronan 1980).
We shall concentrate here on dredges designed to
sample rock outcrop in geological studies of conti-
nental margins, mid-ocean ridge systems and other
sea floor physiographic features.
Geological dredge design has remained very sim-
ple, largely because of the logistical constraints of
operation in rugged sea floor terrain and the high
risk to sophisticated equipment when the primary
aim is to break off and recover in situ material for
study. Three basic designs have had widespread use:
(a) pipe dredges
(b) rectangular dredges Fig. 4. Dredge rig currently used by the Institute of Oceano-
( c) cylinder dredges graphic Sciences, U.K. (after Gaunt and Wilson, 1975): (a) The
complete rig with weak and strong links. (b) The pinger housing
(cutaway drawing).
(a) PIPE DREDGES
These relatively small dredges comprise a metal pipe
linked to the trawl warp (tow-cable) through a metal should collect large numbers of small cobble-size
bail. The pipe frame is generally around 15 cm in samples; and (b) that it should be able to free itself
diameter and 45 cm long (Fig. la). A net collecting from a rocky bottom. The bail linking the dredge
bag completes the dredge, whose total weight is frame to the trawl warp was free to swivel and this
around 25 kg. Although popular for relatively rapid permits the dredge to ride over large boulders or
reconnaissance sampling of sea floor features until massive outcrops without becoming snagged. One
the early 1960s, pipe dredges are now rarely used. side of the Nawalk dredge frame or mouth was
weighted with lead so that this side dragged the
(b) RECTANGULAR ROCK DREDGES bottom. Also a weight was towed at the rear of the
These are the most commonly used dredges. Most system to stabilize the chain bag in deployment and
rectangular-frame rock dredges (Fig. Ib) have towing. Weak-links were used in the system to pre-
evolved from, or have similar basic components to vent the tail weight "hanging up", if it became
that described by Nalwalk et al. (1961). The specific caught in the bottom, and to allow the bail to break
aims of the Nalwalk design (Fig. 4b) were: (a) that it free if the frame became snagged. The weak-link was
134 R. B. KIDD ET AL.

of rope, with a swivel on each end and was rigged

between the ship's cable and the bail. A light chain
was led in parallel from the ships cable end of the
wire rope to the chain bag such that a break at the
weak-link would overturn the dredge and allow its
recovery without dumping the bag and contents. In
this design varying amounts of angular swivel at the Fig. 5. Cylindrical Dredge after Fillippov et. al. (1970). The
base were used depending on how much soft sedi- dredge is I metre in diameter by 1.5 m long.
ment cover was expected in the sampling area.
Sizes of rectangular rock dredge vary from 38 cm it. A 4-ton shearing link was the last line of defence,
to 90 cm mouth-width and total weights vary from situated at the end of the trawl warp. If the dredge
30 to 120 kg. had failed to clear, by sailing the ship on an opposite
A range of modifications on the basic design have course or by shearing of the lower strength links,
been utilized: cutting teeth have been included on the this last link ensured loss of the dredge but kept the
dredge frame mouth (Fig. lb) with a view to break- warp intact. Figure 4 shows a later version of the
ing off outcrop samples. Some systems, including a lOS system, modified for use with a tilt-switch
variation used by the Lamont Doherty Geological pinger, but still incorporating some of the weak-
Observatory (LDGO), include: linked system.

- a lead ball weight placed ahead of the bail of the (c) CYLINDER ROCK DREDGES
dredge, to try to break samples off outcrop prior to
Cylinder dredges are most commonly used by Rus-
the dredge passing over them;
sian research institutions, although they are in use
- nylon nets of varying mesh size used as inner
also by American (e.g. the United States Geological
linings for the metal grommet collecting bag. These
Survey Marine Geology Group) and Japanese insti-
control the size of material collected. The mesh
tutions. The design of the cylinder dredge is again
linings cannot be too fine or they restrict water flow
simple but the frame ahead of the collecting bag is an
through the bag and it soon fills with mud;
elongate cylinder. The design was an attempt to
- heavy anchor chains sometimes used instead of a
circumvent perceived problems with early rectangu-
single tail weight. In some systems two chains are
lar dredges relating to rigid towing bails, short
towed in parallel from a box-shaped grommet bag.
frames and generally low weight, which together
Systems of weak-links in series are frequently a often resulted in these early devices bottoming in a
feature of rock dredging operations because of the non-operating position or adopting an orientation
high incidence of dredges becoming stuck on the during towing in which the mouth would not sample
bottom. These weak-links ensure failure in stages throughout the operation. Later rectangular dredges
should the dredge "hang up" and put the ship's trawl as described above have swivelling bails, weighted
warp at risk of parting. lower frames and release mechanisms, but the simple
The standard rock dredging gear used over the operation of the cylinder dredge is still favoured by
years by the British Institute of Oceanographic Sci- some groups, especially in reconnaissance studies.
ences was an example that used weak links and was The cylinder dredge described by Fillippov et a/.
designed to collect a wider range of sample sizes than (1970) is a typical example (Fig. 5). The main com-
the basic Nalwalk design. Between the frame and the ponent is a 12 mm-thick steel cylinder, 1 m in diame-
towing bridle, a pin that sheared at 1 ton strain ter and 1.5 m long. Two 12 mm-thick rings are
allowed the dredge to tilt at 30° (Burt, 1979). A welded to the cylinder's outside surface, providing
2.5-ton weak link was inserted between the dredge additional strength. Rounded cutting teeth are an
bail and the chain connection to the warp. A 13-mm integral part of the forward end of the cylinder. Two
wire ran as a noose from this chain through the front lugs of round steel rod provide towing attachments,
part of the grommet bag so that when the 2.5-ton welded to the outside of the cylinder forward of its
link broke, the bag would be throttled and the frame centre of gravity. A series of drilled holes at the tail
was turned over from the rear which usually released end provide securing links for the net collecting bag.
 GEOLOGICAL DREDGING TECHNIQUES 135

L
--;SCAlE MARKERS o
W
E
R
I
N
G

PINGER AND DREDGE

.-:....-.... _-.....- - RbRIZONTAl ON BOrrOM o
t
N
B
o
- T
T
o
M
~ •. ' - , (JUMPING OF DREDGE

15·55
r(-I-'- PINGER GOIN
VERTICAL MOD1 )INTO
INDICATING ROUGH
- BOTTOM
. -
r ,/ •
I. .
..,...-
...
. . ..' . . ~
-

PINGER AND DREDCe- ~.-

LEAVE BOTTOM
SLOW
tiAUUNG

.~-

.'- A
I
S
R

I
N
G

Fig. 6. Pinger record from a dredge station where several "bites" were felt. (After Gaunt and Wilson, 1975).

This conical bag is protected against tearing by an were seen to be:

ox-hide cover and narrows to a towpoint for a 50 kg
tear-drop tail weight. The dredge is towed simply I. its smaller contact area with the substrate
from chains connecting the side lugs to the ship's and the relatively large weight applied to this
cable. The overall weight of the system is 650- area to engage or scrape rock projections;
750 kg. 2. its shape which ensures that it lands and is
The advantages of the simpler cylinder dredge over kept in an operating position on the sea
less sophisticated versions of the rectangular dredge floor;
136 R. B. KIDD ET AL.

3. the flexibility of the towing arrangement the depth below the ship are of little use in trying to
which allows it to roll sideways rather than calculate true dredge position.
hang up on a projection; and The key elements in any successful dredging opera-
4. the relatively weak cutting teeth which will tion are navigation of the ship and of the dredge
bend or break off before tensions build system and also monitoring of the behaviour of the
enough to threaten breakage of the ship's dredge at the bottom. It is clearly important not only
cable. to recover rocks but also to have information on
where in the terrain they were sampled from. Even the
The greatest attributes of the cylinder dredge are its
crudest operations depend on a pinger placed near the
simplicity, its strong construction and its reliability.
dredge (Fig. 6) to transmit information from depth
and a tensiometer to record load on the ship's cable
Dredging Operations and from this dredge behaviour and likely sampling
times from the recorded "bites" (Fig. 7).
THE TARGET
Although improvements in navigation and in
Rock dredging operations are carried out in rugged monitoring dredge behaviour have been spectacular,
sea floor terrains where slopes are usually sufficient there still remains a certain mystique to dredging
to ensure only a thin cover of soft sediment. Detailed operations. Thus, even in the early days of tensiome-
regional bathymetric and seismic surveys will have ter measurements, practitioners would still sit on the
been run and the dredging is usually aimed at sam-
pling particular sequences identified as outcropping

I
on seismic profiles. Targets are mainly fault scarps or €
o
o
the eroded sides of canyons and channels. Rock
debris or talus at the base of scarps can often make
'"
up most of the recovery, although freshly-broken
material from the scarps themselves provide the ideal
,
sample. Some investigators involved in submersible
l ocation of
," l ;
I ,~~

~ {} <9
i'

~
"
'1..
operations believe that so much of the material h i p when load _ _ ? 6 4
ell r ecorded 5
recovered in dredging at mid-ocean ridges is from ma or pulls
1 1 98 7 65 4
talus piles that its usefulness in interpretation is sea level

seriously impaired (e.g. Fox and Stroup, 1981). As feet

will be described later, dredging operations in higher '000

latitudes recover additional ice-rafted material from 2
terraces in the terrain and from talus piles. This has
been transported and dropped by melting icebergs. 500
1m
These lithified rocks require careful sorting from in 4
situ material.

OPERATIONS IN GENERAL
' 000 6
Once the targets have been selected, ship operations fm
echoes
bot t om _ _ _ _ ____
are generally designed to time the dredges' arrival at
the bottom such that the ship can then take a course
at right angles to the feature and make the dredge t rue location of dredge
from synchronized pinger's
computed~
haul over the scarp. If at all possible operations are ship-to-dredge range and
_~,

balhykymograph record
carried out head-to-wind, recognizing that the cate-
nary of the ship's tow cable or trawl warp will put 10 8 6
feet x 1000
4 2

the dredge a kilometer or so behind the vessel.

Typically, around 8000 m of cable is paid out for
Fig. 7. (above) Plan view of a dredge run showing the estimated
dredging operations at around 5000 m water depth. "bite" positions. (below) Elevation view showing how the "bite"
However, information on the length of cable out and times are converted into estimated positions (from Aumento, 1970).
 GEOLOGICAL DREDGING TECHNIQUES 137

ship's warp as it passed over the afterdeck in order to dredge to record both absolute depths reached by the
"feel" the "bites" as the dredge was supposedly dredge with time (actually ship-to-dredge range) and
sampling! the instants and depths when it underwent the major
Through the early 1970s the most accurate deep- "jerks" or abrasions. (These were also recorded by
sea dredging operations relied on a system of moored the load cell tensiometer on the ship.) Typical results
radar transponder buoys, operated in conjunction from such an arrangement are compiled in Fig. 7.
with satellite navigation (Loncarevic, 1969) for con- Alternatively, a combination of a regular pinger,
trol of the ship's position. mounted up to 300 m from the end of the ship's
cable to measure water depth, and a tilt-switch
DREDGE "MONITORING"
pinger (Fig. 4) mounted in a protective housing
Information on dredge behaviour in the early 1970s immediately ahead of the dredge was used for moni-
came from a ruggedly built "bathykymograph" (Au- toring (Gaunt and Wilson, 1975). The latter pinger
mento, 1970) or a pinger mounted 20 m ahead of the was designed so that its pulse-repetition rate changed

," '.' . .. ..
. . ..
..
o 5 10
, , , . . ..
Kilometres

Fig. 8. North slope study area for the detailed sampling programme of "Discovery" Cruise 84. Contours are in corrected metres. Shaded
areas represent fault scarp outcrops seen on GLORIA long-range sidescan monographs. Stippling shows the sedimented trough floor below
4500 m stations. Dots indicate core positions; bars represent dredge. Dashes and crosses show the trace of the summit of the ridge; A and
B are the positions of the acoustic beacons. CDEF, GH and PQ are seismic profiles (after Kidd et. al., 1982).
138 R. B. KIDD ET AL.

when its orientation (and that of the dredge) RIA imagery, together with seismic profiles were
changed from vertical to horizontal. The acoustic used to detect outcrops. On sonar images over
signal from this pinger was received on the ship's rugged terrain, fault scarps give rise to sharp linear
hydrophone, and vertical and horizontal modes echoes and shadows, outcrops of volcanic basement
could be determined by changing the facsimile create more diffuse or chaotic echoes and sediment
recorders' scan speed (Fig. 6). The main advantage cover shows up as fields of uniform low-level
of the tilt switch pinger arrangement was that it was backscattering of acoustic energy. Thus with GLO-
no longer necessary to make the assumption from RIA an accurate sketch map of the regional geology
the regular pinger record that, provided the height of could be drawn and constrained by crossing seismic
the pinger off the bottom is substantially less than its profiles. These satellite-navigated survey data can
distance from the dredge, then the dredge must be in then be used to select target dredge sites and the
contact with the bottom. In deeper water it was positioning for bottom transponders to navigate the
possible to misread the facimile recorder. The prac- dredging operations.
tice was to observe the crossing of the direct signal Kidd et al. (1982) used two transponders laid about
and the bottom echo and to count the number of 10 km apart, and 4 km away from a scarp that was to
complete phases, stopping when the pinger suggested be sampled as part of dredging operations in 1977
the dredge was on the bottom. With the pinger away within the King's Trough tectonic complex in the
from the dredge this exercise could be flawed. The Northeast Atlantic Ocean (Fig. 8). The transponders
tilt switch pinger was mounted immediately ahead of were on tethered lines about 200 m above the sea floor
the dredge and it also indicated exactly when the and were "surveyed in" with additional echo sounder
dredge was horizontal and so able to sample. and seismic profiles. An interrogator pinger was towed
Some researchers at this time also attempted to in a protective "fish" 200 m along the trawl warp
monitor their operations with a small bottom cam- ahead of the dredge. The make-up of the rest of the
era/flash unit mounted along the ship's cable. This system followed that of the lOS rectangular dredge
would give important information on where in the arrangement with tilt switch pinger as described by
dredge traverse the material being sampled was talus Gaunt and Wilson, 1975 (Fig. 4). This first attempt
or potentially from outcrops (Aumento and at transponder-navigated dredge sampling in King's
Lawrence, 1968; Laughton et aI, 1971). Most major Trough provided accurate navigation of the ship
developments in the design and deployment of relative to the baseline and to the scarps as identified
dredging equipment had taken place by the mid- by GLORIA, and so provided a high degree of
1970s. Improvements in dredging operations after confidence as to which outcrops and terraces were
that came largely from improved navigation and being sampled (Fig. 8). It paved the way for develop-
from major advances in survey equipment. ment of more efficient bottom navigation of the dredge
gear across areas of GLORIA survey in later years.
DREDGING WITH BOTTOM-TRANSPONDER NAVIGA-
TION AND GLORIA DREDGING WITH SEABEAM:
In the late 1970s bottom transponders began to be Bathymetric contouring with swathe-mapping sys-
used to provide a baseline around which deeply- tems (Grant and Schreiber, this volume) has pro-
towed vehicles and sampling equipment could be vided even more opportunities for precision in
navigated by triangulation of acoustic ranges be- dredge sampling. Satellite-navigated swathe contour
tween the ship and the bottom transponders. Later maps, computer-drawn at a resolution of 10 m or
an interrogator pinger was incorporated on the less for rugged terrain are frequently so reproducible
ship's cable (Robertson, this volume). At about the that it is possible to reposition the operating ship
same time, the British Institute of Oceanographic over a selected fault scarp or outcrop and accurately
Sciences long-range sidescan sonar system (GLO- control a dredging operation to sample it. This sim-
RIA) (Somers et al., 1978) provided a facility which plified mode of operation has become increasingly
displayed the regional morphology and the texture of popular during the 1980s with French and American
the sea floor and provided new opportunities for institutions who operate hull-mounted Seabeam sys-
precise dredging activities. Satellite-navigated GLO- tems. It avoids the time taken to deploy bottom
 GEOLOGICAL DREDGING TECHNIQUES 139

