Vol. 38, No. 1 - Marine Technology Society

Transcription

General Issue
THE INTERNATIONAL, INTERDISCIPLINARY SOCIETY DEVOTED TO OCEAN AND MARINE ENGINEERING, SCIENCE, AND POLICY
VOLUME 38, NUMBER 1, SPRING 2004
Marine Technology Society Officers
EXECUTIVE COMMITTEE
President
Ted Brockett
Sound Ocean Systems, Inc.
President-Elect
Jerry Streeter
JP Kenny, Inc.
Immediate Past President
Andrew Clark
Maritime Communication Services
VP-Technical Affairs
Daniel Schwartz
University of Washington
Secretary-Treasurer
John Head
Prevco Subsea Housings
Director-Budget & Finance
Jerry Boatman
COMNAVMETOCCOM
Director-Publications
Jerry Wilson
Fugro Pelagos, Inc.
Director-Public Affairs
Richard Butler
Aanderaa Instruments
SECTIONS
VP-EASTERN REGION
Robert Winokur
Oceanographer of the Navy
Canadian Maritime
Ferial El-Hawary
B.H. Engineering Systems, Ltd.
New England
James Case
SAIC
Washington, DC
Barry Stamey
Mitretek Systems
VP-SOUTHERN REGION
Sandor Karpathy
Stress Subsea, Inc.
Florida
Doug Briggs
Florida Atlantic University
Gulf Coast
Laurie Jugan
Planning Systems, Inc.
Houston
John Whites, III
Submar, Inc.
VP-WESTERN REGION
Brock Rosenthal
Ocean Innovations
Hawaii
William Friedl
CEROS
Los Angeles
James Edberg
Consultant
Monterey Bay
Mark Brown
MBARI
Puget Sound
Edward Van Den Ameele
NOAA Pacific Hydrographic Branch
San Diego
Harry Maxfield
RD Instruments
Japan
Toshitsugu Sakou
Tokai University
PROFESSIONAL DIVISIONS
& COMMITTEES
ADVANCED MARINE TECHNOLOGY
Autonomous Underwater Vehicles
Justin Manley
Mitretek Systems
Dynamic Positioning
Howard Shatto
Shatto Engineering
Ocean Energy
Open Position
Oceanographic Instrumentation
Kim McCoy
Ocean Sensors, Inc.
Manned Underwater Vehicles
William Kohnen
SEAmagine Hydrospace, Inc.
Remote Sensing
Richard Crout
CNMOC
Remotely Operated Vehicles
Drew Michel
TSC Holdings, Inc.
Underwater Imaging
Donna Kocak
Green Sky Imaging, LLC
MARINE RESOURCES
Porter Hoagland
WHOI
Marine Geodesy
Open Position
Marine Living Resources
Open Position
Marine Mineral Resources
John C. Wiltshire
University of Hawaii
Oceanographic Ships
Open Position
Ocean Pollution
Open Position
Physical Oceanography and Meteorology
Open Position
OCEAN & COASTAL ENGINEERING
Captain Diann Karin Lynn
NFEC
Buoy Technology
Walter Paul
WHOI
Cables and Connectors
Thomas Coughlin
Tomas Coughlin and Associates
Diving
William C. Phoel
Phoel Associates Inc.
Marine Archaeology
Brett Phaneuf
Texas A&M University
Marine Materials
Open Position
Moorings
Open Position
Offshore Structures
Open Position
Ropes & Tension Members
John F. Flory
Tension Technology International, Inc.
Seafloor Engineering
Herb Herrmann
NFESC
MARINE POLICY & EDUCATION
Coastal Zone Management
Open Position
Marine Education
Sharon H. Walker
University of Southern Mississippi
Marine Law and Policy
Myron Nordquist
University of Virginia
Marine Recreation
Open Position
Marine Security
Open Position
Merchant Marine
Open Position
Ocean Economic Potential
Open Position
Ocean Exploration
Paula Keener-Chavis
NOAA Coastal Services Center
STUDENT SECTIONS
Florida Atlantic University
Counselor: Douglas Briggs
Florida Institute of Technology
Counselor: Eric Thosteson
Massachusetts Institute of Technology
Counselor: Alexandra Techet
Roger Williams University
Santa Clara University
Counselor: Christopher Kitts
Texas A&M University—College Station
Counselor: Robert Randall
Texas A&M University—Galveston
Counselor: Victoria Jones
U.S. Naval Academy
Counselor: Cecily Natunewicz
University of Hawaii
Counselor: R. Cengiz Ertekin
University of Rhode Island
Counselor: Chris Baxter
University of Southern Mississippi
Counselor: Stephan Howden
HONORARY MEMBERS
The support of the following individuals is
gratefully acknowledged.
Robert B. Abel
†Charles H. Bussmann
John C. Calhoun
John P. Craven
†Paul M. Fye
David S. Potter
†Athelstan Spilhaus
†E. C. Stephan
†Allyn C. Vine
†James H. Wakelin, Jr.
†deceased
Volume 38, Number 1, Spring 2004
FRONT COVER:
Isometric, profile, and body plan views of a generated
catamaran demihull form (see pages 5-11), images
courtesy of V. Anantha Subramanian and Patrick Joy;
deep submergence research vehicle: operating depth
2000m, surface navigational range 4500NM (see pages
40-51), image courtesy of Japan Deep Sea Technology
Association. Collage design by Michele A. Danoff
In This Issue
3
61
Crosstalk
Development of a Real-time Regional
Ocean Forecast System with Application
to a Domain off the U.S. East Coast
MTS Journal Readers’ Comments
5
A Method for Rapid Hull Form
Development and Resistance
Estimation of Catamarans
BACK COVER:
Examples of ocean forecast products from NOAA’s
East Coast Regional Ocean Forecast System (see pages
61-79), image produced by the Environmental Modeling
Center at NOAA’s National Centers for Environmental
Prediction.
MTS Journal design and layout:
Michele A. Danoff, Graphics By Design
80
12
The Legal Status of Autonomous
Underwater Vehicles
Satellite Data Assimilation for
Improvement of Naval Undersea
Capability
Commentary by Stephanie Showalter
Peter C. Chu, Michael D. Perry, Eric L.
Gottshall, David S. Cwalina
Book Reviews
24
Use of Expert Systems for Optimum
Maintenance of Marine Power Plants
K.D.H. Bob-Manuel
The Science and Technology of
Nonexplosive Severance Techniques
MTS members can purchase the printed Journal for
$25 domestic and $50 foreign. Non-members and
library subscriptions are $120 domestic and $135 foreign.
Postage for periodicals is paid at Columbia, MD, and
additional mailing offices.
POSTMASTER:
Please send address changes to:
Marine Technology Society Journal
5565 Sterrett Place
Suite 108
Columbia, Maryland 21044
Copyright © 2004 Marine Technology Society, Inc.
Laurence C. Breaker, Desiraju B.Rao,
John G.W. Kelley, Ilya Rivin,
Bhavani Balasubramaniyan
V. Anantha Subramanian, Patrick Joy
30
The Marine Technology Society Journal
(ISSN 0025-3324) is published quarterly (spring summer,
fall, and winter) by the Marine Technology Society, Inc.,
5565 Sterrett Place, Suite 108, Columbia, MD 21044.
General Issue
Mark J. Kaiser, Allan G. Pulsipher,
Robert C. Byrd
40
A Design Study of Manned Deep
Submergence Research Vehicles in Japan
Dan Ohno, Yozo Shibata, Hisao Tezuka,
Hideyuki Morihana, Ryuichiro Seki
52
Temperature and Salinity Variability
in the Mississippi Bight.
Sergey Vinogradov, Nadya Vinogradova,
Vladimir Kamenkovich, Dmitri Nechaev
84
Editorial Board
Dan Walker
Editor
National Research Council
The Marine Technology Society
is a nonprofit, international professional
organization established in 1963 to promote
the exchange of information in ocean and
marine engineering, science, and policy.
John F. Bash
Book Review Editor
University of Rhode Island
Kenneth Baldwin
University of New Hampshire
Scott Kraus
New England Aquarium
James Lindholm
Pfleger Institute of Environmental Research
Please send all correspondence to:
The Marine Technology Society
5565 Sterrett Place, Suite 108
(410) 884-5330 Tel.
(410) 884-9060 FAX
E-mail: [email protected]
[email protected]
[email protected]
Online: www.mtsociety.org
Phil Nuytten
Nuytco Research, LTD.
Bruce H. Robison
Monterey Bay Aquarium
Research Institute
Terrence R. Schaff
National Science Foundation
Edith Widder
Harbor Branch Oceanographic Institution
Editorial
Jerry Wilson
Publications Director
Dan Walker
Editor
Amy Morgante
Managing Editor
Administration
Ted Brockett
President
Judith T. Krauthamer
Executive Director
Emily L. Speight
Membership
Circulation Manager
2
MEMBERSHIP INFORMATION
may be obtained by contacting the Marine
Technology Society. Benefits include:
■ Receipt of the online Marine Technology
Society Journal
■ Receipt of Currents, the bimonthly MTS
newsletter covering events, products,
and people in marine technology
■ Expansion of professional opportunities
and knowledge through section,
committee, and conference activities
and announcements
■ Member discounts on MTS publications
■ Reduced rates for attendance and
exhibit space at MTS-sponsored
conferences and exhibitions
Individual dues are $75 per year. Life,
Patron, Student, Emeritus, Institutional,
Business, and two classes of Corporate
memberships are also available.
ADVERTISING
Advertising is accepted by the Marine
Technology Society Journal. Ads are
included to increase awareness among the
readership of capabilities that exist within
the marine technology community as well
as to alert society members to organizations that offer various services. Advertisements are limited to six per issue. First
priority will be given to MTS sponsors.
For more information on MTS advertising, please contact the managing editor at
the address listed above.
COPYRIGHT
Copyright © 2004 by the Marine Technology
Society, Inc. Authorization to photocopy
items for internal or personal use, or the
internal or personal use of specific clients, is
granted by the Marine Technology Society,
provided that the base fee of $1.00 per copy,
plus .20 per page is paid directly to Copyright
Clearance Center, 222 Rosewood Dr.,
Danvers, MA 01923. For those organizations
that have been granted a photocopy license
by CCC, a separate system of payment has
been arranged. The fee code for users of the
Transactional Reporting Service is 00253324/89 $1.00 + .20. Papers by U.S
Government employees are declared works
of the U.S. Government and are therefore in
the public domain. The Marine Technology
Society cannot be held responsible for the
opinions given and the statements made
in any of the articles published.
ABSTRACTS
Abstracts of MTS publications can be found
in both the electronic and printed versions
of Aquatic Sciences and Fisheries Abstracts
(ASFA), which is published by LexisNexis
Academic and Library Solutions, 7200
Wisconsin Avenue, Bethesda, MD 20814.
Electronic abstracts may be obtained through
Geobase’s Oceanbase, Fluidex, and Compendix,
which is published by Elsevier Science,
The Old Bakery, 111 Queen Road,
Norwich, NR1 3PL, United Kingdom.
Microfishe may be obtained through
Congressional Information Services, Inc.,
4520 East-West Highway, Bethesda,
Maryland 20814
CONTRIBUTORS
Contributors can obtain an information and
style sheet by contacting the managing editor
at the address listed above. Submissions that
are relevant to the concerns of the Society
are welcome. All papers are subjected to a
review procedure directed by the editor and
the editorial board. The Journal focuses on
technical material that may not otherwise be
available, and thus technical papers and notes
that have not been published previously
are given priority. General commentaries
are also accepted, and are subject to review
and approval by the editorial board.
Reader’s Comments...
CROSSTALK
A Review of the Effects of Seismic Surveys on Marine Mammals,
by Jonathan Gordon, Douglas Gillespie, John Potter, Alexandros Frantzis,
Mark P. Simmonds, René Swift and David Thompson
(Winter 2003/2004, Vol 37, No 4)
I found this review a useful contribution for managing undersea noise appropriately. Lest we
make the common mistake of only focusing on direct auditory damage, the following quote from
the Gordon et al. paper is appropriately cautionary:
“...changes in animals’ behavior [may] lead to physical damage...[thus, the assumption] that physical
damage will be restricted to limited areas very close to powerful sound sources may have to be
revised...behavioral responses can occur at extended ranges and are often highly variable.”
Unfortunately, this paper failed to mention the deleterious effects seismic noise may have on
whole ecosystems. It is not enough to state that marine mammals’ prey may be impacted. Noise can
conceivably alter the links within food webs. The effects of past whaling, e.g., may explain the
present reduction of the kelp forests in the Pacific Northwest. This perspective should make us
realize just how impossible it will be to fully describe the effects of our acoustic “footprint” on the
oceans and how important precautionary management will be for many years to come.
Also useful would have been to highlight a problem plaguing marine mammal research on the
effects of undersea noise: conflict-of-interest. Because noise producers like the U.S. Navy and oil
companies generally directly fund marine mammal science, results are perceived to be biased and
lose all-important credibility. Preferable would be to have noise producers contribute funds to a
communal “pot” administered by a non-aligned body. Thus, expensive marine mammal research
would be less wasted by being tainted through its funding sources.
LindyWeilgart,Ph.D.
Dalhousie University
Author’s Response, continued on page 4
J O I N T H E C O N V E R S A TION…
If you have a comment or question about this issue of the MTS Journal or a previous
issue that you’d like to submit to Crosstalk, please limit it to 250 words or less and send it
to: [email protected]. Please include your name, affiliation, and contact information
(telephone, fax, and e-mail address), and identify the paper and author to whom your
comments are addressed.
We look forward to hearing from YOU!
Spring 2004
Volume 38, Number 1
3
Reader’s Comments...
CROSSTALK
Authors’ Response
It is heartening to see that our review has already generated interest, and that this includes some
controversy is inevitable given such a highly-charged and difficult subject.
With regard to seismic impact on whole ecosystems, there is but a poor understanding of how
marine acoustic pollution in general propagates through entire ecosystems, let alone impacts of
seismic surveys in particular. The authors recognize that a complete and integrated understanding
needs to take a holistic view and include all components and their interactions. Sadly, the understanding of even the basic components is still at such a primitive level that integrating our knowledge
over entire ecosystems is as yet an unrealistic task that would involve a level of speculation beyond
that appropriate for a review article. Perhaps we will see some progress in this area at the meeting on
seismic impacts on marine mammals organized by the Canadian Department of Fisheries to be held
17-21 May 2004. This does not imply that these authors view seismic impact at the ecosystem level
as negligible, simply too poorly-understood and too difficult to quantify at present to make meaningful review comment.
Dr. Weilgart raises concern about a conflict of interest arising from marine mammal researchers
working on anthropogenic noise being directly funded by bodies responsible for creating “acoustic
pollution.” This is a subject of considerable and general importance as funding mechanisms and the
administration of projects can certainly affect the scientific questions that are asked, the way research
is done, and the research teams that are chosen to carry it out. However, this is primarily a political
question which we feel could not be covered in a scientific review attempting to summarize research,
observations, and our current factual understanding of the issue.
Dr. John Potter.
Associate Director, Tropical Marine Science Institute
National University of Singapore
and co-authors Alexandros Frantsis, Douglas Gillespie,
Jonathan Gordon, Rene Swift, Dave Thompson
4
PAPER
A Method for Rapid Hull Form Development
and Resistance Estimation of Catamarans
AUTHORS
ABSTRACT
V. Anantha Subramanian
Patrick Joy
Department of Ocean Engineering,
Indian Institute of Technology
Catamarans are being built with higher speed capability for transportation of passengers and for other applications. The wide deck area and large transverse stability are
attractive features that account for the high demand for this class of vessels. A rapid hull
form development, combined with assessment of resistance, is presented here. The hull
form development is based on simple inputs for the polygon net, and using the bi-parametric bi-quintic surface in a computer-aided development scheme, a faired 3-D surface
is developed. The hull volume is computed and iteratively matched to the targeted volume. Using the developed form, a formulation based on Michell’s theory for slender vessels, combining wave interference factor and viscous interference factor, is used to arrive
at the total resistance of the combined hulls. The method is demonstrated for a series of
9 catamaran forms. The forms have varied geometric ratios, but a common displacement
of 180t. The method is presented as a rapid development and resistance assessment tool.
INTRODUCTION
I
n recent decades, catamarans have received considerable attention as vessels for
research, transportation of passengers and
cargo. They have large deck area, good sea
keeping qualities when equipped with ride
control features, and manoeuvrability afforded by slender and widely separated hulls.
The design of a high-speed hull form should
be carried out so as to provide an adequate
capacity to carry a given payload at a required
speed. Given the owner’s requirements, the
designer generally aims to provide a
favourable form that gives sufficient internal volume and deck area.
The interference effects between the
demihulls characterize the resistance of catamarans, and this must be considered in addition to the resistance of the demihulls in
isolation. Two types of interference effects
specific to catamarans can be identified:
namely, viscous interference caused by the
asymmetric flow around the demihulls and
its effect on the viscous flow such as boundary layer formation and the development of
vortices, and wave interference originating
from the interactions between the wave systems of the demihulls.
Insel and Molland (1991) have conducted investigations on the components of
resistance of catamarans. According to the
experimental results published by them, the
viscous interference effect component of resistance for high speed catamarans primarily depends upon the L/B ratio and is much
less influenced by the spacing ratio (s/L). The
form factors established from the resistance
tests were found to be considerably higher
than those for monohulls. With this in mind,
the present approach considers a viscous interference factor β.
Couser et al. (1997) have observed that
it is unclear whether the viscous resistance
increase is due primarily to modifcations of
the boundary layer and velocity augmentation between the demihulls or to additional
spray associated with constructive interference of the wave systems, particularly in the
vicinity of the transom. In consideration of
the above, a modified form factor, namely
(1 + βk) as given by them has been used for
obtaining the modified frictional resistance
component. An appropriate value of 1.42
has been used for high speed, round bilge
catamarans (independent of demihull separation) considered here for assessment of
resistance. The wave resistance component
is obtained using Michell’s theory for slender vessels with a modification using a wave
interference factor (Tuck, 1987).
The focus of this paper is the combination of a quick computer-aided scheme for
the catamaran hull form design together with
a theoretical method to estimate the total
calm water resistance of the developed hull
forms. Results are given for a typical common
range of catamaran sizes, which provide a
better understanding of the components of
catamaran resistance including the influence
of hull spacing ratio and length-to-beam
ratio over a wide range of Froude numbers.
The results are useful for a rapid first estimation of the powering of catamarans.
Methodology for Hull Form
Development
The ship hull form design is an ab initio
problem, i.e. the hull surface cannot be formulated entirely in terms of quantitative
criteria but must be resolved by a judicious
combination of computational and heuristic methods. The flowchart for the hull surface generation scheme is given in Fig. 1.
Since the objective here is rapid hull form
generation, the basic data is obtained from
simple freehand description of the hull form
with minimal body plan sections. The raw
offset data points are given in XYZ format
to form the 3-D mesh of polygon points
and are transformed into consistent faired
Spring 2004
Volume 38, Number 1
5
3-D surface data, obtained in any desirable
closely spaced digitized format. Hence starting with a tentative mesh, the program
builds a smooth continuous surface, with
the original points being the weighting functions. Hydrostatic calculations are in-built
in the scheme in order to check the obtained
volume of displacement with the targeted
volume. The basic B-spline technique is described in Rogers and Adams (1997). The
bi-quintic hull surface generation program
described in Subramanian and Suchithran
(1999) is used to generate the hull surface
of catamaran demihulls of varying length but
having a constant volume of displacement.
The number of stations and the number of
points in each station to be obtained are user
defined through screen input. The input data
file consists of four columns giving the coordinates of each point at each station in
the x, y, z, h format. x, y and z denote the
longitudinal, transverse, and vertical coordinates, respectively. The fourth column h
is for homogeneous coordinate representation and the value is kept as unity. The input is to be given for one demihull, stations
starting from the aft to forward. The program uses 2-D B-spline interpolation at every station to regenerate an equal number
of control polygon points. After generation
of the control polygon points, the program
fairs a B-spline surface through them. The
fairness of the curves depends upon the order of the curve and the number of defining
polygon points. A script file is generated as
output containing the output points, which
is formatted for directly interfacing with
AutoCAD. The drawing thus obtained can
be processed further.
A Cartesian product B-spline surface is
defined by,
n +1 m +1
Q (u, w) =
∑∑
i =1 j = 1
Bi,j Ni,k (u) Mj,l (w)
Bi,j’s are the vertices of a defining polygon
net. The indices n and m are the number of
defining polygon vertices in the u and w
parametric directions, respectively. Ni,k(u)
and Mj,l(w) are the B-spline basis functions
in the bi-parametric u and w directions, respectively. k and l are the order of the curves
6
in the u and w directions respectively. The definition for the basis functions is given below.
Ni,1(u) = 1 if ai
≤ u < ai+1
= 0 otherwise
Ni,k (u) = [(u – ai)Ni, k-1(u) / (ai+k-1 – ai)] + [(ai+k – u)Ni+1, k-1(u) / (ai+k – ai+1)]
Mj,1(w) = 1 if bj w < bj+1
= 0 otherwise
Mj,l (w) = [(w – bj)Mj, l-1(w) / (bj+l-1 – bj)] + [(bj+l – w)Mj+1, l-1(w) / (bj+l – bj+1)]
ai and bj are elements of knot vectors. There are three types of knot vectors that can be used,
namely, uniform, open and non-uniform knot vectors. It is not essential to use the same type
of knot vectors in both parametric directions.
The B-spline surfaces have well known properties of ability to be locally controlled in
shape by means of choice of order, choice of knot vectors, modification of local shape by redefining knots by means of knot removal and knot insertion algorithm. The references give
details for implementation of these schemes to rapidly evolve a faired form.
FIGURE 1
Flowchart for surface generation program
Estimation of Total Calm
Water Resistance
The coefficient of wave resistance for
catamarans may be expressed as,
The total calm water resistance of catamarans is due to the two major components,
namely, frictional resistance and wave resistance and, in addition, the interference effects between the demihulls have to be considered. These effects consist of viscous interference caused by the asymmetric fluid
flow around the demihulls and its effect on
the viscous flow such as the formation of
the boundary layer and the development of
vortices, and the wave interference caused
by the interactions between the wave-systems of the demihulls.
The well known ITTC 1957 friction
resistance line can be used to estimate the
equivalent frictional resistance,
CF = 0.075/(log10 Re – 2)2
(1)
Tuck (1987) used a mathematical formulation for the computation of wave resistance for catamarans in deep waters based
on the theory developed by Michell for slender vessels. The formula for the wave resistance of a catamaran in an unbounded sea
can be expressed as,
π/2
RW = (ρg4/πU6)
(P2 + Q2) sec5θ dθ
0
P and Q are the Michell wave functions
defined by,
L/2
0
b(x,z) e(iwox + koz) dz dx
P + iQ = f
CW = RW/(ρAWU2)
(2)
where, r is the density of sea water which is
taken as 1025 kg/m3 at 150 C. AW is the
wetted surface area of the demihull.
The total calm water resistance of catamarans, in the coefficient form may be expressed as,
CT = (1 + βk) CF + CW
(3)
The viscous interference factor (β) depends upon the speed of the vessel, slenderness ratio, and the hull spacing ratio. For a
monohull, β =1. For practical purposes, the
viscous interference factor is combined with
the form factor (1 + k). An average value of
1.42 is assumed for the modified form factor (1 + βk) from experimental results
(Couser et al., 1997).
Description of Generated
Parametric Hull Forms
In principle, the method described can
be used to evaluate resistance for the hull of
any targeted displacement and speed. As an
illustration, the method is demonstrated for
a series of catamaran forms with a targeted
displacement of 180 tonnes and Froude
number up to 1.0. This range has been chosen as a value typical of common high-speed
catamarans being built today, mainly for
passenger transportation. The hull form conforms to the general body plan shapes of the
Nordstrom series, but has been generated
with simple raw inputs as described earlier.
Nine catamaran demihull forms corresponding to three different lengths are generated.
The lengths considered here are 30 m, 40 m
and 50 m. The volume of displacement for
all the forms is kept constant at 90 m3 for
each demihull. For each length, three different breadths are chosen and the draught
is varied so as to maintain the volume of displacement. The isometric view and profile
view of a generated hull form are shown in
Fig. 2. The body plan views of the generated
catamaran demihull forms are shown in Fig.
3 to 5. The hull form parameters for the generated forms are given in Tables 1 to 3. The
coefficient of total resistance is computed for
the generated forms by considering four different hull spacing ratios as given in Fig. 6 to
9. In order to bring out the comparative values of wave resistance coefficient Cw for the
same length of hull, with different spacing
ratios, Fig.10 shows a sample plot for the
case of L=30m and different s/L ratios.
FIGURE 2
Isometric and Profile view of a generated catamaran demihull form (L = 30 m)
–L/2–T
b(x,z) is the local beam of the demihull. θ is
the wave angle. U is the speed of the vessel.
The wave interference factor f for a catamaran is given as,
f = 2 cos(uo.s/2)
s is the spacing between the center-planes of
the demihulls. ko, uo and wo are the circular,
transverse, and longitudinal wave numbers,
respectively, and are given as,
ko = (g/U2) sec2θ
uo = ko sin θ
wo = ko cos θ
Spring 2004
Volume 38, Number 1
7
FIGURE 3
TABLE 1
Body plan views of the generated demihull forms
(L = 30 m)
Hull form parameters (L = 30 m, ∇ = 90 m3)
Parameter
Form 1
Form 2
Form 3
b (m)
3.20
3.60
3.90
D (m)
3.00
3.00
3.00
T (m)
1.78
1.86
1.70
bWL (m)
3.00
3.28
3.56
CB
0.57
0.49
0.49
0.60
0.52
0.67
AW (m )
130.40
131.10
125.20
L/b
9.37
8.33
7.69
b/T
1.79
1.93
2.29
L/∇
6.69
6.69
6.69
Form 4
Form 5
Form 6
b (m)
3.20
3.40
3.60
D (m)
3.00
3.20
3.20
T (m)
1.70
1.65
1.60
bWL (m)
2.70
2.92
3.12
CB
0.48
0.47
0.46
0.67
0.65
0.62
AW (m )
155.50
152.40
151.10
L/b
12.50
11.76
11.11
b/T
1.88
2.06
2.25
L/∇
8.92
8.92
8.92
CP
2
1/3
FIGURE 4
TABLE 2
(L = 40 m)
Parameter
CP
2
1/3
8
FIGURE 5
TABLE 3
(L = 50 m)
Parameter
Form 7
Form 8
Form 9
b (m)
3.20
3.60
4.00
D (m)
3.00
3.00
3.00
T (m)
1.50
1.40
1.25
bWL (m)
2.77
3.06
3.34
CB
0.45
0.43
0.43
0.94
0.92
0.90
AW (m )
173.00
169.20
166.53
L/b
15.62
13.89
12.50
b/T
2.13
2.57
3.20
L/∇
11.15
11.15
11.15
Spring 2004
Volume 38, Number 1
CP
2
1/3
FIGURE 6
Coefficient of total resistance (s/L = 0.15)
9
FIGURE 7
FIGURE 8
10
FIGURE 9
Conclusions
A computer-aided surface development
scheme based on the B-spline surfaces is used
to generate high speed catamaran hull forms.
The scheme has been applied to generate
typical high-speed catamaran hull forms of
displacement 180 tonnes. The user defined
input data (control polygon net) can be refined iteratively to generate hull forms of
targeted displacement and other hydrostatic
particulars. Based on a given range of geometric parameters, nine catamaran demihull
forms are generated. The forms are grouped
into three sets of three forms each. The
length (LBP) is varied over the sets as 30 m,
40 m and 50 m respectively. The breadth to
the draught ratio (b/T) is varied for each
form to maintain the volume of displacement. All forms are symmetric about their
center-planes and have round bilge form.
The fairness of the surface is evaluated in the
sense of being pleasing to the eye of the designer. This is enough for the present purpose of estimating resistance. However a
more objective standard of fairness such as
based on curvature criteria can be set for rigorous development of hull forms.
The total calm water resistance of catamarans is mainly affected by the wetted surface area, the slenderness ratio (L/∇1/3) and
the hull spacing ratio (s/L). The following
inferences can be made from the calculated
values of the total calm water resistance for
the generated hull forms:
1. The total calm water resistance increases
with length and speed of the vessel.
2. For a particular length and speed, the
total resistance decreases with increasing
hull spacing ratio (s/L).
3. For a particular length, speed, and hull
spacing ratio, the total resistance
increases with increasing (b/L) ratio and
(b/T) ratio.
4. Thus when geometric ratios are constraints, the method permits rapid form
development and assessment of resistance
characteristics.
The wave interference influences the
total resistance to a large extent, particularly
at lower speeds (Fn < 0.35). The beneficial
wave interference (hollows) is achieved by
the cancellation of a part of the divergent
wave systems of each demihull, whereas adverse wave interference (humps) arises on
interaction of the transverse wave systems.
Above a particular speed (Fn > 0.5), the wave
interference factor which is dependent on
hull spacing ratio and speed takes a constant
value due to which the wave interference has
little effect on the total resistance at higher
speeds. The present method is useful in the
preliminary design of catamarans and for
rapidly obtaining form ratios and therefore
favourable resistance characteristics.
References
Couser, P.R., A. F. Molland, N.A. Armstrong
and I.K.A.P. Utama. 1997. Calm water
powering predictions for high speed catamarans.
Proceedings International Conference of Fast Sea
Transportation (FAST’ 97), Sydney, July 1997.
Insel, M. and A.F. Molland. 1991 An investigation
into the resistance components of high speed
displacement catamarans. Trans. RINA
Rogers, D.F. and J.A. Adams. 1997. Mathematical
elements for computer graphics. New Delhi:
Tata McGraw Hill Publishing Company Ltd.
Subramainan, V.A. and P.R. Suchitran. 1999.
Interactive curve fairing and bi-quintic surface
generation for ship design. Intl. Shipbuilding
Progress, 46(44).
Tuck, E.O. 1987. Wave resistance of slender
ships and catamarans, Report T8701
University of Adelaide.
Spring 2004
Volume 38, Number 1
11
PAPER
Satellite Data Assimilation for Improvement
of Naval Undersea Capability
AUTHORS
ABSTRACT
Peter C. Chu
Michael D. Perry
Naval Ocean Analysis and Prediction
Laboratory, Department of Oceanography,
Naval Postgraduate School
Monterey, CA
Impact of satellite data assimilation on naval undersea capability is investigated
using ocean hydrographic products without and with satellite data assimilation. The
former is the Navy’s Global Digital Environmental Model (GDEM), providing a monthly
mean; the latter is the Modular Ocean Data Assimilation System (MODAS) providing
synoptic analyses based upon satellite data. The two environmental datasets are taken
as the input into the Weapon Acoustic Preset Program to determine the suggested
presets for an Mk 48 torpedo. The acoustic coverage area generated by the program
will be used as the metric to compare the two sets of outputs. The output presets
were created for two different scenarios, an anti-surface warfare (ASUW) and an
anti-submarine warfare (ASW); and three different depth bands, shallow, mid, and
deep. After analyzing the output, it became clear that there was a great difference in
the presets for the shallow depth band, and that as depth increased, the difference
between the presets decreased. Therefore, the MODAS product, and in turn the satellite data assimilation, had greatest impact in the shallow depth band. The ASW
presets also seemed to be slightly less sensitive to differences than did presets in
the ASUW scenario.
Eric L. Gottshall
Space and Naval Warfare System Command
San Diego, CA
David S. Cwalina
Naval Undersea Warfare Center
Newport, RI
INTRODUCTION
E
ven with all the high technology
weapons onboard U.S. Navy ships today,
the difference between success and failure
often comes down to our understanding
and knowledge of the environment in
which we are operating. Accurately predicting the ocean environment is a critical factor in using our detection systems
to find a target and in setting our weapons to prosecute a target (Gottshall, 1997;
Chu et al., 1998). From the ocean temperature and salinity, the sound velocity
profiles (SVP) can be calculated. SVPs are
a key input used by U.S. Navy weapons
programs to predict weapon performance
in the medium. The trick lies in finding
the degree to which the effectiveness of
the weapon systems is tied to the accuracy of the ocean predictions.
The U.S. Navy’s Meteorological and
Oceanographic (METOC) community
currently uses three different methods to
obtain representative SVPs of the ocean:
climatology, in situ measurements, and
data (including satellite data) assimilation.
12
The climatological data provides the background SVP information that might not
be current. The Generalized Digital Environmental Model (GDEM) is an example of a climatological system that provides long-term mean temperature, salinity, and sound speed profiles. The in situ
measurements from conductivity-temperature-depth (CTD) and expendable
bathythermograph (XBT) casts may give
accurate and timely information, but these
are not likely to have large spatial and temporal coverage over all regions where U.S.
ships are going to be operating. In a data
assimilation system, an initial climatology
or forecast is improved by using satellite
and in situ data to better estimate synoptic SVPs. The Modular Ocean Data Assimilation System (MODAS) utilizes sea
surface height (SSH) and sea surface temperature (SST) in this way to make
nowcasts of the ocean environment (Fox
et al., 2002).
The value added by satellite data assimilation for use of undersea weapon systems can be evaluated using the SVP in-
put data from MODAS (with satellite data
assimilation) and GDEM (climatology
without satellite data assimilation). The
question also arises of how many altimeters are necessary to generate an optimal
MODAS field. Too few inputs could result in an inaccurate MODAS field, which
in turn could lead to decreased weapon
effectiveness. There must also be some
point at which the addition of another
altimeter is going to add a negligible increase in effectiveness. This superfluous
altimeter would be simply a waste of
money that could be spent on a more useful system.
The purpose of this study is to quantify the advantage gained from the use of
data from MODAS assimilation of satellite observations rather than climatology.
The study will specifically cover the benefits of MODAS data over climatology
when using their respective SVPs to determine torpedo settings. These settings result in acoustic coverage percentages that
will be used as the metric to compare the
two types of data.
2. Navy’s METOC Models
and Data
2.1. GDEM
GDEM is a four dimensional (latitude,
longitude, depth and time) digital model
maintained by the Naval Oceanographic
Office. GDEM was generated using over
seven million temperature and salinity observations, most of them drawn from the
Master Oceanographic Observation Data
Set (MOODS). Globally GDEM has a resolution of 1/2º degree. However, in a few select areas, higher resolutions are available.
In order to represent the mean vertical distribution of temperature and salinity for grid
squares, GDEM determines analytical
curves to fit to the individual profiles (Teague
et. al., 1990; Chu et al., 1997, 1999)
Before curves can be fitted to the data,
quality control must be implemented that
removes anomalous features or bad observations. The data is checked for proper range
and static stability, and it is checked to ensure that it has not been misplaced in location or season. Once the data has been inspected for quality, curves are fitted to the
data. From the mathematical expressions
that represent the curves, coefficients are
determined. It is these coefficients that will
be averaged. It can be shown that the coefficients resulting from averaged data are not
the same as the averaged coefficients of the
data. In order to minimize the number of
coefficients necessary to generate smooth
curves, different families of curves are used
for different depth ranges. This necessitates
the careful selection of matching conditions
in order to ensure that no discontinuities in
the vertical gradients occur. Separate computation of temperature and salinity allow
the results to be checked against each other
to ensure stable densities.
modularity allows MODAS to be quickly
and easily modified to handle problems or
new requirements as they arise. MODAS
has varying degrees of resolution starting at
1/2º in the open ocean increasing to 1/4º in
coastal seas and increasing again to 1/8º near
the coast (Fox et al., 2002). To generate
nowcasts and forecasts, the MODAS system uses a relocatable version of the
Princeton Ocean Model (POM). To initialize the POM MODAS temperature and
salinity grids, geostrophically estimated currents, or extracted currents from other
POM’s can be used.
One of the most important features of
MODAS is its use of dynamic climatology
(Fox et al., 2002). Dynamic climatology is
the incorporation of additional information
into the historical climatology in order to
portray transient features that are not represented by the climatology. Two useful quantities that are easily gathered from satellites
are sea surface height (SSH) and sea surface
temperature (SST). While SST from altimeters can be used directly, the SSH, which is
measured as the total height relative to the
proscribed mean, must be converted into a
steric height anomaly in order to be used.
2D SST and SSH fields are generated from
point observations through the use of optimal interpolation.
Optimal interpolation is a process by which
the interpolated temperature or salinity
anomaly is determined as the linear combination of the observed anomalies. Each of
the anomalies is given a weight that accounts
for variation in temporal and spatial sampling. Weights are computed by minimizing the least square difference between the
interpolated value and the true value at the
grid point and by solving the equations
(Gandin, 1965),
Σα µ + λ
N
–2
2.2. MODAS
MODAS is a collection of over 100
FORTRAN programs and UNIX scripts
that can be combined to generate a number
of different products (Fox et al., 2002). A
few examples of MODAS programs include
data sorting, data cross-validation, data assimilation, and profile extension. This
j
ij
αi = µGi ,
(1)
j=1
where αi are the weights, λ is the signal to
noise ratio, µij is the autocorrelation between
locations i and j, and µ Gi is the
autocorrelation between the grid point and i.
For each grid node location matrix inversion
is used to solve the system of N equations
for the N unknown weights. The other parameters are computed using the first guess
field, MOODS profiles, and climatology.
Using this process any new observation can
be interpolated into the appropriate
MODAS grid node.
The first guess field, the prior day’s 2D
SST field, or the weighted average of 35 days
of altimeter data respectively, is subtracted
from the new observations, and the resulting deviations are interpolated to produce a
field of deviation. This is added to the first
guess field to generate the new 2D field. For
the first iteration of the optimal interpolation, climatology is used for SST and the
SSH measurement is assumed to have a zero
deviation. This means that until the field
deviates from the climatology, the extra data
has added no value and MODAS reverts to
climatology.
Once the data is in a useful form,
MODAS begins with the climatology profile and then correlates variations in the SSH
and SST to variations in the subsurface temperature. The regression relationships used
here were constructed by performing a leastsquares regression analysis on archived temperature and salinity profiles. This is a three
step process starting with the computation
of regional empirical orthogonal functions
from the historical temperature and salinity
profiles. The second step is to express the
profiles in terms of an empirical orthogonal
function series expansion. The final step is
to perform regression analysis on the profile
amplitudes for each mode, truncating the
series after three terms. This is possible because of the compactness of the empirical
orthogonal function representation.
Once the subsurface temperatures have
been revised, MODAS adjusts the subsurface salinity profile using the relationship
between temperature and salinity. This
new profile is referred to as a synthetic
profile. Synthetic profiles only utilize these
regression relationships down to a depth
of 1500 m due to the decreasing reliability of the relationships at depth (Fox et
al., 2002).
MODAS is also able to include measurements from in situ CTDs and XBTs. The
first guess field is the field generated by the
Spring 2004
Volume 38, Number 1
13
dynamic climatology, and the in situ profiles are subtracted from it to get residuals.
Optimal interpolation is once again used to
update the temperature field and from the
temperature field the salinity field can be
generated. This salinity field then serves as a
first guess field for the inclusion of the salinity profiles (Fox et al., 2002).
2.3 Satellite Altimetry Data
Assimilated into MODAS
The Navy currently uses satellite altimeters and inferred data to assess the
ocean environment for the naval operations. Of primary interest is mesoscale
variability. Meandering fronts and eddies
can significantly change the temperature
and salinity structure of the ocean. This
importance is clearly seen in sonar dependent operations such as anti-submarine
warfare (ASW). Sonar range can be greatly
helped or hindered by the acoustic environment created by the salinity, temperature, and density. Altimeters also provide
the SSH measurements that MODAS uses
in its optimal interpolation.
While monitoring mesoscale variability
is of prime importance to the Navy, an
emerging secondary role for Navy altimeters
is monitoring continental shelf and coastal
zones. As the Navy conducts more and more
operations in littoral waters, the ability to
predict near-shore parameters will have increasing importance. Altimeter data can be
used to get up-to-date information on rapidly changing near-shore characteristics such
as tides and wave height (Jacobs et al., 2002).
These are important issues for anyone dealing with mine detection, beach operations,
or ship routing.
Altimeters have also been used to measure the flow through important straits, such
as the Tsushima Strait, and to measure largescale circulation. The first of these helps researchers and modelers to develop constraints on local numerical models. Largescale circulation measurements can also help
in the development of models by aiding in
error correction. They also help explain the
local environment that is often affected by
not just local forcing, but large-scale circulation variations as well.
14
Satellite altimeters can provide a great
variety of data, but no single altimeter can
provide measurements on all desired time
and length scales. Different parameters must
be sampled at different frequencies if they
are going to be of any use. For instance, sea
surface height must be sampled every 48
hours while wave height must be sampled
every three hours. While different ocean features all have different time and spatial scales,
only the requirements for observation of
mesoscale features are presented here as an
example (Jacobs et. al., 1999).
In order for an altimeter to efficiently
and accurately sample mesoscale features,
there are several requirements placed on its
accuracy, orbit, and repeat period. A satellite altimeter must produce measurements
that are accurate to within 5 cm, or the errors that propagate down into the temperature and salinity calculations will be unacceptable. With an error of only 5 cm, the
error in the temperature calculation can be
1-2° C. Satellites should also have an exact repeat orbit to maximize the usefulness
of the data collected. Without an exact repeat orbit, the only way to get differences
in sea surface heights is to use only the data
from points where the satellite crosses the
track of another altimeter or itself. An exact orbit is considered to be a 1 km wide
swath of a predefined ground track. Finally, the period of a single satellite should
be greater than the typical 20 day time scale
of a mesoscale feature. If two satellites are
used, then they should be spaced so that a
point on the ground is not sampled more
than once in a 20 day period (Jacobs et.
al., 1999).
As described earlier, systems such as
MODAS rely heavily on the information
provided by these satellites. MODAS uses
interpolation to estimate SSH at points
that the satellite did not cover. If the
ground track spacing is too coarse then
the optimal interpolation scheme of
MODAS will begin introducing errors
into the fields between the tracks. It is
important that the satellites be properly
set up so that a maximum amount of information can be gathered with a minimum amount of error.
3.Navy’s Weapon Acoustic Preset
A Weapon Acoustic Preset Program
(WAPP) is used to get automated, interactive means of generating Mk 48 and Mk 48
Advanced CAPability (ADCAP) acoustic
presets and visualizing torpedo performance.
It combines the Mk 48 Acoustic Preset Program (M48APP) and the Mk 48 ADCAP
Acoustic Preset Program (MAAPP) into a
single integrated package. The Royal Australian Navy as a part of the Collins Class
Augmentation System (CCAS) also uses the
M48APP, and the Royal Canadian Navy has
changed the M48APP for Java. The program
is based around a graphical user interface
that allows the user to enter the environmental, tactical, target, and weapon data. With
these user specified parameters, the program
then performs a series of computations to
generate accurate acoustic performance predictions. The output includes a ranked listset of search depth, pitch angle, LD, and
effectiveness values, an acoustic ray trace, and
a signal excess map (Cwalina, 2002, personal communication).
The Environmental Data Entry Module (EDE) is a simple Graphic User Interface (GUI) that allows the user to enter a
variety of environmental parameters (Fig. 1).
The sea surface fields allow the user to specify
wind speed, wave height, and sea state based
on either the World Meteorological or Beaufort scale conventions. The three fields are
coupled so that an entry into one field will
bring up the appropriate default values for
the others. The bottom condition field allows the user to specify the bottom depth
and to choose the bottom type from a list of
possibilities. The bottom of the GUI is devoted to the water column characteristics and
a sound speed profile. The temperature,
sound speed, and depth are all in the appropriate English units. The volume scattering
strength (VSS) is in dB. The additional fields
include the latitude, longitude, the profile
name, and the table group identifiers.
Once the environmental parameters
have been entered, the user can move on to
the Acoustic Module Preset Display. This
GUI allows the user to specify a number of
parameters about the weapon, the target, and
the way the weapon should search (Fig. 2).
FIGURE 1
Environmental data entry.
The list-set on the right side of the GUI displays a series of search depths, pitch angles,
laminar distances, and effectiveness values.
The effectiveness values for the various presets are based on expected signal excess and
ray trace computations. Both plots can be
viewed from a pull-down menu. These provide a visual representation of the acoustic
performance of the Mk 48.
In addition to automatically computing
the most effective preset combination for a
given set of environmental parameters, the
program also allows the user to manually
examine the effectiveness of any allowable
preset combination via the signal excess and
ray trace plots. The program also allows the
user to save the tactical preset list and the
accompanying environmental data. The data
are stored locally in the weapon module and
can be recalled later or transferred via a network to the combat control system.
4. Statistical Analysis
4.1. Input and Output Difference
FIGURE 2
Acoustic preset module display.
The difference between the two sets of
input (GDEM and MODAS) ψ input , or
between the two sets of output weapon preset data (running using GDEM and
MODAS) ψ output ,
∆ψ (r,t) = ψ Μ (r,t) – ψ G (r,t) , (2)
represents the ocean data update using satellite and in situ observations (input) and
the effect of using satellite and in situ observations on the weapon preset (output). Here
ψ M and ψ G are the variables (either input
or output) using GDEM and MODAS,
respectively. We may take the probability
histograms of ψ M and ψ G to show the
difference between the statistical characteristics.
4.2. Root Mean Square Difference
GDEM and MODAS have different
grid spacing: 1/2º x 1/2º in GDEM and
1/12º x 1/12º in MODAS. For a GDEM
cell, one data is available for GDEM and 36
data for MODAS. The root-mean-square
difference (RMSD),
Spring 2004
Volume 38, Number 1
15
RMSD =
,
(3)
is commonly used to represent the difference in the input and output data. Here, N
(=36) is the total MODAS data number in
a GDEM cell. The RMSD can be computed
for either the input data to the weapon preset model such as the temperature, salinity,
or sound speed, or it can be computed for
the output data such as nondimensional detection area.
Due to the differing resolutions of
GDEM and MODAS, this area provided
117 GDEM profiles and 1633 MODAS
profiles (Fig. 4). Each profile was simply a
text file that consisted of a header row and
columns of data. The header row contained
the number of depths the profile covered,
the file’s name and the latitude and longitude of the profile.
FIGURE 4
The columns corresponded to depth in
feet, the temperature in degrees Fahrenheit,
the sound speed velocity in feet per second,
a volume backscatter value, and salinity in
PSU. Despite the common use of International units in scientific experiments, it was
necessary for the profiles to be set up in the
appropriate English units. The Weapon
Acoustic Preset Program (WAPP), the program used to generate the presets from the
profiles, requires inputs to be in English units.
GDEM and MODAS data points
5. Comparison between
GDEM and MODAS in the
Gulf Stream Region
5.1. Data
In order to make a meaningful comparison of MODAS and GDEM data, a sufficiently large data set had to be obtained. The
Area of Interest (AOI) also needed to be an
area where the ocean environment fluctuated
on a fairly short time scale. The GDEM data
in March and MODAS data on March 15,
2001 was obtained for the area off the North
American coast corresponding to 40°-35° N
latitude and 75°-70° W longitude (Fig. 3).
FIGURE 3
Area of interest.
16
5.2. Difference between GDEM
and MODAS
While GDEM gives the climatological
background ocean environment at a given
place, MODAS is expected to provide more
current and synoptic interpretations of the
environment. The amount of accuracy
MODAS adds is in proportion to the scale
on which ocean parameters vary. For areas
such as the Gulf Stream, where environmental factors are known to vary rapidly on a
relatively small time scale, it is expected that
there would be at least a few areas where the
two data sets differ. It is these areas that are
of particular interest, since the difference in
the weapon presets should be greatest.
On the surface, the GDEM data provided a view of the temperature distribution
that consisted of smooth, uniformly spaced
lines of constant temperature that were consistent with the overall flow of the region
(Fig. 5). The cool water on the shelf gradually gives way to the warm water flowing
FIGURE 5
March surface temperature and salinity distribution from GDEM.
north along the Gulf Stream. The GDEM
generated surface salinity distribution is extremely similar to the surface temperature
distribution and is consistent with the Gulf
Stream region. Fresher water lies inland and
the salinity increases with distance from the
shore. The only variation is in the northeastern section, where there is a slight intrusion of the salty offshore water.
As expected, the GDEM and MODAS
distributions are, overall, fairly similar in both
their range of values and overall distribution.
They are similar to each other in shape, and
both show areas of cool fresh water near the
coast and areas of warm salty water lying offshore. There are, however, a few differences,
with the intrusion of warm salty water in
the northeastern section of the MODAS figure being the most notable. There is also an
area of high temperature in the lower right
corner of the MODAS figure that does not
show up in the GDEM figure. In general,
the MODAS figure shows a sharper front
with the water increasing in temperature and
salinity much more rapidly as the distance
from the coast increases (Fig. 6). The GDEM
figure shows a gradual increase in temperature and salinity starting in the top left corner and continuing almost entirely down to
the lower right corner. The MODAS figure
shows the water reaching maximum tem-
perature and salinity quickly and then staying constant to the lower right corner.
While the GDEM and MODAS data
offer similar ranges of temperatures, salinities, and sound speeds at the surface, the
distribution of the values is quite different.
The histograms in Figure 7 reveal that while
the temperature values reported by both data
sets are similar, the MODAS data has a
higher proportion of profiles located in the
6°-7° C range. The difference in the salinity
graphs is even more drastic with the bulk of
the GDEM values located in the middle of
the range and the MODAS values split between the high and low ends of the range.
The sound speed graph indicates that
MODAS typically reports higher sound
speeds than does the GDEM data. This is
not too surprising, since sound speed in the
upper water column tends to be tied closely
to temperature, and the MODAS data indicates warmer water than the GDEM data.
Increasing depth to 50 m and then 100
m, it is clear to see that, for temperature, the
distribution of the values over the range for
both sets of data is quite similar. There is
still a slight preference in the MODAS
graphs towards higher temperatures, but it
is not as drastic as is seen on the surface.
Salinity is much the same, with the difference in shapes of the two figures more a factor of the small number of GDEM profiles
as compared with the number of MODAS
profiles. Sound speed is the only area where
the two data sets continue to diverge. From
the 50 m and 100 m sound speed figures it
is clear that, with depth, the MODAS data
indicates increasing sound speed and the
GDEM data predicts some sort of sound
FIGURE 6
MODAS generated surface temperature and salinity distribution on March 15, 2001.
Spring 2004
Volume 38, Number 1
17
speed minimum at depth. This is causing
the peak on the MODAS graph and the peak
on the GDEM graph to move away from
each other as depth increases.
By 2000 m the temperature and salinity
histograms for the two data sets are virtually
identical. At this point any perceived difference in the two is solely a factor of the difference in the number of profiles between
the two data sets. For the sound speed figures this is the point of maximum separation. The GDEM data indicates low sound
speeds representative of a deep sound channel, whereas the MODAS data indicates that
the sound speed has increased to this point.
After this point the GDEM values begin rising again to match the MODAS data.
While the distribution of the values over
the range is a useful tool in examining the
inputs, it is the difference between the inputs that is of real importance. Figure 8
shows the RMS difference of the inputs.
From the surface temperature figure in Figure 8, the RMS difference of temperature
peaks out in the lower left corner of the AOI
at about 2° C. Besides the peak, the other
significant area is the ridge starting in the
lower left corner and running to the middle
top of the figure. This corresponds to a narrow region where the GDEM distribution
warmed slower than the MODAS distribution moving from the coast out to sea. The
warm water intrusion is represented by the
gradual increase in height of the ridge. The
salinity difference at the surface is nearly zero
for most of the AOI and reaches its maximum value of 4.5 PSU along the top of the
region. The derived sound speed RMS difference, as expected, is smallest far from the
coast where the difference in temperature
and salinity is smallest and increases towards
the coast.
As depth increases, the RMS difference
in temperature and sound speed changes
slowly, but the difference in salinity drops off
quickly. Neither the temperature nor sound
speed difference change significantly, but by
100 meters the RMS difference for salinity
has gone down to values of less than .8 PSU.
From 100 meters down, the temperature difference begins to decrease slowly, and by 2000
meters the RMS difference for both tempera-
18
FIGURE 7
Comparison between GDEM and MODAS temperature histograms
at (a) the surface, (b) 100 m depth, and (c) 2000 m depth.
FIGURE 8
Horizontal dependence of RMSD at the surface between GDEM and MODAS for
(a) temperature, (b) sound speed, and (c) salinity.
ture and salinity has dropped to negligible
levels for most of the AOI. This is expected
since MODAS reverts to climatology at
depth. Except for the profiles in the northwestern corner of the AOI that did not run
as deep as the other profiles farther from the
coast, all the RMS difference vs depth profiles were remarkably similar. All of the temperature differences showed either a gradual
decrease in the difference down to about 1000
meters or a slight increase in the difference
immediately followed by a gradual decrease
in the difference down to 1000 meters. At
about 1000 meters the temperature differences all rapidly dropped to near zero.
The sound speed profiles all show the
difference increasing down to a maximum
value of 60 m/s at around 2000 meters. After that the RMS difference drops off, and
approaches zero by 3000 meters. The cause
of the maximum at 2000 meters is lack of a
deep sound channel according to the
MODAS data. The MODAS profiles almost all have the sound speed steadily increasing down to the maximum depth
whereas climatology indicates a sound speed
minimum at 2000 m. While there is some
variation in how quickly the salinity differences drop to near zero, they are less than 1
PSU by 200 meters. Shown in Figure 9 is a
representative RMS difference profile.
6. Comparison of Weapon
Acoustic Preset Using
GDEM and MODAS
The raw data was processed by the Naval Underwater Warfare Center (NUWC)
Division Newport. They received the input
profiles, ran them through the WAPP, and
generated the output. Percentage coverage
was calculated based on both surface
(ASUW) and submarine (ASW) scenarios.
The submarine scenario is a low Doppler
scenario consistent with diesel submarine
operations. The coverage percentages represent coverage in the target depth band, either shallow, mid, or deep. The coverage
percentages were also normalized over acoustic modes to produce an output that was
dimensionless.
Spring 2004
Volume 38, Number 1
19
FIGURE 9
Depth dependence of RMSD between GDEM and MODAS for (a) temperature, (b) sound speed, and (c) salinity.
mean coverage for the AOI. The ASW scenario yields similar results except for the fact
that the two means were not even statistically different. While this seems to indicate
that the two data sets are returning similar
results, there are some important differences.
First are the outliers on the GDEM graphs.
Values in the high thirties to low fifties are
extremely rare, yet the GDEM data indicate
that in at least one location for the ASUW
scenario and several for the ASW scenario,
the weapon will perform to this level. The
ASW scenario also has a rather significant
number of GDEM profiles that generate
below average coverage percentages. This
would indicate that GDEM predicts that
coverage will vary greatly with location. In
comparison the MODAS values for both
scenarios tend to be very consistent. Coverage percentage varies little with location due
to the fact that most of the profiles lie within
a very narrow range. Overall GDEM predicts excellent coverage some of the time and
poor coverage the rest of the time. MODAS
data on the other hand, indicates that coverage percentage will not be excellent anywhere
but the expected values will be uniform over
the whole shallow depth band region.
6.1. Output Distributions
The output provided by NUWC from
the WAPP runs consisted of twelve different percentage coverage groups, three depth
bands times two scenarios times the two different types of input data. For the non-SVP
derived WAPP inputs, consistent values were
used throughout the runs to ensure that any
difference in the outputs was a result of differences in the GDEM and MODAS data.
For each of the groups, basic statistics such
as mean, maximum, minimum, and standard deviation were computed and then the
data were constructed into histograms to give
a visual representation of how the data are
distributed.
In the shallow depth band ASUW scenario both MODAS and GDEM yield
mean coverage percentages that are very close
to each other. While statistically the means
are different, in real world applications a few
percentage points difference is negligible (Fig.
10). From a user’s standpoint, this means
that both sets of data predict about the same
20
FIGURE 10
Shallow depth band coverage percentage distributions: upper panels for GDEM and lower panels for MODAS,
left panels for ASUW scenario and right panels for ASW scenario.
FIGURE 11
Medium depth band coverage percentage distributions: upper panels for GDEM and lower panels for MODAS,
left panels for ASUW scenario and right panels for ASW scenario.
scenario has the larger predicted values, but
the values in the ASW scenario are still on
the upper end of what is normal. The
MODAS data predicts performance that is,
while not particularly bad, still much more
pessimistic than the GDEM predictions.
For both scenarios the means of the
GDEM and MODAS derived predictions
are statistically different with the MODAS
data providing the smaller mean in both scenarios. Although the dispersion of the
GDEM data is large in both scenarios, the
data is so heavily weighted towards the upper end that low GDEM coverage percentages are average values for the MODAS data
coverage percentages. The MODAS data
coverage percentages are once again tightly
grouped; the uniformity of the predicted coverage percentages observed in the two other
depth bands extends from the surface down
to the selected maximum operating depth.
6.2. Difference of MK48 Acoustic
Presets Using GDEM and MODAS
The mid depth band yielded results that
were similar in distribution to the shallow
depth band (Fig. 11). Across both scenarios
the mean coverage of the GDEM data and
the mean coverage of the MODAS data are
statistically identical. Outliers are once again
observed in the GDEM data, the larger outlier in the ASUW scenario, and the greater
number of outliers in the ASW scenario. The
wide dispersion of the GDEM derived coverage indicates that weapon effectiveness will
vary depending on location. This is similar
to the predictions for the shallow depth band
and would indicate that GDEM predicts a
water column that has varying coverage values depending on horizontal and vertical
location. MODAS data once again indicates
an overall performance in the region that is
slightly less than the GDEM prediction;
however, the MODAS data is grouped even
more tightly than in the shallow depth band.
The coverage in the ASW scenario in particular varies little about the mean value. This
and the shallow depth band predictions indicate uniform coverage can be expected
even at some depth.
In the deep depth band the graphs take
on a slightly different shape, but they con-
vey much the same meaning (Fig. 12). In
both scenarios the GDEM graphs are
weighted heavily to the right end, predicting
that in the deep depth band coverage will be
very good over most of the area. The ASUW
For the shallow depth band, the RMSD
in the percentage coverage area was small
over most of the AOI, consistent with the
similar means and range of values noted in
the previous section (Fig. 13). The areas
FIGURE 12
Deep depth band coverage percentage distributions: upper panels for GDEM and lower panels for MODAS, left
panels for ASUW scenario and right panels for ASW scenario.
Spring 2004
Volume 38, Number 1
21
computed to have small RMSD coverage
percentages also had small RMSD in temperature and salinity. In the region where
the RMSD in temperature and salinity was
largest, though, a large RMSD in percentage coverage is also observed. These larger
values are likely areas where the GDEM
data generates overly optimistic coverage
percentage predictions. For the surface scenario, RMSDs of up to 25% are shown in
the region around 39° N 73° W, and the
warm salty intrusion observed on the
MODAS data coincides with a second peak
in the northeastern section of the graph.
Overall the ASW scenario shows RMSDs
that are similar to the ASUW scenario, the
only difference being that the values are,
on average, slightly smaller. The notable
exception is the peak located at the top
portion of the graph.
For the mid depth band the percentage
coverage RMS difference for the ASUW scenario is simply a scaled down version of the
shallow depth band ASUW graph (Fig. 14).
This makes a great deal of sense considering
the fact that the coverage percentage distributions for the shallow and mid depth
ASUW scenarios were very similar. The real
difference is in the ASW scenario. The single
exceptional peak at the top of the previous
graph is gone and the observed differences
have become much smaller. Most of the
RMS differences for the mid depth ASW
scenario do not exceed 10%. This is probably due to nearly identical coverage per-
FIGURE 14
RMSD for medium depth band coverage: left panel for ASUW scenario and right panel for ASW scenario.
centage means from both data sets, and the
tighter grouping of the GDEM data coverage percentage predictions in the mid depth
band ASW scenario. The RMSD values are
small even in the areas where the temperature and salinity differences were observed
to be large, such as in the upper section of
the graph.
The RMSD observed in the deep depth
band scenarios are smaller than those of the
shallow depth band, but similar in magnitude to the mid depth band (Fig. 15). For
the ASUW scenario the RMSD peaks near
the northwestern corner of the AOI and then
decreases steadily in steps heading toward
the opposite corner. While the individual
FIGURE 13
RMSD for shallow depth band coverage: left panel for ASUW scenario and right panel for ASW scenario.
22
RMSD values seen are not as large as some
of the ones in the other depth bands, more
of the area has a non-negligible RMSD. The
cause of this can be seen from the percentage coverage distribution for the deep depth
ASUW scenario.
The GDEM data results in values that
are almost all larger than the largest
MODAS derived values. This overly optimistic prediction means that over a large
portion of the AOI, the RMSD is going to
be non-zero. The RMS difference in the
ASW scenario changes very little from the
mid depth band save for the fact that the
values in the lower right corner are smaller.
The coverage distributions for the deep
ASW scenario are similar to the ASUW
case, but the separation between the two
means is not so pronounced. The result is
a larger region where the RMSD is small
or zero.
Both of these graphs match the pattern
that has so far been observed in the other
depth bands. The ASUW scenario has the
higher RMS difference values, with areas of
both high temperature and salinity differences corresponding to peaks on the graphs.
The RMS difference values also approach
zero moving toward the top left or bottom
right corners. Also, as depth increases, the
difference between the two data sets decreases causing the difference between the
coverage percentages to decrease.
FIGURE 15
Acknowledgments
RMSD for deep depth band coverage: left panel for ASUW scenario and right panel for ASW scenario.
This work was jointly supported by the
Space and Naval Warfare System Command
and the Naval Postgraduate School. The
authors would like to thank Charlie Barron
at the Naval Oceanographic Office for providing MODAS data set.
References
Chu, P.C., E. Gottshall and T.E.
Halwachs.1998. Environmental effects on
Naval warfare simulations. Institute of Joint
Warfare Analysis. Naval Postgraduate School.
Technical Report, NPS-IJWA-98-006. 33 pp.
Chu, P.C., S.K. Wells, S.D. Haeger, C.
Szczechowski and M. Carron. 1997. A
parametric model for the Yellow Sea thermal
variability. J Geophys Res. 102:5655-5668.
Conclusions
By looking at the RMSD in the temperature and salinity fields generated from
the GDEM and MODAS data, it is possible to look for areas where the data differ
significantly. It is at these points that the difference in the preset effectiveness should be
the greatest. This is observed for both scenarios at all depth bands. The percentage
coverage is the most different at points where
both the temperature and salinity RMS differences are large. This is especially true for
the shallow depth band where differences
of 25% are observed for both scenarios. It is
of interest to note that even at the surface
the RMS differences for the temperature and
salinity are never more than a few degrees or
PSU. Even with only this slight increase in
the accuracy of the inputs, a large increase
in the accuracy of the prediction of the
weapon effectiveness occurred. This seems
to imply that the sensitivity of the presets to
changes in the inputs is quite high.
From the output distributions it becomes
clear that the GDEM derived coverage percentages indicate that weapon effectiveness
should vary not only in the horizontal but
also in the vertical. The implication is that
in some areas coverage will be very high and
in others the coverage will be very poor, but
the tendency is for the coverage to be high
for any given area. The MODAS derived
percentages reveal that the exact opposite is
true. The coverage will be consistent no
matter what the horizontal location or depth
band. This is an important result since prediction of weapon effectiveness is vital to
mission planning and execution. In this case
an unrealistic expectation in the weapons
effectiveness would have resulted from the
use of the GDEM data to predict the coverage percentages in the water column. The
MODAS data also would have given the user
the freedom to operate anywhere in the region knowing that their weapon would function about the same no matter the location.
The most obvious limitation of this work
was the limited data set. Any future work
should include data that covered a wider
number of areas and times. Areas of strong
thermal and salinity contrast are of particular interest. Various combinations of the user
inputs into the WAPP should also be studied. The effects of variables such as bottom
type and position (upslope/downslope) need
to be addressed. Another avenue of study is
the determination of how the number of
altimeters affects the accuracy of the outputs. It has been determined that the presets are sensitive to the addition of satellite
data. However, the effect of the number of
satellite inputs still remains to be determined.
Once this is done an optimal number of altimeters can be determined based on minimizing cost and maximizing preset accuracy.
Chu, P.C., Q.Q. Wang and R.H. Bourke.
1999. A geometric model for Beaufort/
Chukchi Sea thermohaline structure. J Atmos
Oceanic Technol. 16:613-632.
Fox, D.N., W.J Teague, C.N Barron, M.R.
Carnes and C.M. Lee. 2002. The Modular
Ocean Data Assimilation System (MODAS).
J Atmos Oceanic Technol. 19:240-252.
Gandin, L.S. 1965. Objective analysis of
meteorological fields. Israel Program for
Scientific Translation. 242 pp.
Gottshall, E. 1997. Environmental Effects on
Naval warfare simulations. MS Thesis in
Physical Oceanography, Naval Postgraduate
School, (Monterey). 92 pp.
Jacobs, G.A., C.N. Barron, M.R. Carnes,
D.N. Fox, H.E. Hurlburt, P. Pistek, R.C.
Rhodes, W.J. Teague, J.P. Blaha, R. Crout,
O.M. Smedstad and K.R. Whitmer. 1999.
Naval Research Laboratory Report NRL/FR/
7320-99-9696. Navy Altimeter Data
Requirements. 25 pp.
Teague, W.J., M.J. Carron and P.J. Hogan.
1990. A Comparison between the Generalized
Digital Environmental Model and Levitus
Climatologies. J Geophys Res. 95:7167-7183.
Spring 2004
Volume 38, Number 1
23
PAPER
Use of Expert Systems for Optimum Maintenance
of Marine Power Plants
AUTHOR
ABSTRACT
K.D.H. Bob-Manuel
Rivers State University of Science
and Technology
Port Harcourt, Nigeria
The reliability and economical operation of marine power plants depend upon the
design quality, capability, and skill of the technical staff that operate the plant. The normal
breakdowns of marine machinery and the frequency of planned maintenance are based
mainly on the subjective judgement of the operators who ensure that such breakdowns
are minimized. To achieve such objectives, an optimum maintenance goal must be adopted
using various types of computer software for expert systems with high processing speed,
which have been developed to aid fault or failure diagnosis. Such systems are at different
stages of application. In this paper, a typical expert system for condition monitoring,
budgeting, and spare part management to enhance the optimum management of a marine
power plant is presented. The use of this technique with known models will substantially
reduce downtime and fully utilize the technical crew onboard and ashore. From the author’s
experience in the management of ferries, there is an optimum amount of maintenance
effort for any given condition.
INTRODUCTION
M
arine power plants are in a period
of intensive research and development. Energy conservation and manning are some of
the factors contributing to escalating operational cost and forcing rapid changes in long
established design and operational practices.
The unmanned engine rooms concept and
automated control systems of power plants
are being adopted by most shipping companies having large fleets of vessels. These
new and changing requirements have greatly
increased the complexity of engine design
to give optimum operating characteristics
within given limits.
It is widely recognized that maintenance cost
is the most controllable element in ship operation. The maintenance of power plants depends
on the ease with which a plant can be stripped
down and repaired, the skill and number of the
technical crew, and the cost of the components
(Inozu and Karabakal, 1993). Hence, ship operators believe that the potential saving in an
optimized maintenance system can be substantial. The saving also depends on the maintenance technique and the level of interaction
among the working party. The use of condition monitoring systems for the main engines
of ships and equipment on offshore rigs is now
very common. It is now possible to transmit
24
information on marine equipment from computers in control rooms on board a ship or offshore rig to control departments at shore base
using an interactive system (Goto and Kaibara,
1993). This has been made possible by the remarkable development of microelectronics and
artificial intelligence. The economic use of this
system depends on the size of the fleet and the
skill of the technical staff.
This paper presents the merits and weaknesses of three different maintenance policies:
(1) a breakdown/replacement, (2) preventive
(planned) only, and (3) optimal (conditionbased and predictive) maintenance schemes.
Maintenance Policy
The normal breakdown and planned
maintenance systems are based solely on
subjective judgement, which ensures that
breakdowns are kept to a minimum. The
most economical system will be one that has
a better balance between the breakdown and
the planned maintenance systems.
Breakdown/Repair-only
Maintenance Policy
A breakdown/replacement policy considers a component that has suffered a major
failure and will cost a substantial amount to
repair or replace. The repair or replacement
of a component after a major breakdown of
a power plant can be defined as maintenance
work which involves the rectification of a
failed component that is brought back to
working condition by repairing or replacing
the damaged part. During the major repairs
or replacement, checks and other minor repairs would be carried out. However, it
would not be very obvious if other components which are in fairly good working condition should be replaced or not at the same
time. A very crucial decision has to be made
if a component is worn to an extent that it
either becomes ineffective and causes trouble
in the near future or the part in question
costs little to replace. A wise decision is to
replace the component immediately to avoid
an unexpected major breakdown.
Replacement of Worn Components
During a major overhaul, replacement
and repair of failed or worn components
have to be carried out by competent crew or
given out on contract to the manufacturers.
It is always a wise decision to allow the manufacturers to install a major component to
attract a guarantee on it. Whether to replace
a worn component immediately to avoid
catastrophic failure or defer it until the next
maintenance schedule is usually a very difficult decision to make. If the ship owners
decide to wait until a significant portion of
the component is worn out to save cost, they
will be exposed to the cost associated with
decreased performance and material degradation as well as unanticipated breakdown.
This could put the safety of personnel and
other equipment at risk in addition to the
cost and replacement of the component. For
example, if during a routine maintenance
and inspection, a component was found to
be cracked or worn, the possible decision
concerning the component, which is still
serviceable, will be one of the following:
■ Replace the component immediately.
■ Wait until the component finally fails.
■ Wait and replace the component at the
next scheduled maintenance.