I
I
: 8,7 % In Situ 1100% In Situ
,, I
Volcanic Asn
.
,t1Y.lloCI •• tih

Hyaloclastite

Fig. 9. Classification of material from two dredge hauls from the North Atlantic showing the variability of ice rafted content of dredge
material (after Huggett and Kidd, 1984).

transponders and to survey in their baseline, which Dredge Haul Recovery

can run to a full day of shiptime.
There are a number of factors that may affect the
More advanced versions of long-range sidescan
amount and type of recovered material in an individ-
and swathe bathymetry systems are coming into use
ual dredge haul.
as the decade closes. An advanced GLORIA system
with quantitative bathymetry is mooted (M . Somers,
1989, pers. comm.) utilizing principles of signal mea- IN SITU AND EXOTIC MATERIAL

surement similar to those developed in Seamarc I The types of recovered material are broadly divided
and its derivatives. Hydro sweep (Grant and into the "in situ" and the "exotic". In situ rocks are
Schreiber, this volume) and other derivatives of the either freshly-broken samples from outcrop or are
Seabeam systems are now operational. All are likely rocks recovered from talus or scree accumulating at
to have a profound effect on any future advances in the bases of scarps. Exotic material on the ocean
dredging technique, since all these survey tools are floor principally comes from floating ice (glacial
destined to be used in conjunction with the Global erratics) and is much less confined to high latitudes
Positioning System (GPS) (Robertson, this volume). than had been previously thought. Generally, hauls
Already the order-of-magnitude improvement in at higher latitudes than 40° should now be expected
ship's navigation provided by GPS, to within 1 m to contain glacial erratics (Kidd and Huggett, 1981),
accuracy, can mean that, if dredging is carried out but investigators should note that erratics have been
over GLORIA or Seabeam-surveyed terrain, the op- recovered from seamounts within 30° of the equator.
erator has considerable control over which scarps are Other exotic material can be transported to the sea
being sampled, even in the absence of a bottom floor attached to help hold fasts (Emery and
transponder net. Tschundy, 1941), by sea mammals (Emery, 1963), as
 .......
~
Absc>lute age IMal o to 20 30 40 50 60 70 80

Chlcn 1 22A 33A 44A SA S8SCSOSES SA SC 77AS 9 to 1112 13 15'61718 '920 2' 22 232' 2S 28 212829303 ' J2 33 3'
MagnetiC ChronOlogy _ . i . _. 1M • - -- _. . --
and time scale
(Berggren el al
.98Sa , bl
CampanIan
Eocene

104 520 8
o
Mag Anom 21
Trachyles Sasalts
v
. v
Klng's TullS
Trough
H ---v" H I 09572 ~
Foraminiferal ZeolltlC
AXIS sand sand
- r---t
H H
09574 1 09571 wi 09567
...
o
Mag Anom 24
AmphlbOhtes
(retrogradedl Amphtbolites
AI •
V
Palmer Winnowed
Ridge chalks
!"'
~ H ·vvvv - ---, ~ Iwvv l ~
:><:
H
05971 8
Age range 01 recovery In
dredges and at drill sIte ~
o Age lange 01 rock cores
>
Mag Anom 18
reco'Yered
r
King'S
I- -I Age range 01 absolute dates
Trough ~ I I on dredged rocks
flank
Site 608
vvvv ...
I WWWVVV Sasatt
o Age of ocean crust Irom
I~ = :~ magnellc anomaly number
? H,atus
WWWVVWV VolcaniC ash mpul

... Braafudosphsars occurrence

EVidence 01 sediment instability

Age range 01 hIatus at S,te 608

Biostratigraphic age gap

I I TOlal Klng's Trough samphng

Fig. 10. Comparison of the age ranges of dredged materials from two North Atlantic study sections (Kings Trough and Palmer Ridge) with
the drilled stratigraphy at a Deep Sea Drilling Project site (DSDP site 608) from the same area. Ages for rock cores taken in the dredging
locations are also shown. Note that the dredging recovery partly encompasses a time gap represented by a Hiatus at Site 608. (See key for
further detail).
 GEOLOGICAL DREDGING TECHNIQUES 141

ballast and clinker discharged from sailing and has also highlighted the effect of ship's track in
streamships, as modern man-made litter or as float- faulted terrain. Hauls taken along-strike frequently
ing volcanic pumice. Thus location of the sampling contain more material but this is dominated by
area in relation to glaciated land areas, past and manganese-coated talus and erratic material as com-
present shipping lanes, volcanically active areas and pared to the small hauls taken across strike that can
sea-mammal migration routes can be important in contain a high proportion of freshly broken material.
determining the content of a dredge haul.
Huggett and Kidd (1983/84) have constructed a
Comparison with Drilling
set of criteria for the often-difficult identification of
glacial erratics in dredge hauls. The criteria are in Prior to the advent of scientific deep ocean drilling in
two categories. The first includes observations based 1968, geological studies of sea floor features were
on direct evidence visible on individual clasts, for heavily dependent on dredging. Typically, an area or
example glacial striations and facetting. The second feature would have been extensively surveyed
category includes inferred evidence drawn from com- geophysically with echo-sounding, seismic reflection
plete dredge hauls, for example, variations in the profiling, gravity, magnetics and seismic refraction
thickness of manganese coating and groupings of the before detailed station work with rock coring, bot-
rocks into assemblages. Typical results of this ap- tom photography and dredging. The deep-sea
proach on two North Atlantic dredge hauls are drilling vessel Glomar Challenger allowed investiga-
shown in Fig. 9. tors to precisely test interpretations of the geophysi-
cal data and to calibrate seismic sections by drilling
DREDGE DESIGN AND OPERATIONS RECOVERY reflectors at depth. With steady improvement in
As outlined in the Section, Types of Dredges, dredge drilling technique through the Deep Sea Drilling
design and operations probably have the most sig- Project and, since 1986, through the Ocean Drilling
nificant effects on haul recovery. The size and shape Program aboard DV JOIDES Resolution, a continu-
of the dredge mouth will control maximum clast size ously-cored section that establishes the stratigraphy
collected, while the size of links in the collecting bag of an area became possible using HPC + XCB +
or of any finer mesh netting liner will control mini- RCB drilling (see Storms, this volume) and this has
mum clast sizes retained. Although it is assumed that become the standard. Although dredging is fre-
most dredges collect all the loose material they cross quently used during site survey to establish the scien-
that is small enough to enter the mouth, some tific case for drilling a particular feature, the
dredges are designed to ride over (e.g. the rectangu- technique has become recognized as a poor substi-
lar dredge of Nalwalk et al., 1961) or sideways (e.g. tute for a drilled section.
the cylindrical dredge of Filippov et al., 1970) past Dredging cannot provide the investigator of base-
obstructions. Not all dredges have substantial cutting ment petrology with entirely unaltered material and
teeth that are capable of breaking outcrop (like the this has provided a strong case for carrying out
rectangular dredge of Aumento, 1970). With excep- drilling on basement features that may be suspected
tionally large hauls some dredges may lift off and as being composed of easily altered material such as
spill some of their contents, thus it is sometimes serpentinite in diapirs (Boillot et al., 1987; Kastens et
difficult to link haul size to distance travelled over al., 1987).
the bottom (Kidd and Huggett, 1981). A study which was able to directly compare the
results of dredging and rock coring with drilling data
SHIP'S TRACK RECOVERY
and which favours dredging was reported by Kidd
Careful survey and planning of ship's track during and Ramsay (1986). A drilling site occupied late in
dredging can improve the quality of material recov- the DSDP programme at King's Trough in the north
ered. Scree-like accummulations of talus at the foot east Atlantic was designed to test a hypothesis on the
of fault scarps can result in large hauls but these are formation of this tectonic complex that had been put
also the sites that, in high latitudes, commonly pro- forward based on a range of geological and geophys-
duce high proportions of glacial erratics (up to ical data including GLORIA sidescan and bottom
45%). Precise dredging arounq bottom transponders transponder navigated dredging (Kidd et ai., 1982).
142 R. B. KIDD ET AL.

A continuously cored section, DSDP Site 608, was tion is likely to provide the most powerful investiga-
drilled into Eocene oceanic basement at 515.4 m tive approach.
sub-bottom on one of the flanks of the tectonic
complex, penetrating reflectors that were identified
The Future
on regional seismic profiles. One of the reflectors is
represented by an Eocene/Oligocene hiatus of Many marine geologists are recognizing, at the be-
around 9 m.a. in duration. All of the rock types ginning of the 1990s, a move towards conducting
recovered in the drilling had already been recognized deep-ocean geological studies at the outcrop scale
and dated in dredge hauls. Some of the dredge rather than solely through regional remotely-sensed
materials spanned part of the period of hiatus devel- investigations. We have the "Landsat and aerial
opment and some haul material from the east of the photography" level tools in the derivatives of GLO-
complex pre-dated the rocks immediately above the RIA and Seabeam with which to select sea floor sites
basement at the drill-site (Fig. 10). Kidd and Ram- and set about studying the geology at the outcrop
say (op. cit.) were able to use the drill-site as a level. With the precision afforded by GPS and bot-
reference stratigraphic section and to combine this tom transponder nets dredging probably will experi-
with the dredge haul data and thus refine the tectonic ence an upsurge of interest, at least in reconnaissance
model and timing of formation of this intraplate studies of newly explored regions. Subsequent de-
feature. tailed studies of such areas, however, are likely to be
The major features of the model were, however, conducted by remotely-controlled combined camera
originally defined on the basis of a series of ~ruises and sampling systems (e.g. rock drills), submersibles
that conducted dredging and, although an early drill- and/or by dynamically-positioned and video-moni-
site may have resulted in the same conclusions, some tored drilling. On the other hand, we can also foresee
of the information on timing and the character of dredging being used to extend information gained at
regional tectonic events would not have come avail- reference drillsites.
able without a number of sites being drilled. We
conclude that dredging can be a useful additional
tool, even when the drilling of sequences is attain- References
able. Additional regional information might be Aumento, F., 1970, Improved Positioning of Dredges on the
derived by extension of the geology away from a Seafloor, Can. J. Earth Sci. 7, 534-539.
drilled reference section using dredging. There may Aumento, F. and Lawrence, D. E., 1968, Photographic Control
of Deep Sea Dredging, Ceol. Surv. Can. Paper 68 (9), pp.
be many locations like Kings' Trough that have been 1-3.
drilled but would benefit from at least a fresh analy- Boillot, G., Winterer, E. L., and Meyer, A. W. et al., 1987, Proc.
sis of regional dredges hauls. ODP 'Initial Reports' 103: College Station TX (Ocean Drilling
Program) 809-828.
Burt, R. G., 1979, Handling Oceanographic Equipment: Notes
and Sketches from a Netman's Log, Institute of Oceanographic
Comparison with Submersible Studies
Sciences Report 34, 62 pp.
Cronan, D. S., 1980, Underwater Minerals, Academic Press, Lon-
Fox and Stroup (1981) stress that there is no substi-
don, 362 pp.
tute in complex mid-ocean ridge terrain for close-up Emery, K. 0., 1963. Organic Transportation of Marine Sedi-
geological observation using submersibles. However, ments, in M. N. Hill (ed.), The Sea, Vol. 3. Wiley and Sons,
the amounts of selected samples that can be collected New York, pp. 776-789.
Emery, K. O. and Tschundy, R. H., 1941, Transportation of
in submersible dives are very limited. Hoillot et al. Rock by Kelp, Ceol. Soc. Amer. Bull. 52, 855-862.
(1987) very successfully combined initial dredge Fillippov, L. A., Krausch, A., Barash, M. S., Laurov, V. M. and
haul results on the Galicia continental margin with Dmitriyea L. V., 1970, A Large Cylindrical Dredge, Oceanol-
ogy, 140-142.
both deep-sea drilling and follow-up submersible
Fox, P. J. and Stroup, 1981, The Plutonic Foundation of the
observations. Clearly the general availability and Oceanic Crust, in Emiliani, C. (ed.), The Oceanic Lithosphere,
the expense involved in both submersible operations Vol VII, The Sea: Ideas and Observations on Progress in
and drilling mean that many investigators may the Study of the Sea, J Wiley and Sons, New York, pp.
119-218.
have to confine their efforts to dredge activity; how- Gaunt, D. I. and Wilson, J. B., 1975, Acoustic Monitoring of
ever, the combination of all. three types of opera- Dredge Behaviour on the Sea Floor, Deep Sea Res. 22, 91-97.
 GEOLOGICAL DREDGING TECHNIQUES 143

Huggett, Q. J. and Kidd, R. B., 1983/84, Identification of Ice- T., and Handa, K., 1975, Technical Notes on Deep Sea Bottom
Rafted and Other Exotic Material in Deep Sea Dredge Hauls, Sampling, in Mizuno, A. et al., Cruise Report No.4. Deep Sea
Geo-Marine Letters 3, 23-29. Mineral Resources Investigation in the Eastern Central Pacific
Kastens, K. A., Masc1e, J., Auroux, C. et al., 1987, Proc. ODP Basin. Geol. Surv. Japan, 49-61.
Initial Reports 107: College Station TX (Ocean Drilling Pro- Laughton, A. S. et al., 1971, RRS Discovery Cruise 33, N.I.O.
gram). Cruise Report No 33, National Inst. of Oceanography, Worm-
Kidd, R. B. and Huggett, Q. J., 1981, Rock Debris on Abyssal ley, Surrey, 23 pp.
Plains in the Northwest Atlantic: A Comparison of Epibenthic Loncarevic, B. D., 1969, Buoy Plot as a Survey Aid. Trans.
Sledge Hauls and Photographic Surveys, Oceanologica Acta 4, Applications of Sea-Going Computers Symposium, Mar. Tech.
99-104. Soc., Washington D.C., pp. 27-33.
Kidd, R. B. and Ramsay, A. T. S., 1986, The Geology and Nalwalk, A. J., Hersey, J. B., Reitzel, J. S., and Edgerton, H.,
Formation of the King's Trough Complex in the Light of Deep 1961, Improved Techniques of Deep-Sea Rock-Dredging, Deep
Sea Drilling Project. Site 608 Drilling, in Ruddiman, W. F., Sea Res. 8, 301-302.
Kidd, R. B., Thomas, E. et al., Init. Repts. of Deep Sea Drilling Somers, M. L., Carson, R. M., Revie, J. A., Edge, R. R., Barrow,
Project, Vol. 94, Washington, pp. 1245-1261. B. J., and Andrews, A. G., 1978, GLORIA II-An Improved
Kidd R. B., Searle, R. c., Ramsay, A. T. S., Pritchard, H., and Long-Range Sidescan Sonar, Proc. IEEE/IERE Sub-Conference
Mitchell, J., 1982, The Geology and Formation of King's on Offshore Instrumentation and Communications, OCEANOL.
Trough, Northeast Atlantic Ocean, Marine Geology 48, 1-30. INT. 1978, Tech. Sess. J. BPS Publications Ltd., London, pp.
Kinoshita, Y., Maruyama, S., Honza, E., Yamakado, N., Usami, 16-23.
The Use of Sediment Traps in High-Energy Environments

J. WHITE
Department of Civil Engineering, University of Southampton, Southampton S09 5NH, UK

(Received 27 April, 1989; 1 September, 1989)

Key words: sediment traps, high-energy environments, suspended Only recently have experiments been conducted to
sediment sampling.
assess sediment trap behaviour in flow velocities
above 0.1 m/s. This chapter provides an assessment
of sediment trap designs, and reviews recent develop-
Abstract. A sediment trap is a container deployed in the water
column with the aim of providing a representative sample of the ments in the use of traps in high-energy environ-
material settling through that water column before it passes to a ments.
greater depth and ultimately to the seabed or lake bottom. A
review of the previous literature shows cylinders and baffled
funnels to be the most efficient sediment trap design in flows less
than 0.1 m/s. For flow velocities above 0.1 m/s recent evidence Sediment Trap Shape
suggests upwelling from the trap base, and possible undercollec-
tion. The degree of undercollection depends on the flow velocity, Since the first recorded use of a sediment trap by
the type of trap, the height: diameter (aspect) ratio of the trap,
Heim (1900) a variety of designs have evolved to
and the type of sediment. Recent experiments suggest that cylin-
ders with an aspect ratio of ~ 3 may be efficient collectors in suit individual needs, the designs commonly being
velocities up to 0.2 m/s. The use of cylinders is not recommended based on intuitive assumptions of trap behaviour
in velocities above 0.2 m/s. For unbaffied asymmetric funnels a rather than tested models. Gardner (1980a)
lower limit of 0.12 m/s is suggested.
classified the designs into the five broad categories
below:
i) Cylinders;
ii) Funnels;
Introduction
iii) Wide-mouthed jars;
Over the past fifteen years sediment traps have be- iv) Containers with bodies much wider than the
come an increasingly popular tool for investigating mouth (e.g. Flasks and Tauber Traps);
particulate flux in oceanic and lacustrine environ- v) Basin/tray-like containers with width much
ments. The aim of a sediment trap is to provide a greater than height.
representative sample of the material settling The different shapes are illustrated in Fig. 1.
through the water column, before it passes to a The earliest studies to investigate the particulate
greater depth and ultimately to the seabed, or lake flux values determined from traps of different shapes
bottom. Most of the early sediment trap studies were were simple field comparisons. Pennington (1974)
undertaken in environments in which the current found that the sedimentation rate inferred from cylin-
velocities were below 0.1 m/s. Laboratory experi- drical traps deployed in Lake Windermere agreed
ments have verified that in such conditions certain closely with known rates from core samples, palaeo-
trap designs provided an accurate estimate of the magnetic evidence and Pb 210 dating. Funnel traps
vertical flux. The subsequent deployment of sediment deployed simultaneously, however, tended to give
traps in a wider range of conditions, such as the sediment accumulation rates of 0.3 to 0.5 of the
continental slope, submarine canyons, estuaries, and expected value (i.e. to "undercollect" sediment). This
the nearshore zone, meant that assumptions about supported the earlier work of Johnson and Brinkhurst
trap performance were made beyond the hydrody- ( 1971) who reported differences in the collection
namic conditions for which they had been tested. efficiencies of cylinders and funnels deployed in the

Marine Geophysical Researches 12: 145-152, 1990.

© 1990 Kluwer Academic Publishers.
146 J. WHITE

Flaw yglocl'Y_
o 0.01. - O· 01; 5 mls
• 0·0/' -0·0/'5 m/s (trap rotat~d)
180 o 0.09-0095 m/s

>-
~ 160 s..gmW1t~d Jar Ftask Taub<zr trap

.. o~
'!
~
Box
1/, 0
0
~
Ii! o net
tested
'0 120 o
..'"
v ~

..
c
C
v
100 - - - - - - -0-- - -0- - - -
8
~- - - t --- -- --- --------
0

~-
0

-~- ~-

.f o •• • • 0
80
•
60

Fig. 1. The collection efficiency of different sediment trap shapes (adapted from Gardner, 1980a).

Bay of Quinte and Lake Ontario which appeared to and Burns ( 1980), Reynolds et al. ( 1980) and
depend upon the size of the funnel or cylinder used. Blomqvist and Hakanson (1981) have recognized
Tauber traps tested by Pennington (op. cit.) tended cylinders to be the most efficient sediment trap
to "over collect" sediment at an average of 2.3 shape.
times that collected in cylinders. When compared The experiments of Butman (op. cit) have verified
against basin/tray traps however, Reynolds and that in flows up to 0.1 m/s baffled funnels also
Godfrey (1983) found Tauber traps in Lake Winde- provide a good estimate of particulate flux. Baffled
mere to collect up to 25 times more sediment. In funnels have been used extensively in deep oceanic
the oceanic environment, Dymond et al. (1981) ob- environments, for example by Honjo (1980) in the
served only a factor of two variation in the amount Sargasso Sea and E. Hawaii Abyssal Plain and Jick-
of sediment collected by cylindrical, funnel and ells (1984) also in the Sargasso Sea. Funnels have
basin/tray traps in the Santa Barbara Basin. the distinct advantage of concentrating the collected
It is only through laboratory investigations that a material in a sample container at the funnel base
reliable measure of the response characteristics of from which resuspension is unlikely during retrieval.
different types of sediment trap has been obtained.
The turning point in sediment trap methodology
Sediment Trap Size
came with the work of Gardner (1977), later sum-
marized in Gardner (1980a and b). Models of the The earliest investigations into the effect of sediment
five different sediment trap shapes were tested in a trap size on collection efficiency considered different
recirculatory flume for flow velocities in the range sediment trap designs. Davis (1967) found in labora-
0-0.095 m/s. The trapping efficiency was calculated tory experiments that the amount of material col-
by comparing the sediment flux measured in the lected in cylindrical jars with openings varying
trap (mass/cm 2 of trap opening/per unit time) with between 25 and 100 cm2 , was directly proportional to
the sedimentation rate on the flume bed. The results the area of the trap mouth (Fig. 2). Field experi-
(Fig. 1) show cylinders, segmented boxes and ments using funnels by Watanabe and Hayashi
baffled funnels to be the most efficient trap shapes. ( 1971) in a lake environment yielded similar results.
The effect of rotating the traps through 180°, 45°, These results indicate that for fixed relative dimen-
and 135° during the experiments, to simulate a sions of the trap, i.e. height: diameter ratio or "as-
change of current direction is also shown in Fig. 1. pect ratio", the cross-sectional area of the trap will
Since the work of Gardner (1977) subsequent labo- not affect the amount of sediment per unit area.
ratory tests, e.g. those of Hargrave and Burns The collection efficiency of traps is significantly
(1979) and Butman (1986), together with other re- affected, however, if the relative dimensions (aspect
views of existing data, such. as those of Bloesch ratio) are changed. Most of the work in this field has
 SEDIMENT TRAPS 147

/
/
/
I

//}/ /
10
// / ",/
~ /
/
",'"
0
3 // /'"
Q.
0
"-
//
/ /.~ ,/

..
",/
.S
c / / / / "'----- Expecled number
~~
I / //
0">
1/ '"

/ '"
C
~ /
0 5 / /
Q. / //~

'" '"
/: / '" '" 95% confidence limils

/ /
/
/' / /
/./ /'"
/ /
/,0'"
.%/
~/

20 40 60

Cross-sectionol oreo of lrop opening (cm 2 )

Fig. 2. Relationship between amount of material collected and trap cross-sectional area (adapted from Davis, 1967).

been concerned with defining the optimum aspect which depends upon the prevailing hydrodynamic
ratio for cylinders in any given hydrodynamic region. regime (Fig. 3).
The flume experiments of Gardner (1980a) over the Cylindrical sediment traps are thought to colleCt
velocity range 0-0.095 cmls showed the aspect ratio sediment by:
to have no apparent influence on the trapping (i) Particles falling directly into the trap; and
efficiency for cylinders with aspect ratios of 1, 1.1 (ii) Particles being carried into the trap by trap-
and 2.3. In field conditions near the Woods Hole induced turbulence.
Oceanographic Institution, however, where the ve- The asymptotic relationship between collection
locity occasionally reached 0.5 mis, an increase in the efficiency and aspect ratio is thought to mark the
material collected was observed with increasing as- dominance of the process of particles being carried
pect ratio, and Gardner (1980b) suggested an opti- into the trap by trap-induced turbulence. The critical
mum aspect ratio of between 2 and 3. This aspect ratio marks the point at which a quiescent
relationship was also investigated in the laboratory zone is formed at the base of the trap. Above this
by Hargrave and Burns (1979) at velocities of 0.04- aspect ratio there is very little change in the amount
0.05 mls for aspect ratios of 1.2, 2.6, 3.6, 5 and 20.4 of material collected; below this limit however, ed-
and by Blomqvist and Kofoed (1981) in the Baltic dies may resuspend material from the trap base.
Sea for ratios of 0.5, I, 2, 3, 4, 6, 8 and 10. The
results indicate that the apparent flux rate of material
The Relationship between Trap Collection Efficiency
into the trap (i.e. the amount of material trapped per
and Flow Velocity
unit area, per unit time) increases with the aspect
ratio up to a value of about 3 (Blomqvist and All the aforementioned work was conducted in flows
Kofoed, 1981) or 5 (Hargrave and Burns, 1979). less than 0.1 m/s. The encouraging results from sedi-
After this the flux rate tends to a constant value ment trap deployments in these environments lead to
148 1. WHITE

,.. 25
0
"0

;
No..
'e
.2 20

I
0
E
15
""
..
;:

<;
3
~0 10
0..

~
"0

"0
"0 5
r

Aspect Ratio

Fig. 3. Evidence for asymptotic relationship between amount of material collected in trap and aspect ratio, for cylindrical traps (after
Blomqvist and Kofoed, 1981).

their use in a wider range of hydrodynamic condi- the oil droplets stayed or escaped from the traps,
tions. Parmenter et al. (1983) used traps in flows Lau (op. cit.) determined the aspect ratio at which
with a mean speed of 0.3 mls off Georges Bank, upwelling would occur for any given hydrodynamic
Gardner et al. (1983) deployed traps in the "Hebble" conditions (Fig. 4). Unfortunately, the range of con-
area where mean current speeds ranged between ditions and aspect ratios tested by Lau (op. cit.), and
0.08 mls and 0.32 mis, and recently Gardner (1989) the use of oil droplets rather than sediment particles,
investigated resuspension in the Baltimore Canyon limits the extent to which these results can be applied
where the mean velocities were up to 0.19 mis, with to natural sedimentary environments.
maximum velocities of 0.8 m/s. It is only very re-
cently however, that the relationship between collec-
tion efficiency and flow velocity has been fully I
I

10 o 0
o 0 a>eG,'f
examined.
The first investigation of the relationship between 9 I
I
I

STAV
trap efficiency and aspect ratio under a wider range cPo.
of hydraulic conditions was undertaken by Lau
B o o 0 Jr--
I

(1979), who considered the aspect ratio, hid, in .Q 7 o ClIO Gf'

i5 ESCAPE
relation to the trap Reynolds number R t , Udlv, a: 6 00 o 0 o~'
•
U o o.
where; 0
,
h = height of trap
g5
<t
o <><>;SrI>- •
o ~/••• e ,

d = diameter of trap mouth

,,
4
o stay
U = velocity of fluid at trap mouth <> borderline
J
v = kinematic viscosity. • escape

In a series of flume experiments the motion of oil 10.:1 104

droplets at the trap base was observed over velocities Trap Reynolds number

between 0.03 and 0.75 mls in cylinders with aspect

Fig. 4. The influence of aspect ratio and trap Reynolds number
ratios of between 4.7 and 10 i.e., a range of R t values on oil droplet movement from a cylinder's base. Line indicates
between 2 x 103 and 3 x 104 • By observing whether stay/escape boundary (after Lau, 1979).
 SEDIMENT TRAPS 149

Subsequent experiments exammmg the relation-

25

..... -- -...
ship between hydrodynamic conditions, aspect ratio x moor~d FAST 2
and collection efficiency have used the R/ value to % 20 /
/ o moored FAST 5
"0
characterize the flow. In flow visualization experi- E
• drilling FAST 5

ments Gardner (1985) observed a tranquil zone at 2> a moored FAST 6

":J • drotting FAST 6
the base of cylindrical sediment traps with aspect :: 10

ratios of 5, in velocities of up to 0.22 m/s. The '"

£'" 5
corresponding R/ value was 8.4 x 10\ which com-
pares favourably with the value of 8 x 103 in Lau's OI+--.---~;;;;';+;;;;;;;;;;4
0·2 0·1. 0·6
Current speed (m/s)
experiments. Recent flow visualization experiments
by Hawley (1988) have shown that upwelling of a Fig. 6. Amount of material collected in asymmetric funnels at
layer of dye at the trap base in cylinders with aspect different velocities (from Baker et al. 1988).
ratios of 5 starts at R/ = 4.9 X 103, and is almost
continuous at R/ = 8.5 X 103 •
At lower aspect ratios the upwelling in cylinders -; 160 x • 'esulls of Butman 1966
~ for cyhnders
occurs at lower R/ values. The results of Butman (op.
cit.) suggest that cylinders with aspect ratios of 3,
accurately collect sediment at R/ = 2.2 x 10\ but at
4.6 x 103 significantly less sediment is collected (Fig. '"c:~ 80

g
5). Hawley (op. cit.) has shown that in a cylinder with . 40

an aspect ratio of 3, upwelling starts at R/ = 3.5 X 103

.
~
:g
and is almost continuous at R/ = 5.1 X 103• a: 0
10l 10'
r rop Reynolds number
Information regarding the performance of other
trap designs at velocities above 0.1 mls is scant. For Fig. 7. Decrease in relative trapping efficiency with increasing trap
R/ values of 1.0 to 1.2 X 103 Butman (op. cit.) showed Reynolds number for an asymmetric funnel (from Baker et al.
that wide-mouth jars overcollected sediment and fun- 1988).
nels undercollected sediment as compared to cylin- tions in Colvos Passage, Puget Sound. A Flow Acti-
ders, supporting the earlier work of Gardner vated Sediment Trap (FAST) was developed, capable
( 1980a). In the same experiments, baffled funnels and of partitioning the collections according to the veloc-
cylinders collected similar amounts. ity regimes in which the collection occurred. The
Recently Baker et al. (1988) tested an unbaffled velocities were <0.12 mis, 0.12 - 0.3 mis, 0.3-
asymmetric funnel over a range of velocity condi- 0.5 mis, and 0.5 m/s. The results were compared to
similar free drifting sediment traps deployed simulta-
neously, considered to give an accurate estimate of
180
the vertical particle flux since there is little velocity
x
shear across the trap mouth. Figure 6 shows the
160 x
x
results of the moored traps and the free drifting traps
.§ 140 plotted against the velocities. Clearly less sediment is
U
collected at velocities above 0.12 mls in the moored
"8 .,. 120
!:!
traps. Taking the collections at 0.12 mls to represent
100 - -----.I<-----·x - ---- ------ ----- ----
100% efficiency Baker et al. (op. cit.) have shown the
80 "x drastically reduced efficiency of asymmetric funnels at
""
'Ie
lie
60 " " higher R/ values (Fig. 7).
"
40 "
10 3 104
Trap Reynolds number Laboratory Experiments of Trap Efficiency
versus Velocity
Fig. 5. Relative particle collection efficiency vs trap Reynolds
number for cylinders with aspect ratios of - 3 (from Butman, As a precursor to the deployment of cylinders in an
1986). estuarine environment where the current velocities
ISO ], WHITE