The cost of immediate replacement or
waiting until the component finally fails entails the cost of a new component, and labour
in stripping the old component and fitting
the new one. The decision to wait until the
next scheduled maintenance imposes some
level of uncertainty since the component may
fail before the scheduled period.
FIGURE 1
Weibull curve and probability of failure (Shields et al., 1995 )
Assuming that the fraction of the old
component life that has undergone a certain
period of usage is xn, the remaining life of
the component at the time of replacement is
(1 – xn). The failure rate during this period
can be determined by the component design factor and the operational condition. An
increased rate of failure should be expected
during the wear-out of the component.
Remnant Life Assessment
Using a normal theoretical Weibull distribution curve shown in Fig.1 can assess the
useful life remaining in an old component
before replacement. In this paper, the distribution will consider failure due to wear-out
only. The height of the Weibull curve is defined (Shields et al., 1995) as:
(1)
where
β > 0 determines the shape of the
distribution.
η > 0 is the characteristic life of the
component.
t0
determines the origin of the
distribution time factor.
Preventive (Planned)
Maintenance Policy
In the preventive only or planned maintenance policy, work is usually carried out
to cut down the amount of downtime by
observing wear in the component and conducting condition monitoring to forestall
unexpected breakdown. A preventive or
planned maintenance policy is also known
as a calendar-based or scheduled maintenance system. In this case, maintenance is
programmed with an inspection system
based on either plant operating hours or calendar interval regardless of the duty cycle of
the component or its condition. The maintenance intervals are almost conservative as
stipulated by the manufacturers. With this
policy, wear in a component is frequently
observed and replacement made to forestall
unexpected failure. The engineer may not
foresee the actual service conditions and so
(s)he operates the policy with the unexpected
failure that may occur due to design fault,
which cannot be overruled.
A large portion of preventive maintenance can be planned to coincide with the
period when a ship is in port or when repair work is being carried out on other
components. It could also be done when
components are opened up for survey by
classification society. However, the merit of
the policy depends on the man-hours spent
for inspection and the reduction in the frequency of breakdown.
The risk of not keeping to a scheduled
maintenance period can be observed in Table
1. It shows that most of the parameters
marked (*) are off specification. This was as
a result of the tugboats not being taken in
for normal planned maintenance at the
specified period due to operational demand.
When eventually the engine in tugboat C
was dismantled for servicing, it was found
that the piston rings were excessively worn
with a burnt cylinder head gasket and the
big-end bearings were found to be worn.
One major drawback of this policy is that
it assumes that all engine hours are equal,
yet factors such as running speed and load,
fuel quality, ambient condition, etc. which
have influence on the wear rate of the engine or accelerate the deterioration of the
Spring 2004
Volume 38, Number 1
25
TABLE 1
Analysis of lubricating engine oil from 3 (three) tugboats using Spectrometric Oil Analysis Procedure
(SOAP) (Chugbo, 1999)
PARAMETERS
Density @15
LUB-OIL SPEC.
TUGBOAT A
TUGBOAT B
TUGBOAT C
0.916
0.9253*
0.9026*
0.8887*
Nil
0.40
0.15
—
182.0
150.23*
125.23*
68.45*
14.4
33.49*
13.21*
8.95*
16.0
Too dark to observe
2.56
0.43
1.09
0.46
ˆC(kg/l)
Water Content (% vol.)
Kinematic Viscosity
@ 40ˆC (cst)
Kinematic Viscosity
@ 100ˆC (cst)
Total acid No.
(Mg KOH/gr)
Sulphated Ash % Wt
colour change
2.0
1.73
* Values are off specification
lubricating oil are not considered. However,
excessive maintenance is not cost effective
since engine downtime and the risk of exposure of the engine to maintenance-induced failure would increase.
Optimum Maintenance Policy
The procedures proffered in breakdown/repair only or preventive (planned)
maintenance policy cannot provide the solution to minimize overall maintenance
cost. The policy which gives an optimal
solution is one with a specified degree of
planned and breakdown/replacement maintenance policies as illustrated in Fig.2 and
is expected to result in the minimum cost
of maintenance and downtime. The determination of the level of preventive (planned)
maintenance requires mathematically oriented analysis and analytical ability. The
application in practice can be involved and
complex requiring the use of competent
skilled engineers. Another option is to introduce an integrated maintenance policy
that capitalizes on the advantages of the optimum policy while minimizing the disadvantages of other policies earlier discussed
(Haller and Kelleher, 1999). This policy has
been proven to identify component degradation before failure occurs.
FIGURE 2
Constituent and total maintenance cost for breakdown and planned maintenance policies (Shields et al., 1995)
26
Use of an Expert System in
Fault Diagnosis
The conventional method of fault diagnosis of a running engine is the detection of
the change in performance using the sense
of hearing, sound detection rod, lubricating
oil analysis, etc. The accuracy of detection
depends upon the experts’ skill and experience, hence the use of expert systems. Vibration measurement is the most widely used
method for fault diagnosis of rotating and
reciprocating machineries. The principle is
based on structure-borne noise, which is a
high frequency vibration caused by rotating
and reciprocating forces that occur at microscopically small unsteady points. The
general state of a plant will alter structureborne noise characteristics as a result of deterioration of the components.
Fig. 3 shows spectrums of vibration levels at different frequencies from a bearing
housing of an electricity generating plant
(Bob-Manuel, 1999). The result can be
scanned to computer memory for diagnosis
when displayed on the visual display unit
(VDU). Measurements of the vibration level
are usually taken by placing an accelerometer on the bearing housing at different positions to find the maximum amplitude of
vibration level.
Frequency analysis of acoustic signals obtained for normal and faulty components
during operation as displayed in Fig. 4 can
be collected from the exhaust tail pipe to determine the waveform in different load and
speed conditions. Linear regression by means
of the least square method can be applied to
each region of the signal and the decay or
the rise of the amplitude approximated by
straight lines to determine the characteristic
tendency of the fault (Hikima et al., 1993).
In diagnosing a component that has
failed using an expert system, the computer
will be programmed to make the judgement.
In Fig. 5, the input variable would determine the operational condition of the component. Comparing the actual and the reference output variables is usually the starting point for diagnosis. However, there
would be an engineer–computer interaction
i.e. the diagnosis of the failed component
FIGURE 3
Computer display of vibration spectrum for diagnosis (Bob-Manuel, 1999)
being displayed on the VDU of the computer
and showing the functional content of the
component. Possible causes of the failure and
the history of the unit will be displayed. On
the diagnostic item menu, the results of the
measured values that indicate eminent failure which are marked with codes would be
displayed. The result and solution guidance
can be printed for detailed study. With the
present communication system, if failure occurred on board a ship or offshore rig and the
replacement is required urgently, the result of
the diagnosis can be sent directly to management at shore-base through the Internet.
The engine signature analysis in an expert system involves recording various engine-operating signatures during loaded condition. This includes combustion pressure,
cylinder vibration, lubricating oil analysis,
oil and turbocharger boost pressure, oil and
air filter pressures drops, excessive exhaust
temperatures, etc. Component failure does
not usually exhibit symptoms, which are
readily apparent in engine performance until significant damage has occurred. The author observed on vibration analysis that engine performance did not deteriorate for several hours of operation, yet from vibration
FIGURE 4
Spectrum of acoustic noise levels taken at exhaust pipes (Hikima et al., 1993)
Spring 2004
Volume 38, Number 1
27
FIGURE 5
Component model for monitoring and diagnosis
signature analysis, fault was identified (BobManuel, 1999). When such data are diagnosed from a vibration test, the computer
can be programmed to predict when certain performances parameters or wear limit
will be exceeded.
Illustrated on the left side of Fig. 6 is an
ultrasonic signature taken when unexpected
noise occurred during compression and
power stroke of a four-stroke diesel engine.
It was suspected that blow-by gases from the
cylinder were responsible for the trends that
are circled. After the cylinder liner and piston rings were changed, the blow-by gases
ceased as confirmed by the signature shown
on the right side of the figure. Subsequent
inspection of the rings and liners confirmed
that excess liner wear caused the blow-by
(Haller and Kelleher, 1999). It is important
to note from the pressure diagram that firing pressure has not changed since the blowby has not yet impaired performance, yet
the vibration signature analysis has identified the excess wear.
FIGURE 6
Ultrasonic signature before and after cylinder liner replacement for a medium-speed diesel engine
(Haller and Kelleher, 1999)
28
Spare Part and
Store Management
Expenditure on spare parts for marine
power plants accounts for a substantial
amount of the vessel’s annual operating costs.
One vital problem is that unavailability of
one relatively low cost component can prevent the sailing of a ship for a considerable
period thereby increasing operational cost.
Therefore, the effective management of spare
parts in any maintenance scheme is of paramount importance. Utilization patterns will
vary because the demand may be fast or slow.
In most cases, an unexpected demand
for spare parts not in stock can be made but
if enough notice is not given the cost arising
from downtime can be exorbitant. The procurement section should not be shortsighted,
as it is often the case when considering the
cost of spare parts. Factors such as invoiced
cost, availability, quality, transportation, and
personnel cost have to be taken into account.
Correct decisions have to be made to enable
parts to be supplied at the lowest possible
cost. The storekeeper faces the problem of
knowing when spare parts would be needed
and the availability of funds to execute the
order. Hence, there is a need to define groups
of spares in stock and evolve codes for all
equipment and spares. The problem of
which stock should be carried at any particular time is complex and requires statistics of the utilization of spares. Standard soft-
ware for the management of spares exists and
has facilities for labeling individual spare
parts for easy identification. It may also be
necessary to scan essential drawings of spares
along with part reference numbers for storage in the computer memory.
Budgeting for
Maintenance Cost
Effective control of maintenance cost
involves adequate feedback by an efficient
management structure. Shore-based and
shipboard management staffs require effective coordination on budget control. A consultation process, which could lead to far
reaching decisions, should be instituted. For
shipboard management in particular, head
office has to be committed to the principle
that the technical crew onboard have an
opportunity to contribute significantly to the
maintenance policy, procedure, and regulation under which they operate. Such a system must be established before the controlled events begin to occur. All relevant
events must be reported and compared with
the approved plan and budget.
One of the basic essentials for preparing
a meaningful account is a good ordering system. Such a system should enable tracking
of cash flows and commitments of the budget to be kept. Budgeting should take account of technical work that falls due at the
scheduled planned maintenance period and
a contingency allowance made for breakdowns based on past experience. In an expedient planning and budgeting system, the
chief engineer onboard is expected to itemize all work that needs to be done at a specified time during the year, taking cognizance
of the annual survey period. (S)He does not
decide alone the time and place for dry-docking and even the planned maintenance period can be affected by management decision. The cost of the system can be appraised
on a yearly basis and updated. The reduction in the cost of optimum maintenance
and the repair cost can take care of the cost
of the diagnosis system. Therefore, effective
control of budgeting is a collective responsibility of shipboard and management staffs.
Conclusion
The economical operation of marine
vessels can be realized if there are ingeniously
skilled manpower and reliable power plants
to minimize maintenance cost. The evolution of computer technology with rapid development of expert systems and condition
monitoring over the past years has enabled
optimal maintenance policies to be adopted
for marine power plants. This has resulted
in the reduction of maintenance cost rather
than providing alarm signals only when a
condition has exceeded set limit for safety.
An efficient diagnostic system has the
potential to aid ship operators to utilize a cost
effective maintenance policy for marine
power plants and facilitates a clear presentation and storage of crucial data. The implementation of an expert system to achieve
optimal maintenance should be seen as a
challenge and an opportunity for engineers
to use technical and computer skills with
current scientific and management principles.
Hikima, T., T. Katagi, T. Naka, N. Ohyama
and T. Hashimoto. 1993. Diagnostic of marine
diesel engine faults by pattern recognition of
acoustic sound. Paper 16. ICMES 93 I.
Mar.E. Conference on Marine System Design
and Operation. pp.16.1-16.9.
Goto, T. and M. Kaibara. 1993. Remote
maintenance system by a programmed expert
knowledge network between ship and land.
Paper 29 ICMES 93, Conference on Marine
System Design and Operation. pp. 1- 29.11.
Inozu, B. N. and Karabakal. 1993. Replacement and maintenance optimization of marine
system under budget constraints. Paper 30
ICMES 93, Conference on Marine system
Design and operation. pp. 1-30.7.
Shields, S., K. J. Sparshott and E. A. Cameron.
1995. Ship maintenance—a qualitative
approach. I.Mar.E. London: Marine Media
Management Ltd.
References
Bob-Manuel, K.D.H., H.I. Hart, and E. A.
Ogbonnaya. 1999. Computer-based condition
monitoring of an electricity generating plant.
I.Mar.E. Conference on Computers and Ships
- From Ship Design and Build, through
Automation and Management and on to
Support, pp. 87-95.
Chugbo, J.O. 1999. Project report on
condition monitoring of pusher tugs main
engines used for creek operations. Department
of Marine Engineering, Rivers State University
of Science & Technology, Port Harcourt,
Nigeria, p.11.
Haller, C.L. and E. P. Kelleher. 1999. Practical
integrated maintenance and diagnostic for
medium and slow speed diesel engines.
I.Mar.E. Conference on Computers and Ships
from Ship Design and Build, through
Automation and Management and on to
Support, pp. 103-128.
Spring 2004
Volume 38, Number 1
29
PAPER
The Science and Technology of Nonexplosive
Severance Techniques
AUTHORS
ABSTRACT
Mark J. Kaiser
Allan G. Pulsipher
Center for Energy Studies,
Louisiana State University
A variety of cutting technology is used in support of decommissioning offshore structures in the Gulf of Mexico. The purpose of this paper is to describe the science and technology of nonexplosive severance techniques, and to examine their environmental, physical,
safety, and activity requirements. Abrasive water jet, diamond wire, diver torch, mechanical
methods, and sand cutters are the primary nonexplosive cutting techniques applied in the
Gulf of Mexico. The technology of nonexplosive removal techniques has not changed dramatically over the past decade, but technological progress continues to be made, most
notably in abrasive water jet and diamond wire technology. A review of each nonexplosive
cutting technique will be described within and across each stage of decommissioning.
Robert C. Byrd
Twachtman Synder & Byrd, Inc.
INTRODUCTION
T
here are currently around 4000 or so
structures in the federally regulated offshore
waters of the Gulf of Mexico (GOM) associated with oil and gas production. The structures vary widely according to function and
configuration type and since 1947, when offshore production in the Gulf first began,
roughly 6,000 structures have been installed
in the federal offshore waters and 2,000 structures have been removed. On average, about
100 or so structures are removed in the
GOM per year, and over the past 10 years,
the number of removals have ranged between
79 and 168 (Kaiser et al, 2002).
Structures are installed to produce hydrocarbons and when the time arrives that
the cost to operate a structure (maintenance,
operating personnel, transportation, fuel,
etc.) outstrips the income from the hydrocarbons under production, the structure exists as a liability instead of an asset. The economic lifetime of a structure can be extended
if the operating cost can be reduced (by divestment, unitization, or more efficient production practices) or hydrocarbon throughput increased (by investment). In the GOM,
federal regulations require that all structures
on a lease be removed within one year after
the lease is terminated. Typically, a lease is
terminated when production on the lease
ceases, but if the operator intends to re-work
wells or is pursuing activity on the lease, or
30
the lease contains an active pipeline, conditions may warrant granting an extension of
the lease termination.
Decommissioning activities in the GOM
are driven by economics and technological
requirements and governed by federal regulation. Decisions about when and how a
structure is decommissioned involve issues
of environmental protection, safety, cost, and
strategic opportunity, and the factors that
influence the timing of removal as well as
the manner in which the structure is severed from the seabed are complicated and
depend as much on the technical requirements and cost as on the preferences established by the contractor and the scheduling
of the operation.
Decommissioning occurs in stages and
typically over disjoint time frames as illustrated in Figure 1. Greatly simplified, following project engineering and cost assessment, federal and state regulatory permits
for well plugging and abandonment, pipeline abandonment, structure removal, and
site clearance verification must first be obtained. Wells are plugged and the facility is
prepared for removal, including flushing and
cleaning process components, installing
padeyes, etc. Pipelines are pigged and/or
flushed riser-to-riser or riser-to-subsea tiein, detached from the structure, capped, and
normally left in place. The topside facilities
are prepared by removing all traces of hy-
drocarbon, and then the deck is cut and removed, and the conductors and piles cut and
pulled. Heavy lift vessels bring the jacket
ashore for recycling, sale, or scrap, or the
operator may participate in a reefing program. After the jacket has been removed, the
site is cleared with a trawling vessel or divers
deployed with scanning sonar, and then
clearance is verified with a trawler.
For readers requiring more specific information on the activities involved throughout
a decommissioning project, the case studies
in Hakam and Thornton (2000), Kirby
(1999), and Thornton (1989) are a good starting point. (See also Dodson, 2001, and
Twachtman et al., 1995.) Detailed descriptions of each stage of the decommissioning
process can be found in Byrd and Velazquez
(2001), Manago and Williamson (1997),
Pulsipher (1996), Twachtman et al. (2000),
and National Research Council (1996).
The decommissioning of offshore structures is often a severing intensive operation.
Cutting is required throughout the structure
above and below the waterline and mudline
on braces, pipelines, risers, umbilicals, manifolds, templates, guideposts, chains, deck
equipment and modules. More significant
cutting operations are required on elements that
are driven into the seafloor, such as multi-string
conductors, piling, skirt piling, and stubs which
need to be cut at least 15 feet below the mudline, pulled, and removed from the seabed.
FIGURE 1
Decommissioning is a Severing Intensive Operation
A variety of technology exists to perform
severance operations. These include abrasive
water jet, diamond wire, diver torch, explosive charges, mechanical methods, and sand
cutters. For severing operations that occur
above the waterline, the cutting technique
selected is usually dictated by the potential
for an explosion. Cold cut methods are used
when the potential for an explosion exists;
otherwise hot cuts are employed. Cutting
in the air zone is considered conventional
since it involves methods that are regularly
used for dismantling onshore industrial facilities. Below the waterline cutting is more
specialized. Divers can perform cuts on
simple elements such as braces and pipeline,
and for shallow water structures such as caissons, diver-cutting is often the preferred severance method. In deep water, remotely
operated vehicles and automated diving systems are sometimes deployed with abrasive
and diamond wire cutters and often with
explosive charges. Major cutting operations
on conductors, piling, and stubs normally
employ mechanical, abrasive water jet, and
explosive charges. Mechanical and explosive
methods are primarily used for conductors
with abrasive water jet and explosives predominately used for pile severance.
A large number of factors are potentially
involved in selecting the severance technique
for a specific job with cost, safety, risk of
failure, and technical feasibility the primary
factors that are considered when alternative
options are available. Many different severance operations are required during decommissioning, and depending upon the job,
more than one alternative may be available.
Variables that drive the cost and risk associated with a specific severance technique are
numerous and involve factors such as the
location and nature of the site, sensitivity of
the marine habitat, structural characteristics,
the amount of pre-planning involved and
the schedule of the operation, salvage/reuse
decisions of the operator, marine equipment
availability, operator experience and preference, contractor experience and preference,
the number of jobs the contractor is scheduled to perform, the weather at the time of
the procedure, and market conditions. In
general, cutting techniques are expected to
Spring 2004
Volume 38, Number 1
31
be reliable, flexible, adaptable, safe, cost effective and environmentally sensitive (National Research Council, 1996). If a cutting
technique fails with respect to one or more
of these factors, or if an operator has more
than one “bad experience” with a particular
method, then chances are that the technology will gain neither in popularity or acceptance among contractors.
The purpose of this paper is to review
the various nonexplosive cutting techniques
used in support of offshore decommissioning, and to describe the technological, environmental, physical, safety, and activity requirements associated with each method.
The outline of the paper is as follows. In
section 2, the cutting activities that occur
across the main stages of decommissioning
are outlined, and in section 3, the science
and technology of mechanical, abrasive water jet, diamond wire, and diver torch methods is described. In section 4, the cutting
systems and activity requirements associated
with each method are discussed, and in section 5, the paper concludes with a brief summary of the environmental and physical
impact and safety issues associated with nonexplosive cutting technology.
2. Decommissioning is Often
a Severing Intensive Operation
The basic aim of an offshore platform
decommissioning project is to render all wells
permanently safe and remove surface/seabed
signs of production activity. A site should be
returned to such a state as to allow the use of
the site by other marine users, such as commercial fishermen and shrimpers. Cutting
operations occur throughout each stage of
decommissioning except the first (permitting)
and last (site clearance and verification) stage.
2.1. Well Plugging and Abandonment
A well abandonment program is carried
out by injecting cement plugs downhole to
seal the wellbore to secure it from future leakage. The purpose of plug and abandonment
(P&A) is to destroy the permeability within
the formation and stabilize the wellbore and
its associated annuli until geologic forces can
re-establish the natural barriers that existed
32
before the well was drilled. Techniques associated with P&A are based on industry
experience, research, and conformance with
regulatory standards and requirements
(Englehardt et al., 2001; Manago and
Williamson, 1997). Federal regulations
specify the minimum requirements that
must be performed in P&A operations.
A traditional approach begins by “killing” the well with drilling fluids heavy
enough to contain any open formation pressures. The Christmas tree is then removed
and replaced by a blowout preventer through
which the production tubing is removed.
Cement is placed across the open perforations and squeezed into the formation to seal
off all production intervals and protect aquifers. The production casing is then cut and
removed above the top of the cement and a
cement plug positioned over the casing stub.
The remaining casing strings are then cut
and removed close to the surface and a cement plug set across the casing stubs.
Mechanical methods of cutting and sand
cutters are primarily associated with P&A
activities. After wells are plugged and casing
tubing cut and pulled, a sand cutter or mechanical cutting tool may be run downhole
to cut the conductors, or depending on the
preference of the operator/contractor and
configuration of the platform, abrasive or explosive severance methods may be contracted.
In a typical mechanical operation, the tubing
and production casing is first cut using a jet
cutter—a small explosive blast that utilizes
less than 5 pounds of explosive—and then
the strings are cut using a mechanical cutter.
The general philosophy during decommissioning is to get as much cutting done as possible off the critical work path and before the
arrival of the derrick barge if the activity can
be performed in a cost effective manner.
All well conductors and casings are required to be removed to a depth of at least
15 feet below the mudline, or to a depth
approved by the District Supervisor. The
District Supervisor may approve an alternate
removal depth if (1) the wellhead or casing
would not become an obstruction to other
users of the seafloor or area, and geotechnical
and other information demonstrate that erosional processes capable of exposing the ob-
struction are not expected; (2) the use of
divers and seafloor stability pose safety concerns; or (3) the water depth is greater than
800 m (2,624 ft). The requirement for removing deepwater subsea wellheads or other
obstructions may be reduced or eliminated
when, in the opinion of the District Supervisor, the wellheads would not constitute a
hazard to other users of the seafloor.
2.2. Topside Equipment and
Deck Preparation
Topside preparation and deck removal
is made with cold and hot cuts. Cold cuts
are generally made with pneumatic saws,
diamond wire methods or abrasive techniques. Hot cuts—torch cutting and arc
gouging—are used to cut steel when there
is no risk of explosion. Arc gouging, essentially an arc welder, is used to remove welds
between steel connections. Diamond wire
methods have occasionally been employed
in the GOM to cut the deck from the jacket.
2.3. Jacket Preparation
Several severing methods are used below
the waterline. Underwater burning torches
work on the same principle as an arc-gouger,
where a burning rod, usually magnesium, is
arced with the member to be cut. Cuts made
to the jacket bracing and trimming, flowlines,
umbilicals, and manifolds are typically performed with divers using burning torches, hydraulic saws, or abrasive technologies. Intermediate cuts may be required to separate the
jacket into vertical sections if the lift weights
exceed the capacity of the derrick barge.
2.4. Pipeline Abandonment
Federal regulations allow decommissioned OCS pipelines to be left in place
when they do not constitute a hazard to
navigation, commercial fishing, or other uses
of the OCS. Pipelines will generally be removed offshore through the surf zone and
capped. Onshore pipelines may be removed
completely, or some sections may be abandoned in place if they transition through a
sensitive environment. The pipeline end seaward of the surf zone is capped with a steel
cap and jetted 3 feet below the mudline.
Most pipelines in the GOM are abandoned
in place after cleaning, cutting, capping and
burying the ends.
The methodology for cutting a pipeline
depends on the manner the pipeline is to be
recovered. The protective coatings typical of
most pipeline sections must first be removed
in order to cut the pipe with an arc torch. If a
pipeline crosses or is adjacent to an “active”
pipeline, chances are it will not be disturbed
due to the potential damage that could result
if complications arise in the removal. Diamond wire methods, abrasive water jet, and
pneumatic saws deployed with diver or ROV
have been used to cut pipeline.
2.5. Pile and Conductor Severing
Pile and conductor severing is the most
critical and typically the most expensive of
all the severance operations required on the
structure with direct cost generally falling
between 1-3% of the total cost to decommission a structure (Kaiser et al., 2003). The
indirect cost of pile and conductor cutting,
primarily the expected cost of failure to cut,
is an important criteria in selecting a severance method. Piles are steel tubes welded
together and driven through the legs of the
jacket and into the seabed to provide stability to the structure, while conductors conduct the oil and gas from the reservoir to
the surface. Piles and conductors must be
cut and removed 15 feet below the mudline
unless a special waiver is granted.
Conductor severing and recovery may be
completed as part of well plugging and abandonment activities unless the platform configuration, equipment availability or scheduling of the activities prevent the operation.
Conductors are cut and pulled, if possible,
early in the decommissioning process to
avoid delay when the barge is on-site. Depending on the number of structures to be
decommissioned, the type of structure and
the sequencing of the activities, a small spread
may be sent to pre-cut the conductors. This
saves derrick barge time if the conductors
are successfully severed, but also costs additional money to dispatch the cutting crew
and necessary support vessels. To verify a
complete cut, a jacking spread may be used
to lift the conductor after the severing attempt. To jack the conductors (“prove” the
cut), the platform must have the structural
capacity to provide a point to jack against
and have a crane large enough to set the cutter, jacks, and load spreading beams. Mechanical casing cutters and abrasive water jet
(AWJ) cutters can be used to perform the
cut if a crane is available on the platform for
the deployment of the tool. With a derrick
barge on-site, mechanical and AWJ cutters
are rarely deployed due to the time-consuming and inefficient nature of the operation.
The physical characteristics that describe
piles and conductors are also important since
they determine the technical feasibility of
severance options. Conductors are configured in various diameters and wall thickness
and are characterized by the number of inner casing strings, the location of the strings
relative to the conductor (eccentric vs. concentric), and the application of grout within
the annuli. Grouted annuli are usually easier
to cut than annuli with voids since voids
dissipate the energy/focus of the abrasive and
explosive cutting mechanisms. Conductors
are usually cut with mechanical or explosive
charges. Mechanical methods are applied
during P&A activity and if conductors are
cut when the barge is on-site, then explosive
charges will probably be employed.
To sever jacket legs and piles, abrasive
cutters and explosive techniques are effective. Cutting piles is usually a much simpler
operation than conductor cutting, and in
such cases, AWJ cutting may be used with a
derrick barge spread. In principle, mechanical cutting could be used to cut piling, but
in practice it is rarely used because piles are
only open when a barge is on-site, and with
a barge on-site, mechanical cutting is neither a cost-effective nor efficient way to sever.
Explosives are deployed down the piling and
below the mudline while abrasive cutters can
be deployed internally or mounted externally
if the soil is excavated from around the outside of the piling. Obstructions within the
pile will necessitate additional operation or
deployment of an external cutter. Internal
cutting is usually the preferred approach with
water jet technology since it does not require
the use of divers to set up the system or soil
jetting operations to access the required cutting depth below the mudline.
Cutting the piles and conductors is the
most critical and important part of a decommissioning project since if the piles and conductors are not cut properly, a potentially
dangerous condition could arise during the
lift. The cuts on jacket members, piles and
conductors must be “clean” and “complete”
to allow for a safe operation. A dangerous
situation occurs when an element is not completely cut (lease “hangers”) and “lets go”
after the crane vessel has applied a significant pulling force.
3. The Science & Technology
of Nonexplosive Removal
Methods
For the most part, the resources required
in decommissioning involve standard,
readily available technology and tools which
have been available for some time. Mechanical pipe cutters and diver torches are roughly
the same today as they were ten years ago,
and while AWJ and diamond wire applications have seen a moderate increase in the
frequency of usage, a variety of factors continue to limit their application. The physical limitations associated with AWJ systems
(e.g., water depth, reliability) are significantly
better today than they were a decade ago,
and although the routine application of diamond wire methods is still several years away,
the technology has made strides due to product development and the steady influx of
new contractors into the GOM. The following discussion can be considered an update of the National Research Council’s 1996
report on cutting techniques (NRC, 1996).
3.1. Mechanical Methods
Cutting mechanisms that use hydraulically actuated, carbide-tipped tungsten
blades to mill through tubular structures are
called mechanical cutters. Figure 2 shows a
schematic of this type of system. The mechanical casing cutter is perhaps the oldest
method for cutting well conductors. The
casing cutter is deployed on a drill pipe string
and lowered into an open pile or well. The
cutting tool has 3 blades that fold up against
the drill pipe. When hydraulic pressure is
applied to the tool, the blades are forced
Spring 2004
Volume 38, Number 1
33
outward as the tool is rotated by a power
swivel. The carbide-tipped blades cut
through the strings of the well while centralizers on the tool keep it concentric inside the tubular member. Drillers watch the
pressure to determine when the cut is complete and cut indication can be made after
the tool is recovered by observing the marks
of penetration of the blades (Manago and
Williamson, 1997).
FIGURE 2
Schematic of a Mechanical Cutting System
(Courtesy of Hydrodynamic Cutting Systems)
Mechanical cutters are frequently used
for cutting shallow-water, small-diameter
caissons with individual wells and well protector platforms with vertical piles
(Pulsipher, 1996; NRC, 1996). Mechanical
cutting is rarely used in conjunction with a
derrick barge spread, however, since the operation is time-consuming and inefficient
and rig-up and rig-down time may be considerable. After wells are plugged and casing
tubing cut and pulled (see Fanguy, 2001),
the contractor may run a mechanical cutting tool (or sand cutter) downhole to cut
the conductors, or depending on the preference of the operator and configuration of
the platform, may subcontract for abrasive
water jet or explosive severance methods.
A number of limitations are associated
with mechanical cutting. For conductors
with casing string that is not cemented, lateral movement of the string may cause uneven cutting of the next casing. If strings are
pulled after each cut, lifting equipment is
required which adds to the time to remove
and reinstall the tool. For cemented strings,
trips in and out of the well may be required
to replace worn blades, which add to the
time to complete the cut. Realignment of a
partial cut after re-entry is problematic and
eccentricity of the casing strings may result
in incomplete cuts forcing the deployment
of divers to perform the operation. Mechanical cutting is also problematic for tubular
members at a batter (angle). To cut piling
with mechanical cutters, the piling must be
open at the surface to accommodate the
power swivel. Thus, the deck of the platform must be removed prior to the operation using a derrick barge, and after the deck
is removed, it is highly unlikely for a barge
to stay on-site when the mechanical cutters
operate. Remobilizing a derrick barge, however, is usually not an option and would represent a significant cost increase in the operation. Mechanical cutting is therefore
rarely used for piling.
3.2. Abrasive Methods
Mechanisms that inject cutting materials into a water jet and abrasively wear away
steel/concrete are called abrasive cutters.
Abrasive technology has a long history of
application in industrial and manufacturing
processes, and has been used in shipyards
for many years. Several different systems of
abrasive cutters exist.
Abrasive cutters can be classified as:
(1) Low pressure/high volume systems
(sand\cutters), or
(2) High pressure/low volume systems (AWJ).
Cutters that use sand or slag mixed with
water at low pressure (4000-10,000 psi) and
34
FIGURE 3
Abrasive Water Jet Cutting Tool for Conductors and Small Piles
(Courtesy of Circle Technical Services)
FIGURE 4
Abrasive Water Jet Cutting Tool for Piles or Caissons from 30 to 72 inches Diameter
(Courtesy of Circle Technical Services)
FIGURE 5
Grounded Conductor Cross-section, Concentric
high volume (80-100 gal/min) are called
sand cutters, while cutters that use garnet
injected at the nozzle at high pressure
(50,000-70,000 psi) and low volume (5080 gal/min) are commonly referred to as
abrasive jet cutters (NRC, 1996). The abrasive provides the force for cutting and is introduced at the cutting nozzle and sent down
a hose with air pressure or through a waterbased solution. The abrasives typically used
are garnet and copper slag.
Sand cutters use a turning mechanism
(or power swivel) like a mechanical cutter.
The power swivel is connected to the top of
an open pile and as the drill string turns, the
cutting nozzle cuts the pile through the abrasive action of the water jet. Abrasive jet cutters produce a jet of water mixed with garnet under high pressure and directed through
a diamond orifice.
The minimum inside diameter that can
be accessed with abrasive cutters is approximately 7 inches, and beyond 200-250 feet,
some of the abrasive cutting technology employed in the GOM is not effective. Improvements to the systems over the past decade, especially with the influx of North Sea
technology, has allowed abrasive cutters to
work in deeper waters than in previous years.
Figure 3 shows a cutting tool intended for
use in conductors or small piles. Figure 4
shows an AWJ tool capable of cutting piles
and caissons to 72 inches. Air delivery systems are limited to shallow water application,
while systems with a fluid delivery have been
used in water depths exceeding 600 feet (Brandon et al, 2000). Abrasive cutting has also
been deployed by ROVs to depths exceeding
1,000 feet. Casing strings that are eccentric
and with void areas remain problematic since
the void dampens the energy of the water jet
and may cause an incomplete cut. (See Figure 5 for a concentric cross-section).
There also exists the problem of verifying that the cut has been made when using
an internal abrasive cutter. Unlike explosives,
the conductor or pile does not drop, confirming that the cut was successful. With an
abrasive tool, the width of the cut is small
and when combined with the soil friction, a
visual response does not occur. To verify the
cut, the conductor is pulled with either the
Spring 2004
Volume 38, Number 1
35
FIGURE 6
Diamond Wire Cutting System Applied to a 72 inch Deck Leg on the Surface
(Courtesy of CUT USA, Inc.)
platform crane or hydraulic jacks, and the
lift force must overcome the conductor
weight and the soil friction. For an unsuccessful cut, the abrasive cutting tool is either
re-deployed to make another complete run
or explosives are used to complete the cut.
3.3. Diamond Wire Methods
FIGURE 7
Diamond Wire Cutting System Deployed from a ROV for use in Pipeline or Small Member Cutting
(Courtesy of CUT USA, Inc.)