TABLE 1 a current of 0.38 mls in the channel and sweeping the

entire length of the channel bed with a weighted
Dimensions of the cylinders tested in laboratory calibrations
domestic broom. The mean mid-channel velocity was
Aspect adjusted to the desired setting, monitored with an
ratio electromagnetic current meter, and the cylinders
External Internal Internal Internal
were placed within the flume.
diameter (ED) diameter (ID) height (heightf
(mm) (mm) (mm) diameter Concentration profiles were measured at either end
of the test section, at the beginning and end of each
100 94 282 3 run. Since there were no closed spaces in the system,
100 94 188 2
100 94 194 I and it was assumed that the flow velocity in the
75 69 276 4 pumps and pipes was too great for particles to settle,
75 69 207 3 the difference between the two sets of concentration
75 69 138 2
75 69 69 values gives a measure of the total amount of sedi-
50 44 220 5 ment to have settled either on the flume bed or in the
50 44 176 4 traps.
50 44 132 3
Each trap was removed after 24 hours, the trapped
50 44 88 2
50 44 44 sediment was filtered through a preweighed glass
microfibre filter, dried at lOsoC, desiccated, and then
weighed to determine the dry sediment weight in the
reached up to 0.4 mls White (1989), conducted a trap. This was then corrected for the amount of
series of laboratory experiments to investigate the material remaining in suspension within the trap at
collection efficiency of 12 cylinder types (Table I) at the time of its retrieval, to specify the total dry
velocities of 0.1, 0.2, 0.3, and 0.38 m/s. The labora- weight of sediment collected on each trap base. This
tory experiments were conducted in a 22.S m long, was then compared to the calculated amount of
1.37 m wide and 0.6 m deep recirculatory flume. The material collected on the flume bed per unit area.
freshwater of the flume was "seeded" with natural For a 100% efficient trap in any given flow, the
sediment taken from Port Hamble Marina, Hamble amount of sediment collected on the trap base and
UK, to a concentration of approximately 60 mgli. that collected on the flume bed (per unit area) would
For each experiment, 24 cylinders were tested simul- be equal. The procedure was similar to that per-
taneously, as shown in Fig. 8 with the mouth of each formed by Gardner (1980a).
cylinder at 0.3 m above the bed in a water depth of Figures 9a and b show the collection efficiencies
0.S2 m. Before the start of each run the sediment was plotted against velocity for cylinders with different
stirred into suspension within the flume by producing aspect ratios at 0.1 and 0.2 mls respectively. The

Concqnlraloon rrod,ngs WIth a

Concqnlratoon rqadlngs.
hond hqld somplqr 01 0. 1 m
Pump somplqr o..3m aboyq bqd IOlqrvals above 1M b¢d
Hand hqld sompl~r 01 Olm

,nlqrvals abovq lhq bqd,

E
N
III
o
Cyl,nd<los placqd on ,nd'Ylduol slands WIth ~ach
' 035m 067m 035m Irap mouth o.·3m abov<> 1hq flu"", bod .
downflum. spacing of traps 1m

Fig. 8. Flume layout for experiments testing the effect of velocity on the collection efficiency of cylinders.
 SEDIMENT TRAPS 151

11.0 resuspension of material from the flume bed at these

m9.0n and standard dqytQlIon of all rt2'odlngs velocities.
120

;!.
~IOO Conclusions
~
i.. 80 Laboratory experiments and field investigations have
c:
Q suggested that cylinders and baffled funnels are the
~ 60 most efficient sediment trap designs for estimating
"0
u
the vertical flux in velocities up to 0.1 m/s. At higher
40+---~----'------r-----r----~
5 velocities (or R/ values) recent work suggests that
ASpoc t Rat,a
upwelling from a trap base may occur although the
Fig. 9( a). The collection efficiency of cylinders with different point at which particle resuspension and hence un-
aspect ratios at 0.1 mls (from White, 1990). dercollection occurs is still unclear. The point at
which resuspension occurs depends upon the trap
type, the trap aspect ratio, the ambient velocity, and
the sediment type. Recent laboratory experiments
140
suggest that cylinders with an aspect ratio of 2 may
- 120
be efficient collectors in velocities up to 0.2 m/s. As a
t precaution it is suggested that an aspect ratio of at
>-
u
~ 00 least 3 and preferably 5 is used in deployments in
v
-::. such environments. The use of cylinders in flows
.~ 60 above 0.2 mls is not recommended. Un baffled asym-
~ m<lan and standard d"",ation ot all r~adings metric funnels have been shown to seriously under-
.3 60
collect sediment at velocities above 0.12 m/s.
40+---.-----,-----,------,-----,-
5
Further work is needed to investigate resuspension
of particles from the trap base, and any biasing
Fig. 9(b). The collection efficiency of cylinders with different
effects that resuspension may have on the composi-
aspect ratios at 0.2 mls (from White, 1990). tion of the particles collected. If sediment traps are
to be used in high-energy environments such as the
continental slope, estuaries, and the nearshore zone,
results suggest that at velocities up to 0.2 mls cylin- the limitations outlined in this paper must be consid-
ders provide a reasonably accurate estimate of verti- ered when the results are interpreted.
cal flux.
The flow visualization experiments from previous Acknowledgement
literature suggest that for the range of R t values
tested (given on Figs. 9a, b) upwelling should have Figures in this paper are reproduced from the
occurred from virtually all traps, particularly at "Marina 89" Conference proceedings, published by
0.2 m/s. The fact that undercollection did not appear Computational Mechanics Publications, UK, with
significant in cylinders with aspect ratios greater than the kind permission of the publishers.
1 suggests that upwelling does not necessarily pro-
duce particle resuspension. References
Butman (op. cit.) also found that traps with aspect
ratios of2.7 collected efficiently at 1 x 104 R" whereas Baker, E. T., Milburn, H. B., and Tennant, D. A., 1988, Field
Assessment of Sediment Trap Efficiency under Varying Flow
Hawley's experiments suggest upwelling should be
Conditions, J. Mar Res. 46, 573-592.
complete at 5.1 x 103 R, for this aspect ratio. Clearly Bloesch, J. and Burns, N. M., 1980, A Critical Review of Sedi-
further work is needed to verify the limit at which mentation Trap Technique, Schweizerisch Zeitschrift fiir Hy-
drologie 42, 15-55.
resuspension occurs rather than upwelling of fluid.
Blomqvist, S. and Hakanson, L., 1981, A Review of Sediment
Unfortunately the results of White (1990) for Traps in Aquatic Environments, Arch Hydrobiol, 91, 101-132.
flows of 0.3 and 0.38 mls are inconclusive due to Blomqvist, S. and Kofoed, c., 1981, Sediment Trapping- A
152 J. WHITE

Subaquatic in situ Experiment, Limnology and Oceanography 26, see, Vierteljahresschrift Naturforschenden Gesellschaft in Zurich.
585-590. AS, 164-182.
Butman, C. A., 1986, Sediment Trap Biases in Turbulent Flows: Honjo, S., 1980, Material Fluxes and Modes of Sedimentation in
Results from a Laboratory Flume Study, J Mar Res. 44, the Mesopelagic and Bathypelagic Zones, J. Mar. Res. 38,
645-693. 53-97.
Davis, M. B., 1967, Pollen Decomposition in Lakes as Measured Jickells, T. D., Deuser, W. G., and Knap, A. H., 1984, The
by Sediment Tarps, Geol Soc Am Bull. 78, 849-858. Sedimentation Rates of Trace Elements in the Sargasso Sea
Dymond, J., Fischer, K., Clauson, M., Cobler, R., Gardner, W., Measured by Sediment Trap, Deep Sea Res. 31, 1169-
Richardson, M. J., Berger, W., Soutar, A., and Dunbar, R., 1178.
1981, A Sediment Trap Intercomparison Study in the Santa Johnson, M. G. and Brinkhurst, R. 0., 1971, Benthic Community
Barbara Basin, Earth and Planetary Sci Lett. 53, 409-418. Metabolism in Quinte Bay and Lake Ontario, J. Fisheries
Gardner, W. D., 1977, Fluxes, Dynamics and Chemistry of Partic- Resource Board of Canada 28, 1715-1725.
ulates in the Ocean, PhD Thesis MIT/WHOI Joint Program in Lau, Y. L., 1979, Laboratory Study of Cylindrical Sedimenta-
Oceanography, pp. 405. tion Traps, J. Fisheries Resource Board of Canada 36,1128-1291.
Gardner, W. D., 1980a, Sediment Trap Dynamics and Calibration: Parmenter, C. M., Bothner, M. H., and Butman, B., 1983, Charac-
A Laboratory Evaluation, J. Mar. Res. 38, 17-39. teristics of Resuspended Sediment from Georges Bank Collected
Gardner, W. D., 1980b, Field Assessment of Sediment Traps, J. with a Sediment Trap, Estuarine and Coastal Shelf Sci. 17,
Mar. Res. 38, 41-52. 521-533.
Gardner, W. D., 1985, The Effect of Tilt on Sediment Trap Pennington, W., 1974, Seston and Sediment Formation in Five Lake
Efficiency, Deep Sea Res. 32, 349-361. District Lakes, J. Ecology 62, 215-251.
Gardner, W. D., 1989, Baltimore Canyon as a Modern Conduit of Reynolds, C. S., Wiseman, S. W., and Gardner W. D., 1980, An
Sediment to the Deep Sea. Deep Sea Res. 36, 323-358. Annotated Bibliography of Aquatic Sediment Traps and Trap-
Gardner, W. D., Richardson, M. J., Hinga, K. R., and Biscaye, P. ping Methods, Freshwater Biological Assoc. Occasional Publica-
E., 1983, Resuspension Measured with Sediment Traps in a tion 11.
High Energy Environment, Earth and Planetary Sci Lett. 26, Reynolds, C. S. and Godfrey, B. M., 1983, Failure of a Sediment
262-278. Trapping Device, Limonology and Oceanography 28, 172-176.
Hargrave, B. T. and Burns, N. M., 1979, Assessment of Sediment Watanabe, Y. and Hayashi, H., 1971, Investigation on the Method
Trap Collection Efficiency, Limnology and Oceanography 24, for Measuring the Amount of Freshly Precipitating Matter in
1124-1136. Lakes, Japanese J. Limnology 32, 40-45.
Hawley, N., 1988, Flow in Cylindrical Sediment Traps, J. Great White J., 1990, The Use of Sediment Traps to Monitor Marina
Lakes Res. 14, 76-88. Siltation, PhD Thesis, Civil Engineering Department, Univ. of
Heim, A. 1900, Der Schlammabsatz am Grund des Vierwaldstatter Southampton (pending).
Pore Pressures in Marine Sediments: An Overview
of Measurement Techniques and Some Geological and
Engineering Applications
P. J. SCHULTHEISS·
Schultheiss Geotek, Fern Cottage, Marley Lane, Haslemere, Surrey GU273RF, UK

(Received 27 April, 1989; accepted I September, 1989)

Key words: in-situ measurements, pore pressure, marine sediments. neers and total stresses are controlled by equilibrium
measurement techniques, hydrology, sediment properties. effective conditions. Hence pore pressures are necessary to
stress, geotechnical. estimate effective stresses from calculated total stresses
and thus to allow the rational interpretation and/or
prediction of the response of soil masses. Practically,
Abstract. Pore pressures in the seabed are extremely sensitive to any
the pore pressure in the soil is often easier to measure
imposed stress because of the low permeabilities commonly exhib-
ited by marine sediments. Consequently, the measurement of than other equally meaningful aspects of soil behaviour
sediment pore pressures can be used to infer either the nature of because it exhibits no directional dependence. This is
the imposed stress (if the sediment properties are known) or the especially so in field situations involving anisotropic
physical properties of the sediment (if the imposed stresses are and non-uniform stressing (or straining) of the soil."
known). Stresses of many different types may be exerted on the
seabed either through hydrostatic forces (e.g. tidal and wave Not only is the pore pressure a fundamental soil
effects), or directly by lithospheric forces (e.g. tectonic and thermal
forces). Several techniques for measuring in situ pore pressures in
parameter that has to be known if a soil or sediment
the upper few metres of sediments have been developed, and one behaviour is to be understood, it is also often ex-
instrument, the PUPPI, will operate autonomously in water depths tremely sensitive to stresses imposed on the soil.
up to 6000 m. Basic sediment properties and processes can already
Laboratory testing of soils is often accompanied by
be inferred from pore pressure responses using this technique.
However, further application and development could greatly en- detailed pore pressure measurements in order to
hance its capability, especially for long-term monitoring of sedi- understand the soil's behaviour, however, in situ pore
ment conditions. In this Chapter, pore pressure measurement pressure measurements are far less common and are
techniques are briefly reviewed and problems are highlighted. An
outline is given of some of the many ways in which pore pressure even rarer in marine sediments. The high pore pres-
measurements could be used to gain further insight into geological sure sensitivity to imposed stresses in marine sedi-
processes and to determine some of the pertinent sediment proper- ments arises because they are commonly saturated
ties more accurately for engineering applications.
and have relatively low permeabilities. The measure-
ment of sediment pore pressures can be used to infer
Introduction either the nature of the imposed stress, if the sediment
properties are known, or the physical properties of
The importance of knowing the pore pressures within
the sediment, if the imposed stresses are known.
a mass of soil, in order to understand the soil's
In marine sediments the measurement of in situ
behaviour, has been recognized for many years.
pore pressures can be complicated by the relatively
Baligh (1986) succinctly summarizes this importance:
inaccessible environment and high ambient hydro-
"Porewater pressures occupy a central position in static pressures. However, once these difficulties are
modern soil mechanics for conceptual and practical overcome, the marine environment offers some dis-
reasons. Conceptually, effective stresses control most tinct advantages for pore pressure measurements over
soil behaviour aspects of interest to geotechnical engi-
the land environment. These include saturation of the
* Previously at: Institute of Oceanographic Sciences, Deacon Lab- sediments in most marine environments (this im-
oratory, Wormley, Godalming, Surrey GU8 5UB, UK proves the response time of the measurement system),

Marine Geophysical Researches 12: 153-168, 1990.

© 1990 Kluwer Academic Publishers.
 154 P. 1. SCHULTHEISS

important for very accurate measurements) and nat- pressure within the sediments (the pore pressure) is
urally occurring dynamic stresses (e.g. hydrostatic also linear and equivalent to hydrostatic pressure.
tidal cycles) that can be used in in situ experiments to However, the pore water pressure within the sedi-
determine sediment properties. Above all there is a ments, U, is not necessarily equal to hydrostatic (in
host of applications for accurate pore pressure mea- Fig. 1, U is shown as being greater than the hydro-
surements in marine sediments in all water depths static pressure). The differential pressure, AU (often
that could be used for strategic, commercial or called the "excess pore pressure" when it is greater
purely scientific ends. than the hydrostatic pressure) is given by U - Uh •
The stresses that largely control the strength and
Pore Pressures deformation behaviour of unconsolidated sediments
are the effective stresses. Terzaghi (1943) showed
STATIC AMBIENT FORE PRESSURES that the vertical effective stress, a~, is given by
In order to understand the critical importance of
a~ = a v - U.
pore pressure measurements to engineering and hy-
drogeological applications it is essential to appreciate AU and a~ are depicted in Fig. 1 which illustrates
the concept of effective stress as used in soil mechan- that the important effective stress is independent of
ics. Figure I is used to illustrate the changes in the water depth.
pressure and vertical stress within a water-sediment If pore water is flowing vertically within the sedi-
column. For simplicity, it is generally assumed here ment column, then the pore pressures are not hydro-
that the sediment beneath the overlying water static and are driven by the differential pore pressure,
column is fully saturated (no free gas) and that there AU. This differential pressure gradient can be either
are no density or permeability gradients within the positive (from an upward flow of pore water) or
sediment. It is also assumed that the total vertical negative (from a downward flow of pore water).
stress, a v , at any depth is caused simply by the Assuming that the water flows according to Darcy's
weight of the overlying sediment and water column. Law, then the seepage velocity, v (m S-I) of an
The hydrostatic water pressure, Uh , increases lin- element of water in the soil is given by;
early with depth in the water column (assuming
v = k AUjpgnz,
constant density), and under conditions of static
equilibrium (no movement of pore water) the water where k is the hydraulic conductivity of the sediment
(m S-I), AU is the differential pressure (Pa) at a
STRESS/PRESSURE
depth z (m), n is the fractional porosity, g is the
acceleration due to gravity (m S-2) and p is the
density of the fluid (kg m- 3).
The consequence of this water flow and differential
pore pressure gradient is to either increase the effec-
WATER tive stress (from negative differential pore pressures)
or to decrease the effective stress (from positive
differential, or excess, pore pressures). It is therefore
obvious that to determine the in situ state of the
vertical effective stress, both the total stress and the
total pore pressure need to be determined. The total
vertical stress can be calculated from the density
profile determined from a core sample (assuming
.. ' .. :. Uh - hydrosta tic prealur.
that the total stress is simply related to the overbur-
DEPTH
Differential pore pres.ur. L:l.U 2 U -Uh
den) but the only direct method of determining the
Vertical effective stress 0 v ' : Ov- U
pore pressure is by in situ measurement.
There is a natural upper limit to the magnitude of
Fig. I. Generalized profiles of pressure, total stress and effective excess pore pressure that can exist in an unlithified
stress in a water-sediment column. sediment and hence an upper limit to the seepage
 PORE PRESSURES IN MARINE SEDIMENTS 155

velocity of pore water. Above this limit the sediment period surface gravity waves, such as tides, can be
will liquefy and behave as a fluid. Cohesion less sedi- considered as standing waves, even in full oceanic
ments (silts and sands) will liquefy when the effective depths (Hurley, 1989). The pressure amplitude re-
stress is zero, i.e. when the pore pressure equals the sponse caused by surface standing waves in a water-
total stress. In cohesive sediments (muds and clays) sediment column is shown as a function of depth in
it is possible for small negative effective stresses to Fig. 2a. This pressure response in the sediments is
exist without the sediment liquefying. It should be governed by the elastic properties of the sediment. If
noted that there is no corresponding maximum for the sediment is fully saturated then the response is
downward advection velocity as negative pore pres- governed primarily by the sediment shear modulus,
sure Increase the effective stress and the sediment G, and the sediment permeability, k, as the other
stability. sediment parameters are essentially constant (Hur-
ley, 1989). In Fig. 2a the limiting cases are for a
SEDIMENT PORE PRESSURES
sediment with zero rigidity (G = 0, i.e. a suspension)
Dynamic pore pressure effects in sediments can occur where the pore pressure is a maximum and remains
as a result of any stress wave. In this chapter only equal to the hydrostatic pressure (Uh ) and for an
surface water waves are considered. They are of con- impermeable sediment (k = 0) where the pore pres-
siderable interest and fall into two categories: stand- sure is a minimum Umin • A real sediment with finite
ing waves and travelling waves. Only standing waves values of permeability and shear modulus will have a
and very long-period travelling waves such as tides pore pressure profile, U, which lies between these two
have any effect in the deep ocean because the dynamic limiting cases. Figure 2a is somewhat simplified as it
hydrostatic pressure does not attenuate with depth. In assumes constant values of G and k with depth. The
shallow water, travelling surface waves can have a differential pore pressure, AU, measured between
significant effect on the sediment pore pressure. some point in the sediment and the overlying water is
given by Uh - Umin •
STANDING WAVES

Hydrostatic pressures from surface standing waves TRAVELLING WAVES

do not attenuate with depth; consequently their The hydrostatic pressure caused by travelling surface
effects occur even in deep ocean sediments. Long waves attenuates exponentially with depth in a free

0) Standing Waves b) Travelling Waves

Normalized
o pressure amplitude o

.. _-_ _-------_..,. -------

.. -'---'~'---'---
Water
._------------_ ..-
Sediment

l>U Um;n U Uh
(K-O) (G~O)

Urnin
U

Fig. 2. Normalized pore pressure amplitude, U, in a water-sediment column caused by: a) standing waves and b) travelling waves. Uh and
Umin are the limiting cases when the shear modulus, G, and the permeability, k, are zero.
156 P. J. SCHULTHEISS

water column. Consequently ocean waves, with peri- vective water flow. At the same time the heat flow
ods of about 2-10 s, only influence sediments in from the crust decreases as it cools and more of the
relatively shallow water. Figure 2b depicts the pres- heat loss occurs by conduction rather than by con-
sure amplitude profile through the water-sediment vection. Heat flow surveys (e.g. Becker and Von
column for surface travelling waves. The pore pres- Herzen, 1983) have indicated that this free convec-
sure in this case attenuates from two processes: (a) tion may occur in cells that are in some cases evenly
the exponential hydrostatic pressure decay, and (b) spaced and in other cases controlled by large-scale
the decay caused by the elastic properties of the features such as faults. When the crust has cooled
sediments and its permeability. It should be noted sufficiently and the resistance to fluid flow has in-
that a differential pore pressure measurement made creased, through both the decrease in permeability of
between some depth in the sediment and at the the basement and the increasing thickness of sedi-
sediment/water interface is not in this case I1U ments, the heat loss occurs only by conduction and
(Uh - U). Therefore, the pressure attenuation caused major fluid circulation ceases. This is probably the
by the sediment alone cannot be measured directly general case for most ocean basins.
with a single differential pressure transducer. 3. Hydrologic flow occurs at continental margins
where continental rainfall raises the groundwater
level above that of the surrounding seawater. Conse-
Geological Applications
quently, it can be expelled through aquifers directly
Hydrogeological processes in the ocean crust are into the ocean. This pressure-driven circulation can
major factors influencing the exchange of heat and be enhanced by density-driven circulation where the
elements between the crustal rocks and the oceans. groundwater passes through evaporites which raise
Water circulating through the crust will influence not the density to levels above that in seawater. Passive
only the processes of plate tectonics but will also margins can also accumulate sediments of sufficient
have a major effect on global geochemical budgets. A thickness to heat organic constituents and produce
recent paper (Working Group 3, 1987) produced at oil and gas. Fluid flow will occur upwards as a result
the Second Conference on Scientific Ocean Drilling of positive bouyancy of these products and by com-
(COSOD II) clearly summarizes the importance of paction of the thick layers of sediment.
hydrodynamic fluid circulation within the crust for 4. Tectonically induced physical compaction of
the further understanding of the mechanisms of sediments occurs at active ocean margins, causing
crustal tectonics and the interrelationships, through water to be expelled in the process of subduction and
the global geochemical budget, between the litho- accretion. Although the patterns of pore water expul-
sphere, the hydrosphere, the atmosphere and the sion within accretionary complexes are not known in
bio-sphere. Four primary areas of fluid circulation in any detail, it is thought that dewatering at the toe of
the ocean crust were identified: such complexes is pronounced (Westbrook and
1. At the ridge axis spreading centre, where forced Smith, 1983). It is currently assumed that water is
convection occurs at the ridge axis caused by large expelled through a variety of features such as thrust
temperature gradients in the surrounding crust. It is faults and mud volcanoes as well as in a pervasive
this vigorous forced convection that causes the well manner through the bulk of the sediments.
known 350°C black smokers. At sediment-free ridge These four hydrodynamic flow regimes are shown
axes the low permeability sediment layer which in- schematically in Fig. 3 (after Working Group 3,
hibits water circulation at sedimented ridge axes is 1987), which also shows a very rough estimate of the
absent. As a result the residence time available for possible total water discharge for each zone. It is
fluids to chemically react with the sediments and interesting to note that while the ridge axes have the
rocks is greater at sediment-covered ridge axes. most vigorous circulation, the biggest contribution
2. Free convective circulation is thought to occur to the total discharge is from the ridge flank zone.
on the ridge flanks, possibly between a few kilome- Here the free convection is much slower but is
tres and as much as 1000 km away from the ridge cumulatively greater because of the much larger area
axis. Sedimentary cover generally increases away over which it occurs.
from the ridge axis, increasing the resistance to con- In areas where pore water advection is taking
 PORE PRESSURES IN MARINE SEDIMENTS 157

600
E"':'
05.. 500
............
lL...,
..,E
. - ,:,< 400
-=.....-
lL. ....
.., .,
(II ;J
300
~U 200
o c:
.s::: 0
~ (II 100
.- u
00
0

Ridges Flanks Bosin Morgin

<i?
-------j ~

<::>
Forced
Convection
Free
Convection
I No Flow I Compoctionol
Expul sion
Hydrologic
Flow
'-

Fig. 3. Hydrodynamic zones and fluid flow processes, with estimates of total global discharge rates, on and around an oceanic plate (after
Working Group 3, COSOD II).

place by a slow pervasive flow through the whole of transformation. It is not inconceivable that pore
the sediment body, in situ pore pressure and perme- pressures other than hydrostatic do exist in ocean
ability measurements in the surficial sediments will basin sediments and that the driving force may not
enable the flow patterns to be defined and flow rates be thermally induced but may be caused by other
quantified. In some areas it is likely that the flow is processes.
much more vigorous, causing significant amounts of It was noted by Richards (1984) that pore pres-
water to be either entering or undergoing expulsion sures in excess of hydrostatic have always been mea-
from the sediments and ocean crust through faults, sured in situ in cohesive soils, having a grain-size
mud volcanoes or other small-scale features within smaller than most sands at depths below a metre or
the sediments. Accurate pore pressure measurements two into the seabed. This generalization was based
in the upper 10 m of sediments in these types of areas on only a small number of in situ piezometer mea-
could still enable patterns of water movement to be surements, primarily in soils on the continental shelf.
identified within these hydrogeologically active areas. Subsequent pore pressure measurements in the
However, they would only be able to quantify the Madeira Abyssal Plain (Schultheiss and McPhail,
pervasive component of the overall flow pattern and 1986; Schultheiss and Noel, 1987) have shown that
other measurements or assumptions would still be pore pressures in this region are generally hydro-
needed to estimate the total flux of water through a static. Nevertheless, there is some evidence to indi-
given area. cate that excess pore pressures may exist (at least to
some depth) in many marine sediments as a result of
the sedimentation and compaction processes them-
Ocean Basins and Abyssal Plains
selves.
The ocean basins and abyssal plains are defined It is now generally accepted that many cohesive
hydrogeologically by Working Group 3 (1987) as surficial sediments in the upper 4 m exhibit an appar-
regions of diffusive flux where pore fluid convection ent overconsolidation at high porosities. Apparent
in the oceanic crust has ceased. However, the transi- overconsolidation is not caused by the removal
tion between ridge flank and ocean basin is a gradual or erosion of any overlying material and has been
158 P. J. SCHULTHEISS