A diamond wire cutting system uses a
diamond embedded wire on a chain sawlike mechanism to cut steel, concrete, or
composite material above and below the
waterline. A diamond embedded wire is
veered onto hydraulically driven pulleys resembling a band saw and mounted on a
frame. The system can be configured to cut
virtually any structural component and is
not limited by size, material, or water depth
as long as the cutting tool can be fixed to
the cut member. Diamond wire has been
used to cut caissons, conductors, risers, and
pipeline. Diamond wire cutting has been
used since the early 1990’s in the North,
Adriatic, and Red Seas; but in the GOM,
diamond wire has only been applied on a
few jobs. Figures 6 and 7 show diamond wire
tools for different applications.
The cutting machine is hydraulically
clamped or manually strapped to the structure, and a surface-activated motor activates
the tool. The diamond wire is driven at high
speeds and depending on the material and
thickness, wire speeds are maintained to produce the cut. Even under large axial compressive loads, tubular members can be cut
with diamond wire, and one of its strengths
lies in its ability to cut large wall section thickness (Brandon, 2000). The operator monitors the progression of the cut and makes
adjustments at the surface to improve the
efficiency of the machine.
3.4. Diver Torch Methods
Underwater diver cutting is virtually the
same as land-based cutting but the torch
used is somewhat different. In underwater
arc cutting, an outside jet of oxygen and
compressed air is needed to keep the water
from the vicinity of the metal being cut. A
tube around the torch tip uses air and gas
pressure to create a gas pocket. This will induce an extremely high rate of heat at the
36
work area since water dispels heat much
faster than air. As the water depth of the cut
increases, higher air pressure is required to
form the gas pocket (DOE, 1994).
Underwater cutting can also be accomplished with an oxy-hydrogen torch. Hydrogen is typically used instead of acetylene
because of the greater pressure required in
making cuts at increased depths. Oxyacetylene may be used up to 25 feet while depths
greater than 25 feet require the use of hydrogen gas (DOE, 1994). Underwater cutting is generally limited to caissons, pilings,
bracing, or other structural components, but
not wells. In shallow water and for simple
structures such as caissons, diving is sometimes the preferred method.
4. Nonexplosive Cutting
Systems and Activity
Requirements
4.1. Mechanical Systems
A mechanical cutting system requires a
tool to be lowered into an open pile or well
with a crane or drill rig. If the cutting tool is
operated by a drilling rig, diesel-fired engines
drive generators which provide electrical
power to motors, which rotate the turntable,
which turn the drill string, which mills the
tubular element. In addition, a water pump
capable of up to 5000 psi pressure provides
the force to keep the mechanical blades extended as the cutting progresses. If a drilling
rig is not used, the rotary table is replaced
by a power swivel, which is driven by a hydraulic power pack. The independently
driven cutting tool does not approach the
rig-based tool in cutting capacity, however,
and thus may take a considerably longer
period of time to cut. Mechanical cutting
generally requires at least three dedicated
personnel, although this is not easy to define, since a drilling rig requires significantly
more people to function, and an independent operation would also require significantly more personnel to be self supporting. The mechanical cutting operation is
generally only conducted from a platform
without an attending derrick barge, or from
a drilling rig.
4.2. Abrasive Water Jet Systems
A standard abrasive water jet unit consists of a cutting tool or manipulator to control the positioning and movement of the
nozzle, the abrasive mixing or dispensing
unit, high pressure water pump(s) and hydraulic power unit, control panels and cut
monitoring systems. The total weight of the
AWJ system may range from 5-15 tons and
have a footprint of 200-400 ft. Several different AWJ systems are commercially available with prices ranging from $250,000$500,000 for a complete system.
In a conventional internal pile cutting
operation, the cutting tool is lowered into
the pile from a wire line winch (deployment
frame) or by a construction vessel crane. The
arms of the tool’s centralizing system stabilize the tool and the cutting nozzle is positioned against the pile wall. A diesel-driven
water pump supplies the high pressure water stream to the cutting nozzle and the pressure required is determined by the cut parameters (e.g., wall thickness, cut configuration, abrasive mixing system, etc.). The cutting speed, direction of travel, and nozzle
position is controlled and monitored by the
operator at the surface control station (Brandon, 2000). External cutting operations on
legs, piles, and brace members are carried out
using diver or ROV installed tracks. Subsea
video equipment, lights, and audio systems
for cut observation and monitoring are common for both internal and external cutting.
The surface personnel required for 12
hour operations are generally two operators
and two roustabouts. External underwater
AWJ systems need to be placed either by
divers or ROVs. The operation can be supported from any work platform that has sufficient lifting capabilities; i.e., derrick barge,
platform with a capable crane, lift boat, etc.
4.3. Diamond Wire Systems
Diamond wire cutting systems are typically composed of a clamping frame, cutting frame with wire drive pulleys and motors, wire feeding system, wire tensioning
system, cut wedging system, underwater
power unit, umbilical assembly, and diamond wire cable (Brandon, 2000). The
power to the system can be provided from
the surface, by means of a dedicated subsea
power pack, or by a work-class ROV power
unit. Monitoring of the cutting progress is
provided by video cameras mounted on the
machine frame, ROV, or by divers from a
safe distance. The cutting machine is hydraulically clamped or manually strapped to the
structure, and a drive mechanism is either
remotely controlled by an operator at the
surface or configured for automatic operation by an ROV or diver.
4.4. Diver Torch Systems
Arc torches for underwater cutting are
produced in a variety of types and forms and
are constructed to connect to oxygen-air
pressure sources. Electrodes may be carbon
or metal and they are usually hollow in order to introduce a jet of oxygen into the
molten crater surrounding the arc. The current practice is to use direct current for underwater cutting and welding.
The torch used in underwater cutting is a
fully insulated celluloid underwater cutting
torch that utilizes the electric arc-oxygen cutting process using a tubular steel-covered, insulated, and waterproofed electrode. It utilizes the twist type collect for gripping the
electrode and includes an oxygen valve lever
and connections for attaching the welding
lead and an oxygen hose. The arc is struck
normally and compressed oxygen or air is fed
through the electrode center hole to provide
cutting. The burning electrode tip is shielded
from the surrounding water by the rapidly
expanding gas from the combustion process.
5. Environmental and
Safety Issues
5.1. Environmental and
Physical Impact
Energy is required to do work and all cutting operations require the expenditure of
energy. As work is performed, energy is transferred and transformed which may have an
impact on the ocean environment where the
operations are performed. The power requirements of a cutting spread are approximately
the same as a small offshore fishing vessel (less
than 200 horsepower, or 150 kW), but unSpring 2004
Volume 38, Number 1
37
like the typical sport fishing vessel, the cutting spread is fully self-contained with no
marine discharges. Nonexplosive cutting
methods do not create the impulse and
shockwave-induced effects which accompany
explosive detonation and are therefore considered to be an ecological and environmentally sensitive severance method. Environmental and physical impact data of nonexplosive methods are quite limited in the academic and trade literature and no record of
negative environmental impact for nonexplosive cutting methods has been found.
In mechanical, abrasive water jet, and diamond wire removals, diesel fueled mechanical systems are employed in the operation
which result in vibrations, the emissions of
CO2 and other gases to the atmosphere, and,
potentially, low frequency sound waves into
the ocean environment. Abrasive water jet
cutting involves using sea water and garnet
or copper slag (grit), and so there is the question of the impact of the fluid and garnet on
the marine environment. Since the fluid involved in abrasive cutting is water and the
garnet is essentially inert, the environmental
impact is believed to be inconsequential.
Garnet is an inert rock material that poses
no environmental consequences that have
been reported. The level of copper present
in the slag is very low and there are currently
no restrictions on its use or reported environmental issues. The noise level of the supersonic cutting jet is safe for divers and is
not considered harmful to marine life. The
direct products of the processes are water,
metal cuttings, and abrasive grit particles.
5.2. Safety Issues
Offshore oil and gas operations involve
a number of distinct phases—exploration,
development, production, and decommissioning—and present a continuing risk of
accident and injury to the personnel involved
in the operations. Drilling operations involve
moving heavy equipment into place (e.g.,
pulling or hauling pipe) and the continual
adjustment of controls and rotary equipment. Production operations involve the
maintenance of process equipment as well
as activities associated with changing
flowrates and reservoir depletion. Develop-
38
ment and decommissioning activities involve
the lifting and moving of heavy loads and
numerous other manual tasks such as rigging and welding. Decommissioning also
often involves significant cutting operations
above and below the waterline.
Drilling, installation, production, and
decommissioning operations are all personnel intensive, but the exposure time involved
with drilling and production operations are
several orders-of-magnitude greater than with
decommissioning activities, and so if all operations in the offshore environment are assumed to be “equally hazardous,” we would
expect no significant safety issues to be associated with decommissioning projects since
the time for a possible occurrence is so small,
and indeed this is the case. Injuries and accidents that may occur on decommissioning
are difficult to reliably detect relative to the
exposure time involved in the activity.
On a drilling facility the crew size consists of about 20 people per 12-hour shift,
while on a production platform, the crew
size varies with the number of wells and the
complexity of the equipment (NRC, 1990).
More than half the structures in the GOM
are unmanned and serviced from a central
platform with 20 or fewer people. In the
Western GOM, where gas fields are widely
scattered and platforms smaller, crew sizes
also tend to be smaller (2-10 people). The
average crew size on platforms that have
more wells per platform and more equipment than average, are expected to have
larger crew sizes.
In decommissioning operations, the
number of personnel required on the job is
determined by the size of the equipment
used. A small decommissioning project on
a single platform in shallow water may require 14-20 personnel and 3-7 days to operate the marine equipment spread. A moderately sized project with multiple platforms
in shallow water may require 35-50 personnel spread out over 30-45 days. A deep water decommissioning project with large
equipment may require in excess of 100-200
personnel over a number of months.
All GOM leaseholders are required to
notify MMS of all serious accidents, any
death or serious injury, and all fires, explo-
sions, or losses of well control connected with
any activities or operations on the lease. This
data is reported to MMS and processed during each calendar year. An event refers to a
reported happening that may involve more
than one incident. An incident refers to a
category of accident that occurred during
an event. From 1995-2000, 80 percent of
the reported events occurred during development/production activities and 20 percent
occurred during exploration activities (NRC,
1990). The breakdown of incidents according to welding/cutting-related and cranerelated incidents can be found in NRC,
1990. Welding and cutting operations
caused no deaths in the GOM, but they did
cause injury and accidents causing fire.
The basic safety issues with respect to
mechanical, abrasive, and diamond wire
cutting methods are somewhat comparable.
Mechanical cutting tools and safety precautions are familiar to any drilling crew. The
abrasive water jet system involves high pressure hydraulics but the cutting spread area
is considered a restricted work zone with
safety barriers and warning signs posted. The
cutting manipulators and hydraulic power
units incorporate high pressures ranging
from 5,000 psi (350 bar) to 50,000 psi
(3,500 bar). The tools, hoses, winches or
power unit could cause injury if damaged
or mishandled. The diamond wire methods
may require a diver to be deployed which
also presents special risks.
Acknowledgements
Many industry and government personnel provided generously of their time and
expertise for this study. We would especially
like to thank Sim Courville and Jeff Cole,
Hydrodynamic Cutting Services, and Aly
Hakam, ChevronTexaco. Nick Jones, Oil
States MCS; Svein Solversen, Norse Cutting
and Abandonment; Michael Leska, Superior Energy Services; Jeff Steele, Rodrigue
Consultants, Inc.; Tommy Broussard, Jeff
Childs, and Vicki Zatarian, Minerals Management Service, also provided critical information and useful feedback for this study.
This paper was prepared on behalf of
the U.S. Department of the Interior, Minerals Management Service, Gulf of Mexico
OCS region, and has not been technically
reviewed by the MMS. The opinions, findings, conclusions, or recommendations expressed in this paper are those of the authors,
and do not necessarily reflect the views of
the Minerals Management Service. Funding for this research was provided through
the U.S. Department of the Interior, Minerals Management Service.
References
Brandon, J.W., B. Ramsey, J. Macfarlane and
D. Dearman. 2000. Abrasive water-jet and
diamond wire-cutting technologies used in the
removal of marine structures. OTC 12002,
pp. 301-304. Houston, TX, 2000.
Byrd, R.C. and E.R. Velazquez. 2001. State of
the art of removing large platforms located in
deep water. OTC 12972. Houston, TX, May 2001.
Dodson, R.C. 2001. Meeting today’s plug and
abandonment/decommissioning demands.
World Oil. September 2001:67-69.
Englehardt, J., M.J. Wilson and F. Woody.
2001. New abandonment technology new
material and placement techniques. SPE
66496. Houston, TX, February 2001.
Fanguy, D.J. 2001. Coiled-tubing-conveyed
hydromechanical pipe cutting: A safe, effective
alternative to chemical and explosive severing
methods. SPE 68365. Houston, TX, March 2001.
Hakam, A. and W. Thornton. 2000. Case
history: Decommissioning reefing, and reuse of
Gulf of Mexico platform complex. OTC
12021. Houston, TX, May 2000.
Kaiser, M.J., D.V. Mesyinzhinov and A.G.
Pulsipher. 2002. Explosive removal of offshore
structures in the Gulf of Mexico. Coastal and
Ocean Management. 45(8):459-483.
Manago, F. and B. Williamson (eds.). 1997.
Proceedings: Public Workshop, Decommissioning and Removal of Oil and Gas Facilities
Offshore California: Recent Experiences and
Future Deepwater Challenges. MMS OCS
Study 98-0023. September 1997.
National Research Council (NRC), Marine
Board, Committee on Techniques for
Removing Fixed Offshore Structures. 1996.
An assessment of techniques for removing
offshore structures. Washington, D.C.:
National Academy Press.
National Research Council (NRC), Marine
Board, Committee on Alternatives for
Inspection of Outer Continental Shelf
Operations. 1990. Alternatives for inspecting
outer continental shelf operations. Washington,
D.C.: National Academy Press.
Pulsipher, A.G., ed. 1996. Proceedings: An
International Workshop on Offshore Lease
Abandonment and Platform Disposal:
Technology, Regulation, and Environmental
Effects. New Orleans, LA, April 15-17, 1996.
Thornton, W.L. 1989. Case history: Salvage
of multiple platforms and pipelines offshore
Texas. OTC 6074. Houston, TX, May 1989.
Twachtman Synder & Byrd, Inc. 2000. State
of the art of removing large platforms located
in deep water. U.S. Minerals Management
Service, November 2000.
Twachtman Snyder & Thornton, Inc. 1995.
Project management, a checklist. Supplement
to Petroleum Engineer International
(July 1995):12-13.
U.S. Department of Energy (DOE). 1994.
Decommissioning Handbook. DOE Office of
Environmental Restoration, March 1994.
Kaiser, M.J., A.G. Pulsipher and R.C. Byrd.
2003. Decommissioning cost functions in the
Gulf of Mexico. ASCE Journal of Waterways,
Ports, Harbors, and Ocean Engineering.
129(6):1-11.
Kirby, S. 1999. Donan field decommissioning
project. OTC 10832. Houston, TX, May 1999.
Spring 2004
Volume 38, Number 1
39
PAPER
A Design Study of Manned Deep Submergence
Research Vehicles in Japan
AUTHORS
ABSTRACT
Dan Ohno
Japan Deep Sea Technology Association
This paper covers the results of a design study recently completed in Japan on manned
submergence research vehicles equipped with Autonomous Underwater Vehicles (AUVs)
and/or Remotely Operated Vehicles (ROVs). The primary features and general overview of
the vehicle designs are described, and some of the major items to be examined in each
study are introduced.
At the outset of this study, the opinions of many domestic scientists and scholars
were collected in order to identify the most important subjects of future scientific research in the deep ocean.
This study was carried out by the “ad hoc Committee” organized by the Japan Deep
Sea Technology Association.
Yozo Shibata
Kawasaki Shipbuilding Corporation
Hisao Tezuka
Mitsubishi Heavy Industries, Ltd.
Hideyuki Morihana
Tokai University
Ryuichiro Seki
Japan Marine Science and Technology Center
INTRODUCTION
T
he mission of deep submergence research vehicles built to date—such as
“SHINKAI 2000,” “SHINKAI 6500,”
“Alvin,” “Nautile,” and others—has been to
provide a means of enabling researchers to
safely access and directly observe the deep
ocean at a time when it was difficult for a
man to venture to the deep seafloor. In the
past few years, the above-mentioned purpose has been realized, and it is now not
unusual to explore the deep ocean bottom
at depths of several thousands of meters.
(Busby Associates, 1990; Rona, 1999-2000).
On the other hand, the needs for investigation and research in the deep ocean continue to increase. Looking ahead 10 to 20
years, it is obvious that those needs will become more complicated and more advanced.
(Walsh, 1998; Morr, 2001; Jones, 2001;
Sagalevitch, 1997 & 2001). The increased
level of demand for investigation and research expected will be satisfied neither quantitatively nor qualitatively with the few deep
submergence research vehicles that are currently available in the world.
When we study manned submersibles
needs for the future, we should not neglect
the role of unmanned vehicles such as remotely operated vehicles (ROVs) and au-
40
tonomous underwater vehicles (AUVs),
which utilize highly developed sensors and
remote sensing technologies. Today, successful operations and investigations are being
achieved with such unmanned vehicles. Favorable opinions for development of deep
ocean scientific research systems using unmanned vehicles are growing. More productive operations will be accomplished in the
future as more effective and efficient unmanned vehicles are developed.
However, man is still important. In
natural science field studies, when man
goes down in situ into the deep ocean environment he fully utilizes the “human
sensor,”—i.e. the five sensory organs:
“looking at with the eyes,” “listening to with
the ears,” and “feeling with the body.”
Man’s presence will continue to be an important element for expanding ideas and
generating innovative knowledge and hypotheses, along with establishing effective
operating procedures for the use of unmanned vehicles. It is considered that
manned deep submergence research vehicles will remain essential tools as well as
a central component in future development and use of deep ocean scientific research systems.
Present and future areas of deep ocean
scientific research are interdisciplinary and
global. Though they are too numerous to
list fully here, they span such earth studies
fields as:
■ The atmosphere and oceans
■ Studies of the correlation between fluid
movement and seabed movement on the
deep ocean floor
■ Earthquake prediction
■ Chemosynthetic ecosystems on the deep
ocean floor
■ Overall earth scientific studies focusing
on geophysics phenomena from a
biological perspective
To do this work it has become necessary to carry out wide range, three-dimensional, and simultaneous measurements.
The nature of scientific observation itself
has been changed; consequently it will be
necessary to ensure that new manned submergence vehicle designs respond to such
changes. Manned submersible development should not remain on the periphery
of conventional research concepts and
methods. To bring them to the forefront,
it will be necessary to introduce new concepts in design and technology.
To address this situation, an “ad hoc
Committee” organized in the Japan Deep
Sea Technology Association first itemized the
needs for future deep ocean scientific research, then studied the design of manned
deep submergence research vehicles.
Prof. H. Morihana exercised overall
control of the study, and was joined by two
companies, Kawasaki Shipbuilding Corporation and Mitsubishi Heavy Industries,
Ltd., representing manufacturers of
manned deep submergence research vehicles in Japan.
Development and building of manned
submergence vehicles for future deep ocean
scientific research will be an enterprise of
national importance. It must be actively promoted for the noble purposes of contributing to mankind’s well-being today and to
ensure prosperity of the future as international collaboration advances ocean exploitation technologies.
The Needs for Future Deep
Ocean Scientific Research
In recent years, the requirements of deep
ocean scientific research have led to increasingly larger areas to be investigated while the
mesh of investigations has become finer and
finer, and the depths have become deeper
and deeper.
Scientific research in the deep ocean is a
recent branch of wide-ranging marine research, and has been made possible by newly
evolving technologies. In the deep seas the
scientist ventures into an unlimited treasure
house of knowledge; it is truly a voyage into
the unknown (JAMSTEC, 2002).
In this regard, efforts to study the needs
for scientific research in the deep ocean over
the next 10–20 years will be critical for the
development of both manned submergence
vehicles and AUV/ROV systems for future
deep ocean scientific research.
The Japan Marine Science and Technology Center (JAMSTEC) requested the Japan Deep Sea Technology Association to
undertake an “Investigation and study for
the outlook of deep ocean scientific research
for the 21st century.” The study was done
during fiscal years 1999 –2000.
The Association accomplished the investigation and study by organizing an expert
committee comprised of members selected
from wide-ranging fields covering marine
geophysics, chemistry, and biology in Japan.
Seven themes shown in Table 1 were selected
as key fundamental items for new deep ocean
scientific research.
Studies for Future Deep Ocean
Scientific Research Systems
For the success of future deep ocean
scientific research in the 21st century, it will be
necessary to clearly ascertain the needs and specifications of that research and to design manned
submergence vehicles together with unmanned
systems that will satisfy the scientific needs.
As mentioned earlier, the areas of the
ocean to be surveyed should include the intermediate layer, deep layer, seafloor, and
sub-seafloor. Therefore, the submersible research systems must be able to function
throughout broad, three-dimensional areas.
Some of the capabilities for investigation
and observation that will be required from
this system are multiple points simultaneous
measurements, online observation, and repeat observation through a network. The
system should be equipped with functions
such as observation capabilities on the seafloor, a means to collect and transfer the data
stored in the observation facilities, and the
ability to freely move the observation facilities from point to point.
Future deep ocean scientific research
work will require new manned deep submergence research vehicles to incorporate
unconventional features that will advance
existing concepts of this kind of technology.
It is critical not only to adopt emerging technologies that are currently undeveloped but
also to seek development of ultramodern
original technologies that are far out of the
realm of conventional technologies
(Allmendinger, 1990).
It is also necessary for submersible operational capabilities to be greatly improved.
For example, endurance times of existing
manned deep submergence research vehicles
are too short: 5 hours on the seafloor for
Nautile (maximum operating depth 6000
m); 10 hours submerged for Alvin (max.
operating depth 4500 m), and maximum of
9 hours underwater cruising (with 3 hours
on the seafloor) for SHINKAI 6500 (max.
operating depth 6500 m). Such short mission times result in inefficient operations requiring multiple dives with repeated launching/recovery from the support ship in order
to cover the wide range necessary to meet
the needs for investigation and observation.
One of the major limitations of deep
submersibles is on-board power. Existing
types of batteries will not provide the needed
long duration mission times. In recent years,
however, fuel cells have attracted attention
as a power source, and they may be expected
to significantly extend the underwater operating time of manned deep submergence
research vehicles.
It is also important to consider manned
submersible operations equipped with AUV/
ROV in order to achieve wider observations
and include many more items of investigation within a limited period. For these purposes, it will be necessary to study the functions to control and manipulate AUVs/
ROVs remotely from manned vehicles.
There are numerous other items to be
considered and studied to improve manned
submersible capabilities. Better instrumentation, sensors and sampling devices; more
capable and powerful manipulators; advanced imaging systems, and higher power
external lighting systems are examples of
these items. It is also an important objective to provide the capabilities for precision
as well as heavy work in the deep ocean, as
such needs are required in the future. Also,
it is important to improve abilities for in situ
(direct) observations, the most essential capability of manned deep submersibles, by
rearranging the position, increasing the number, and expanding the size of the viewports
to permit a greater viewing area from various points in the pressure hull.
When designing improvements for
manned deep submergence research vehicles
it is important to pay sufficient attention to
the opinions and experiences of all operators
of manned deep submersibles that have been
used for oceanographic research. Special consideration should be given to JAMSTEC’s
Spring 2004
Volume 38, Number 1
41
TABLE 1
Fundamental items for new deep ocean scientific research
Study item
Integrated studies of the
deep environment in the
ocean plate subduction zone
Integrated crustal studies by
whole mid-oceanic ridges
survey
Global dynamics studies
Integrated deep sea life
environmental studies
Development studies of deep
sea survey. direct-connecting
social and economic
activities
Contents
Studies are aimed to understand the following phenomena:
■ plate motion in the trench region
■ crustal deformation, active fault activities and submarine
landslide occurrence
■ distribution and evolution for the ecosystem in the ultra abyssal
zone, the bottom layer flow and water seeping flux
Surveying mid-oceanic ridges of the whole earth with
high accuracy, studies are aimed to understand
the following phenomena:
■ distribution of hydrothermal vents
■ submarine volcanic activity
■ crustal deformation
■ ocean floor spreading process
Studies are aimed to understand global dynamics of the deep
sea environment; monitoring plate motion, seismic activities,
volcanic activities, crustal deformation, deep current dynamics
and ecosystem changes of deep seafloor.
Main observation and measuring item
■ geomorphology and geological structure (by using precise
survey instrument, and ultra abyssal online network system)
■ pore water seeping (by using vehicle and seafloor observatory)
■ crustal deformation (by setting datum point and using vehicle)
■ observation of benthos
■ sampling the drilling core
■ geomorphology, geomagnetic and gravimetric data
■ CTD
■ sea water composition etc.(by repeat observation with high
solution)
Investigating deep sea ecosystem throughout intermediate
layer, deep layer, seafloor and under seafloor, studies are aimed
to understand life evolution, oceanic environmental changes to
protect global ecosystem.
Studies are aimed to develop deep sea survey techniques
corresponding to mineral and energy resources survey and
sampling, CO2 management and rescue for important deep sea
accidents etc.
■
■
■
■
■
■
■
■
■
■
■
■
Deep seafloor carbon cycles
studies
Development studies of
explorative techniques on
underground biosphere
beneath the deep seafloor
Studies are aimed to determine global physical-chemical flux
on boundary layer between seawater and deep seafloor.
Studies are aimed to develop innovative deep sea exploitation
techniques for resolving ecosystem of under seafloor biosphere
and discovering new species of deep sea microorganisms etc.
experiences with “Shinkai 6500” and “Shinkai
2000” (Takagawa, 1995).
An optimum design study of manned
submergence vehicles for future deep ocean
scientific research should be carried out based
upon all of the above-mentioned considerations (Brown et al., 2000).
Studying Results for Next
Generation Manned
Submergence Vehicles for
Deep Ocean Scientific Research
In selecting the characteristics for
manned submergence vehicles for future
deep ocean scientific research, the operating
depth and time which will be required to
42
■
■
■
seismic wave
crustal strain
pressure
temperature (by using integrated deep sea floor monitoring
system and borehole measuring system)
deep sea microorganism (by long term deep sea environment
monitoring and sampling)
deep sea organisms (by continuous observation and
sampling specimen)
geomorphology and geological structure
heat flow
water seeping
magnetic and gravimetric data
benthos
drilling core samples
hydrothermal flux (by using hydrothermal event monitoring
system on and under seafloor long-term monitoring system)
drilling core sampling technology (by using a non-contamination
type drilling machine for surface layer of deep seafloor)
deep sea environmental monitoring technology
meet the needs for scientific research mentioned above were finalized first. These are
summarized in Figure 1, which shows required operating depths varying widely because needs for scientific research are diverse, and the operating depths overlap according to the various research themes.
Furthermore, maximum times on the
seafloor, or undersea cruising times, are also
vary corresponding to the operating depths.
Since designing manned deep submergence
research vehicles that meet all these required
conditions (depth and time) involves very
difficult technical issues, it could be anticipated this could require some trade-offs, such
as a cutback of the functions, and producing a vehicle that is inconvenient to use.
Therefore, after examining thoroughly both
the functions to be provided to meet the
needs of the scientific research, and the technical problems encountered in achieving
them, it has been decided to provide several
plans for manned submergence vehicles that
have different maximum operating depths
and maximum cruising times.
The manned submergence vehicles under consideration at this time are of five
depth classes:
(1) 11000 m (full depth)
(2) 6500 m
(3) 4000 m
(4) 2000 m
(5) 500 m
TABLE 2
Results of conceptual designs of submergence vehicles
(1)
(2)
(3)
(4)
(5)
Principal Items of Submersibles
Max. operating depth
(1) Dimensions (L×B×H)
(2) Weight in air
(3) Cruising speed
(4) Navigation range (surface)
(5) Bottom operating time
(6) Crew
(7) Pressure hull
(8) Power source
11,000m (Full depth)
11.7 × 2.7 × 3.5 m
36 t
1.0 kt (Max. 3.0 kt)
¯¯¯¯
3 hours
2 (1 pilot)
1 sphere, 1.9m i.d., Ti-alloy
Li-ion secondary batteries
(2) Weight in air
(3) Cruising speed
(6) Crew
(7) Pressure hull
(8) Power source
6,500m
12.0 × 2.7 × 3.4 m
38 t
1.0 kt (Max. 3.0 kt)
¯¯¯¯
21 hours
3 (2 pilot)
Intersecting 2 spheres, 2.0m i.d., Ti-alloy
(2) Weight in air
(3) Cruising speed
(6) Crew
(7) Pressure hull
(8) Power source
4,000m
13.5 × 5.5 × 4.8 m
72 t
2.0 kt (Max. 3.0 kt)
¯¯¯¯
92 hours
4 (2 pilot)
Intersecting 3 spheres, 2.5m i.d., Ti-alloy
(2) Weight in air
(3) Cruising speed
(6) Crew
(7) Pressure hull
(8) Power source
2,000m
30.0 × 5.5 × 8.8 m
300 t
4.0 kt (Surface 10 kt)
4,500 NM
7 days (168 hours)
8 (Max. 16)
1 cylinder, 3.9m i.d., Ti-alloy
Li-ion secondary batteries,&
Diesel electric generator
500m
37.0 × 5.5 × 8.8 m
350 t
4.0 kt (Surface 10 kt)
4,500 NM
14 days (336 hours)
8 (Max. 16)
1 cylinder, 3.9m i.d., High tensile steel
Fuel cell & Diesel electric generator
(2) Weight in air
(3) Cruising speed
(6) Crew
(7) Pressure hull
(8) Power source
Major Features ofSubmersible System
(1) Capable of full depth underwater
operation
(2) Efficient ascent / descent speed
by minimizing fluid resistance
(3) High performance automatic
maneuvering / 1 pilot cont. system
(4) Large capacity of payload
(5) Wide range of vision for TV camera
& view ports
(6) High efficient buoyancy material
(1) Extended operating time on the
sea floor compared with “Shinkai
6500” (3h→21h)
(2) Cross over ocean ridges
(3) Reserved inner utility space by
2 spheres
(4) Capable of AUV / ROV operation
(Launch, recovery & supervising)
(6) Wide range of vision for TV
camera & view ports
sea floor compared with the
existing submersibles
(2) Cross over ocean ridges
(3) Reserved inner utility space by
3 spheres
(4) Capable of AUV / ROV operation
(6) Wide range of vision for TV
camera & view ports
(1) Capable of ocean-going /no need
of support ship
(2) Long term & large area
(3) Comfortable large space, habitable
for a long term operation
(4) Introduction of various
automatic systems
(5) Capable of AUV/ROV operation
(6) Very large capacity of payload
(1) Capable of ocean-going /no need
of support ship
(2) Long term & large area
sea floor by using FC-AIP system
(4) Comfortable large space habitable
for a long term operation
(5) Capable of AUV/ROV operation
(6) Very wide range of vision with
large acrylic spherical view port
Spring 2004
Measuring/OperationCapabilities
(1) Sampling of sea water & bottom
layer
(2) Measurement of CTD, O2,
current, magnetism, etc.
(3) Geological monitoring &
measurement of seafloor
(4) Monitoring & sampling of bottom
organisms
(5) Setting & recovering of
instruments
(1) ~(5); Do.
(6) Research on the dangerous
points /narrow courses by
AUV/ROV
(1) ~(5); Do.
AUV/ROV
(1) ~(5); Do.
AUV/ROV
(1) ~(5); Do.
AUV/ROV
Volume 38, Number 1
43
Design studies for planning each of these
vehicles were carried out. The characteristics of each system, their operational and
measurement capabilities, and principal particulars for each class of manned submergence vehicles are summarized in Table 2.
The profiles of five types of submersibles are
shown in Figs.2 (1)–(5).
In the planning of the five submergence
research vehicles, the most important items
in each vehicle were studied. The major items
are as follows.
■
■
■
■
■
■
Descending and ascending method.
Reduction of vehicle weight and etc.(for
11000 m class)
Study of battery capacity required for both
cruising speed and on site operation time.
Study of two connecting spherical hulls,
and etc.(for 6500 m class)
Comparison study of multiple intersecting
spheres and cylindrical hull, and etc.
Study of underwater launching and recovery system from a support vessel, and
etc. (for 4000 m class)
FIGURE 1
The operating depth and time which will be required by scientific needs
Study of weight and capacity of various
types of AIPs (Air Independent Power
systems), and corresponding studies of
numbers of cruising days and distance
to be available under restricted weight
and capacity.
■ Comparison study of secondary battery
with the AIPs , and etc.(for 500 m class
and 2000 m class)
Taking the following two items from the
above listing, the outline of the study results
shall be introduced in the two sections below.
■ Descending and ascending methods of
the full depth (11000 m class) manned
deep submergence research vehicle
■ Size of the submergence research vehicle
when an AIP is incorporated in a 500 m
class manned deep submergence research
vehicle
■
Descending and Ascending Methods
of the 11000 m (Full depth) Class
Manned Deep Submergence
Research Vehicle
The preliminary study concluded that the
time of one dive to the depth of 11000m
would be 10 hours. This includes three hours
each for descending and ascending which
would require an average speed is 61.1m/min.
This is approximately 1.4 times as fast as the
average speed of 43.3m/min for descending
and ascending of “Shinkai 6500” (time for
descending and ascending to and from
6500m depth is 2.5 hours respectively). It
requires approximately 2 times the descending force in order to increase the speed up to
as much as 1.4 times faster. This speed is obtainable by increasing the weight of the steel
ballast for descending and ascending. However, in order to minimize the total weight of
the vehicle and to keep enough payload, a
means to keep the vehicle at an inclined attitude during descending and ascending was
studied, instead of the conventional method
of keeping the vehicle at horizontal attitude.
With the submersible pitched down there will
be less drag in an advancing direction compared to the drag in a vertical direction.
If the angle is too small when a vehicle
descends and ascends at an inclined attitude,
the kinetic energy is consumed in horizontal movement and the speed in the vertical
44
FIGURES 2 (1) - (5): “The profiles of five vehicles”
FIGURE 2 (1)
11,000m class (full depth) vehicle
FIGURE 2 (2)
6,500m class vehicle
Spring 2004
Volume 38, Number 1
45
FIGURE 2 (3)
FIGURE 2 (4)
46
FIGURE 2 (5)
500m class vehicle
direction becomes rather slower than those
obtained at a horizontal attitude. From these
circumstances, to increase the speed with an
inclined attitude, the vehicle should take a
considerably large trimmed angle and also
have a small drag coefficient in the horizontal direction.
The study results for the effects of a
trimmed angle on the speed of descending and
ascending with an inclined attitude are shown
in Figure 3. Changes of speed of descending
and ascending against the trimmed angle are
indicated with change of apparent drag coefficient Cda, and the vertical axis of the illustration takes the ratio to drag coefficient Czo in
the case of descending and ascending in attack
angle of 90 degrees (descending and ascending in a vertical direction while keeping the
vehicle horizontal). As shown in this illustration, exceeding 20 degrees of trimmed angle,
effectiveness in increasing speed could be observed; and around 30 degrees might be considered as the most effective trimmed angle.
In addition, we observe that changing
the ratios of drag coefficients between an
advancing direction and a vertical direction
of the vehicle produces an improvement in
speed. When the ratio is 1:10, the apparent
drag coefficient at 30 degrees of trimmed
angle becomes about half as low as that of
vertical descending and ascending, so that
the effectiveness is significantly large. A
smaller drag coefficient causes reduction of
the weight of the steel plate ballast to be
stored on the vehicle, resulting in a reduction of the overall weight of the vehicle. In
the case of the full depth manned deep submergence research vehicle examined in this
study, the above-mentioned descending and
ascending with an inclined attitude achieves
a reduction of the weight of the steel plate
ballast to be stored by as much as about 1
ton, and a reduction of overall vehicle weight
by as much as about 1.4 tons.
Study of an AIP System for a
500m Class Deep Submergence
Research Vehicle
For submersible vehicles to continue active sub-sea operations for many days, a
power source such as an AIP system using a
fuel cell is considered an effective solution.
(Perry Jr, et al., 1990 ; Meyer, 1993 ;
Baumert, 1993). This is based on the fact
that naval submarines and submergence vehicles equipped with an AIP system using a
Stirling engine have already been put into
service. As for fuel cell systems, it is reported
that German submarines equipped with AIP
systems using fuel cells have been commissioned in 2004. (Strasser, 1995 ;
Hammerschmidt, 2001 ; Hauschildt, 2001).
Also smaller cells will become a better and
more practical type of AIP system in the near
future, as a result of the current state of their
technical development for use in automobiles. Generally, AIP systems have complicated organization and heavy weight and
require a large space, in comparison with
conventional batteries.
Improved AIPs will bring the advantage
of endurance to large sized submergence
vehicles, such as the 500m class, over small
ones. This section shows the results of the
study of the required weight and volume of
an AIP system for use in a 500m class vehicle. In this study, to evaluate the effectiveness of the AIP system simply in compariSpring 2004
Volume 38, Number 1
47
FIGURE 3
Effectiveness in improving speed for descending and ascending with inclined attitude
son with the battery, we have calculated the
necessary weight and volume of both AIP
system and battery for the same required
performances.
Table 3 shows the design conditions of
the study with some of the values in the table
being estimated. For the AIP system, we have
assumed a PEFC (Polymer Electrolyte Fuel
Cell) type fuel cell to be used, and for the
battery, a lithium-ion secondary battery
which has the highest energy density of the
presently available ones. The performances
of the lithium-ion secondary battery are
improving every year and are expected to be
48
higher than the present ones several years
from now. Therefore, the present performances and expected targeted ones for the
future are both indicated. Since the fuel cell
performance is expected to be steadily improved, the target values in the future are
indicated. As for the hydrogen supply, the
methanol reforming method is selected from
among potential ones expected to be developed in the near future.
Figure 4 compares the weights (left) and
volume (right) of the lithium-ion secondary battery and the PEFC-AIP system. According to these results, it is apparent that
the AIP system is lighter than the lithiumion secondary battery, but in volume, there
is little difference between the AIP system
and battery (target values).
Figure 5 shows the breakdown of weight
(left) and volume (right) of the PEFC-AIP
system. It is seen that more than 70 % of
the weight is attributed to oxygen-related
items such as liquefied oxygen (LOX),
LOX-tank and cold box. To reduce the AIP
system weight further, therefore, it will be
especially important to reduce the weight
of the LOX-tank and cold box (through
use of new materials, optimization of the
structure and required inner pressure of
LOX-tank, etc.), in addition to reducing
oxygen consumption by improving the
power generation efficiency. In volume
breakdown it is also seen that oxygen-related items, as with the weight breakdown,
take up a greater percentage.
However the percentage covered by
methanol and pure water generation is
greater than in the case of weight breakdown. To reduce the volume of AIP system, it is important to improve the power
generation efficiency to reduce the quantity of methanol, pure water, etc., in addition to taking the same means as in the
weight reduction case.
On the assumption that a lithium-ion
secondary battery (target value) and PEFCAIP system as studied above are installed
in a 4 m diameter cylindrical hull as a battery room and AIP room, respectively, we
have studied the weights and dimensions
of these rooms. The results are shown in
Figure 6. Since the LOX-tank of the AIP
system is installed outside of the hull, it is
not included in the required room length
but added separately in the total length
(room length + LOX-tank length). According to these room length results, when the
endurance is short, the AIP room is longer
than the battery room. This is mainly because in the AIP system, a diesel generator
is installed to obtain the surface speed of
about 10 knots. Regarding the weight, the
battery room is far heavier than the AIP
room. That seems to be caused by the light
weight of the LOX-tank that is designed
for a 5 MPa pressure tank.
TABLE 3
Design conditions of AIP study
Items
Deep submersible research
vehicle
Lithium-ion secondary battery
PEFC system(Fuel cell)
Liquefied oxygen tank
Required electric power
Endurance of submergence
Power source
Operational depth
Weight energy ratio
Volume energy ratio
Power generation efficiency (Output)
Power generation capacity
Weight (package)
Volume (package)
Oxygen consumption
Methanol consumption
Pure water generation
CO2 generation
Type
Material
Design pressure
Buoyancy adjustment tank
Outer pressure vessel
Inner pressure vessel
Outer pressure vessel
Inner pressure vessel
Material
Conditions
40 kW(on running speed of about 4 knots)
7 to 28 days
Lithium-ion secondary battery or PEFC
500 m
150 Wh/kg (target value)100 Wh/kg (present value)
300 Wh/lit. (target value)200 Wh/lit. (present value)
50 % (target value)
46 kW(Target: 6 kW for CO2 compressor)
31 kg/kW
67 lit./kW
0.6 kg/kWh (target value)
Double-hull horizontal cylindrical (Vacuum insulation between
outer and inner pressure vessels)
Titanium alloy
9 % Ni-steel
5 MPa
5 MPa
Titanium alloy
FIGURE 4
Comparison in weight and volume between Li-ion battery and PEFC-AIP for 500m class vehicle
Spring 2004
Volume 38, Number 1
49
FIGURE 5
Breakdown of weight and volume of PEFC-AIP system for 500m class vehicle
FIGURE 6
Comparison in weight and length between Li-ion battery room and PEFC-AIP room for 500m class vehicle
50
Conclusions
References
It has been fifteen years since “Shinkai
6500” was launched in 1989, providing the
world’s greatest depth capability for a
manned submersible. During this same time
period remotely operated vehicles and autonomous unmanned vehicles were developed worldwide and became available with
advanced controlling and sensing technologies. Today a lot of investigations and operations in the deep sea are utilizing these
unmanned systems. However, it is still very
important that humans work in situ to directly observe and act; this capability will be
necessary at any time and in any period. Even
if simulation technologies are advanced in
the field of natural science, the necessity of
fieldwork will not be diminished.
Efficient and highly capable manned
submergence vehicles that are responsive to
the future needs of scientific research in the
deep ocean must be used in collaboration
with various kinds of unmanned vehicles.
In summary, the study concludes that
manned vehicles will continue to assume the
most important role in the whole system.
The study team expects that there will
be lively discussions among scientists, submergence vehicle operators, and engineers
on the best means to support effective future deep sea scientific research using the
manned deep submergence research vehicle
concepts described in this paper.
Allmendinger, E. E. 1990. Submersible
Vehicle Systems Design. The Society of Naval
Architects and Marine Engineers.
Baumert, R. and Epp, D. 1993. Hydrogen
Storage for Fuel Cells Underwater Vehicles.
Proceedings of Oceans ’93 (Vol.II), pp. 166-171.
Brown, R. S., Foster, D. & Walden, B. 2000.
Manned Submersible Improvement Options –
Summary Report, WHOI. http://
www.marine.whoi.edu./ ships/Sea Cliff/report.htm.
Busby Associates. 1990. Undersea Vehicles
Directory—1990-91. Busby Associates, Inc.
Hammerschmidt, A. E. 2001. PEM Fuel Cells
– An Attractive Energy Source for AIP
Independent Propulsion Systems. Proceedings
of UDT 2001-EUROPE, Session 6c.
Hauschildt, P. 2001,. Hydrogen Storage and
Hydrogen Generation on Board Modern
Submarines. Proceedings of UDT 2001EUROPE, Session 6c.
Japan Marine Science and Technology Center
(JAMSTEC). 2002. Toward Life and the Whole
Earth—Ocean Science’s New Direction. The
30th Anniversary Commemorative Issue,
JAMSTEC.
Morr, B. 2001. Autonomous Submarines
Broadening Industry’s Horizons. Sea Technology
(Sep.), pp. 26-33.
Perry Jr., J. P., Alessi Jr, D. P., Misiaszek, S. M.
and Person, A. 1990. Application of a Proton
Exchange Membrane Fuel Cell (PEFC) to an
Existing, Man-Rated, Small Submarine.
Techno-Ocean ’90, pp. 543-551.
Rona, P. A. 1999-2000. Deep-Diving Manned
Submersibles. MTS Journal. 33(4):13-20.
Sagelevitch, A. M. 1997. 10 Years Anniversary
of Deep Manned Submersibles MIR-1 and
MIR-2. Proceedings of Oceans ’97, pp. 59-65.
Sagalevitch, A. M. 2001. Second Discovery of
the Bismarck Wreck. Sea Technology. (Dec.),
pp. 31-35.
Strasser, K. 1995. Air-Independent Propulsion
with PEM Fuel Cells. Proceedings of
SUBCON ’95, German Submarine Technology, pp. 32-33.
Takagawa, S. 1995. Advanced Technology
Used in Shinkai 6500 and Full Ocean Depth
ROV Kaiko. MTS Journal. 29(3):15-25.
Walsh, D. 1998. Deep Diving for Fun and
Profit. Sea Technology (Dec.), pp. 47-52.
Jones, T. N. 2002. The Investigation and
Excavation of a Deepwater Shipwreck in the
Gulf of Mexico. MTS Journal. 36(3):51-54.
Meyer, A. P. 1993. Development of Proton
Exchange Membrane Fuel Cells for Underwater
Applications. Proceedings of Oceans ‘93
(Vol.II), pp. 146-151.
Spring 2004
Volume 38, Number 1
51
PAPER
Temperature and Salinity Variability
in the Mississippi Bight
AUTHORS
ABSTRACT
Sergey Vinogradov
Nadya Vinogradova
Vladimir Kamenkovich
Dmitri Nechaev
Department of Marine Science,
University of Southern Mississippi,
Stennis Space Center
Conductivity-temperature-depth (CTD) profile data from five surveys performed by
the R/V Pelican in the Mississippi Bight in February, May, and November 1999; and January-February and August-September 2000 have been analyzed. The data were collected
within the framework of the Northern Gulf of Mexico Littoral Initiative (NGLI). The analysis of the T-S diagrams demonstrated substantial seasonal changes. Some estimates of
the spatial variability at different scales were suggested. The analysis of the T-S data
obtained at time-series stations revealed some interesting effects such as along-shelf
intrusion of deep water into the coastal system and fine vertical T-S structures in shallow
passes between the barrier islands.
INTRODUCTION
T
he Northern Gulf of Mexico Littoral
Initiative (NGLI) is the most recent and
comprehensive study of the littoral zone of
the Mississippi/Alabama coast, focusing on
the Mississippi Sound area. The goal of
NGLI was to develop a sustained
nowcasting/forecasting system for the coastal
areas suitable for coastal resource management (Asper et al., 2001). Starting in 1999,
NGLI has gathered a unique and significant
database of hydrological observations that
lends itself to various statistical and variability analyses.
The area of the Mississippi Sound is
2130 km2 and the mean low water depth is
only 3.0 m (Kjerfve, 1983). The series of
sandy barrier islands (including Cat, Ship,
Horn, Petit Bois, and Dauphin Islands) separate the shallow coastal waters from the
Northern Gulf of Mexico. Fresh water inflow comes directly from major rivers such
as the Pearl, Biloxi, and Pascagoula along
with diffuse land drainage and small bayous, and indirectly from the Mississippi
River and Mobile River. Saline waters from
the Gulf are transported through the passes
and straits between barrier islands.
Some estimates of temperature and salinity variability were performed by
Eleuterius (1976a; 1976b) who suggested
preliminary daily patterns of salinity distribution at various depths for the period from
June1973 to February 1975 and determined
52
monthly trends of salinity and temperature
distribution at several stations for the same
period of time. Three years later, Eleuterius
and Beaugez (1979) presented a description
of the climatic hydrographic conditions in
the Mississippi Sound. It gave some estimates
of the distributions of several hydrological
parameters (temperature, salinity, oxygen,
etc.) at various depths and at the bottom,
but it was based on a small number of stations. Kjerfve (1983) collected and analyzed
meteorological data, fresh water discharge,
tides, currents, and temperature and salinity data from several mooring sites for the
period of April-October 1980.
All these studies were based on rather
scarce data available before NGLI and therefore they cannot be used for validating an
operational numerical model of the area.
NGLI surveys have provided a much larger
number of sampling locations for the period of 2 years, higher vertical resolution,
more repeatable stations, and better geographic positions of the stations than any
previous hydrographic observations in the
Mississippi Bight. Based on NGLI data an
opportunity presents itself to perform some
reasonable quantitative estimates of spatial
and temporal variability of temperature and
salinity in the area.
The purpose of this paper is to present
an analysis of NGLI temperature and salinity data aimed at the comparison of observed
and simulated data in the area. No attempts
have been undertaken to determine the circulation in this area.
Oceanographic Observations
a. Data Acquisition, Processing
and Editing
The measurement procedure for the
CTD (conductivity-temperature-depth)
profiler consists of lowering it to depth and
raising it back to the surface. The seawater
properties are measured during both lowering and raising, and the data sets obtained
by the sensors are called downcasts and
upcasts, correspondingly. Two sensors installed for both temperature and salinity result in two data sets—primary and secondary—so each CTD station consists of 4 data
sets for both temperature and salinity.
Preliminary data processing was performed by the U. S. Naval Oceanographic
Office. The editing of data that we have carried out was aimed at the creation of a special database useful for the assessment of a
model performance, rather than at the generation of a universal database of all measured data. First, a set of primary data has
been chosen for all stations. The analysis has
shown a better quality of upcast data in the
upper layers (0-5m), compared to downcast
data, while both upcast and downcast data
looked reliable in deeper layers at most stations. A more homogeneous flow and less
bubbles around the sensing elements, and a
more stable upward motion of the CTD
rosette may be the determining factors of a
better quality of the upcasts. The data at the
surface (0m) appeared very unreliable which
is why they were not included in the database. Then the obvious outliers such as negative or very high values of temperature and
salinity have been removed. No data filtering or smoothing has been performed.
A time-series group is a set of stations
done at the same location for a period of up
to 25 hours (diurnal cycle). Normally the
sampling interval was 1 hour, with a few
exceptions. There are 7 time-series stations
in the February 1999 survey, 2 in May 1999,
6 in November 1999, 4 in January-February 2000, and 2 in the August-September
2000 survey, for total of 21 time-series stations (see Table 1).
A transect group is a sequence of stations
done along some straight line. Most of the
CTD stations were done within the Mississippi Sound, including short transects along
ship channels off Pascagoula, Gulfport, and
Biloxi. The transect along the ship channel
in Mobile Bay was performed in all surveys
(locations along the median of Mobile Bay,
between the Mobile River delta and the Mobile Bay Pass; see maps on Figs 1.1a – 1.5a).
Time-series groups were often done in or near
the passes and straits of the Mississippi Sound
or between Dauphin Island and Mobile Point.
Offshore data consisted of long transects seaward of the barrier islands. The duration of
the surveys ranged from 10 to 15 days.
FIGURE 1
Location of CTD stations and T-S diagrams
(1.1a) Location of February 1999 CTD stations
(1.1b) T-S diagrams for February 1999 survey
(1.2a) Location of May 1999 CTD stations
(1.2.b) T-S diagrams for May 1999 survey
(1.3a) Location of November 1999 CTD stations
(1.3b) T-S diagrams for November 1999 survey
poral and spatial scales of the whole survey
(10-15 days, coverage of more than 30,000
km2). It allows the analysis of large-scale features of the whole Mississippi Bight, with
the separate consideration of the shallow and
deep water contributions. At the same time,
if needed, any particular water mass can be
easily ‘detached’ from the total T-S diagram
for examination on smaller scales.
The February 1999 T-S diagram
(Fig.1.1b) shows that, in wintertime, the
seaward waters are warmer but still saltier
b. General Analysis
To estimate the total range of salinity and
temperature variations for the whole survey,
T-S diagrams have been plotted. T-S diagrams reproduce not only the hydrological
picture of the survey, but also allow a qualitative estimation of the data variability.
Figs.1.1a – 1.5a show the location of CTD
stations in 5 cruises of the R/V Pelican. T-S
diagrams of these surveys are shown in
Figs.1.1b – 1.5b. The T-S diagram for each
survey includes both shallow and offshore
stations, so it captures the variability on tem-
Spring 2004
Volume 38, Number 1
53
(1.4a) Location of January/February 2000 CTD stations
(1.4b) T-S diagrams for January/February 2000 survey
(1.5a) Location of August/September 2000 CTD stations
(1.5b) T-S diagrams for August/September 2000 survey
TABLE 1
Synopsis of analyzed oceanographic data.
CTD Surveys
Total number Duration of Number of time- Number of
Number of
of stations survey (days) series stations spatial groups transect groups
February 1999
245
13
7
4
5
May 1999
219
10
2
6
5
November 1999
312
14
6
6
4
January/February 2000
385
15
4
5
16
August/September 2000
178
15
2
3
7
Total:
1339
67
21
24
37
than coastal ones. The main curve on the TS diagram bulging down from the right to
the left side reflects this fact. A smaller group
of data points builds up a secondary curve
with almost constant temperatures (e.g.,
Pascagoula transect). A scattering of data
points around the two distinct curves on the
T-S diagram gives a clue about scales of temperature and salinity variability. Based on
54
such an analysis we conclude that salinity gradients persist while the variation of temperature remains small. The T-S diagram for the
May 1999 survey (Fig.1.2b) reveals a late
spring vertical distribution when the surface
waters become warmer by 6-7º C than deep
layers. Again there are two T-S curves, but
now they both are bulged down from the left
corner to the right, which corresponds to the
change in direction of the temperature gradient (towards land). Large changes in the
heating regime in the springtime lead to
higher variability of temperature values as
compared to the February 1999 data.
For the November 1999 survey
(Fig.1.3a), a single mean T-S curve (most
concentrated data) was considered. This
curve consists of two parts: a vertical group
of points and a set of points bulged down
from the right side to the left (see Fig.1.3b).
The latter reveals subsurface gradients of
both salinity and temperature while the
former is caused by a sharp temperature gradient at depths from 60 m to 80-90 m. This
sharp gradient is presumably the effect of
upwelling along the shelf and shelf break.
The January-February 2000 survey
(Fig.1.4a) has the largest number of stations
(385). The T-S diagram (Fig.1.4b) represents
a winter pattern where the seaward waters
can be warmer than the waters in the nearcoast zone. Here we have two distinct T-S
zones: a vertical group of points, corresponding to a substantial temperature gradient in
the 100–500 m layer, and two curves bulged
down from the right side to the left. All T-S
groups tend to 8º C temperature value—
the surface value in the Mississippi Sound
and the value in the deep continental slope.
As compared to the February 1999 survey,
the overall mean temperature of shallow
coastal zones decreased by 4-5º C. The variability analysis shows that temperature variations at the surface layer 0.5–8 m are the
highest among all processed surveys (at spatial scales more than 24 km). This is due to
large differences in temperature (up to 10º
C) between cold coastal surface waters and
warmer Gulf waters.
The August-September 2000 survey
(Fig.1.5a) includes 6 transects performed
seaward of the barrier islands to depths as
great as 480 m, the Mobile ship channel
transect, and observations within the Mississippi Sound. It has been used for preliminary assessment of the performance of the
Estuarine and Coastal Ocean Model
(ECOM) (Vinogradova et.al., 2004). Small
variations in temperature and salinity values from the mean vertical profiles have been
observed in this survey. The average T-S
curve (Fig.1.5b) with a very small data scatter is readily seen on the T-S diagram. The
horizontally aligned group of points corresponds to the river inflow (mostly Mobile
Bay data) in shallow homogeneously
warmed waters. The almost vertical group
of points represents the 50–480 m portions
of the Gulf transects where the temperature
monotonically decreases with depth from
30ºC to 7ºC at the bottom while salinity
has the mean value of about 36 (mean salinity of the Gulf of Mexico waters).
Spatial Variability within
the Mississippi Bight
In this section, we estimate the variability of the T-S fields for different horizontal
scales at various depth intervals and for different surveys. A rather simple approach is
used to obtain a preliminary estimate. The
amount of data available precludes the use
of any sophisticated methods, like cluster
method (see, e.g., Hur et al., 1998).
Towards the objective we select special
groups of stations. A group is suitable for
spatial variability analysis if (1) it is not a
transect; (2) stations within a group are located in a region with similar bottom depths,
current intensity, etc.; (3) the number of stations is not too small (otherwise the statistical methods would be inapplicable); and (4)
the stations within a group should have been
performed within a relatively short time period. All non-transect or non-anchored
groups are considered as potential spatial
groups, and, if they satisfy the requirements
described above, they were used in the analysis. The total number of selected spatial
groups is 24. The radius of the spatial group
is defined as a half of the distance between
the two most distant stations within the
group. Fig.2a illustrates one example of spatial groups.
Estimation of spatial variability was performed by using the following procedure:
1. The spatial radius was calculated for each
spatial group.
2. The average values of temperature and
salinity were calculated within a group
for each depth levels from 0.5 m to 8.0 m.
This depth range is determined by the
fact that most spatial groups are in shallow
regions in the vicinity of barrier islands.
3. The standard deviations of temperature
and salinity (within a group) were calculated for each depth level from 0.5 m
to 8.0 m.
As a result, the plots shown in Figs. 2b and
2c have been constructed. These plots have
three major axes: the depth in meters; spatial
radius in kilometers; and standard deviation,
in ppt, for salinity and in ºC for temperature.
In general, the variability decreases with
depth and increases with spatial scale. Note
that there are pronounced sharp-gradient
layers for some salinity and temperature vertical profiles in the upper subsurface zone of
2.5–6.0 m, which are caused by fresh water
inflow, diurnal variations, and vertical mixing (see Temporal Variability section for further discussion). Due to these gradients, the
deviation of salinity or temperature from
their mean values increases with depth
within some spatial groups.
Thus the overall maximum salinity variability is 8.0 ppt (February 1999 survey,
spatial scale of about 22.7 km, 5.0 m depth),
and the maximum temperature variability
is 3.0 ºC (January-February 2000 survey,
spatial scale of about 23.8 km, 1.0 m depth).
FIGURE 2
(2a) The illustration of spatial group definition. This is group 17, February 1999 survey, performed between
Gulfport ship channel and Biloxi ship channel in the Mississippi Sound. Total duration (a time scale) of this
group is 5 hours 24 minutes. Spatial scale is determined by the radius of a circle that barely covers all stations
in the group. It is 8.83 km for this group, based on coordinates of stations 153 and 168, the most distant among
all stations in this group. Standard deviations versus spatial scale for (2b) salinity and (2c) temperature. Vertical
bars on these plots represent standard deviations computed for each spatial group. It gives an estimate of
characteristic variability for a particular spatial scale, season, and depth associated with spatial groups.
(2b)
(2a)
(2c)
Spring 2004
Volume 38, Number 1
55
Seasonal trends are also seen on these two
plots. The winter surveys (February 1999,
November 1999, and January-February
2000) have similar high amplitudes of salinity deviations as compared to the May 1999
and August 2000 surveys on equivalent spatial scales. At all spatial scales, the smallest
variations in temperature correspond to the
August-September 2000 survey data whereas
the highest temperature deviations correspond to the January 2000 survey.
FIGURE 3
Along-shelf intrusion. (3a) Location of 4 selected time-series stations. Diurnal mean vertical profiles and standard
deviations for salinity (3b) and temperature (3c) at 4 stations. Diurnal changes in temperature (3d) and salinity
(3e) at 60, 75 and 80 m depths, station A.
(3a)
Temporal Variability in the
Mississippi Bight
A time-series group represents a set of
observations performed at the same location
for a period up to 25 hours with a sampling
period from 30 min to 2.5 hours. The analysis based on these CTD stations gives information on the temporal variability of temperature and salinity fields for scales up to
the diurnal cycle. The total number of processed time-series groups is 20.
We use two types of temporal variability
representation. First, mean vertical temperature and salinity profiles are plotted with
horizontal “error” bars equal to standard
deviations at the corresponding depth levels. Second, graphs of temperature and salinity values versus time are constructed at
different depths.
Time-series stations performed in November 1999 at 50–90 m depths in the vicinity of the shelf break reveal significant
temperature drops in the near-bottom layer
(Fig.3c). The upper layer (0–65 m) at station A with a total depth of 81 m is strongly
homogeneous (25.5°C, 36.3 ppt, see Figs.
3b and 3c). At 65 m, the temperature drops
dramatically while salinity increases slightly.
Mean temperature and salinity at the bottom are 22.1° C and 36.4 ppt respectively.
These gradients are stable on a diurnal time
scale, with a slight change in the thickness
of the near-bottom layer.
The presence of a stable, significant temperature gradient indicates the intrusion of
another water mass into the homogeneous
system. The intruding water mass is much
colder and slightly saltier than the local wellmixed system. Temperature changes at 60,
56
(3b)
(3c)
FIGURE 3
Along-shelf intrusion, continued.
(3d)
(3e)
FIGURE 4
Time-series stations located near passes and straits. (4a) Location of 3 selected time-series stations. Vertical
profiles of temperature and salinity observed at station 2 (4b), station 6 (4c) and station 14 (4d). Diurnal
changes in temperature and salinity at 3 characteristic depth levels: station 2 (4e), station 6 (4f) and station 14 (4g).
(4a)
75, and 80 m depth levels at station A
(Fig.3d) clearly show a large amplitude fluctuation at 75 m around the mean value that
implies a strong and stable source of this
intrusion. Similar bottom temperatures for
different locations (as seen in vertical profiles of stations B, C, and D on Fig.3d) possibly identify the same water mass. An alongshelf upwelling is assumed to be a source of
this water intrusion.
Station B (Fig.3a) was performed at 51m
depth two days later than station A. It represents similar characteristics of the nearbottom layer, with a little less intensity. Shallow time-series station C (30 m depth) shows
very homogeneous profiles from surface to
bottom while the station D done at 88 m
reveals a more sophisticated, stair-like structure for the temperature profile (Fig.3c).
The persistence of these gradients for
more than two days, and the pronounced
direction and intensity of intrusion prove
the stable character of this phenomenon.
The geographical location of described stations gives some indication of the horizontal scales of near-bottom intrusion. Data
from this survey show the presence of a nearbottom water mass forming at about 40 m
depth in the areas open to the shelf break.
Time-series stations located in passes between the barrier islands provide some interesting material for studying the temperature and salinity vertical structure. The February 1999 survey reveals different types of
diurnal variability in the vertical structure
of both the salinity and temperature fields.
Station 02 (Fig. 4a) is located south of
Mobile Bay, in the area of extensive fresh
water outflow from Mobile Bay into the Gulf
of Mexico. All temperature and salinity profiles at this station (Fig. 4b) show the existence of two distinct layers: lower (5–10 m)
vertically homogeneous layer with no
changes in time and upper layer (0-5m)
highly variable both in time and in depth.
Salinity changes from 2 to 10 ppt at the surface, and, starting from 5 m, gradually increases with depth up to 35 ppt. Low surface values of salinity may be caused by intense fresh water transport from Mobile Bay.
Temperature changes from 21 to 17.5° C at
the surface, while the subsurface vertical graSpring 2004
Volume 38, Number 1
57
FIGURE 4
Time-series stations located near passes and straits, continued.
(4b)
dient at 2–3 m remains almost unchanged.
Subsurface temperature gradients are developed due to the tidal motion and diurnal
warming/cooling of the upper layer. Fig.4e
represents diurnal changes in temperature
and salinity at 1.0m, 3.0m and 9.0m depths.
Station 06 south of Horn Island Pass
represents another type of diurnal variability. There is no strong evidence of subsurface temperature gradients (Fig. 4c); and
subsurface vertical salinity gradients are not
as strong as at the previous location. This
difference can be explained by smaller transport out from the Mississippi Sound through
this pass. Fig.4f shows diurnal changes of
temperature and salinity at 1.0 m, 3.0 m
and 9.0 m depths.
Time-series station 14 located south of
Cat Island Pass shows very small changes in
both salinity and temperature within 25 hours
(Fig. 4d) and it could be considered almost
homogeneous in the vertical. Diurnal variations of temperature and salinity at 1.0 m,
3.0 m and 7.0 m depths are shown on Fig.4g.
Discussion
The NGLI CTD database appears to be
useful for a preliminary study of the climatology of the Mississippi Sound. The comparison of surveys performed over one year
from February of 1999 to January-February
of 2000 provides some material for studying
seasonal variability of temperature and salinity fields in the area. The comparison of February 1999 and January/February 2000 surveys have demonstrated a possibility of strong
interannual variability. The results of the
analysis are important for assessing the skill
of hydrodynamic operational models of the
region (Vinogradova et al., 2004).
High diurnal variability has been disclosed, based on the analysis of time-series
stations. For the coastal waters, stratification
can change dramatically within a day in some
areas. In once vertically homogeneous waters,
high subsurface vertical salinity gradients can
develop within a few hours and can also vanish within a short time due to tidal advection
through passes between the barrier islands.
(4c)
58
FIGURE 4
(4d)
Special diagrams have been suggested helpful in estimating the spatial variability.
The late fall offshore stations located in
the vicinity of the shelf edge reveal stable
near-bottom vertical gradients in temperature caused by some intensive intrusion of a
colder water mass. We hypothesize that
along-shelf upwelling causes this intrusion
although additional data and further analysis are needed to validate this assumption.
Acknowledgements
The authors are very grateful to John
Blaha, head of the NGLI project, for useful
suggestions and to Carl Szczechowski for providing us with initially processed data and
many helpful comments. The paper was significantly improved due to the detailed and
helpful guidelines provided by anonymous
reviewers. This work was funded by the Commander, Naval Meteorology and Oceanography Command through N62306-01-DBOO01-0002 to support the Northern Gulf
of Mexico Littoral Initiative.
(4e)
Spring 2004
Volume 38, Number 1
59
FIGURE 4
References
Asper V., J.P. Blaha, C. Szczechowski, C. Cumbee,
R. Willems, S. Lohrenz, D. Redalje, and A-R.
Diercks. 2001. The Northern Gulf Of Mexico
Littoral Initiative (NGLI): A Collaborative
Modeling, Monitoring, And Research Effort.
J Miss Acad Sci. 46(1):60.
(4f)
Eleuterius, C. K. 1976a. Mississippi Sound
Salinity Distribution and Indicated Flow
Patterns. Technical Report. MississippiAlabama Sea-Grant Publication 76-023.
Eleuterius, C. K. 1976b. Mississippi Sound
Temporal and Spatial Distribution of
Nutrients. Technical Report. MississippiAlabama Sea-Grant Publication 76-024.
Eleuterius, C. K., and S. L. Beaugez. 1979.
Mississippi Sound, a Hydrographic and
Climatic Atlas. Technical Report. MississippiAlabama Sea-Grant Publication 79-009, 135 pp.
Hur, H. B., G. A. Jacobs and W. J. Teague.
1998. Monthly Variations of Water Masses in
the Yellow and East China Seas, November 6,
1998. Journal of Oceanography, 55:171-184.
Kjerfve, B. 1983. Analysis and Synthesis of
Oceanographic Conditions in Mississippi
Sound, April Thru October 1980. Technical
Report. Army Corps of Engineers, Mobile
District, 438 pp.
(4g)
Vinogradova, N., S. Vinogradov, D. Nechaev,
V. Kamenkovich, A. Blumberg, Q. Ahsan, and
H. Li. 2004. Validation of the Northern Gulf
of Mexico Littoral Initiative (NGLI) Model
Based on the Observed Temperature and
Salinity in the Mississippi Bight Shelf.
J Atmos Ocean Tech. (submitted)
60
PAPER
Development of a Real-Time Regional Ocean Forecast
System with Application to a Domain off the
U.S. East Coast A B S T R A C T
1
AUTHORS
Laurence C. Breaker
Moss Landing Marine Laboratories
Desiraju B.Rao
Environmental Modeling Center,
National Centers for Environmental
Prediction
John G.W. Kelley
Coast Survey Development Laboratory,
National Ocean Service
Ilya Rivin
Bhavani Balasubramaniyan
Science Applications International Corporation
This paper discusses the needs to establish a capability to provide real-time regional
ocean forecasts and the feasibility of producing them on an operational basis. Specifically, the development of a Regional Ocean Forecast System using the Princeton Ocean
Model (POM) as a prototype and its application to the East Coast of the U.S. are presented. The ocean forecasts are produced using surface forcing from the Eta model, the
operational mesoscale weather prediction model at the National Centers for Environmental Prediction (NCEP). At present, the ocean forecast model, called the East Coast-Regional Ocean Forecast System (EC-ROFS) includes assimilation of sea surface temperatures from in situ and satellite data and sea surface height anomalies from satellite altimeters. Examples of forecast products, their evaluation, problems that arose during the
development of the system, and solutions to some of those problems are also discussed.
Even though work is still in progress to improve the performance of EC-ROFS, it became
clear that the forecast products which are generated can be used by marine forecasters if
allowances for known model deficiencies are taken into account. The EC-ROFS became
fully operational at NCEP in March 2002, and is the first forecast system of its type to
become operational in the civil sector of the United States.
INTRODUCTION
he population of coastal regions around
the continental U.S. has increased dramatically over the past 60 years and is expected
to continue to increase in the foreseeable
future. Over 50% of the U.S. population
now resides along our coastlines. Populations
in a majority of coastal counties from Texas
through North Carolina have increased almost fivefold between 1950 and 1990
(Pielke and Pielke, 1997). The greatest increase in population occurred in Florida
where the increase was over 500%. By the
year 2025, nearly 75% of all Americans are
expected to be living and working in coastal
areas (Hinrichsen, 1998). Such increases in
human population are affecting the coastal
oceans more profoundly and more rapidly
than is global climate change (Hay and
Jumars, 1999). The pollution problem due
to terrestrial, atmospheric, and in situ sources
continues to degrade the quality of coastal
waters surrounding the U.S. Over two trillion gallons of partially treated sewage plus
more than 2 million tons of chemical wastes
are discharged into U.S. coastal waters each
year (Hinrichsen, 1998).