attributed to the effects of high interparticle bonding, Madeira Abyssal Plain, where pore pressures have
ageing and cementation, which result in an intrinsic been found to be generally hydrostatic, the sediments
strength sufficient to prevent normal consolidation have OCRs of just over unity (Silva and Mairs,
and resembling an overconsolidated state in the labo- 1987), indicating a normally consolidated or appar-
ratory (e.g. Hamilton, 1964; Richards and Hamilton, ently slightly overconsolidated state. However, there
1967; Noorany and Gizienski, 1970, Silva and is no evidence of underconsolidation which is in
Jordan, 1984). agreement with the pore pressure data.
The technique normally used to evaluate the con-
solidation state of soil samples is the one dimen- Engineering Applications
sional consolidation test on high-quality samples.
TOXIC WASTE DISPOSAL
From the e -log (j ~ (void ratio-log vertical effective
stress) curves, the maximum past preconsolidation Oceanic disposal of all kinds of waste products has
stress ((j~) can be estimated and compared with the been, and continues to be, considered and used.
calculated in situ vertical effective stress assuming Ocean sediments have been considered as the pri-
hydrostatic pore pressures (obtained by integrating mary barrier to the release of high-level radioactive
the sediment density to the required depth). The waste (HLRW) into the ocean waters as part of an
overconsolidation ratio (OCR) is given by (j~/(j~ and internationally co-ordinated programme investiga-
is an indication of the stress history/consolidation ting the feasibility of sub-seabed disposal (see Fran-
state of the sediment (Silva and Mairs, 1987). OCRs cis, 1984 for references). For safe burial of wastes
around unity indicate a normally consolidated sedi- into deep-sea sediments it is necessary to obtain a
ment, whereas values significantly greater or less detailed understanding of the physical and geotechni-
than unity indicate over or under consolidation cal properties of the host sediments for the emplace-
respectively. ment phase of the waste. This can only be fully
Richards (1984) discusses the overconsolidation achieved with in situ measurements, among which in
ratio profile found in many marine sediments which situ pore pressure measurements are an important
often decreases from a value greater than unity in the part as they can provide strength, consolidation and
upper few metres to unity, or significantly below permeability data. In addition to the short-term site
unity, as the depth increases. This underconsolida- assessment requirements, the long isolation periods
tion at depth in sediments which have apparently required for HLRW (tens to hundreds of thousands
been deposited continuously or episodically without of years) requires that a detailed knowledge of any
any erosion over long periods of time, appears to be naturally occurring pore water flow patterns is ob-
real and can be explained by the existence of excess tained. The most direct method of determining pore
pore pressure within the sediments. A recent example water flux rates is by measuring any naturally occur-
of a detailed study investigating the stress history of ring pore pressure gradients and the sediment perme-
sediments in the Nares Abyssal Plain is that of Silva abilities. It has been estimated that flow rates as low
and Mairs (1987), who showed that the OCRs were as 1 mm/year need to be resolved for safe disposal of
less than unity between 5 m and 25 m below the HLRW. Pore pressure measurements in fine-grained
seabed with minimum values of around 0.4 at about sediments are capable of an accuracy in that order in
15 m below the seabed. It would seem from this fine grained sediments where the permeability is very
study that if the hypothesis of excess pore pressures low (Schultheiss and McPhail, 1986).
is used to explain the data, then using a simple
SITE INVESTIGATION AND SLOPE STABILITY
one-dimensional model of fluid flow requires that the
excess pore pressures must exist not only in the As seabed exploration and exploitation moves into
deeper sediments, where the OCRs are less than deeper waters, foundation problems associated with
unity, but also in the upper few metres, where the slopes and different sediment types may be encoun-
sediments may exhibit an apparently overconsoli- tered. Slope failure may become a major hazard.
dated state. It is clear that to test this hypothesis in Some site investigation techniques may change, with
situ pore pressure measurements are needed in the more emphasis placed on in situ measurements. Pore
same area. It is interesting to note that in the pressure measurements will undoubtedly figure
 PORE PRESSURES IN MARINE SEDIMENTS 159

prominently because of their fundamental impor- One of the primary problems associated with an
tance to soil behaviour. analysis of the slope stability of sedimented areas of
Submarine slope instability probably occurs over the sea floor is an evaluation of the in situ effective
the entire spectrum of offshore environments from stress. To estimate this, it is essential to measure in
shallow, near shore zones to the continental slopes situ pore pressures (a component of the effective
and beyond to the deep ocean floors (Prior and stress) within the sediments to the depth of interest,
Coleman, 1984). Slope failures have been docu- as well as the density profile (obtained from sam-
mented in offshore petroleum areas worldwide (nota- pling).
bly in the Mississippi Delta), and must be reckoned It can be assumed that some types of sediments
with in any seabed construction or extraction opera- may be relatively unstable as a result of excess pore
tion, particularly in areas with moderate-to-high sed- pressures. In fact all slope failures are probably
imentation rates, and where significant seabed slopes initiated as a result of an increase in pore pressure.
or gassy sediments prevail. Ambient excess pore pressures can be caused by, for
While it is known that sediment failures occur on example, fast sedimentation rates, gas production or
even very gentle submarine slopes (less than 1°) the by naturally occurring advection caused by thermal
precise cause of the instability, or the cause and gradients. For a failure to occur, the pore-pressure
timing of the failure, is rarely known. In addition, has to exceed a critical level. This critical level may
there are currently no proven methods for evaluating be exceeded through either these slow changes in the
the susceptibility of a specific area to likely failure or ambient conditions or rapid transient causes such as
for monitoring changes in seabed conditions that severe storm conditions or seismic activity. The
may lead to failure. identification of areas at risk is of considerable

. - : :;::: : .. -.....

Fig. 4. Impression of a real-time monitoring system of pore pressures in the seabed. Data are transmitted acoustically from the sea floor
to a data buoy and then via satellite to a monitoring station.
160 P. J. SCHULTHEISS

importance to any seabed engineering project. This is with one successful test being made at a water depth
especially true as seabed exploration and hydrocar- of 278 m. The probe had an overall length of 4.9 m
bon extraction extends into deeper waters and into and weighed 570 kg in air. It used a differential
areas that have significant slopes. pressure transducer constructed using a boudon tube
Future investigations into slope stability may re- and vibrating wire with an accuracy of + / - 6.2 kPa.
quire pore pressures to be monitored over long peri- A maximum excess pore pressure of 59 kPa was
ods of time, to look for example at gas build-up in measured 3.2 m below the sea floor, caused by the
the sediments, changing pore pressure from rapid insertion of the probe, which decayed to 9.8 kPa
deposition of sediments, seasonal effects and/or the after 5 hours and did not decay significantly further
effects of storms. Indeed in some areas it is conceiv- after a total of 10 hours.
able that real-time monitoring of pore pressures may Submarine piezometer technology and data aquisi-
be required purely for safety purposes. A schematic tion were most significantly advanced as a result of
diagram of how this can be achieved, even in very the SEASWAB experiments. SEASWAB (Shallow
deep or very remote areas is shown in Fig. 4. Pore Experiment to Assess Storm Waves Affecting the
pressure instruments could be strategically located in Bottom) was one element of the US Geological
and around the area of interest or potential danger. Survey's broader study of sea floor instability known
The data would be transmitted to shore, rig or ship as the "Delta Project" and both of the SEASWAB
anywhere in the world via an underwater acoustic experiments took place in an area known as "East
data-link to a surface bouy and then to a satellite Bay" in the Mississippi Delta. SEASWAB I (Gar-
system. The technology for this type of system al- rison, 1977) was conducted in a water depth of 19 m,
ready exists, and it would not be difficult to imple- whereas SEASWAB II (Hottman et aI., 1978) took
ment such a system if the need was perceived. place 1830 m to the northeast of the SEASWAB I
location in a water depth of only 13 m.
The piezometers used during the SEA SWAB ex-
Review of Pore Pressure Measurement Technology
periments were specifically designed for shallow-
The potential value of having in situ pore pressure water applications. They took advantage not only of
data from submarine sediments for engineering pur- the shallow water but also of the nearby oil produc-
poses has been recognized for many years. More tion platforms to directly record the pressure data
recently, an awareness has developed of the impor- via undersea cables connected to the instruments on
tance of the role of fluid flow in many geological the bottom.
processes. Despite these facts very few instruments
have been built to make these measurements and SEASWAB I
there is generally a lack of reliable data. Some of the The two piezometers used during the SEASWAB I
instruments that have been successfully built and experiments were on the sea floor for six months in
used will be briefly discussed here. This review is 1975/76, and consisted of a large instrument built
restricted to those devices that remain static in the and operated by the National Oceanic and Atmo-
seabed. It excludes, therefore, a number of devices spheric Administration (NOAA) and a smaller in-
that measure pore pressures as part of another test strument, similar to the NGI-UI piezometer,
such as the piezocone test. It also excludes some operated by Lehigh University. NOAA's piezometer
piezometers that have been buried at very shallow was 17.12 m long and constructed from 120 mm
depths to measure the effects of surface waves only. diameter pipe. It used variable reluctance, absolute
The first known reported uses of a submarine pressure transducers to measure both the pore pres-
piezometer were by Lai et al. (1968) and Richards et sure at two locations (8 m and 15 m below the
al. (1975). These authors describe a differential mudline) and the hydrostatic pressure at two loca-
piezometer, built at the Norwegian Geotechnical In- tions (1 m and 15 m below the mudline). The Lehigh
stitute (NGI) and the University of Illinois (UI) that piezometer (Hirst and Richards, 1977) was about
was designed to operate in water depths of up to 7 m long and constructed from 54 mm diameter
500 m. This NGI - UI probe was deployed several pipe. It used interconnected, fluid filled, bourdon
times in 1967 in the Wilkinson Basin, Gulf of Maine, tubes connected to vibrating wire strain gauges as a
 PORE PRESSURES IN MARINE SEDIMENTS 161

differential pressure transducer to measure the pore inhibits, by capillary effects, direct communication
pressure at a depth of 6.4 m below the mudline. The with free pore-gas (Bennett and Faris, 1979). Conse-
piezometers were installed about 6 m apart by at- quently, the pressures measured through the corun-
taching weights and then lowering to the bottom by dum stones may contain a contribution from high
a crane. After installation the weight sections were pore-gas pressure, whereas those pressures measured
removed with the aid of divers. through HAE stones should not (Vaughn, 1973).
The details of the pore pressure data from the The Lehigh piezometer was similar to that used
SEASWAB I experiment are not easy to explain previously except that it used a high air entry stone
fully. Hurricane Eloise passed over the site two days and a semiconductor-type pressure transducer. The
after insertion. This was probably prior to the com- NOAA piezometer was also similar to the one used
plete dissipation of the insertion pressures and un- previously except that it used a long slender tip to
doubtedly complicated the interpretation of the data. study pore water dissipation, and a different combi-
However, based on data extrapolation it was esti- nation of transducers was used. Coarse stone corun-
mated that ambient excess pore pressures were dum pressure ports were used at depths of 6.5 m,
11.7 kPa and 44.1 kPa at depth of 8.4 m and 15.2 m 12.6 m and 15.6 m below the mudline and an HAE
respectively below the mudline from the NOAA pressure port was used at 15.6 m below the mudline.
piezometer (Bennett, 1977; Bennett et a/., 1976 and The transducers were all absolute sensors apart fom
1982). The data from the Lehigh piezometer indi- the one at 15.6 m with the corundum stone, which
cated an ambient excess pore pressure of 32 kPa at used a differential pressure transducer.
6.4 m below the mudline. In addition to the Lehigh and NOAA piezometers,
One of the primary problems encountered from a third piezometer was installed during the
the piezometer measurements made during SEASWAB II experiment. This piezometer (T AMU-
SEASWAB I was identifying the effect of free gas on USGS) was similar in construction to the NOAA
the pressure measurements in the sediments. Did the instrument except that it used thin-film strain gauge
ambient pressure measurements determine the pore transducers to measure pore gas pressure through
water pressure, the pore gas pressure or a combina- corundum pressure ports at depths of about 3 m and
tion of both? Did the contribution of water and gas 12.5 m below the mudline and to measure pore water
pressure change with time? How much gas had mi- pressure using an HAE port at about 12.5 m below
grated into the pore pressure measuring system? the mudline.
What was the effect of this gas on the dynamic Overall, the pressure data from the SEASWAB II
pressures recorded? Did it simply increase the re- experiment is confusing for several reasons. Both the
sponse times of the pressure measuring system to Lehigh and the TAMU-USGS piezometers were lo-
short-period pressure transients (wave effects) or cated in a deformational feature and needed driving
were there other effects? All the porous pressure into the sediments which may have caused some
ports used during SEASWAB I were relatively coarse damage to the instuments and hampered the determi-
porous corundum stones (porosity = 40-50%, per- nation of initial insertion pressures. The deforma-
meability = 1-3 * 10- 2 cm/s) (Bennett and Faris, tional feature is also believed to have moved
1979). significantly during a storm in late February 1977.
Both of these piezometers sank into the mud during
SEASWAB II the course of the experiment; the TAMU-USGS
SEASWAB II had basically the same objectives as instrument sank approximately 1.6 m and the Lehigh
SEASWAB I but an effort was made to increase the instrument sank approximately 1.8 m. The Lehigh
number of sensors and to measure separately both piezometer malfunctioned for most of the time be-
the pore water pressure and the pore gas pressure cause of a cable problem, and on recovery it was
(Dunlap et a/., 1978). In an attempt to measure pore found that the HAE stone had cracked and sediment
water pressure, fine-pore, high air entry (HAE) was filling the cavity leading to the pressure trans-
stones were used (porosity = 35-38%, permeabil- ducer. The NOAA piezometer was installed outside
ity = 1-3 * 10- 7 cm/s). It is generally accepted that the deformational feature and was free of these
the use of small pore-size filters as pressure ports problems. On recovery it was found that there was
162 P. J. SCHULTHEISS

hardly any fluid left in the cavities behind the porous from the piezometers described previously in that it
stones irrespective of whether they were corundum is completely self-contained, stores the pressure data
or high air entry (Dunlap et al., 1978). The NOAA in an onboard solid-state memory and can telemeter
instrument provided the best data, showing high these data acoustically to a surface vessel. GISP was
excess pressures increasing with depth for all the designed to operate in water depths of up to 450 m.
corundum ports (11 kPa at 6.5 m to 43 kPa at The Data Gathering System (DAGS) on the sea
15.6 m). However, the data obtained from the trans- floor and the Command and Recording Subsystem
ducer connected to the HAE port at 15.6 m were on the surface (CARS) were developed for both the
erratic. A few days after insertion the excess pressure GISP and a Seafloor Earthquake Measuring System
increased dramatically to a level to that recorded by (SEMS) (Reece et al., 1979). The 10.5 m GISP probe
the transducer with the corundum port at the same was constructed from 108 mm diameter O.D. steel
level. This was interpreted by Dunlap et al. (1978) as pipe containing absolute pressure transducers
being caused by gas slowly diffusing through the (quartz crystal, oscillating beam type) at 3.32 m,
HAE stone. However, after 40-50 days this pressure 6.42 m and 9.52 m below the mudline, with coarse
dissipated back to a much lower level and Bennett corundum stones as pressure ports. A reference hy-
and Faris (1979) concluded that "further research is drostatic transducer was mounted 0.36 m above the
necessary to understand the significance of the large mud plate. The GISP instrument, first tested in 1980,
differences observed between HAE and corundum revealed several design problems. The modified GISP
stones and the anomolous character of the HAE was tested in 1981 at the SEASWAB I site in a water
pore pressure record". The same data presented by depth of 19 m. Two identical probes were installed,
Bennett et al. (1982) considered the HAE excess 120 m apart, by lowering them into the sediment
pressure data, which rose and fell during a period of under their own weight (1725 kg) at a rate of
about 65 days, to be anomolous. The NOAA instru- 0.15 mls in August 1981. The ballast weight
ment also provided data on the pressure response (1180 kg) was lifted from the GISP instrument after
from surface activity. Pressures caused at depth installation by triggering a release motor. The two
within the sediment as a result of both tidal and instruments were recovered in January 1982. Recov-
wave activity were recorded (Bennett and Faris, ery was accomplished by acoustically releasing a
1979). bouyant reel of kevlar cable which was picked up
The NOAA piezometer was used again in 1978 in and used to haul the whole instrument out of the
a water depth of 44 m in Block 73 of the Mississippi mud and back to the surface. Peak insertion pres-
in an area known as "Main Pass" (Bennett et al., sures of up to 63.4 kPa were recorded at 9.5 m.
1982, 1985). Pressure transducers were again located Despite some difficulties in interpreting the data it
at depths of 6.5 m, 12.6 m and 15.6 m below the was concluded that ambient excess pore pressures
mudline with a combination of both corundum and increased with depth up to 17.9 kPa at a depth of
HAE ports. Excess pore pressures measured at this 9.52 m below the mudline. This is in reasonable
site were considerably less than in the SEASWAB agreement with some of the data from NOAA
experiments in the East Bay area (14.5-17.9 kPa at piezometer used during the SEASWAB I experi-
15.6 m and 15.8 kPa at 12.6 m). However, it is not ments.
clear whether there was less free gas in the sediments
in this area or not. The insertion pressures were OXFORD UNIVERSITY