As one of the tools to manage environmental problems created by the above mentioned causes, interest in developing a capability to provide short-term forecasts of
coastal ocean conditions is now rapidly growing. This is, however, a formidable task since
the coastal oceans represent some of the most
challenging marine environments for modeling in the world (Haidvogel and
Beckmann, 1998). The time and space scales
of interest associated with short-term coastal
circulation may be as short as a few hours
and as small as a few tens of meters or less.
Irregular coastlines and steep and variable
bottom topography near the coast (and at
the shelf break) can create highly complex
patterns of flow. Circulation on the continental shelf is primarily governed by factors
such as winds, tides, buoyancy fluxes,
throughflow ( i.e. the permanent and seasonal alongshelf currents), and cross-shelf
forcing by basin scale processes, etc. (e.g.,
Johnsen and Lynch, 1995). Within this
framework, many (but not all) coastal processes occur. Wind forcing produces both
surface and internal waves, and contributes
to surface flow directly through wind drift,
Ekman transport, and Stokes drift. Tidal
forcing, in addition to the depth-independent barotropic processes, also includes internal tides which are often generated at the
shelf break (Wiseman et al., 1984). Coastal
waters are particularly sensitive to major atmospheric events which may occur frequently (Brink et al., 1990). Fresh water discharge from various bays and estuaries along
the coast add buoyancy fluxes which further complicate the water motions locally.
Also, in coastal areas, water mass integrity
breaks down and the property relationships
which characterize these water masses in the
deep ocean often do not apply in shallow
coastal areas where the effects of local mixing often destroy their coherent nature. As
noted by Mooers (1976), however, the situation is not hopeless since the circulation,
although complex, is not simply an unstructured, incoherent, noise-like turbulence, but
rather can be interpreted (and thus modeled)
in terms of (albeit many) simple processes.
1
Spring 2004
OMB Contribution No. 206
Volume 38, Number 1
61
Interest in the scientific and technical
challenges involved in predicting the state
of the coastal ocean started with a series of
workshops on coastal physical oceanography starting ca.1974. In particular, a workshop was held in 1989 to determine system
requirements, and the research and development needed to establish an initial operational coastal ocean prediction system by the
year 2000 (Mooers, 1990a; 1990b). Also,
the National Research Council (1989) recommended that the nation establish an “operational capability for nowcasting and forecasting oceanic velocity, temperature, and
related fields to support coastal (and offshore) operations and management.” In
1993, NOAA developed a strategic plan for
establishing a Coastal Forecast System “to
create and maintain an effective coastal forecast system that meets today’s requirements
and that can be rapidly updated and enhanced as new requirements, knowledge,
and technologies emerge” (U.S. Dept. of
Commerce, 1993). The long-range goal of
such a Coastal Forecast System is “to improve our ability to measure, understand,
and predict coastal environmental phenomena that impact public safety and well-being, the national economy, and environmental management.”
As a result, the National Weather Service (NWS) and the National Ocean Service (NOS), together with Princeton University, have initiated development activities
within NOAA for a “Coastal Ocean Forecast System” (COFS)—the word “ocean” is
specifically introduced to distinguish it from
atmospheric prediction systems. Now the
generalized name Regional Ocean Forecast
System (ROFS) has been adopted to reflect
the fact that (i) such systems can be deployed
anywhere in the global ocean, and (ii) the
domain covered by a forecasting system extends far beyond what might strictly be considered a coastal area in several cases in order to take into account the influence of
dominant ocean features and processes occurring closer to the coast, such as the Gulf
Stream off the U.S. East Coast. The purpose of this paper is to describe the needs
for a real-time ROFS, briefly present the
history of the numerous efforts that have
62
been undertaken to develop such a capability, and to describe one system that has now
become operational at NCEP for the East
Coast of the U.S. in March, 2002.
2. The Needs for Real-Time
Ocean Forecasts
A brief description of the needs for realtime forecasts of ocean conditions, particularly in coastal areas, and relevant issues are
discussed in Brink et al. (1992). These issues deal with ecosystem management,
coastal hazards, navigation, recreation,
coastal meteorology, mineral exploitation,
defense requirements, fisheries, anthropogenic inputs, etc. in many of which an operational ocean forecasting capability would
play an important role.
(a). Marine Transportation and Search
and Rescue Operations: The amount of
cargo transported by ships traversing coastal
waters on their way into ports is expected to
increase substantially in the near future placing greater stress on the coastal environment.
Forecast products (ocean currents, water levels, and water temperatures) from an ocean
forecast system can play a critical role in
ensuring the safety of, and providing optimum routing for, ships at sea. For example,
vessels leaving U.S. East Coast ports and
heading to Europe can increase their average speed significantly over a major portion
of their route by knowing the location of
the Gulf Stream axis. Knowledge of water
temperature can be important for tankers
transporting crude oil. As water temperature increases, the viscosity of oil decreases,
making it easier to pump out the oil when
the vessel arrives in port. In certain coastal
areas, particularly on the East Coast, water
level forecasts are critical for safe and economic operations of marine transportation.
Forecasts of currents and temperature are
vital to all hazardous material spill containment efforts and search and rescue (SAR)
missions conducted by the Coast Guard in
U.S. coastal waters. Surface current information is required to estimate the direction
and extent of spreading of a spill or for the
direction and movement of downed planes
and incapacitated vessels prior to search and
rescue operations. Water temperatures are
needed to estimate survival times for those
who are lost at sea and exposed to hypothermal conditions. For example, following the
TWA flight 800 disaster off southern Long
Island on July 17, 1996, information on local water conditions, particularly near the
bottom where the search operations were
taking place, would have been very helpful
during the search activities which took place
in and around the crash site because only
four days earlier, Hurricane Bertha had
passed through this area and stirred up the
ocean, reducing visibility throughout the
water column.
(b). Coastal Flooding: Storm surges and
the subsequent potential for coastal flooding
are ever present dangers in low lying coastal
areas. One of the most devastating flooding
events in history that resulted from storm surge
occurred in Galveston, Texas in 1900. A storm
surge of 20 feet was estimated and as many as
12,000 people were killed (Rappaport and
Fernandez-Partagas, 1995). In 1957, Hurricane Audrey produced a storm surge of over
12 feet along the Gulf coast which extended
25 miles inland in Louisiana, killing almost
400 people (Pielke and Pielke, 1997). According to Ho et al. (1987), the coastal areas that
are at greatest risk of hurricane encounter lie
between south Florida and Texas. In addition
to threatening life and property, coastal flooding also causes detrimental changes in beach
morphology and increases erosion. Water levels predicted by a ROFS could provide storm
surge forecasts directly if its domain were extended to the coast.
(c). Boundary Conditions for Other
Forecast Models: Operational, high-resolution regional ocean forecast models could
provide initial conditions and boundary conditions to support oil spill models, estuarine
circulation models, and coupled ocean-atmosphere hurricane models (Bender and
Ginis, 2000). At the present time, lack of
real-time information on initial conditions
for the ocean represents a serious limitation
in our ability to produce quality forecasts
for both oil spill and coupled hurricane forecast models Also, models to predict currents,
water levels, salinity, and temperature in a
number of estuaries around the coastal U.S.
are being developed by NOS such as those
for the Chesapeake Bay (Gross et al., 1999),
the Ports of New York and New Jersey (Wei,
2003), and Galveston Bay (Schmalz, 2000).
Water levels in Chesapeake Bay, for example,
are of particular interest to large vessels which
usually have only a small clearance above
the bottom and thus are susceptible to the
danger of running aground. One of the critical pieces of information that these estuarine forecast models require is the oceanic
forcing where they interface with the coastal
ocean (i.e., bay mouths).
(d). Offshore Construction and Operations: A knowledge of ocean currents is important for designing offshore structures, for
planning marine construction, and for conducting marine operations at sea (Wiseman
et al., 1984). During extreme events, current speeds at the time of peak surface waves
are especially important since their combined effect determines the maximum forces
that are experienced by oil rigs and other
fixed structures deployed offshore. Also, it
is important to know current speeds at
deeper levels where drilling and construction activities often take place because currents may still be strong even though the
effects of surface waves will have decreased
significantly. Knowledge of current speeds
near the ocean bottom in coastal areas is also
important since vigorous currents in this
region can lead to sediment erosion,
resuspension, and transport. When vigorous near-bottom currents exist, resuspension
of bottom sediments can produce major reductions in visibility as occurred during the
SAR operations following the Flight 800
disaster. Vigorous bottom currents, through
the redistribution of bottom sediments, can
also fill in navigation channels leading to the
need for subsequent dredging operations.
(e). Input to Ecological Forecasts: A predictive modeling capability for U.S. coastal
waters would provide useful information for
a number of important ecological problems.
The health of our coastal ecosystems is declining according to several recent studies
(Raloff, 1999). The rise in marine-related
diseases along the U.S. East Coast, the Gulf
of Mexico, and the Caribbean suggests that
conditions conducive to illness are wide-
spread, and that if present trends continue,
the health of our ecosystems could be significantly degraded, resulting in large economic losses for the fishing industries
(Epstein, 1998). Off the East Coast, river
runoff containing high levels of phosphorous and nitrogen have been linked to ailing
sea grass beds which provide important nurseries for a variety of fish. Pollution from
untreated sewage, industrial wastes, and agricultural runoff during the early 1990’s was
primarily responsible for the closure of over
50% of America’s shellfish beds along the
Atlantic and Pacific coasts, and nearly 60%
along the coast in the Gulf of Mexico
(Hinrichsen, 1998). A survey conducted by
the Natural Resources Defense Council
(NRDC, 1996), found that 29 coastal states
and territories had over 3500 beach closings
and pollution advisories in 1995, a 50%
increase from 1994, and that most of the
closures were related to high coliform counts
linked mainly to partially treated or untreated sewage, storm runoff, and other
municipal wastes. The fate of pollutants
which are being discharged into coastal waters can be predicted based on forecasts from
a regional ocean forecast system. Point source
pollutants, for example, could be routinely
tracked and their movements predicted.
Such a system could also provide inputs of
temperature, salinity, and water transport to
ecosystem models which have been, or are
presently being developed. In the bottom
waters which reside on the continental shelf
off Louisiana and Texas, hypoxic conditions
frequently arise during summer, which result in a so-called “dead zone”. A close relationship exists between the outflow from the
Mississippi River, river borne nutrients, net
productivity, and bottom water hypoxia in
this region (Rabalais et al., 1994). The physical characteristics and space-time structure
of this recurring feature could be tracked
through the application of a coastal ocean
forecast capability. In the New York Bight, a
cold pool of water forms each year in the
spring as the surface waters warm up and
isolate the deeper waters below (e.g.,
Aikman, 1984). This feature is bounded
offshore by the Slope Water near the continental margin and inshore by warmer wa-
ters in the shallow regions adjacent to the
coast. When fully developed, this water mass
can extend from Cape Cod to Cape
Hatteras. In the fall, increased winds and
reduced heating combine to destratify the
water column leading to increased vertical
mixing and the subsequent disappearance
of the cold pool. Because of the seasonal isolation of the waters that form the cold pool,
species of fish which inhabit this region are
effectively trapped until the seasonal breakdown of this water mass occurs. As in the
case of the dead zone in the Gulf of Mexico,
the capability to forecast the onset, spatial
extent, and demise of this unique ocean feature is clearly important.
Over the past 25 years or so, there has
been a significant increase in the incidence
of Harmful Algal Blooms (HABs) in U.S.
coastal waters. Also, the nature of the HAB
problem has changed recently, and larger
geographic areas, including most coastal
states, are now threatened by more than one
harmful or toxic species (Boesch et al., 1997).
One type of HAB, for example, is caused by
high concentrations of a toxic algae called
Gymnodinium breve (Gb). Gb occurs naturally in warm coastal waters, and with a certain combination of temperature, salinity,
and nutrients, massive increases in Gb, often referred to as red tides, can occur. Red
tides frequently originate in the Gulf of
Mexico and are then transported toward
shore and along the coast according to the
prevailing winds and currents. NOS has
begun experimental HAB forecasts in the
Gulf of Mexico based on satellite-derived
ocean color data and real-time wind observations (Stumpf et al., 1998). A regional
ocean forecast system could predict the trajectories and arrival times at specific locations of these harmful algal blooms
(f). Fisheries: For the purpose of fisheries management, model-generated fields of
temperature, salinity, and transport will be
of great value for applications where it is
necessary to recreate oceanic conditions for
past events that lead to changes in fish behavior and/or unexplained movements of
specific fish populations. Forecasts of surface and subsurface temperatures could be
used by commercial fishermen to make their
Spring 2004
Volume 38, Number 1
63
operations at sea more efficient by rapidly
locating areas which are potentially fish-productive. Maps of analyzed sea surface temperatures (SST) have been used by fishermen for many years for this purpose. Bottom temperatures are also of interest to the
fishermen since they influence the reproduction and recruitment of certain fish which
spend part or all of their life in this environment. Just south of Cape Hatteras, for example, lies an area called Big Rock where
local upwelling contributes to the abundance
of marine life. Several bottom fish including snapper and grouper are plentiful in the
Big Rock area, and, as a result, both commercial and recreational fishing take place
there. Marine aquaculture, or mariculture,
is another activity that could benefit from
information provided by a coastal ocean forecast system. To successfully culture marine
fish and shellfish commercially, information
on the local ambient water conditions is required. If changes in temperature and/or
salinity are too large or too rapid, the fish
under cultivation may be harmed or killed.
(g). Protected Marine Areas: At the
present time there are 11 National Marine
Sanctuaries located in U.S. coastal waters.
These sanctuaries are managed by NOAA
for protecting a variety of selected marine
habitats. This mission includes restoring and
rebuilding marine habitats or ecosystems to
their natural condition as well as monitoring and maintaining areas which are presently in good health. In order to accomplish
these goals, information on the existing environmental conditions in these sanctuaries
from an operational coastal ocean forecast
system would be beneficial. For example,
ecosystem models for diagnosing the health
of the biological communities which inhabit
these sanctuaries will require information on
their physical state, including temperature,
salinity, and currents.
(h). Additional Factors: With respect to
the evolution of ocean forecasting, the development of Rapid Environmental Assessment
(REA), whose purpose is to provide environmental information in coastal waters on
time scales of use in producing “tactical” forecasts, is becoming an increasingly important issue for naval operations (Robinson and
64
Sellschopp, 2002). Although this development is primarily related to naval requirements (Curtin, 1999), it has direct application to civilian environmental assessment and
thus coastal ocean forecasting. A prime example of the need for rapid environmental
assessment in the civilian sector is the ability
to determine the initial state of the ocean
immediately following an oil spill.
In a cost/benefit analysis involving only
commercial shipping, recreational boating,
and fishing sectors of the marine community, Kite-Powell et al. (1994) estimated that
the total expected benefits from an improved
marine forecasting capability will exceed the
costs of developing and implementing an
operational coastal ROFS by more than an
order of magnitude. When benefits to other
marine users (such as offshore gas and oil
industry, the marine scientific and recreational community, and federal, state, and
local coastal resource managers) are taken
into account, the overall benefits relative to
the estimated costs become even greater.
Table 1 summarizes the requirements for a
ROFS capability and reflects the comments
and suggestions provided by many individuals and sources.
3. Historical Evolution of
Ocean Forecasting
(a). Development of Ocean Circulation
Models: Smagorinsky (1963) recognized
the need to develop ocean circulation models to better understand atmosphere–ocean
interactions on time scales suitable for climate studies. Subsequently, the development
of ocean circulation models has received a
great deal of attention (see, for example,
Sarkisyan, 1962; Bryan and Cox, 1967; and
McWilliams, 1996).
In ocean forecasting, it is necessary to
distinguish between short (on the order of a
few days), and long-range (on the order of
seasonal to interannual) forecasting. General Circulation Models (GCM) for the
oceans seem to have been successful to some
extent in making long-range forecasts in the
tropics because the dominant dynamical
processes have much larger temporal and
spatial scales than their counterparts at mid-
latitudes and so can be explicitly resolved
(Philander, 1990). In short-range forecasting, however, the events of interest are frequently transient and tend to have time scales
of variability as short as an hour or less. This
makes them more difficult to forecast than
the signals associated with long-range forecasting. Also, as a general rule, the spatial
resolution for ocean forecasts needs to be
very fine since the energetic spatial scales of
interest for the ocean are small compared to
the atmosphere. Coastal areas require even
higher spatial resolution than the deep open
ocean because they possess inherently complex processes influenced by details of
bathymetry, shore line configuration, fresh
water discharges, and open ocean boundary
forcing. In such areas, spatial resolution of a
km or less may be needed and makes computer resources a critical factor.
The number of regional ocean models
available for predicting the state of the ocean,
particularly for coastal areas, has proliferated
in recent years. Haidvogel and Beckmann
(1998) evaluated fifteen coastal ocean models. All models are based on the primitive
equations and are fully nonlinear. But the
models differ in some details such as the use
of the rigid lid approximation instead of a
free surface, different vertical coordinate systems, different numerical approximations,
different time stepping schemes, and different sub-grid-scale closure schemes. Not surprisingly, when the results from various
models are compared, they often differ. In
particular, the combined effects of stratification and steep bottom topography typically encountered in the coastal ocean
present a particularly difficult problem for
most ocean models. Consequently, the problem of model selection is nontrivial and
clearly depends on the intended application.
(b). Development of Real-Time Forecast
Systems: A limited number of ocean models have been developed for operational use
in forecasting the state of the ocean on a real
time basis. During the 1980’s, Harvard
University developed an ocean model based
on quasi-geostrophic dynamics called the
Harvard Open Ocean Model which was
used to predict the path of the Gulf Stream
(Robinson et al., 1996). As a complete fore-
TABLE 1
User Requirements for Ocean Forecasts
Activity
Temperature
1
4
Forecast Variable
Salinity
Currents
Search & Rescue
X
Oil Spill Models
X
X
X
Estuarine Forecast Models
X
X
X
Ecosystem Models
X
X
X
Mariculture
X
X
X
X
Commercial Fishing
X
X
Commercial Shipping
X
X
Recreational Boating
X
X
Oil & Gas5
X
Fisheries Mgmt. & Research
X
Ship Routing
X
X
X
surface
ECD
hourly
0-48 hrs
X
EWC
WA3
4/day
0-72 hrs
X
EWC
WA
1/day3
UA4
EWC
WA
1/day
UA
X
X
X
X
X
0-48 hrs7
surface
ECD
4/day
0-72 hrs
EWC
ECD
4/day
0-72 hrs
EWC
ECD
4/day
0-72 hrs
surface
ECD
2/day
0-48 hrs
EWC
ECD
4/day
0-72 hrs
X
EWC
WA
4/day
0-72 hrs
EWC
ECD
Variable
Retrospective
X
surface
ECD
4/day
0-72 hrs
X
X
Forecast
Period
hourly
X
X
Forecast Characteristics
Area of
Frequency
Interest
ECD2
X
X
Depth
EWC1
X
Marine Weather Forecasting
Salvage & Mining
Water
Level
Military Applications
X
X
X
X
EWC
Littoral Zone
4/day
0-72 hrs
Coastal Zone Management
X
X
X
X
EWC
ECD
1/day
UA
Marine Science Community
X
X
X
X
EWC
WA
Variable
Variable
2
3
Entire water column ; Entire coastal domain for continental U.S.; Our best estimate
Unavailable; 5 Platforms/drilling; 6 Primary interest out to 200m depth; 7 Plus retrospective applications.
casting system, the model included an observational data network and statistical models which, together, provided the necessary
initial conditions to run the model operationally. This model was primarily intended
to forecast the evolution of the Gulf Stream
and its associated eddies, and not the circulation over the shelf. Consequently explicit
surface forcing from an atmospheric model
was not required. The system produced
weekly, seven-day forecasts between 1986
and 1989. A different forecast system, the
Great Lakes Forecasting System (GLFS), has
been developed by the Ohio State University and the Great Lakes Environmental
Research Lab/NOAA (Schwab and Bedford,
1994) to provide nowcasts and short range
forecasts of the physical conditions of some
of the Great Lakes. The primary components
of the GLFS are the Princeton Ocean Model
(POM) (Blumberg and Mellor, 1987) and
a wave model (Bedford and Schwab, 1994).
Because each of the Great Lakes is essentially a closed system with no open boundaries, the problem of prescribing lateral
boundary conditions does not arise. This
system is currently in the process of becoming fully operational at NOAA with NCEP
taking responsibility for wave forecasting,
and NOS for the circulation component.
A POM-based operational forecast system, forced by the Canadian Meteorological Center’s atmospheric forecast model, is
being developed to forecast the state of the
waters off the east coast of Canada
(Bobanovic and Thompson, 1999). Boundary conditions for the model’s open boundaries are obtained from a large-scale storm
surge model. The model domain includes
the Gulf of Saint Lawrence and the Scotian
shelf and has a horizontal resolution of 1/
16°x 1/16°in latitude and longitude. The U.
K. Meteorological Office (UKMO) has
implemented a global ocean circulation
model called the Forecasting Ocean-Atmosphere Model (FOAM). The model is based
on the primitive equations and has 20 layers in the vertical. It is forced by the UKMO’s
operational atmospheric forecast model (Bell
et al., 2001). Since the horizontal resolution
is 1° x 1°, only the general features of the
circulation can be represented in coastal ar-
eas. A coastal ocean forecast system called
SOPRANE (Système Océanique de
Prévision Régionale en Atlantique Nord Est)
is used by the French (Giraud et al., 1997)
as part of their ongoing SOAP (Système
Opérationel d’Analyse et de Prévision) program. The system is based on a 1/10° quasigeostrophic model of the Northeast Atlantic from 24°N–54°N, 35°W to the coast,
and terminating at the 200m isobath (but
not including the Mediterranean Sea). The
system runs every week providing a 2-week
forecast of the ocean circulation and thermal structure. A coastal ocean forecast system is also being developed by the Norwegians for the North Atlantic and Nordic Seas
with enhanced resolution in the European
coastal zones (e.g., Guddal, 1999). The primary purpose of this effort is to develop an
advanced data assimilation system to be used
with a coupled primitive equation ocean circulation model together with a marine ecosystem model for the regions indicated.
More recently, the European community has
developed a COupled Hydrodynamical
Ecological model for REgioNal Shelf seas
Spring 2004
Volume 38, Number 1
65
(COHERENS). This model is fully threedimensional and is intended for use in coastal
and shelf seas. A full description can be obtained at: http://www.mumm.ac.be/
~patrick/mast/. Although this state-of-theart coastal circulation model is presently
being used primarily for research, it is likely
that it will find operational use in the near
future. Finally, the U.S. Navy is extensively
involved in ocean forecast model development, implementation, and data assimilation. Recent summaries of this work can be
found in Oceanography (2002).
4. Description of East Coast
- Regional Ocean Forecast
System (EC-ROFS)
An earlier version of ROFS, which
started as a joint effort between NOAA’s
NWS, NOS, and the Princeton University
was called the Coastal Ocean Forecast System (COFS), and was described in Aikman
et al. (1996). However, many changes to
the system have taken place since then, including its name (now EC-ROFS).
(a) Selection of Forecast Domain: An
area off the East Coast of the U.S. was chosen as the pilot domain to test the feasibility
of producing real-time coastal ocean forecasts. The model domain extends from approximately 26.5° to 48°N, and from the
U.S. East Coast out to 50°W (Fig. 1). The
choice of the U.S. East Coast was made because the Gulf Stream (GS), which covers a
major portion of the domain, provides a
robust signal and may be somewhat better
understood compared to the California
Current System off the West Coast. Also,
the quality of the atmospheric forcing is better determined off the East Coast because
of the large number of upstream (i.e., continental) meteorological observations that are
available for assimilation into the atmospheric forecast model. The model domain
covers approximately 4.27x106 km2 and contains one landward boundary and two open
boundaries, one along its southern and the
other along its eastern extremities. It was
recognized from the outset that the task of
specifying the open boundary conditions
along its southern and eastern boundaries
66
would be problematic. The model domain
encompasses a number of major ocean features such as different water masses, currents,
frontal zones, and river plumes. Some of the
features are permanent, but transient features
such as eddies associated with the GS also
occur in the region. The circulation off the
East Coast domain is also significantly influenced by outflows from the major rivers
and bays. The five largest outflows of low
salinity water along the east coast are produced by the St. Lawrence River, Connecticut River, New York Harbor, Delaware Bay,
and the Chesapeake Bay. Low salinity waters are discharged from each of these sources
producing plumes which may extend 50km
or more offshore. In the case of the St.
Lawrence River the outflow extends across
a region which is approximately 100 km
wide. For the Connecticut River, the outflow is discharged initially into Long Island
Sound where it spreads to the east past the
tip of eastern Long Island and onto the continental shelf. These plumes of low salinity
water add buoyancy to the shelf waters and
are affected by the earth’s rotation. Discharge
from rivers further south along the east coast
such as the Santee River in South Carolina
and the Savannah River in Georgia also pro-
duce plumes but are smaller in scale and so
are not well-resolved in the model at the
present time.
(b) The Model: The Princeton Ocean
Model (POM) is used to generate forecasts
produced by EC-ROFS. The POM is a
three-dimensional ocean circulation model
based on the primitive equations and employs a free surface. It uses a terrain-following sigma coordinate in the vertical, and a
coastal-following curvilinear grid in the horizontal. The model has 19 levels in the vertical with higher resolution in the mixed layer
and the upper thermocline. The spatial resolution increases from 20 km offshore to10
km near the coast. The coastal boundary
corresponds to the 10 m isobath. The model
bathymetry is based on the U.S. Navy’s Digital Bathymetric Data Base with 5-minute
resolution (DBDB-5). Improvements to the
DBDB-5 bathymetry have been incorporated over the continental shelf and slope
using recently acquired bathymetric data
from NOS at 15-second resolution (Wei,
1995). The momentum equations are fully
nonlinear with a variable Coriolis parameter and a second order turbulent closure
submodel to parameterize vertical mixing.
Horizontal diffusion is based on the pa-
FIGURE 1
EC-ROFS domain including the horizontal grid, the major bathymetry, inflow and outflow boundary conditions
along the open boundaries, and the rivers that currently discharge fresh water into the model domain.
rameterization of Smagorinsky (1963). The
prognostic variables are temperature, salinity, and the horizontal components of velocity, and the free surface. For a complete
description of the POM and its numerical
schemes, see Blumberg and Mellor (1987).
(c) Surface Forcing and Lateral Boundary Conditions: NCEP’s operational Eta
model (see http://www.nco.ncep.noaa.gov)
provides the surface fluxes of heat, moisture,
and momentum every three hours. The forecast parameters from the Eta model are available at a height of 10 m above the surface.
The specific parameters which are extracted
are the sensible and latent heat fluxes, the net
shortwave and downward longwave radiation
fluxes, friction and wind velocities, and the
precipitation minus evaporation. The current
version of EC-ROFS includes astronomical
tidal forcing along the open boundaries and
body forcing within the model domain for
six tidal constituents: three semi-diurnal components (M2, S2, and N2) and three diurnal
components (K1, O1, and P1). A least squares
optimization technique was developed to
determine the tidal forcing on the open
boundaries using tidal constants within the
model domain (Chen and Mellor, 1999).
The model is driven along its open
boundaries using climatological estimates of
temperature and salinity from the Navy’s
Global Digital Environmental Model
(GDEM), and volume transport which is
specified separately. Along the southern
boundary (Fig. 1), inflows totaling 58.25
Sverdrups (Svs) and an outflow of 36.25
Svs are prescribed and are distributed horizontally in accordance with measurements
made during the SubTropical Atlantic Climate Studies (STACS) program (Leaman
et al., 1987). Along the eastern boundary at
50°W, 90 Svs exit the domain between 37°
and 40°N reflecting the expected transport
associated with the Gulf Stream at that location. Inflow north of the GS represents
the estimated transport associated with the
Labrador Current Extension (38 Svs), and
inflow to the south represents inflow associated with the subtropical recirculation gyre
(30 Svs). Temperature along the open
boundaries is based on the monthly GDEM
climatology whereas salinity is based on the
annual GDEM climatology. For additional
details concerning the specification of the
open boundary conditions, see Kelley et al.
(1999). Fresh water inputs are specified for
16 rivers, bays and estuaries along the U.S.
East Coast and are based on a stream flow
climatology by Blumberg and Grehl (1987).
The locations of rivers and bays that discharge fresh water into the model domain
are shown in Fig. 1, and monthly mean values from this climatology are used to prescribe the fresh water that is discharged.
(d) Operations: In order to start a forecast each day, a hindcast cycle is used to produce new initial conditions for that day using the following procedures. Sea surface
height anomalies (SSHA) from satellite altimeters are first assimilated into the initial
conditions from the previous day. Then starting with these modified model fields, the
POM is integrated forward to the current
time using analyzed fluxes during the last 24
hours provided by the Eta Data Assimilation
System (EDAS) while also assimilating SST
data obtained during the last 48 hours. (Methods used to assimilate SSHAs and SST’s will
be described later.) This completes the
hindcast cycle and provides new initial conditions for starting the forecast cycle each day.
Since January 1997, the output fields
from the model including both oceanic
nowcasts and forecasts as well as the atmospheric flux fields from the Eta model were
automatically transferred to National
Oceanographic Data Center (NODC) for
rapid online access to outside users. The results are available online at NODC for up
to three months and then are transferred to
permanent storage media for archiving. The
model output fields are also available from
the archives upon request (contact http://
polar.ncep.noaa.gov for information).
(e) Sample Forecast Products: The basic
forecast fields from EC-ROFS that are produced and examined routinely are nowcasts
and 24-hour forecasts of SST, temperature
at 200 meters, bottom temperature over the
continental shelf, sea surface salinity, salinity at 200 meters, surface currents, currents
at 200 meters, and finally, surface elevation.
Here we present several examples of these
forecast products.
In Fig. 2, an SST forecast valid at
0000UTC on August 26, 2003 is shown in
the upper left-hand corner. Cooler waters over
the continental shelf and warmer waters in
the Gulf Stream and Sargasso Sea are evident.
In the lower left-hand corner of the figure, a
FIGURE 2
Sample forecast products from EC-ROFS: SST (upper left), sea surface salinity (lower left), bottom temperature
over the continental shelf (upper right), and surface elevation (lower right).
Spring 2004
Volume 38, Number 1
67
forecast of surface salinity valid at 00 UTC
on August 26, 2003 is shown. The influence
of fresher waters from the Gulf of St.
Lawrence, and in the Gulf of Maine can be
seen. Although river discharge plumes of low
salinity water are often produced in the model
(Chesapeake Bay, for example), these plumes,
when they can be detected, are not expected
to be realistic since monthly climatological
streamflows are presently employed in the
model. Work is currently underway to replace
the climatological streamflows with daily observed streamflows from the U.S. Geological
Survey (USGS). Bottom temperatures from
the lowest level in the model for August 26,
2003 are shown in the upper right-hand corner of the figure. Because the vertical coordinate system in EC-ROFS is terrain-following (i.e., sigma coordinate), no conversion is
needed to display any of the output fields in
the bottom layer of the model. Bottom temperatures are of interest to fishermen and
marine biologists because many species of fish
reproduce and live at least part of their existence on, or near, the ocean bottom. Bottom
temperatures do not change rapidly on the
shelf but changes of several degrees can occur
over periods of several months. In this figure,
model-predicted bottom temperatures are
almost 10°C higher over the shelf south of
Cape Hatteras than they are north of Cape
Hatteras. Not surprisingly, model-predicted
bottom temperatures are very difficult to
verify since most in situ temperature observations do not reach the bottom. In the lower
right-hand corner of Fig. 2, a 24-hour forecast of surface elevation over the model domain is shown for August 26, 2003. A large
increase in surface elevation occurs across the
North Wall of the Gulf Stream. The surface
rises by as much as 70 cm proceeding from
the Slope Water, across the Gulf Stream, and
into the Sargasso Sea. Higher elevations are
also seen in the Gulf of Maine.
5. Evaluation of EC-ROFS
Forecast Fields
This section deals with a number of comparisons between model generated forecasts
and observations. The comparisons are limited to temperature and water levels since
68
FIGURE 3
Seven-day changes in SST from EC-ROFS (dashed), compared with seven-day changes in SST from buoy
44028 (solid) located in shallow water just off of Buzzard’s Bay, Massachusetts for October through December
1995, prior to data assimilation (top). Comparison of nowcast SSTs from EC-ROFS (dotted) with observed
SSTs from buoy 44138 off the Grand Banks for a 150-day period from June to December 1997(after SST data
assimilation was implemented - bottom). (COFS3.2n was an earlier version of EC-ROFS)
observations of currents and salinity are generally not available. For temperatures, most
comparisons are made before and after introducing data assimilation into the model.
Comparisons of water levels at the coast are
made before and after tidal forcing was introduced into the model.
(a) Assimilation of Sea Surface Temperature (SST) Data: Prior to the assimilation of
SST data in EC-ROFS, large differences
existed between model generated forecasts
or nowcasts and observations. As an example
of the SST variability without data assimilation, Fig. 3 (top panel) shows a comparison
of SST differences between buoy observations and model predictions. In this panel,
7-day changes, stepping through the record
one day at a time, are compared at a near
coastal location off Buzzard’s Bay, Massachusetts. The comparison covered a period of
80 days from October–December, 1995.
Even though the general pattern of change
is remarkably similar, the variability is clearly
much greater in the observed changes than
it is in the predicted changes.
In order to minimize the temperature
differences between the model and the observations, assimilation of SST data, from
in situ and satellite observations received
during the most recent 48 hours, has been
implemented in EC-ROFS. The in situ observations are obtained from U.S. and Canadian fixed buoys, drifting buoys, CoastalMarine Automated Network (C-MAN) stations, and ships participating in the Voluntary Observing Ship (VOS) program.
Within the model domain there are 27 fixed
buoys and C-MAN stations and 5-10 drifting buoys which report SST on any given
day. The remotely-sensed observations consist of multi-channel SST (MCSST) retrievals derived from the Advanced Very High
Resolution Radiometer (AVHRR) on board
NOAA’s operational polar-orbiting satellites.
Each retrieval represents approximately an
8 x 8 km area. The number of retrievals in
the domain on a given day, depending on
cloud cover, ranges from 400 to 7000.
The data assimilation scheme is based on
three steps. In the first step, an SST correction
field is obtained using an equivalent variational
formulation. In the second step, the surface
FIGURE 4
Comparison of EC-ROFS 24-hour forecasts of SST before (dotted) and after (dot-dash) SST data assimilation
was implemented, with observed SSTs from buoy 44008 off Nantucket Island for a 12-day period in March
1997 (top). Comparison of EC-ROFS 24-hour forecasts of SST before (dotted) and after (dot-dash) SST data
assimilation was implemented, with observed SSTs from the C-MAN station off the mouth of Chesapeake Bay
(36.9°N, 75.7°W) for the same 12-day period in March 1997 (bottom). (COFS3.1 and 3.2 were earlier versions
of EC-ROFS. COFS3.1 was without data assimilation, and COFS3.2 was with data assimilation)
Spring 2004
Volume 38, Number 1
69
correction field is projected downward into the
mixed layer following the method of Chalikov
and Peters (1997) to create a 3-D correction
field for temperature. Finally, a nudging procedure is used to slowly apply the 3-D correction field through the model’s mixed-layer. See
Kelley et al. (2002) for details.