higher, which is believed to have been caused by the A differential piezometer has been built and used by
higher shear strengths of the sediments in the Main Sills at Oxford University (OU) (Sills, pers. comm.
Pass area. and Sills and Nageswaran, 1984). The OU piezome-
ter is constructed of stainless-steel pipe with a diame-
GISP ter of 44 mm, increasing at the top to 76 mm to
Another piezometer was built and used in shallow house some instrumentationa. Ports are located at
water in the Mississippi Delta by Sandia National approxmately 2.3 m and 3.5 m below the mudline.
Laboratories (Prindle and Lopez, 1983). GISP (Geo- The differential pressure transducers use a common,
technically Instrumented Seafloor Probe) is different air-filled, back pressure reference system and have a
 PORE PRESSURES IN MARINE SEDIMENTS 163

high resolution (approx. 30 pa). Two absolute pres- port of the US sub-seabed disposal programme
sure transducers measure the common reference examining the feasibility of emplacing high-level nu-
pressure and the seabed hydrostatic pressure. In situ clear waste in the deep seabed. The titanium probe
pressure data have been obtained from three areas in tip is only 8 mm in diameter, has a sharp cone (angle
shallow water around the UK, between 1978 and approximately 5°), and contains a porous stone as a
1981, investigating primarily the effects of gas on the pressure port that leads via a pipe to a variable
measured pressure response caused by tidal cycles. reluctance differential pressure transducer (range =
Tidal cycles with amplitudes of up to 14 m were + / - 69.8 kPa, precision less than 340 Pa). The
monitored for periods of up to 60 hr. The data piezometer was initially tested using the submersible
showed differential pore pressure amplitudes that ALVIN in 1450 m of water off the US Atlantic
ranged from 0% to about 12% of the tidal cycle at continental slope (Lambert, 1982). However, the
depths of up to 3.5 m below the mud line, indicating pore pressure sensor (a differential semi-conductor
the presence of free gas in the sediments with an type was used at that time) failed completely soon
estimated degree of saturation between 97% and after insertion. The probes were subsequently used
98%. The OU piezometer is currently being used to during a laboratory experiment designed to simulate
investigate other sites around the UK. the environmental conditions for the ISHTE experi-
ment. The probes were inserted into a tank (approx.
BGI I m3) filled with a fine-grained sediment at a hydro-
A multiple piezometer was built by the Belgium static pressure of 55 MPa and at 4°C in November
Geotechnical Institute, BGI (Carpentier and Ver- 1981, to depths of 16.9 cm and 26.4 em below the
donk, 1986; De Wolf et al., 1983) to measure pore mudline. Maximum insertion pressures of 6.6 kPa
water pressures caused by waves and tides in the and 12.9 kPa were recorded before the pressure
foundation soils of a new outer harbour at Zee- smoothly decayed to near zero values in times of
brugge. The diameter of the piezometer probe varied 20 min. and 61.5 min. respectively.
from 60 mm at the tip to 193 mm at the top and Abbott et al. (1984) have measured pore pressures
contained absolute, strain gauge, pressure transdu- in deep-water sediments using differential pressure
cers, with a 0.5 MPa range, located at 7 levels along transducers mounted on a heat flow probe in the
its length with the lowest sensor positioned about Gautemala Basin. They postulated a maximum
13 m below the mudline. Very coarse filters made downward pore water advection velocity of between
from perforated nylon discs were used as pressure 30 and 300 mm/year in those areas with thinest
ports on the probe. The piezometer was jacked into sediment cover. However, their residual pore pres-
the sandy and clayey seabed in a mean water depth sure values were based on extrapolating the decay
of about 8 m with a tidal range of about 5 m in curve of the transient pressure pulse caused by inser-
Oct/Nov 1981, and was removed in June 1983. Data tion, using an asymptotic computer-generated model
were collected at different times throughout the de- with less than 10 min. of data. To eliminate the
ployment and were checked at least once per week. uncertainties of theoretical extrapolation, much
Dynamic pore pressures from both waves and tides longer-term deployments are needed in order that the
were measured at depths up to 13 m below the insertion pressure decays more completely before the
seabed, although it was considered difficult to draw residual pore pressure is measured.
any firm conclusions from the data (De Wolf et al.,
1983). PUPPI
An autonomous Pop-Up-Pore-Pressure-Instrument
NORDA DEEP OCEAN PIEZOMETER (PUPPI) has been designed and constructed to accu-
A piezometer probe capable of working in the deep rately measure in situ pore pressure gradients in
ocean (up to 6000 m) has been developed at the deep-sea sediments (Schultheiss et a/., 1985; McPhail
Naval Ocean Research and Development Activity and Schultheiss, 1986). The PUPPI measures differ-
(NORDA) (Bennett et al., 1985). The development ential pore pressures at two ports with a resolution
was an integral part of the In Situ Heat Transfer of + / - 15 Pa on a lance up to 6 m long. The PUPPI
Experiment (lSHTE) which was conducted in sup- is a free-fall instrument that can be ballasted to
164 P. J. SCHULTHEISS

- - - ----
: .
. .... ~ .
. '.
, .
. ..
'

: .', , . ... , .. '"

,
..
- . I

' \ : ' ..

. ' -. , ..
• I"

Fig. 5. The Pop-Up-Pore-Pressure-Instrument (PUPPI) embedded in the sediments, after free-faIling through the water column, and after
the upper buoyant section has been acoustically released.

penetrate a range of sediment types in water depths lected from the Barbados Accretionary Complex in-
of up to 6000 m. Pore pressures are measured rela- dicating significant pervasive flow of pore water.
tive to hydrostatic at the ports on the lance by using A generalized diagram of the pore pressure data
differential pressure transducers connected between obtained from the PUPPI is shown in Fig. 6. It
the ports and open sea water. A programmable illustrates the pertinent features of the pore pressure
solid-state logger is used to record the pressure data record that can be used to infer some in situ proper-
at a maximum rate of 0.5 samples/sec. A vertically ties of the surrounding sediment. These features are
mounted accelerometer enables the penetration event discussed below.
to be examined and allows the depth of penetration (a) The maximum insertion pressure, Umax , gener-
to be calculated. Recovery is accomplished using an ated at the probe surface has been modelled by
acoustic command which activates a release mecha-
nism that simultaneously cuts the pipe connecting
the pressure port to the transducer. The lance, ballast ""MAXI~U~ INSERll0N PRESSURE
w
weight and cone assembly are left on the sea floor a::
::::J
(I)

while the buoyant instrument package ascends to the (I)

w
a::
surface for recovery (Fig. 5). a.
w CURVE
a:: ~ECAY
To date, 40 deployments have been made using o
a. AnENUAlEO 110AL CYCLES
PUPPI systems in water depths between 3500 m and --'
« /
F
5500 m. Most of the deployments have been in the zw
a::
Madeira Abyssal Plain (Schultheiss and Noel, 1987), ~
u.
where differential pore pressures were found to be o
generally absent. However in the Mariana back arc TIME

basin, significant pore pressure gradients were mea-

Fig. 6. Generalized pore pressure record from the PUPPI showing
sured, indicating local upward flow of pore water the four main pressure features that are used to calculate some of
(Leinen et al., in prep.). Data have also been col- the in situ properties of sediments.
 PORE PRESSURES IN MARINE SEDIMENTS 165

Randolf et al. (1979). Esrig et at. (1977) suggested measurement devices for use in. terrestrial soils by
that some simple generalizations could be made Hanna (1985). One of the major problems of calibra-
for different soil types to predict the in situ tion has been overcome on the PUPPI by remotely
undrained shear strength, Cu , from Uimax ' For lean cutting the pipe leading to the pressure port, immedi-
inorganic soils of moderate-to-high sensitivity ately before recovery. In this way both sides of the
(typical of many deep sea clays) this resulted in the differential transducer are open to the sea water
simple expression: pressure, and consequently a very accurate in situ
zero calibration is obtained.
For dynamic measurements, the time-lag associ-
(b) The dissipation of pore pressure (the decay ated with the pressure measurement system is of
curve) can be modelled as radial consolidation critical importance. A perfect pressure transducer
around a cylindrical probe. Bennett et al. (1985) would instantaneously and accurately measure a
uses the solutions of Soderberg (1962) and Ran- change in pore pressure within the sediments no
dolf et al. (1979) to predict the coefficient of matter how sudden. In practice, a finite flow of water
horizontal consolidation, Ch , from the time taken must move into or out of the transducer before
for 50% of Uimax to dissipate (tso) using the ex- equilibrium is reached and the transducer registers
pression: the true change in pressure. Hence there will always
be a time-lag between a pore pressure change in the
Ch = r2Trso/t so ,
sediments and the transducer reaching its equi-
where r is the radius of the probe and Trso is a librium state. The time-lag of a piezometer system in
dimensionless time factor which is approximately a sediment depends on several factors, including the
equal to one for this case. Having obtained Ch , the permeability of the surrounding sediments.
permeability, k, can be determined using appropri- Hvorslev (1951) has given a solution to the time-
ate values of the constrained modulus, D, and the lag for any shape of piezometer in an incompressible
unit weight of water Yw from the expression: soil. However, the assumption of an incompressible
k = ChYw/D. soil can lead to appreciable errors as recognized by
Penman (1960). A solution for the time-lag of spher-
(c) The attenuation of dynamic hydrostatic pres- ical tip piezometers in compressible soils has been
sure in the sediments caused by tidal cycles can be derived by Gibson (1963), using the coefficient of
modelled as standing waves, as discussed previ- consolidation of the soil.
ously. This modelling enables in situ values of In practice, the most important factor to consider,
permeability and shear modulus to be predicted. in order to minimize the time-lag of a piezometer, is
(d) The residual differential pore pressure that to ensure that it is fully saturated. Small amounts of
exists after the complete decay of the insertion air, or gas, in any part of the system (pressure port,
pressure is the ambient in situ pressure. This affects pipes or sensor body) can significantly increase the
the in situ state of effective stress and will cause overall volume compliance, and consequently the
water to flow in the permeable sediments. Conse- amount of water flow into and out of the system.
quently it can be used to determine the natural With differential pressure transducers the effects can
pervasive flux rates of pore waters. be even more dramatic. If the overall compliance
(diaphragm, pipework and trapped air) is not the
same on both sides of the sensing element, then the
Pore Pressure Measurements-Practical Problems
pressure readings will be seriously in error both in
There are many potential problems associated with amplitude and phase.
the accurate measurement of pore pressures in sedi-
MEASUREMENTS IN GASSY SEDIMENTS
ments such as time-lag, resolution, accuracy and
calibration all of which can change with time, tem- It is obvious from the above comments that for
perature and pressure. A discussion of most of these accurate dynamic measurements of pore pressure in
problems is beyond the scope of this chapter, but the sediments the pressure measurement system must be
reader is referred to a useful review of pore pressure saturated. When working in deep water achieving
166 P. J. SCHULTHEISS

saturation is not difficult, as any small amounts of Summary and Conclusions

gas left in the system will be dramatically reduced
Pore pressures are not only a fundamentally impor-
in volume and will finally be forced into solution as
tant parameter for understanding the behaviour of
the instrument descends in the water. With good
soils, they are also extremely sensitive to imposed
design and care, full saturation of a measurement
stress, especially in soils with low permeabilities. In
system can be achieved even in shallow water. Prob-
situ measurements of both static and dynamic pore
lems really begin to occur when the sediments them-
pressures can provide crucial information concerning
selves are not saturated.
the rate at which pore fluids move, and important
The pore pressure measurements made during
data on the in situ properties of sediments. This type
the SEASWAB experiments clearly demonstrate
of information has significant applications in both
the problems of intepreting pressure measure-
geological and engineering studies of the sea floor.
ments in gassy sediments. As previously discus-
Many of the perceived problems in the field of
sed, even the use of HAE pressure ports did not
marine hydrogeology can, at least partially, be ad-
prevent gas diffusing through into the pipework and
dressed from measurements of pore pressure at rela-
sensor.
tively shallow depths within the sediments. These
In fine-grained sediments that contain free gas it
range from investigating major fluid flow processes
can reasonably be assumed that the pore water is within the ocean crust to studying the sedimentation
saturated with gas, that the gas bubbles are proba- and consolidation phenomena of sediments in the
bly much larger than the sediment pore sizes and relatively static oceanic basins. For engineering
that these bubbles are relatively stable within the projects that require detailed site characterization, in
sediment framework. Under these conditions the situ effective stresses must be calculated; hence, pore
free gas cannot easily migrate from the sediment pressures must be known. They must also be known
through the porous pressure port and into the if the dynamic effects of storms and tides on slope
pipework and sensor directly. However, the gas stability are to be assessed more accurately. Hydro-
concentration gradient will cause gas in solution to dynamic models are available that enable important
diffuse through the pressure port and into the sen- parameters such as permeability and shear modulus
sor. The fluid inside the sensor will eventually be- to be calculated from the dynamic pore pressure
come saturated in gas and will consequently be in response of the sediments to tidal cycles.
chemical equilibrium with the pore waters in the The technology for very accurately measuring
sediment. If this is the case, then under conditions differential pore pressures in the upper few metres of
of static ambient pore pressures, the pressure regis- soft sediments now exists even for use in full oceanic
tered by the sensor would accurately reflect the pore water-depths. Simple techniques for measurements at
pressure in the sediments. If, however, the gas in greater sub-bottom depths and in harder substrates
the sensor were to come out of solution, then seri- are not available. One of the largest outstanding
ous errors could result if the gas replaces the water problems in soft sediments lies with measuring pore
over a significant length of the pipe. If the pressure pressures in gassy sediments over any significant
in the sediments is varying dynamically, as in the period of time.
case where waves and tides have an effect, then the Much more data are required from different sedi-
errors caused by even small amounts of gas could mentary and tectonic environments to validate the
be very large indeed. It may be the dynamic effects models, to appreciate further the nature of some of
of pressure and/or temperature that enhance the the outstanding measurement problems and to en-
diffusive flux of gas into the sensor and cause the hance our understanding of the varied, and probably
gas to come out of solution after it has reached complex, hydrogeological processes occurring in the
saturation (T. R. S Wilson, pers. comm.). There- marine environment.
fore, any pore pressure measurements made in gassy
sediments will probably be very difficult to interpret
Acknowledgements
accurately after a couple of days. The author is
unaware of any satisfactory simple solution to this I would like to thank my colleagues at the Institute
problem. of Oceanographic Sciences for their encouragement
 PORE PRESSURES IN MARINE SEDIMENTS 167