Fig. 3 (bottom panel) compares buoyobserved SSTs with model produced SSTs
at a buoy located near the Grand Banks after data assimilation was introduced. The
period covers approximately 150 days between June–December, 1997. Although the
patterns of change in SST are similar, there
are systematic differences between the observed and predicted values throughout the
record that might be missed if only the mean
difference was considered.
Fig. 4 shows comparisons of SST between two National Data Buoy Center
(NDBC) buoys and the model, the first off
Nantucket Island (top panel), and the second, off of the mouth of the Chesapeake
Bay (bottom panel). In both cases, the same
12-day period during March 1997 was employed. At the location off Nantucket Island,
the impact of data assimilation is to improve
the agreement between the model and the
observations by 2-3°C. Off the mouth of
the Chesapeake Bay, the improvement is
even more striking. In this case the improvement is closer to 5°C. At most buoys
throughout the model domain improvements of ~1°C or more were observed.
Next we compare vertical profiles of temperature from the model with observed profiles acquired using eXpendable
BathyThermographs (XBTs) (Fig. 5, top and
bottom panels). The profiles shown in the
top panel are located just beyond the continental shelf at approximately 38°N, 73°W
and were acquired on March 1, 1997. Even
though marked improvement is shown in
the profile with data assimilation when compared to the XBT profile, the vertical structure of the “improved” profile still does not
agree well with the in situ data. The profile
shape and the depth of the mixed layer are
clearly not in close agreement with the observed temperature profile. In the bottom
panel of Fig.5, temperature profiles in deep
water (> 4000m) south of Nova Scotia for
70
FIGURE 5
Vertical profiles of temperature from XBTs (solid line) compared with profiles from EC-ROFS before (shorter
dashes), and after (longer dashes) SST data assimilation was implemented, for a location just beyond the shelf
break at approximately 38°N, 73 °W, for March 1, 1997 (top). Vertical profiles of temperature from XBTs (solid
line) compared with profiles from EC-ROFS before (shorter dashes), and after (longer dashes) SST data assimilation was implemented, for a deep water location at approximately 41°N, 63.5°W, for March 4, 1997 (bottom).
(COFS3.1 and 3.2 were earlier versions of EC-ROFS. COFS3.1 was without data assimilation, and COFS3.2 was
with data assimilation)
March 4, 1997 are compared. Again, much
better agreement with the observed profile
is seen for the case where data assimilation
has been included even though the lack of
agreement at deeper levels still persists. Some
of the disagreement may be attributed to
factors such as parameterization of mixing
and lateral boundary conditions.
(b) Assimilation of Sea Surface Height
Anomalies (SSHA): SSHA’s obtained from
the altimeter aboard the TOPEX/
POSEIDON satellite are assimilated into the
ocean model using a method developed by
Ezer and Mellor (1997). The SSHA’s are
calculated from a three-year mean surface
elevation field. Optimal interpolation is used
to interpolate the SSHA’s along the satellite
tracks horizontally onto the EC-ROFS grid.
The assimilation technique assumes that the
SSHA and subsurface temperature and salinity are related. Using the POM as a basis,
correlations between SSHA’s and the vertical structures of temperature and salinity are
calculated for each grid point in the model
domain where bottom depth > 2000 m.
These correlations are seasonally dependent
and this dependence has been taken into
account in establishing the SSHA/subsurface temperature and salinity relationships.
These correlations are used as the basis for
assimilating the TOPEX altimeter data into
EC-ROFS. Because the technique only addresses the baroclinic structure, it can not
be used in shallow shelf areas where
barotropic contributions to sea surface elevation play an important role.
A control run without altimeter data
(Fig. 6a), and a parallel run with altimeter
data (Fig. 6b) were made for a period from
May through July 1999. The most recent
10 days of SSHA data from TOPEX are assimilated into the model. Fig.6b shows an
anticyclonic eddy near the GS at approximately 39.5°N, 65°W in the surface velocity field in the parallel run (with TOPEX
assimilation) that does not appear in the control run shown in Fig. 6a (without TOPEX
assimilation). The existence of this feature
was verified with imagery from the GOES8 satellite acquired at the same time which
showed a GS meander about to pinch off at
this location (not shown - see CMDP, 2001).
FIGURE 6
(a) A nowcast of surface currents from EC-ROFS for June 3, 1999 without the assimilation of TOPEX altimeter
data or Gulf Stream path data (Version II), during the CMDP (see text for details). (b) A nowcast of surface
currents from EC-ROFS for June 3, 1999 with the assimilation of TOPEX altimeter data and Gulf Stream path
data (Version III), during the CMDP. In the second case, (i.e., with data assimilation), a Gulf Stream eddy
appears at 39°N, 65°W, whose existence was verified independently with imagery from the GOES-8 satellite.
Spring 2004
Volume 38, Number 1
71
FIGURE 7
Comparison of an observed XBT profile with EC-ROFS using only SST assimilation (dotted line), and then using
SST plus SSHA assimilation (dashed line). The location is 37.3°N, 52.1°W for May 17, 1999. (CFS3.2 includes
only SST data assimilation, whereas CFS3.4 includes both SST and SSHA data assimilation. Both are predecessors of EC-ROFS)
FIGURE 8
Model and observed subtidal (30-hour low-pass filtered) water level time series at two stations (Eastport, ME
and Atlantic City, NJ) for July and August, 1996. The solid line is the observed data, the dashed line was for the
current version of EC-ROFS at this time (i.e., Version 3.0), and the large, dotted line is the nowcast/forecast
simulation.
72
Because the assimilation scheme is based on
the correlation between SSHA and subsurface temperature and salinity, it was anticipated that improvements might be obtained
in the subsurface temperature structure.
Fig. 7 shows a comparison of an observed
temperature profile (solid line) with profiles
from EC-ROFS, with only SST data assimilation (dotted line), and with both SST and
SSHA data assimilation (dashed line) for
May 17, 1999 at 37.3°N, 52.1°W. Some
improvement in the agreement between the
observed profile and the model-generated
profile that includes SSHA assimilation is
apparent. However, such improvement in
the vertical temperature (and salinity) structure was not true in all cases. Part of the reason for this could be related to the fact that
assimilation of SSHA data depends on using model-generated vertical correlations
between the surface elevation anomaly and
temperature (and salinity) and such correlations not only can not be expected to be accurate due to the inherent deficiencies of any
given model but can also contribute to deteriorating the intrinsic value of an otherwise valid observation.
(c) Water Levels: EC-ROFS has shown
considerable skill in predicting water levels
at the coast with, and without, tidal forcing
(e.g., Aikman et al., 1998). The highest skill
has been achieved for the subtidal water levels which are strongly influenced by the wind
forcing provided by the Eta model. In this
section, we present an evaluation of the
coastal water level forecasts produced by the
model with wind forcing only and with both
wind and tidal forcing using the pre-spring
2001 version of the model that used the previous day’s 24-hour forecast as the initial condition for the next day’s forecast. We also
show a forecast using the present 24-hour
hindcast cycle, as discussed in section (4d),
to generate initial conditions. Data from
NOS’s National Water Level Observation
Network (NWLON) gages along the North
American East Coast are used to evaluate the
coastal water level forecasts. Fig.8 shows a
comparison of water level observations with
forecasts of subtidal water levels at Eastport,
Maine, and Atlantic City, New Jersey, using
initial conditions from a hindcast simulation
versus the conditions from the previous day’s
24-hour forecast for the period from June
through August 1996. These results indicate
that the 24-hour hindcast cycle presently
being used further improves the model’s
subtidal response by about 20%. Wind generated responses are well represented in the
forecasts even though there are some occasional disagreements in phase and amplitude.
Tidal forcing was introduced into the
forecast system in May 1996. Both astronomical tidal forcing along the open boundaries and astronomical body forcing within
the model domain are included. A leastsquares optimization technique was devised
to solve for the boundary tidal forcing (Chen
and Mellor, 1999), wherein the boundary
forcing is represented by a series of modes
which are coupled to the model through a
response function that is determined by running the model. The optimal boundary forcing coefficients are obtained by minimizing
the error between the model and observations at tidal stations within the domain.
Twelve months of experimental results indicate that the tides improve the model
subtidal response at the coast, reducing RMS
errors by more than 10%.
(d) Evaluation of Results from the
Coastal Marine Demonstration Project: The
CMDP was a two-year program, initiated
in 1998, whose purpose was to demonstrate
the state-of-the-art in coastal marine forecasting. The program was sponsored by the
National Ocean Partnership Program. As a
partnership, eight organizations, including
the federal government, academia, and the
private sector, worked together to plan, prepare for, and conduct the CMDP. The study
area for this project included the Chesapeake Bay and the surrounding coastal
ocean (32°- 42°N, and from the coast out
to 70°W) and falls completely within the
EC-ROFS domain. The demonstration
consisted of two phases. The first phase took
place during June-July of 1999 and the second during February-April, 2000. A broad
cross-section of the marine community was
selected to evaluate the various nowcast/
forecast products that were generated and
distributed in real-time during these demonstration periods. Forecasters from
NCEP’s Marine Prediction Center (MPC,
which is now called the Ocean Prediction
Center), and NOAA’s Coastal Services Center (CSC) in Charleston, South Carolina
had the responsibility of evaluating specific
products from EC-ROFS for the CMDP.
During the first phase of the CMDP, the
following EC-ROFS-related products were
provided: SST, surface salinity, and surface
currents. During the second phase, two
additional EC-ROFS forecast products were
included: temperature at 50m, and bottom temperature. Only a summary of the
CMDP results are given below (see Szilagyi
et al., 2000 for details).
The MPC evaluated all of the products
that were generated from the ocean model
for the CMDP and noted several deficiencies. EC-ROFS had difficulty in predicting the correct location of the GS and its
associated eddies. In particular, unrealistic
behavior was observed just beyond Cape
Hatteras where an anomalous meander often developed. Also, SST gradients just
north of the Gulf Stream were too weak,
compared to independent analyses and
observations. Surface currents, particularly
over the continental shelf, often did not
reflect the prevailing background flow
which was to the southwest. Evaluations
by the CSC were based on comparisons
with AVHRR imagery received on site. The
most significant problems were the inability of EC-ROFS to reproduce the high thermal gradients associated with the North
Wall of the GS, and the anomalous behavior of the GS just beyond Cape Hatteras,
in agreement with the findings of MPC.
On the positive side, CSC indicated that
although significant problems in locating
the position of the GS did exist, these deficiencies were generally systematic so that
forecasters could make allowances for them
in their forecasts in a manner similar to the
way they normally handle known deficiencies in numerical weather prediction models. Due to the lack of data, salinity fields
were not quantitatively evaluated, but it was
noted that the freshwater plumes emanating from major bays and estuaries along
the east coast appeared to respond to wind
forcing in a realistic manner.
6. Problems: Past and Present
As indicated in section 1, the development of a system to forecast the state of the
coastal ocean is one of the most difficult tasks
that faces the modeling community. Consequently it should come as no surprise that
numerous problems have arisen during the
course of developing EC-ROFS. Some of
the problems are clearly related to the prescription of outer boundary conditions,
some are related to the lack of sufficient
ocean data and optimal data assimilation
techniques to improve the initial conditions
in the model, and others are related to deficiencies in model resolution, numerics,
parameterizations, physics, and the imposed
external atmospheric forcing. As discussed
in this section, some of these problems have
been resolved and some still remain to be
resolved (see Breaker and Rao, 1998, for
additional details).
(a) Anomalous Increase in SST: In the
early stages of evaluating COFS, the predecessor to EC-ROFS, a large positive bias in
SST developed over the model domain with
temperatures at least 5°C higher than observed values. This problem was traced to
the significantly higher values of net surface
heat flux from the Eta model compared to
surface heat fluxes from the Comprehensive
Ocean Atmosphere Data Set (COADS) climatology (Woodruff et al. 1987). The latent and sensible heat fluxes, and the incoming short wave radiation in the Eta model
were much higher than those normally expected over a wide range of atmospheric
conditions. As a result of these findings, several refinements have been made to the heat
flux parameterizations in the Eta model to
reduce the net heat flux (Black et al., 1997).
For the incoming short wave radiation, several new features were added including the
introduction of atmospheric absorption by
ozone and aerosols, and the replacement of
a circular orbit for the earth by an elliptical
orbit. The inclusion of these factors reduced
the incoming short wave radiation by approximately 10%. Certain other adjustments
were also introduced into the model to keep
the magnitude of the net surface heat fluxes
consistent with the expected climatological
values. Such uncertainties in the fluxes proSpring 2004
Volume 38, Number 1
73
vided by an atmospheric forecast model have
emphasized the need to carefully evaluate
the surface fluxes derived from NWP models before using them in an ocean model.
This experience has clearly demonstrated
that ocean models can highlight deficiencies in certain parameterizations in atmospheric models that might have otherwise
gone undetected.
(b) Specification of Lateral Boundary
Conditions: The model domain for ECROFS has large open boundaries along its
southern and eastern extremities. Adoption
of a limited-area model was dictated by the
need for relatively high spatial resolution inside the model domain and computational
constraints imposed by available resources.
However, adequate specification of the required open boundary conditions (OBCs)
along these boundaries has been, and continues to be, a serious problem (e.g.,
Westerink and Gray, 1991). Numerous methods have been used to address this problem
with varying degrees of success. See Johnsen
(1994) for an overview of these methods.
At the present time, climatological values of temperature (monthly), salinity (annual), and volume transport are used to
specify the OBCs in the model domain. For
temperature and salinity, the GDEM climatology has been employed. Estimates of the
volume transport into and out of the model
domain have been obtained from various
sources (see, for example, Hogg, 1992). Unfortunately, climatological values of temperature, salinity, and transport are not representative of the actual conditions and do not
contain the important mesoscale structure
and high frequency variability characteristic
of real-time ocean processes. As mentioned
earlier, the model domain was chosen to be
large to prevent boundary generated errors
from propagating into the areas of interest—
namely, the coastal region. However, in an
operational environment, the model runs
every day and errors from unrealistic OBCs
will eventually propagate into the coastal
region and effect the quality of the results in
spite of attempts to nudge the model towards
reality through data assimilation.
One of the problems evident in the forecast fields produced by the model is the con-
74
sistent lack of flow to the southwest over the
shelf and inner slope region that lies between
the Gulf Stream and the coast. This deficiency is almost certainly related to the
boundary conditions prescribed along the
eastern extremity as well as the fresh water
inflows on the landward boundary of the
model domain. Historic Eulerian current
meter data and Lagrangian trajectories from
drifters in this region consistently indicate
flow to the SW at speeds of up to 10 cm/
sec. Sensitivity studies were conducted to
determine if persistent flow to the SW could
be produced by modifying inflow conditions
along the eastern boundary north of the Gulf
Stream. As transport across the boundary
was increased, most of the additional inflow
which initially entered the domain, turned
to the south and then to the east, finally exiting the domain just south of the region
where it had been injected, i.e., just north of
the Gulf Stream. This experiment showed
that intuition does not always lead to the
desired results!
An alternate approach to specifying the
OBCs is to embed or nest the regional model
within a basin scale model. Oneway or
twoway coupling between the models along
their common boundaries will provide the
regional model with the required real-time
information on lateral forcing. As discussed
in Warner et al. (1997), however, model
nesting also has a number of limitations generally related to mis-specification of the lateral boundary conditions. They include
changes in spatial resolution at the boundary between the models, poor initial information from the global model, differences
in the process parameterizations between the
models that can lead to spurious property
gradients at the boundary interface, and, finally, the generation of transient disturbances at the interface that may interact with
the desired solution on the interior of the
regional model domain. However, following the example of model nesting in numerical weather prediction, efficient nesting techniques need to be introduced to develop limited area circulation models for the coastal
ocean. Such an effort is currently underway
at NCEP using the Hybrid Coordinate
Ocean Model (HYCOM) system as the ba-
sis (see Bleck, 2002 for details on the
HYCOM system).
(c) Freshwater Influxes and Coastal Salinities: Along the landward boundary of the
EC-ROFS, 16 bays, rivers, and estuaries discharge fresh water into the model domain
that have a major impact on the distribution of salinity near the coast. As a result, in
many coastal areas, the circulation may be
primarily governed by salinity and not by
temperature. This was clearly shown to be
the case for the low salinity plume off the
Chesapeake Bay (Breaker et al., 1999), for
example. Improved freshwater fluxes along
the coastal boundary of the model domain
are essential to describe salinities and the
primary circulation characteristics near the
coast in a more realistic manner.
At this time, the specification of freshwater discharge for 16 coastal entry points
is based on the monthly climatology of
Blumberg and Grehl (1987) which does not
contain information on major episodic
events such as tropical storms and hurricanes,
or periods of drought, deficiencies that may
lead to significant departures from the climatology. In order to improve this situation,
efforts are underway to replace the monthly
climatological outflows used presently in the
model with observed daily values from the
USGS’s network of gages that measure
streamflows for all of the major rivers in the
U.S. In some cases, readings from one gage
may be representative of the actual outflow
into the model domain. However, in cases
like the Chesapeake Bay, estimating the total outflow at the mouth of the bay is problematic since at least nine rivers discharge
waters into the bay, and the time required
for these waters to circulate through the bay
is difficult to estimate. In some cases,
groundwater contributes to the outflow, further complicating the problem.
(d) Ocean Data Assimilation: An accurate specification of the initial conditions is
a necessary pre-requisite to produce reliable
forecasts from any model. This is accomplished through the incorporation of advanced data assimilation techniques into the
nowcast/forecast system. At the present time,
SST’s from satellite retrievals and from in
situ reports are being assimilated and their
influence is projected down through the
mixed layer. The vast majority of SST data
comes from satellites and so their availability depends on cloud cover. In the GS region, a primary area of interest, cloud cover
is a persistent problem. The time scales of
variability for the Gulf Stream are as short
as 2 - 3 days, and frequently, several days or
more elapse before new coverage can be obtained in this region. Hence, the fact that
the distribution and density of available satellite-derived SSTs is cloud cover-dependent
presents a major problem for ocean data assimilation. It is also important to assimilate
data at deeper levels, particularly in the area
of the GS in order to reproduce realistic surface (and subsurface) flow fields. The only
sources of available subsurface data are from
XBT’s and ARGO type floats. But, unfortunately, data from these sources are sparse.
Often less than 10 XBT’s are available within
our model domain on any given day and
their distributions are usually unfavorable for
resolving the features of interest. Nevertheless, methods are being tested to assimilate
data from XBT’s and ARGO and PALACE
floats. In assimilating XBT data, making
corrections to subsurface temperature alone
is not necessarily sufficient to bring the
model fields closer to reality. It is also necessary to make corresponding adjustments to
the associated salinity field to prevent potential gravitational instabilities (Chalikov et
al., 1998).
Surface elevation anomalies from altimeter
data, as discussed earlier, are being assimilated to correct the subsurface temperature
and salinity structure. There are problems,
however, with the existing data, and the assimilation scheme for application to highresolution, real-time regional ocean forecast
models, particularly in coastal areas. For the
TOPEX/POSEIDON satellite, for example,
adjacent track lines are approximately 250
km apart with a repeat cycle of 10 days. With
a track spacing this coarse, many mesoscale
ocean features are missed, and with a repeat
cycle of 10 days, it is difficult to consider
these data suitable for real-time forecast applications. Perhaps even more serious problems relate to how the data are being assimilated. In particular, using vertical correlations
generated from an imperfect model to project
the SSHA into the model interior to correct
the baroclinic part of the model dynamics is
likely to produce undesirable effects. It is necessary to develop methods to use the
altimetric data so that the assimilation procedure includes corrections to the barotropic
contributions, as well, which play a significant role in the circulation of the coastal
waters on the continental shelf.
Since in situ measurements of ocean currents are costly and time consuming to acquire, adequate ocean current data are practically non-existent for assimilation purposes.
Periodically, a few research sites may provide current measurements over some regions but they only operate for limited periods of time and thus are not suitable for
operational models. There are now plans in
progress to deploy comprehensive ocean
measurement networks, including currents,
along the coastal areas of the U.S. under the
aegis of programs such as the Coastal Ocean
Observing System (COOS; e.g., Seim,
2003). When these programs are fully established and become operational, they
would be invaluable sources of data for assimilation into, and improvement of, ocean
forecast models. In the meantime, satellite
feature tracking procedures could be used
to produce ocean surface currents from the
AVHRR and ocean color imagery which is
now available from a number of operational
satellites. The feasibility of producing such
information on an operational basis has already been established (Breaker et al., 1996).
Unfortunately, however, there is no ongoing effort to produce surface current information from these data sources.
The availability of salinity data from direct measurements would be extremely helpful in near-coastal areas. But again, there are
currently few, if any, observations of surface
salinity available anywhere around the world
on a realtime basis. As a result, new approaches to acquiring information on salinity are required. Remote sensing techniques
using microwave sensors may offer at least a
partial solution to this problem (Miller et
al., 1998). A second possibility is through
the use of Color Dissolved Organic Matter
(CDOM), which can be derived from ocean
color satellite data and related to salinity (see,
for example, Carder et al., 1993). Although
such a relationship has only been verified in
certain coastal regions, and will most likely
be location-specific, it may be possible to
use ocean color from Sea Viewing Wide
Field-of-View Sensor (SeaWiFS) to derive a
proxy for salinity in areas where such relationships can be established and validated.
When the COOS is fully implemented, salinity data in coastal regions around the U.S.
may become available for use in EC-ROFS.
Several mathematical techniques exist for
assimilating data into ocean models but they
require information on the error statistics
and spatial covariance structures for the
model-minus-observation increments for
each ocean parameter of interest. Unfortunately, this information is poorly known for
most models at the present time. As a result,
parallel model runs need to be initiated to
determine the sensitivity of the model to variants of the default values which are presently
being used to represent these statistics and
which may lead to improvements in the existing assimilation procedures. Finally, the
Global Data Assimilation Experiment
(GODAE) is a project intended to make
better use of various remotely sensed and in
situ data, and to develop effective data assimilation techniques which may be of benefit to operational coastal circulation models such as EC-ROFS in the near future
(http://www.bom.gov.au/bmrc/ocean/
GODAE/).
(e) Reproducing a Realistic Gulf Stream:
A problem in the GS separation occurs frequently in the EC-ROFS off Cape Hatteras.
A persistent anticyclonic meander develops
just north of the Cape centered at approximately 36°N and 74°W. This problem arises
in most ocean circulation models. Although
SST data assimilation appears to significantly
reduce this artifact, the unrealistic meander
gradually reforms when SST data are not
available in this region for several days. Several factors may contribute to this behavior
(Dengg et al., 1996). Model speeds in the
core of the GS, and also elsewhere in the
GS, are usually lower than observed (up to
50% lower in some cases). In the region off
Cape Hatteras, the 10 km spatial resolution
Spring 2004
Volume 38, Number 1
75
of the model may not be fine enough to
maintain the necessary jetlike structure.
Without sufficient resolution, there may be
a tendency for the Gulf Stream to spread laterally which could contribute to the formation of the anomalous meander. The GDEM
climatology used to spin up the model does
not contain a realistic GS, which could also
contribute to the separation problem. The
bathymetry is very complex near Cape
Hatteras, and higher resolution bathymetry
may be required locally to provide the correct topographic influence in this region.
(f) Shelf Circulation in the Mid-Atlantic Bight: Surface flow on the continental
shelf between Cape Hatteras and Long Island is generally to the South (e.g., Beardsley
and Boicourt, 1981). However, as indicated
earlier, the Coastal Marine Demonstration
Project showed that EC-ROFS-generated
surface flows in this region were generally to
the North. There are several possible explanations for the general lack of equatorward
flow along the shelf between Cape Hatteras
and Long Island in the model. The existence of an alongshore pressure gradient has
long been postulated as the primary cause
of southerly flow on the shelf along the U.S.
East Coast. Because the source of this alongshore pressure gradient may lie outside the
model domain, the model itself may not be
responsible for producing incorrect flow
along the shelf. However, other factors may
contribute to this problem. Buoyancy fluxes
along the east coast may be too small. We
know, for example, that several of the lesser
rivers along the East Coast are missing from
the model, and less fresh water on the shelf
may have an impact on the cross-shelf density gradient. Circulation in the cyclonic gyre
that lies in the Slope Water region between
the continental shelf and the GS may influence the flow on the shelf itself, and the expected circulation in the Slope Water region
is poorly reproduced in the model. Also,
when the anomalous meander just north of
Cape Hatteras is well developed, it may act
to block equatorward flow along the shelf.
Finally, boundary forcing along the eastern
boundary of the model domain may be incorrectly specified resulting in flow along the
shelf which is likewise incorrect.
76
Concluding Remarks
Some successes and a certain number of
problems have occurred during the development of EC-ROFS. Model performance
near the coast, at least in terms of water level,
was found to be good because of the
barotropic nature of water level variations.
Observations have verified this expectation.
For some of the problems which have been
identified, solutions or at least partial solutions have been found or are close at hand.
Problems related to the specification of the
lateral boundary conditions along the two
large open boundaries, for example, may be
significantly reduced by prescribing more
realistic boundary conditions provided by
using a basin scale model. Replacing the
existing monthly streamflow climatology
with daily observed streamflows from the
USGS should improve predicted salinities
and currents near the coast. In this regard,
better methods need to be developed to estimate inflows into the domain from the
connecting rivers and estuaries. As a case in
point, there is currently no simple way to
estimate the outflow from the Chesapeake
Bay based on the inputs from the major rivers which discharge waters into the bay.
Since the availability and distribution of
oceanographic data are poor compared to
the atmosphere, increased efforts are needed
to develop effective ocean data assimilation
techniques. For real time applications, the
only data types that are routinely available
are SSTs, vertical temperature profiles from
XBTs and ARGO and PALACE type floats,
and altimeter data. The availability of satellite-derived SSTs depends on cloud cover,
and the number of XBTs that are available
are usually small in number and poorly distributed. The utility of altimeter data for
assimilation into EC-ROFS is still open to
question with regard to how the anomalies
in surface elevation are defined, and the
space/time coverage that is presently available. Salinity data to be used for assimilation are very sparse and the possibility of
extracting information on salinities from
ocean color satellite data is exciting and
should be pursued. The newly-developed
Scanning Low Frequency Microwave Radiometer that infers surface salinity from low-
flying aircraft should be used routinely in
coastal areas around the continental U.S.
where the technique can be applied. Advanced three-dimensional multivariate
analysis techniques must be developed to assimilate all types of available ocean observations to improve the initial conditions for
EC-ROFS and similar models.
The CMDP demonstrated that forecast
products from EC-ROFS can be used by forecasters by taking into account certain model
deficiencies because these deficiencies are
known and systematic in nature. This situation is very similar to the atmospheric case
where forecasters generally use the model forecasts (with their known biases and deficiencies) together with observations that may not
have gotten into the model and their own
experience in producing a final forecast. The
same approach could be used by the marine
community in making ocean forecasts.
The development of EC-ROFS has been
a truly collaborative effort involving numerous individuals, groups, and organizations.
The path toward operational implementation has been long and at times circuitous.
Further improvements in ocean model development will most likely be slow and, at
times, painful, similar to the experience in
atmospheric forecast model development.
Just as in the case of the early days of Numerical Weather Prediction, “further improvements will be a slow and generally
unspectacular process” (Thompson, 1983).
The EC-ROFS development described here
is a first step in providing real-time forecasts
on the physical state of the coastal ocean and
in the transfer of techniques from research
to operations. Future improvements to ECCOFS will include extension of its coverage to include all U.S. coastal areas, running
the model every 12 hours to support operational estuarine circulation models, extending the model forecasts out to longer time
periods (up to several days), and interactive
coupling to other NWP models, wave models, and sea ice models. In closing, although
EC-ROFS is still a work in progress, it became fully operational in March 2002, and
is the first forecast system of its type to become operational in the civil sector of the
United States.
Acknowledgments
We thank George Mellor and his colleagues at Princeton University, and present
and former NCEP and NOS personnel for
their contributions to EC-ROFS. In particular, we are grateful to Dmitry Sheinin and
Lech Lobocki for their contributions in
implementing EC-ROFS at NCEP in the
early days of this activity. Finally, funding
from the National Ocean Partnership Program to conduct the CMDP has given us
the opportunity to demonstrate the present
capability in forecasting the state of the
coastal ocean.
References
Aikman, F., 1984. Pycnocline development
and its consequences in the Middle Atlantic
Bight. J Geophys Res. 89:685-694.
Aikman, F., L.C.Breaker, T. F. Gross, J.G.W.
Kelley, J.T. McQueen, G.J.Szilagyi and H.J.
Thiebaux. 2001. The Coastal Marine
Demonstration Project (CMDP): Experimental
Marine Forecast Models for the Chesapeake
Bay and Contiguous Coastal Region.
Submitted to the Bull Amer Meteor Soc.
Aikman, F., G.L. Mellor, T. Ezer, D. Sheinin,
L.C. Breaker, K. Bosley, P. Chen, and D.B.
Rao, 1996. Toward an operational nowcast/
forecast system for the U.S. East Coast. In:
Modern Approaches to Data Assimilation in
Ocean Modeling. eds. P. Malanotte - Rizolli,
and D. Halpern. Elsevier Oceanography Series.
61:347-376.
Aikman, F. III, E.J. Wei, and J.R. Schultz.
1998. Water level evaluation for the Coastal
Ocean Forecast System. In Second Conference
on Coastal and Atmospheric Prediction and
Processes. American Meteorological Society
75th AMS Annual Meeting, 11-16 January
1998, pp. 1-6.
Beardsley, R.C. and W.C. Boicourt, 1981. On
estuarine and continental-shelf circulation in
the middle Atlantic Bight. In: Evolution of
Physical Oceanography. eds. B.A. Warren and
C. Wunsch. Pp. 198-233. Cambridge, MA:
The MIT Press.
Bender, M.A. and I. Ginis, 2000. Real case
simulations of hurricane-ocean interaction
using a high resolution coupled model: Effects
on hurricane intensity. Mon Wea Rev.
126:917-946.
Black, T.L., 1994. The new NMC mesoscale
Eta model: description and forecast examples.
Weather and Forecasting. 9:265278.
Black, T.L. et al., 1997. Changes to the Eta
forecast systems. TPB No. 441, NWS, NOAA,
U.S. Department of Commerce, 6 pp.
Bleck, Rainer, 2002. An oceanic general
circulation model framed in hybrid isopycnicCartesian coordinates. Ocean Modelling.
37:55-88.
Blumberg, A.F. and B.J. Grehl. 1987. A river
flow climatology for the United States Atlantic
coast. Mahwah, New Jersey: Hydroqual, Inc.
6 pp. [Available from Hydroqual, Inc., I
Lethbridge Plaza, Mahwah, N.J. 07430]
Blumberg, A.F. and G.L.Mellor, 1987. A
description of a three dimensional coastal ocean
circulation model. In: Three Dimensional
Coastal Ocean Models, Vol. 4. ed. N.Heaps.
American Geophysical Union, pp. 116.
Bobanovic, J. and K.R.Thompson, 1999.
Prototype operational ocean forecast system
for the East Coast of Canada. In: Research
Activities in Atmospheric and Ocean
Modelling. ed. H. Ritchie. Report No. 28,
WMO/TD-No.942, pp 8.6, 8.7.
Boesch,D.F., D.M.Anderson, R.A.Horner,
S.E.Shumway, P.A.Tester and T.E.Whitledge,
1997. Harmful algal blooms in coastal waters:
options for prevention, control, and mitigation.
Science for Solutions, NOAA Coastal Ocean
Program, Decision Analysis Series No. 10,
National Oceanic and Atmospheric Administration, 49 pp.
Bosley, K. T. and K. W. Hess, 1997. Development of an experimental Nowcast/Forecast
system for Chesapeake Bay water levels. In:
Proceedings of the Fifth International
Conference of Estuarine and Coastal
Modeling. eds. M.L. Spaulding and A. F.
Blumberg. Pp. 413-426. American Society of
Civil Engineers, N. Y.
Breaker, L.C. and D.B. Rao, 1998. Experience gained during the implementation of
NOAA’s Coastal Ocean Forecast System.
Proceedings of the Ocean Community
Conference ’98, Vol. 1. pp. 235-241.
Baltimore, MD: Marine Technology Society
Annual Conference.
Breaker, L.C., J.G.W. Kelley, L.D. Burroughs,
J.L. Miller, B. Balasubramaniyan and J.B.
Zaitzeff, 1999. The Impact of a high discharge
event on the structure and evolution of the
Chesapeake Bay plume based on model results.
Journal of Marine Environmental Engineering
5:311-349.
Brink, K.H. et al. 1990. Coastal Ocean
Processes (CoOP): results of an interdisciplinary
Workshop. Woods Hole Oceanographic
Institution Contribution No. 7584. Woods
Hole, MA. 51 pp.
Brink, K.H. et al. 1992. Coastal Ocean Processes:
A Science Prospectus. Technical Report,
WHOI-92-18, Woods Hole Oceanographic
Institution, Woods Hole, MA. 88 pp.
Bryan, K. and M.D.Cox, 1967. A numerical
investigation of the oceanic general circulation.
Tellus. 19:54-80.
Carder, K.L., R.G. Steward, R.F. Chen, S.
Hawes and Z. Lee. 1993. AVIRIS calibration
and application in coastal ocean environments:
tracers of soluble and particulate constituents
of the Tampa Bay coastal plume. Photogrammetric Engineering & Remote Sensing.
59:339344.
Chalikov, D. and C. Peters. 1997. The NCEP
experimental ocean forecast model. In:
Research Activities in Atmospheric and
Oceanic Modelling, ed. A. Staniforth. Report
No. 25, WMO/TDNo. 792, 8.11.
Chalikov, D., L.C.Breaker, V.M. Krasnopolsky,
and D.B. Rao. 1998. Revisiting the question of
assimilating temperature alone into a full
equation of state ocean model. Ocean
Modelling. 116:13-14.
Chen, P. and G.L. Mellor. 1999. Determination
of tidal boundary forcing using tide station
data. In: Coastal Ocean Prediction. ed.
C.N.K. Mooers. AGU/CES series. pp. 329-351.
Spring 2004
Volume 38, Number 1
77
Coastal Marine Demonstration Project,
2000. Science and engineering of the Coastal
Marine demonstration Project. Project Report,
National Oceanographic Partnership Program,
managed by ONR in collaboration with the
Univ. of MD’s Center for Environmental
Science, Princeton University, Univ. of Rhode
Island, Litton-TASC, Inc., Weather Services
International, National Weather Service,
National Ocean Service, Office of Oceanic and
Atmospheric Research, and the U.S. Navy. 236
pp. [available from Coast Survey Development
Laboratory, National Ocean Service, Silver
Spring, MD 209103282] Curtin, T.B. 1999.