and help with the development and use of the Axial Capacity for Driven Piles in Clay, 9th Offshore Technol-
PUPPI. In particular, I wish to thank Steve McPhail, ogy Conference Proceedings, paper 2943, pp. 495~501.
Francis, T. J. G., 1984, A Review of lOS Research into the
Dave Gunn, Tim Francis, Mike Hurley and Richard Feasibility of High-Level Radioactive Waste Disposal in the
Babb for their particular skills. Thanks are also given Oceans, The Science of the Total Environment 35, 301 ~323.
to Ernie Hailwood who carefully reviewed the origi- Garrison, L. E., 1977, The SEASWAB Experiment, Marine Geo-
technology 2, 117 ~ 122.
nal manuscript. Gibson, R. E., 1963, An Analysis of the System Flexibility and its
Effect on the Time Lag in Pore Water Pressure Measurements,
Geotechnique 13( I), I ~ II.
References Hamilton, E. L., 1964, Consolidation Characteristics and Related
Properties of Sediments from Experimental Mohole
Abbott, D., Menke, W., Hobart, M., Anderson, R. N., and
(Guadalupe Site), J. Geophys. Res. 69(20), 4257~4269.
Embley, R. W., 1984, Correlated Sediment Thickness, Tempera-
Hanna, T. H., 1985, Pore Water Pressure Measurement Devices,
ture Gradient and Excess Pore Pressure in a Marine Fault Block
in Hanna T. H. (ed.), Field Instrumentation in Geotechnical
Basin. Geophys. Res. Lett. 11(5), 485~488.
Engineering, Clausthal-Zellerfeld Trans. Tech. Publications, 843
Baligh, M. M., 1986, Undrained Penetration, II: Pore Pressures,
pp, Series on Rock and Soil Mechanics 10, 121~204.
Geotechnique 36(4) 487 ~501.
Hirst, T. J. and Richards, A. F., 1976, Excess Pore Pressure in
Becker, K. and Von Herzen, R. P., 1983, Heat Transfer through
Mississippi Delta Front Sediments: Initial Report, Marine Geo-
the Sediments of the Mounds Hydrothermal Area, Galapagos
technology 1(4), 337~344.
Spreading Center at 86°W, J. Geophys. Res. 88, 995~ 1008.
Hirst, T. J. and Richards, A. F., 1977, In situ Pore Pressure
Bennett, R. H., Bryant, W. R., Dunlap, W. A., and Keller, G. H.,
Measurements in Mississippi Delta Front Sediments, Marine
1976, Initial Results and Progress of the Mississippi Delta Geotechnology 2, 191 ~204.
Sediment Pore Water Pressure Experiment, Marine Geotechnol- Hottman, W. E., Suhayda, J. N., and Garrison, L. E., 1978,
ogy 1(4), 327~335. SEASWAB II (Shallow Experiment to Assess Storm Wave
Bennett, R. H., 1977, Pore Water Pressure Measurements: Missis- Effects on the Bottom), Offshore Technology Conference Pro-
sippi Delta Submarine Sediments, Marine Geotechnology 2, ceedings, 2, paper 3169, 1059~ 1066.
177~189.
Hurley, M. T., 1989, Biot's Theory of Poroelasticity Applied to
Bennett, R. H. and Faris, J., Ambient and Dynamic Pore Pres- Geoacoustic and Hydrodynamic Models of the Sea-bed, Unpub-
sures in Fine-grained Submarine Sediments: Mississippi Delta, lished PhD Thesis, University of Wales.
Applied Ocean Research 1(3) 115~ 123. Hvorslev, M. J., 1951, Time Lag and Soil Permeability in Ground
Bennett, R. H., Burns, J. T., Clarke, T. L., Faris, J. R., Forde, E. Water Observations. Bulletin No. 36, Waterways Experimental
B., and Richards, A. F., 1982, Piezometer Probes for Assessing Station, Corps of Engineers, U.S. Army.
Effective Stress and Stability in Submarine Sediments, in Marine Lai, J. Y., Richards, A. F., and Keller, G. H., 1968, In-Place
Slides and other Mass Movements (Saxov, S. and Nieuwenhuis, Measurement of Excess Pore Water Pressure of Gulf of Main
J. K., eds), Plenum Press, London, pp. 129~ 161. Clays (Abstract), American Geophysical Union Transactions 49,
Bennett, R. H., Huon, L., Valent, P. J., Lipkin, J., and Esrig, M. 221.
I., 1985, In Situ Undrained Shear Strengths and Permeabilities Lambert, D. N., 1982, Submersible Mounted in situ Geotechnical
Derived from Piezometer Measurements, in Chaney, R. C. and Instrumentation. Geo. Marine Letters 2, 209~214.
Demars, K. R. (eds.), Strength Testing of Marine Sediments: Leinen, M., Schultheiss, P. J., Dadey, K., Becker, K., and
Laboratory and In Situ Measurements Special Technical Publica- McDuff, R., 1990, Ridge Flank Hydrothermal Activity near the
tion 883, American Society for Testing and Materials, Philadel- Mariana Trough Spreading Center (in prep.).
phia, pp. 88~ 100. McPhail S. D. and Schultheiss, P. J., 1986, PUPPI-A Free-Fall
Bennett, R. H., Burns, J. T., Nastav, F. L., Lipkin, J., and Seabed Piezometer for Geotechnical Studies, Advances in
Percival, C. M., 1985, Deep-Ocean Piezometer Probe Technol- Underwater Technology, Ocean Science and Offshore Engineer-
ogy for Geotechnical Investigations, IEEE Journal of Oceanic ing, Vol. 6: Oceanology, Graham and Trotman Ltd., pp. 249~
Engineering 10( I), 17 ~ 22. 258.
Carpentier, R. and Verdonk, W., 1986, Special Pore Water Pres- Noorany, I. and Gizienski, S. F., 1970, Engineering Properties of
sure Measuring System Installed in the Seabed for the Con- Submarine Soils-A State-of-the-Art Review, Proc. American
struction of the New Outer Harbour at Zeebrugge, Belgium, Society of Civil Engineers 96, (SM5) 1735~ 1762.
International Conference on Measuring Techniques, London, Penman, A. D. M., 1960, A Study of the Response Time of
paper C3, pp. 121 ~ 135. Various Types of Piezometer, Proceedings, Conference on Pore
De Wolf, P., Carpentier, R., Verdonk, W., Boullart, L., De Rouk, Pressure and Suction in Soils, Institution of Civil Engineers,
J., de Saint Aubain, T., and De Voghel, J., 1983, In-situ Pore London, pp. 53~58.
Water Pressure Measurements for the Construction of the Prindle, R. W. and Lopez, A. A., 1983, Pore Pressures in Marine
Breakwaters of the New Outer Harbour at Zeebrugge, Proceed- Sediments-1981 Test of the Geotechnically Instrumented
ings 8th International Harbour Conference, Antwerp, pp. 1203~ Seafloor Probe (GISP), Offshore Technology Conference Pro-
1215. ceedings, Vol. I (4463), 173~180.
Dunlap, W. A., Bryant, W. R., Bennett, R., and Richards, A. F., Prior, D. B. and Coleman, J. M., 1984, Submarine Slope Instabil-
1978, Pore Pressure Measurements in Underconsolidated Sedi- ity, in Brunsden, D. and Prior, D. B., (eds.) Slope Instability,
ments. 10th Offshore Technology Conference Proceedings, 2, John Wiley & Sons Ltd., pp. 419~455.
paper 3168, pp. 1049~1058. Randolf, M. F., Carter, J. P. and Wroth, C. P., 1979, Driven Piles
Esrig, M. I., Kirby R. C., and Bea, R. G., 1977, Initial Develop- in Clay-The Effects of Installation and Subsequent Consolida-
ment of a General Effective Stress Method for the Prediction of tion, Geotechnique 29(4), 361~393.
168 P. J. SCHULTHEISS

Reece, E. W., Ryerson, D. E., Kestly, J. D., and McNeill, R. L., Soil, Oceanology International, Brighton, UK, Paper 012.6/1.,
1979, The Development of in situ Marine Seismic and Geotech- 18 pp.
nical Instrumentation Systems. POAC 79, in Proceedings of 5th Silva, A. J. and Jordan, S. A., Consolidation Properties and Stress
International Conference on Port and Ocean Engineering under History of Some Deep Sea Sediments, in Denness, B., (ed.),
Arctic Conditions, 1, The Norwegian Institute of Technology, Seabed Mechanics, Graham and Trotman, pp. 25-39.
Trondheim, pp. 331-334. Silva, G. C. and Nageswaran, S., 1984, Compressibility of Gassy
Richards, A. F., 1984, Modelling and the Consolidation of Marine Soil, Oceanology International, Brighton, UK, paper 012.6/1.,
Soils, in Denness, B. (ed.), Seabed Mechanics, Graham and 18 pp.
Trotman, pp. 3-8. Silva, A. J. and Jordan, S. A., 1984, Consolidation Properties and
Richards, A. F. and Hamilton, E. L., 1967, Investigations of Deep Stress History of some Deep Sea Sediments, in Denness, B.,
Sea Sediment Cores, III. Consolidation, in Richards, A. F., (ed.), Seabed Mechanics, Graham and Trotman, pp. 25-39.
(ed.), Marine Geotechnique, University of Illinois Press, Urbana, Silva, A. J. and Mairs, H. L., 1987, Stress History and Geotechni-
pp.93-117. cal Properties of Nares and Madeira Abyssal Plain Sediments,
Richards, A. F., Oien, K., Keller, G. H., and Lai, J., 1975, Marine Geomechanics Laboratory, Dept. of Ocean Engineer-
Differential Piezometer Probe for in situ Measurement of Sea- ing, University of Rhode Island Technical Report, 161 pp.
floor Pore Pressure, Geotechnique 25, 229-238. Soderberg, L. 0., 1962, Consolidation Theory Applied to Founda-
Schultheiss, P. J., McPhail, S. D., Packwood, A. R., and Hart, B., tion Pile Time Effects, Geotechnique 12, 217-225.
1985, An Instrument to Measure Differential Pore Pressures in Terzaghi, K., 1943, Theoretical Soil Mechanics, John Wiley &
Deep Ocean Sediments: Pop-up-Pore-Pressure-Instrument, Sons, New York.
(PUPPI). lOS Report No. 202, 57 pp. Vaughn, P. R., 1973, The Measurement of Pore Pressures with
Schultheiss P. J. and McPhail, S. D., 1986, Direct Indication of Piezometers, in Field Instrumentation in Geotechnical Engineer-
Pore-Water Advection from Pore Pressure Measurements, in ing, John Wiley, NY, pp. 411-422.
Madeira Abyssal Plain Sediments, Nature 320(6060),348-350. Westbrook, G. K. and Smith, M. J., 1983, Long Decollements and
Schultheiss, P. J. and Noel, M., 1987, Evidence for Pore Water Mud Volcanoes: Evidence from the Barbados Ridge Complex
.Advection in the Madeira Abyssal Plain from Sediment for the Role of High Pore Water Pressures in the Development
Pore-Pressure and Temperature Measurements, in P. P. E. of an Accretionary Complex, Geology 11, 279-283.
Weaver and J. Thomson, (eds.), Geology and Geochemistry of Working Group 3, 1987, Fluid Circulation in the Crust and
Abyssal Plains, Geological Society Special Publication 31, 113- Global Geochemical Budget, in Report of the Second Conference
129. on Scientific Ocean Drilling (COSOD II), Strasboug, France,
Sills, G. C. and Nageswaran, S., 1984, Compressibility of Gassy European Science Foundation, pp. 67-86.
List of Contributors

J. GRANT*, Krupp Atlas Elektronik UK Ltd., Hartcran House, Delta C. M. R. ROBERTS*, Department of Geological Sciences, University
Gain, Carpenders Park, Watford, WDI SEZ, UK of Durham, South Road, Durham, DHI 3LE, UK
E. A. HAILWOOD*, Oceanography Department, University of K. ROBERTSON"', NERC Scientific Services, Research Vessel
Southampton, Southampton, S09 SNH, UK Services, No 1 Dock, Barry, South GIamorgan CF6 6UZ
Q. J. HUGGETT*, Institute of Oceanographic Sciences Deacon R. SCHREIBER, Krupp Atlas Elektronic GmbH, Sebaldsbriicker
Laboratory, Brook Road, Wormley, Godalming, Surrey, Heerstr. 23S, D2Soo, Bremen 44, FRG
GUSSUB,UK P. J. SCHULTHEISS"', Geotek, Fern Cottage, Marley Lane,
R. B. KIDD*, Geology Department, University College of Wales, Haslemere, Surrey GU27 3RF, UK
P.O. Box 6S, Cardiff CFl 3XA, UK R. C. SEARLE"', Department of Geological Sciences, University of
T. P. LeBAS, Institute of Oceanographic Sciences Deacon Laboratory, Durham, South Road, Durham, DHI 3LE, UK
Brook Road, Wormley, Godalming, Surrey, GUS SUB, UK G. C. SILLS, Department of Engineering Science, University of
M. R. G. MacCORMACK, Department of Earth Sciences, University Oxford, Parks Road, Oxford, OXI 3PJ, UK
of Cambridge, Bullard Laboratories, Madingley Road, Cambridge, M. C. SINHA"', Department of Earth Sciences, University of
CB30EZ, UK Cambridge, Bullard Laboratories, Madingley Road, Cambridge,
N. C. MITCHELL, Department of Earth Sciences, University of CB30EZ, UK
Oxford, Parks Road, Oxford OXI 3PR, UK M. L. SOMERS, Institute of Oceanographic Sciences Deacon
T. R. E. OWEN, Department of Earth Sciences, University of Laboratory, Brook Road, Wormley, Godalming, Surrey,
Cambridge, Bullard Laboratories, Madingley Road, Cambridge, GUS SUB, UK
CB30EZ, UK M. STORMS"', Ocean Drilling Program, Texas A&M University
W. R. PARKER*, Blackdown Consultants Ltd., 3, Curdleigh Lane, Research Park, 1000 Discovery Drive, College Station, Texas
Blagdon Hill, Taunton, Somerset, TA3 7SH, UK 77S4O, USA
L. M. PARSON, Institute of Oceanographic Sciences Deacon M. J. UNSWORTH, Department of Earth Sciences, University of
Laboratory, Brook Road, Wormley, Godalming, Surrey, Cambridge, Bullard Laboratories, Madingley Road, Cambridge,
GUS SUB, UK CB30EZ, UK
P. D. PATEL, Department of Earth Sciences, University of P. P. WEAVER"', Institute of Oceanographic Sciences Deacon
Cambridge, Bullard Laboratories, Madingley Road, Cambridge, Laboratory, Brook Road, Wormley, Godalming, Surrey,
CB30EZ, UK GUS SUB, UK
Ph. PATRIAT, Institut de Physique du Globe de Paris, Tour 14-1S, 4 J. WHITE"', Department of Civil Engineering, University of
Place Jussieu, 7S230, Paris, France Southampton, S09 SNH, UK
A. T. S. RAMSAY, Geology Department, University College of
Wales, P.O. Box 6S, Cardiff, CFl 3XA, UK
J. M. REYNOLDS*, Department of Geological Sciences, Polytechnic
South West, Drake Circus, Plymouth, PL4 SAA, UK

* =first author.

Marine Geophysical Researchers 12: 169, 1990.