Preface. In: Naval Research Reviews. ed.
T.B.Curtin. Office of Naval Research, 51(2), i.
Dengg,J., A.Beckmann, and R.Gerdes. 1996.
The Gulf Stream separation problem. In:
The Warmwatersphere of the North Atlantic
Ocean. ed. W. Krauss. pp. 253-290. Berlin:
Gebrader Borntraeger.
Epstein, P.R. 1998. Marine ecosystems,
emerging diseases as indicators of change:
health of the oceans form Labrador to
Venezuela. Year of the Ocean Special Report,
Boston, MA, Center for Health and the Global
Environment, Harvard Medical School, Oliver
Wendell Holmes Society, 85pp.
Ezer, T. and G.L.Mellor. 1997. Data assimilation experiments in the Gulf Stream region:
how useful are satellite-derived surface data for
nowcasting the subsurface fields. J Atmos
Oceanic Tech. 14:1379-1391.
Fox, D.N., C.N. Barron, M.R.Carnes, M.
Booda, G. Peggion, and J.Van Gurley. 2002.
The modular ocean data assimilation system.
Oceanography. 15:22-28.
Giraud S., S. Baudel, E. Dombrowsky, and P.
Bahurel, 1997. The SOPRANE project:
realtime monitoring of the Northeast Atlantic
Ocean circulation Nowcast/Forecast for
oceanographic scientific campaigns. Proc. of
the Third International Workshop on
Information Technology: BIWIT: Data
Management Systems, The Biarritz International Symposium.
78
Gross, T., F. Aikman III, J. McQueen, K.
Fuell, K. Hess, and J. Kelley. 1999. Water
Level Model Response to Wind Forcing Over
the Chesapeake Bay During the Coastal
Marine Demonstration Project. Amer. Met.
Soc.’s (AMS) 3rd Conf. on Coastal Atmospheric
& Oceanic Prediction and Processes, 3-5
November 1999, New Orleans, LA.
Guddal, J. 1999: The Norwegian perspective.
In: Coastal Ocean Prediction. ed. C.N.K.
Mooers. Coastal and Estuarine Studies, 56. pp.
513-519. Washington, D.C.: American
Geophysical Union.
Kelley, J.G.W., D.W.Behringer, and
H.J.Thiebaux. 1999. Description of the SST
data assimilation system used in the NOAA
Coastal Ocean Forecast System (COFS) for
the U.S. East Coast Version 3.2. U.S. Dept. of
Commerce, NOAA, NWS, NCEP, OMB
Contribution Number 174, 49 pp.
Kelley, J.G.W., D.W.Behringer, H.J.Thiebaux,
and B. Balasubramaniyan. 2002. Assimilation
of SST data into a real-time Coastal Ocean
Forecast System for the U.S. East Coast.
Weather and Forecasting. 17:670-690.
Haidvogel, D.B., and A. Beckmann. 1998.
Numerical models of the coastal ocean. In:
The Sea, Vol. 10. eds. K.H. Brink and A.R.
Robinson. pp. 457 - 482.
Kite-Powell, H., S. Farrow and P. Sassone.
1994. Quantitative estimation of benefits and
costs of a proposed coastal forecast system.
Marine Policy Center. Woods Hole, MA:
Woods Hole Oceanographic Institution. 38 pp.
Hay, M. and P.Jumars. 1999. Collaborating
‘ocean ecologists’ assess achievements, prepare
for challenges. EOS, Transactions of the
American Geophysical Union. 80:77,81.
Leaman, K.D., R.L.Molinari, and P.S.Vertes.
1987. Structure and variability of the Florida
Current at 27°N: April 1982 - July 1984. Jour
Phys Oceanogr. 17:565-583.
Hinrichsen, D. 1998. Coastal Waters of the
World: Trends, Threats, and Strategies.
Washington, D.C.: Island Press. 275 pp.
McWilliams, J.C., 1996: Modelling the
oceanic general circulation. Ann Rev Fluid
Mech. 28:215-248.
Ho, F.P., J.C.Su, K.L.Hanevich, R.J.Smith,
and F.P.Richards. 1987. Hurricane climatology
for the Atlantic and Gulf coasts of the United
States. NOAA Tech. Memo. NWS-38, Silver
Spring, MD: National Weather Service. 195 pp.
Miller, J.L., M.A. Goodberlet and J.B.
Zaitzeff. 1998. Airborne salinity mapper makes
debut in coastal zone. EOS. 79:173.
Hogg, N.G. 1992. On the transport of the
Gulf Stream between Cape Hatteras and the
Grand Banks. DeepSea Res. 39:12311246.
Johnsen, M. 1994. Open boundary conditions
for finite element models for continental shelf
flow with applications to the Gulf of Maine.
Ph.D. Thesis, Dartmouth College, Hanover,
New Hampshire, (no pagination available).
Johnsen, M. and D.R. Lynch. 1995. Assessment
of a second-order radiation boundary condition
for tidal and wind driven flows. In: Quantitative
Skill Assessment for Coastal Ocean Models.
ed. D.R.Lynch and A.M.Davies. pp. 49-70.
Coastal and Estuarine Studies. Washington,
D.C.: American Geophysical Union.
Mooers, C.N.K. 1976. Overview of the
physical dynamics of the continental margin.
Bulletin of Hydrological Sciences. 21:467-471.
Mooers, C.N.K, Convenor, 1990a: The
Coastal Ocean Prediction Systems Program:
Understanding and Managing Our Coastal
Ocean, Volume I: Strategic Summary, May
1990. Report of a Planning Workshop Held
31 October to 2 November 1989 at the
University of New Orleans. Washington, D.C.:
Joint Oceanographic Institutions, Inc. 116 pp.
Mooers, C.N.K, Convenor ,1990b: The
Coastal Ocean Prediction Systems Program:
Understanding and Managing Our Coastal
Ocean, Volume II: Overview and Invited
Papers, May 1990. Report of a Planning
Workshop Held 31 October to 2 November
1989 at the Univ. of New Orleans, Washington,
D.C.: Joint Oceanographic Institutions, Inc.
354 pp.
National Research Council. 1989. Opportunities
to Improve Marine Forecasting. Committee on
Opportunities to Improve Marine Observations
and Forecasting, Marine Board, Commission
on Engineering and Technical Systems,
National Research Council. Washington,
D.C.: National Academy Press. 125 pp.
Natural Resources Defense Council. 1996.
Testing the waters: who knows what you’re
getting into. New York: NRDC. 116 pp.
Oceanography, 2002: Navy operational
models: Ten years later. Vol.15, No.1. 152pp
Philander,S.G. 1990. El Nino, La Nina, and
the Southern Oscillation. San Diego:
Academic Press. 293 pp.
Pielke, R.A.,Jr. and R.A. Pielke, Sr. 1997.
Hurricanes: Their Nature and Impacts on Society.
New York: John Wiley and Sons. 279 pp.
Rabalais,N.N., R.E.Turner and W.J.Wiseman.
1994. Hypoxic conditions in bottom waters on
the Louisiana-Texas shelf. In: Special NOAA
Report, Coastal Oceanographic Effects of
Summer 1993 Mississippi River Flooding. ed.
M.J. Dowgiallo, U.S. Department of
Commerce, National Oceanic and Atmospheric Administration, 50 - 54.
Raloff, J. 1999. Sea Sickness Marine epidemiology comes of age. Science News. 155: 72-74.
Rappaport, E.N. and J. Fernandez-Partagas.
1995. The deadliest Atlantic tropical cyclones,
1492-1994. NOAA/NWS Technical
Memorandum NWS NHC-47(January).
Coral Gables, FL, 41 pp.
Robinson, A.R., H.G.Arango, A.Warn-Varnas,
W.G.Leslie, A.J.Miller, P.J.Haley and C.J.
Lozano.1996. Real-time regional forecasting.
In: Modern Approaches to Data Assimilation
in Ocean Modeling, ed. P.Malanotte-Rizzoli.
Amsterdam: Elsevier. 455 pp.
Rogers, D.P. 1995. Coastal Meteorology. Rev
Geophys. 33, Supplement. [Available from
American Geophysical Union, 2000 Florida
Avenue, N.W., Washington, D.C. 20009.]
Sarkisyan, A.S. 1962. On the dynamics of the
origin of wind currents in the baroclinic ocean.
Okeanologie. 11:393-409.
Schwab, D. J. and K. W. Bedford. 1994.
Initial implementation of the Great Lakes
Forecasting System: A realtime system for
predicting lake circulation and thermal structure.
Jour Water Pollution Res. 29:203220.
Seim, H., 2003: SEA-COOS: A model for a
multi-state, multi-institutional regional
observing system. J Mar Tech. 7:92-101.
Smagorinsky, J. 1963. General circulation
experiments with the primitive equations, I,
the basic experiment. Mon Wea Rev. 91:99-164.
Schmalz, R.A. 2000. Demonstration of a
nowcast/forecast system for Galveston Bay. In:
Proceedings of the 6th International Conference on Estuarine and Coastal Modeling. eds.
M.L. Spaulding and H.L. Butler. pp. 868-883.
ASCE, 1801 Alexander Bell Drive, Reston, VA
20191-4400.
Semtner, A.J. 1995. Modeling Ocean Circulation.
Science. 269:1379-1385.
Rhodes,R.C., H.E. Hurlburt, A.J. Wallcraft,
C.N. Barron, P.J. Martin, E.J. Metzger,J.F.
Shriver,D.S.Ko, O.M.Smedstad, S.L.Cross and
A.B.Kara. 2002. Navy real-time global
monitoring systems. 15:29-43.
Stommel, H. 1958. The Gulf Stream: A
Physical and Dynamical Description. Berkeley:
University of California Press and London:
Cambridge University Press. 202 pp.
Robinson and Sellschopp. 2002. Rapid
assessment of the coastal ocean environment.
In: Ocean Forecasting Conceptual Basis and
Applications. eds. N. Pinardi and J. Woods.
pp. 199-230. Berlin: Springer.
Szilagyi, G.J., F.Aikman, and L.C.Breaker.
2000. Evaluation of the Coastal Marine
Demonstration. U.S. Dept. of Commerce,
National Oceanic and Atmospheric Administration, National Ocean Service, National
Weather Service, Silver Spring, MD, 39 pp.
Stumpf, R., V. Ransibrahmanakul, K.A.
Steidinger and P.A.Tester. 1998. Observations
of sea surface temperature and winds associated
with Florida, USA, red tides (Gymnodinium
Breve Blooms), harmful algae. Xunta de
Galicia and Intergovernmental Oceanographic
Commission of UNESCO, 4 pp.
Teague, W.J., M.J. Carron, and P.J. Hogan.
1990. A comparison between the generalized
digital environmental model and Levitus
climatologies. J Geophys Res. 95:71677183.
Thompson, P.D. 1983. A history of numerical
weather prediction in the United States. Bull
Amer Meteor Soc. 64:755-769.
U.S. Department of Commerce. 1993. A
Strategic Plan for a Coastal Forecast System.
National Oceanic and Atmospheric Administration, Coastal Ocean Program Office, Silver
Spring, MD 20910, 31 pp.
Warner, T.T, R.A. Peterson and R.E.Treadon.
1997. A tutorial on lateral boundary conditions as a basic and potentially serious
limitation to regional numerical weather
prediction. Bull Amer Meteor Soc.
78:25992617.
Wei, E. 2003. The New Port of New York and
New Jersey operational forecast system. Bull
Amer Meteor Soc. 84:1084-1086.
Wei, E.J. 1995. ECFS refined bathymetry
tests. ECFS Technical Note No. 956, 19 pp.
Westerink, J.J., and W.G. Gray. 1991. Progress
in surface water modeling. Rev Geophys.
29:210- 217.
Wiseman, W.J., C.N.K. Mooers and G.Z.
Forristall. 1984. Ocean current processes.
Ocean Science and Engineering. 8:367-459.
Woodruff, S.D., Slutz, R.J., Jenne, R.L. and
Steurer, P.M. 1987. A comprehensive oceanatmosphere data set. Bull Amer Meteor Soc.
68:12391250.
Yen, C.C., K. W. Bedford and D. J. Schwab.
1992. Nowcast protocol for the Great Lakes
Forecasting System. Proceedings, ASCE 2nd
International Conference on Estuarine and
Coastal Modeling, pp. 140-148. Amer. Soc.
Civil Eng.
Spring 2004
Volume 38, Number 1
79
C O M M E N TA RY
The Legal Status of Autonomous
Underwater Vehicles
AUTHOR
Stephanie Showalter
Director, National Sea Grant Law Center,
University of Mississippi
A
UVs, Autonomous Underwater Vehicles, are the cutting edge of technology
used to explore the world’s oceans. Today,
AUVs can explore areas of the oceans scientists only dreamed about mere decades
ago. These robots provide unprecedented
access to hydrothermal vents and other
mysteries of the deep. AUVs can swim under the polar ice caps and venture into underwater canyons. But scientists are not the
only group benefiting from these machines.
Once the exclusive purview of the United
States Navy and academic institutions, recent advances are bringing AUVs into the
commercial sector. AUVs can search for
offshore oil and mineral deposits, lay submarine cables, and search for mines. Private individuals and corporations can now
purchase AUVs for use in salvage operations, underwater archaeology, or simple
exploration. The possibilities appear limitless and the benefits incalculable.
Unlike tethered and remotely operated
vehicles which are a simple extension of the
research vessel, AUVs are, and legally should
be, considered separate entities. AUVs, as
the name suggests, are designed to operate
freely in the vast oceans. Ideally, AUVs would
be released and tracked from shore, eliminating the need for a costly support vessel.
The AUV’s autonomous nature, however,
creates a regulatory gap. AUVs, as discussed
in more detail below, may or may not be
vessels as defined by U.S. maritime laws. The
use of AUVs is virtually unregulated by the
federal government, mostly due to a combination of the newness of the technology,
difficulties with classification, and the unwillingness of overburdened federal agencies
to incur additional responsibilities.
80
No legal framework currently exists to
regulate the use of AUVs. Permits and licenses are only required in a few narrow circumstances. While there is no indication that
the oceans are in danger of being overrun
by AUVs, their growing availability and
popularity warrant investigation into the
potential regulatory implications of the
widespread use of AUVs. This commentary
examines the current legal status of AUVs
under U.S. law and suggests that a permitting regime may already exist.
Technology often outpaces regulatory
regimes, whose adaptability is hindered by
the legislative process and administrative
agency resources. In general, the international treaties and domestic law governing
marine activities apply only to vessels. While
AUVs are autonomous vehicles that operate on and below the service of the ocean,
the application of U.S. maritime laws, including the International Regulations for
Preventing Collisions at Sea (COLREG), is
unclear because these machines may not be
considered “vessels” under U.S. law.
A vessel “includes every description of
watercraft or other artificial contrivance used,
or capable of being used, as a means of transportation on water.” (1 U.S.C. § 3). The “vessel” test is simple: is the structure “fairly engaged in or suitable for, commerce or navigation and as a means of transportation on
water?” (Hitner Sons Co. v. U.S., 13 Ct. Cust.
216, 222 (1922)). For a boat, barge, or other
floating structure to be considered a vessel,
“it must have some relation to commerce or
navigation, or at least some connection with
a vessel employed in trade.” (Hitner at 222).
The current AUV models have no such
connection to commerce or navigation.
AUVs are used to study and explore the
ocean environment. The majority, due to
their size and design, are unable to be used
as a means of transportation for goods or
people on water. Small AUVs used for scientific purposes are probably not vessels subject to U.S. maritime regulations and need
not comply with the COLREGs.
Some AUVs, however, could be considered vessels and would be required to comply with the COLREGs and other maritime
laws. For example, research is underway to
develop cargo carrying AUVs to “deliver
payloads or cargoes [sonar arrays, underwater cables, scientific instruments, etc.] to
places that manned ships or submarines cannot operator cost-effectively or safely”
(Griffiths, 2003). Already the Canadian
Defense Research Establishment and the
U.S. Office of Naval Research have proved
that AUVs can be used to lay cables. In the
spring of 1996, during a cable laying mission in the Artic, the Theseus AUV laid two
fibre optic cables under the polar ice cap over
a distance of 175 km. (Griffiths, 2003). The
ability of certain classes of AUVs to operate
in commercial activities, such as laying cables
and carrying cargo, significantly alters the
legal analysis of whether AUVs are vessels.
If AUVs are used to carry cargo, a strong
argument can be made that they are also
vessels capable of being used for transportation on the water.
So let’s assume for a moment that AUVs
are vessels. One class clearly would have to
adhere to the COLREG provisions—the
semi-submersibles. A semi-submersible AUV
is “designed to operate like a snorkeling submarine and consequently, is limited to operations near the sea surface” (Griffiths, 2003).
Rule 22 of the COLREGs requires inconspicuous, partly submerged vessels to display
a white all-round light visible up to a minimum of three miles. Vessels are also required
to carry equipment for sound signals which
varies depending on the size of the vessel. Rule
33 states that vessels less than twelve meters
long are not obliged to carry the whistles and
bells required on larger vessels. However, if
the vessel is not so equipped, it must be provided with some other means of making an
efficient sound signal. Semi-submersible
AUVs should, therefore, also be outfitted with
some type of sound signaling device.
Unlike the semi-submersible AUVs, the
majority of AUVs are designed to operate
completely under the water. It is important
to note that the COLREGs are only applicable to vessels operating on the water. There
are no lighting and signal requirements for
underwater operations, unless a vessel on the
surface is engaged in underwater operations,
such as fishing or laying cables. Submarines
only have to display lights when operating
on the surface. There may be situations,
however, when the AUV might operate on
the surface. It may need to surface to send
or retrieve data or as part of its emergency
abort system. Once on the surface, the AUV
would be subject to the COLREGs.
For vessels less than twelve meters in
length, Rule 22 requires a masthead light,
sternlight, and towing light visible up to two
miles; a sidelight visible up to one mile; and
a white, red, green, or yellow all-round light
visible up to two miles. For vessels more than
twelve meters long but less than fifty meters
long, a masthead light, visible up to five
miles, is required unless the vessel is less than
twenty meters long. For vessels between
twelve and twenty meters long, the masthead light need only be visible for three miles.
A sidelight, sternlight, towing light, and a
white, red, green or yellow all-round light
must also be visible for a range of two miles.
Although it is unclear whether AUVs are
subject to the maritime regulations for vessels, to reduce damage and liability concerns,
it is advisable for AUV operators to adhere
to the COLREG provisions dealing with
lighting and signals when the AUV is on
the surface. While an AUV may not be able
to fully comply with these requirements due
to design limitations, comparable lighting
should be incorporated into the design
whenever possible. Failure to adhere to the
international lighting and signal requirements may result in a maximum civil penalty of $5,000 which can be assessed against
both the vessel operator and the vessel itself.
Proactive engineering may facilitate compliance with the COLREGs and actually eliminate the need to determine whether an AUV
is a vessel.
In addition to classification problems,
questions often arise regarding whether an
AUV operator needs to secure permits prior
to commencing research. To reduce user conflicts and minimize environmental impacts,
a permitting regime is necessary. The foundations of a regime are already in place. If
AUVs are to be used in foreign waters, authorizations must be obtained from the foreign nation in accordance with Part XIII of
the UNCLOS. Researchers may also be required to secure temporary export licenses
through the Departments of State and/or
Commerce for research activities in foreign
waters. In addition, federal permits are currently required for AUV activities impacting
the continental shelf, conducted within a
marine sanctuary, or impacting endangered
species or marine mammals.
Activities on the outer continental shelf
and in marine sanctuaries clearly require permits. The waters are much murkier, however,
if a researcher intends to use an AUV to explore U.S. waters outside a marine sanctuary
and without contacting the continental shelf.
While a researcher can restrict an AUV to a
particular area of the ocean, a researcher has
no control over whether animals enter the
designated area during data collection. The
ocean is not a static environment. Endangered
species and marine mammals move freely,
some over great distances.
The remainder of this Commentary focuses on marine mammal interactions. Because of the overwhelming number of legal
questions currently surrounding the use of
AUVs, I chose to limit my Comment to a
discrete area of the law. Marine mammal interactions, however, are not an AUV operator’s
most serious concern. An AUV is much more
likely to collide with a surface vessel or become entangled in a net. A longer article,
which will discuss a variety of AUV legal issues, including vessel collisions, net entanglement, salvage, and liability, is currently in the
draft stages.
AUV operators do need to be aware that
a regulatory regime exists to protect marine
mammals from noise and harassment. If the
use of an AUV will “take” an endangered species or a marine mammal, an incidental take
permit is required from the National Oceanic and Atmospheric Administration
(NOAA) within the Department of Commerce. An incidental take permit may be authorized under either the Endangered Species Act (ESA) or the Marine Mammal Protection Act (MMPA). The MMPA addresses
all interactions with marine mammal stocks,
regardless of their endangered status.
In theory, an AUV could result in the
take of a marine mammal in violation of
the MMPA. The MMPA defines “take” as
“to harass, hunt, capture, or kill, or attempt
to harass, hunt, capture, or kill any marine
mammal.” While a number of worst case
scenarios can be imagined, such as a marine
mammal–AUV collision, it is unlikely that
an AUV will kill or directly injury a marine
mammal. Most AUVs travel rather slowly,
averaging about three to eight knots, which
should allow any marine mammal plenty of
time to avoid the robot. Rather, the question is whether the operation of an AUV
would be considered harassment.
Operational noise is the most likely trigger for a violation of the MMPA. While the
actual AUV makes very little noise, AUVs
are used as sensor platforms and can be
equipped with a variety of scientific instruments, including multi-beam echo sounders, side-scan sonars, and sub-bottom
profilers (Griffiths, 2003). It is this sensory
equipment, not the AUV, which would trigger the application of the MMPA. The impact of anthropogenic (human-generated)
noise on marine mammals is not well documented, but some preliminary studies indicate that marine mammal “behavior responses [to noise] range from subtle changes
in surfacing and breathing patterns, to cessation of vocalizations, to active avoidance
Spring 2004
Volume 38, Number 1
81
or escape from the region of the highest
sound levels.” (National Research Council,
2003). On a basic level, therefore, the noise
generated by the surveying equipment on
an AUV could potentially disrupt the behavioral patterns of marine mammals.
Fortunately, it is not enough to simply
disrupt the behavioral patterns of marine
mammals or every marine activity would
violate the MMPA. To rise to the level of a
violation of the MMPA, the harassment
must involve a direct and significant intrusion on the normal behavioral patterns of a
marine mammal. Unless an AUV generates
a significant amount of noise, it is unlikely
that the use of an AUV would rise to the
level of a direct and significant intrusion.
Precedent does exists, however, for the
delay and/or prohibition of marine research
projects based on noise. In 2002, the District Court for the Northern District Court
of California in Center for Biological Diversity v. National Science Foundation enjoined
acoustical research by the National Science
Foundation (NSF) due to concerns over the
noise that would be generated by air guns.
The testing of sonar systems by the U.S.
Navy has also been delayed based on concerns regarding noise.
While the lack of a regulatory structure
for AUVs operations may not be high on
the federal government’s priority list, it
should be. As increasing numbers of AUVs
are utilized by the private sector and research
institutions, user conflicts and marine mammal interactions are inevitable. AUV operators have a right to be concerned regarding
their potential liability in the event of an
AUV malfunction or collision. While not
all operators will want to obtain permits or
notify NOAA of their activities, prudent
operators may want to consider obtaining
an Incidental Harassment Authorization
from NOAA under the MMPA.
As a general rule under the MMPA, the
Secretary of Commerce may issue permits
authorizing the taking of marine mammals.
Additionally, citizens of the United States
who engage in a specified activity other than
commercial fishing within a specific geographical region may petition the Secretary
to authorize the incidental, but not inten-
82
tional, taking of small numbers of marine
mammals within that region. “Small take”
authorizations, also known as Letters of
Authorization (LOA), may permit the direct taking of marine mammals through
death and/or serious injury. The process to
secure a “small take” authorization is rather
lengthy. Upon receiving an application,
NOAA must provide notice and an opportunity for public comment and issue regulations setting forth permissible methods of
taking and monitoring and reporting requirements. Recently, the United States
Navy utilized this provision of the MMPA
to obtain a “small take” authorization for its
operation of SURTASS LFA sonar systems.
LOAs usually involve the direct taking
of marine mammals through death or serious injury, and for AUV operators, the concern is not death or injury. Initiating the
LOA process for AUV operations is not advisable, therefore, due both to the considerable amount of time involved and low risk
of an actual taking. In fact, because of the
low risk of serious injury or mortality and
the fact that any potential for injury or mortality could most likely be mitigated, an LOA
is not needed. Rather, an AUV operator
should seek an Incidental Harassment Authorization or IHA.
An IHA allows the incidental, but not
intentional, taking of small numbers of
marine mammals of a species or population
stock. Incidental taking means an accidental taking—those takings that are infrequent
or unavoidable. The National Marine Fisheries Service defines “specified activity” as
“any activity, other than commercial fishing, that takes place in a specified geographical region and potentially involves the taking of small numbers of marine mammals”
The Secretary may issue an IHA only if
he or she finds that the harassment will have
a negligible impact on such species or stock
and will not have an unmitigable adverse
impact on the availability of the species or
stock for subsistence uses. A “negligible impact” is “an impact resulting from a specified activity that cannot be reasonably expected to, and is not reasonably likely to,
adversely affect the species or stock through
effects on annual rates of recruitment or sur-
vival.” The authorization must prescribe the
permissible methods of taking by harassment, measures determined by the Secretary to be necessary to ensure no unmitigable
impact, and monitoring and reporting requirements. Most importantly for applicants, the approval process is extremely
streamlined. Within 45 days of receiving an
application for an IHA, the Secretary must
provide public notice and solicit comments
for 30 days. The Secretary is then required
to issue the authorization, with the appropriate conditions, within 45 days of the closure of the public comment period.
For the NMFS “to consider authorizing
the taking of marine mammals incidental
to a specified activity, or to make a finding
that an incidental take is unlikely to occur,”
the applicant must submit a written request
to the Office of Protected Resources and the
Regional Office where the specific activity
is planned. It is the above italicized language
that indicates the IHA process could easily
be used to determine whether AUV operations need permits. To date, most IHAs have
authorized the incidental harassment of
marine mammals through activities involving noise, including sonar and seismic testing. The potential application of the IHA
program, however, is quite broad.
For example, in May 2003, NOAA issued an IHA for construction activities in
Monterey, California. The United States
Coast Guard applied for an IHA for the
possible harassment of small numbers of
California sea lions and Pacific harbor seals
incidental to the installation of a new floating dock. It was estimated that as many as
600 California sea lions and 20 harbor seals
could be affected by the activities at the dock.
The potential effects of the construction
activities included a temporary shift in the
animals’ hearing threshold during pile driving, behavior changes, and temporary cessation of normal activities, such as feeding.
Several mitigation measures were imposed
on the Coast Guard to reduce the potential
for harassment, including time restrictions
for pile driving. The NMFS concluded that
“while behavioral modifications, including
temporarily vacating the haulout, may be
made by these species to avoid the resultant
visual and acoustic disturbance, this action
is expected to have a negligible impact on
the animals.”
The IHA application process is an ideal
avenue to force the issue of AUV regulation.
Government action is too often reactionary,
with agencies waiting until the activities are
already firmly entrenched before taking steps
to regulate. The IHA process is an excellent
opportunity for NOAA and the industry to
investigate the potential impacts of AUV use
on marine mammals and marine habitats.
Researchers and private operators concerned
about the potential impacts of AUV use on
marine mammals should seriously consider
applying for an IHA prior to their next cruise.
If the agency discovers, after processing a few
IHAs for AUV operations, that the risk of
harassment is so minute that permitting is
not necessary, AUV operations can continue
unimpeded. Even if NOAA determines that
harassment is likely, the benefits of securing
approval should outweigh any costs associated with the additional paperwork.
Once in possession of an IHA, an individual is no longer “subject to the penalties
under the [MMPA] for taking by harassment
that occurs in compliance with such autho-
rization.” Besides immunizing an operator
from prosecution under the MMPA for harassment, an IHA could be used to alleviate
the concerns of insurers and institutions
worried about liability and user conflicts.
Through the application process, a researcher
or operator should discover the frequency
in which other activities are conducted in
the area. Any potential user conflicts would
then be avoidable, through either the voluntary actions of the operator or the mitigation requirements imposed by the agency.
Although a regulatory gap currently exists with regard to AUVs, options are available to obtain permission for AUV operations
or at least notify the appropriate federal agencies. By working within existing regulatory
programs, AUV operators can work with the
federal government to make the oceans a safer
place for both humans and animals. This proactive approach may enable the industry to
postpone and even prevent regulation in the
future, saving research institutions and operators valuable time and money. The wealth
of data that AUVs could collect is unfathomable. Hopefully, the use of these little robots
will continue to grow and enrich the scientific knowledge of the world.
References
Griffiths, G. 2003. Technology and Applications of Autonomous Underwater Vehicles.
London: Taylor & Francis. 342 pp.
National Research Council. 2003. Ocean
Noise and Marine Mammals. Washington,
D.C.: National Academies Press. 204 pp.
Spring 2004
Volume 38, Number 1
83
BOOK REVIEW
In Peril: A Daring Decision, a Captain’s Resolve,
and the Salvage that Made History
By Skip Strong and Twain Braden
The Lyons Press, 2003
252 pp. $22.95
Reviewed by Stephanie Showalter, Director
Sea Grant Law Center,
University of Mississippi
I
n November 1994, Tropical Storm Gordon stalled over the Florida Keys, wreaking
havoc on land and at sea. On the evening of
November 14, the tug J.A. Orgeron, adrift
near Bethel Shoal near Fort Pierce, Florida
after experiencing engine problems, signaled
the Coast Guard for assistance. When Skip
Strong, captain of the 688-foot oil-tanker
Cherry Valley, answered the Orgeron’s distress
call he had no way of knowing that he was
about to make maritime salvage history by
saving the $50 million external fuel tank of
the space shuttle Atlantis.
The story behind the rescue of the J.A.
Orgeron and the barge Poseidon, which carried NASA’s external fuel tank, and the subsequent salvage claim by the owner and crew
of the Cherry Valley springs to life in the capable hands of Skip Strong and Twain
Braden. Unfortunately, after catching the
reader’s attention quickly with a tense pretrial scene, the first fifty pages of In Peril bogs
down with an extraordinary amount of space
devoted to the construction of the external
fuel tank and the logistics of towing it from
Louisiana to Cape Canaveral, which did not
seem all that relevant to the rescue itself. In
Peril, however, regains its momentum in Part
II and quickly carries the reader along to its
historic conclusion.
Although the authors assume a high level
of familiarity with nautical terms and references, In Peril, with its simple style and attention to detail, places the reader right in
the middle of the action. The engineers on
the Cherry Valley operate at a frantic pace,
the third mate is stationed in the chartroom
84
ensuring that the Cherry Valley does not run
aground on Bethel Shoal, and the captains
of the Cherry Valley and the J.A. Orgeron attempt to attach lines without endangering
their vessels and men while struggling with
the darkness, wind, and waves. One memorable passage details the first attempt of the
Cherry Valley’s crew to attach lines to the
Orgeron using a line-throwing gun called the
Speedline 250. When the first shot sends
the rocket soaring into the clouds instead of
toward the tug, the second attempt is critical.
“I [Captain Strong] dash back
out to the wing to check our position. We are sliding past the tug but
still in range of the Speedline. Jim
is set up and ready to go with another Speedline after confirming
with the tug that the first one did
not reach them. He aims just over
the tug, spreads open his feet, bracing himself, and then pulls the trigger. Nothing happens. I remember
that the instructions say to hold onto
it for a minute after pulling the trigger to make sure it doesn’t fire late.
I can see Jim give it a fast count –
nothing – before tossing the whole
thing over the rail into the sea.”
The story of the rescue is exciting
enough, but the events that take place once
the vessels are safe and the attorneys get involved are fascinating. Keystone Shipping
Company sought salvage rights from the
owner of the J.A. Orgeron and NASA. De-
spite the fact that the crew of the Cherry Valley
saved NASA upwards of $50 million, the
federal government vigorously fought the
salvage award. In the end, the Fifth Circuit
Court of Appeals awarded Keystone $4.125
million—the largest maritime salvage award
in U.S. history. The crew received
$1,752,642, what remained after paying
interest, costs, and Keystone’s 63% share.
In Peril contains eight pages of photographs, illustrations, and maps, including
nautical charts identifying the position of the
Cherry Valley and the Orgeron during the rescue and tow. One page of diagrams detailing
the actual rescue is especially helpful for landlubbers unable to visualize the rescue maneuvers from words alone. Thoroughly enjoyable, In Peril is an excellent selection for
adrenalin junkies, history buffs, maritime
lawyers, and for anyone curious about what
really goes on during daring sea rescues.
BOOK REVIEW
Handbook of Acrylics for Submersibles,
Hyperbaric Chambers and Aquaria
By Jerry D. Stachiw
Best Publishing Co., 2003,
1066 pp. $195.00
Reviewed by Mark Olsson, President
DeepSea Power & Light
and Will Forman
Undersea Vehicle Consultant
T
his unique book is a “must have” for
anyone designing and building pressure resistant underwater structures or instruments,
as well as anyone designing with acrylic as a
structural material. While the focus of this
book is on acrylic structures under pressure,
this highly visual 1066 page book, with over
500 graphs, drawings, and photographs,
provides a wealth of information about the
design of pressure tolerant structures in general, from submarines to aquaria. Engineers,
designers, manufacturers, fabricators, operators, and inspectors of acrylic windows will
find this book an essential reference.
This is a deeply practical, nuts and bolts
volume filled with both detailed information and over forty years engineering wisdom and insight into the problem of building stuff that works. For example, with a
basis in theoretical principles, the author
researched, designed, and then reduced to
practice the acrylic hulls for the submersibles
NEMO and Johnson Sea Links 1&2, the
first acrylic hull submarines. Dr. Stachiw has
also made important contributions investigating the use of structural ceramics for high
pressure applications (see his website at
www.hydroports.com), and also remains
active in the field of pressure vessels for human occupancy (ASME PVHO).
The book was a team effort, with a wonderful historical introduction, “The quest for
panoramic vision underwater,” written by
Dr. Joan Stachiw. Dr. Jerry Stachiw has a
refreshingly unabashed, direct style and does
not hesitate to offer clearly stated design
guidelines and criteria. For example: “The
crack-free, cyclic fatigue life of acrylic windows and pressure hulls in manned diving
systems is considered to be adequate if it
exceeds 1000 pressure cycles of 4-hour duration at design pressure and temperature.”
The book is well organized with a detailed
table of contents and an extensive index, allowing the reader to easily drill down into
this large reference and find needed specific
information. The book also includes numerous references and bibliographic citations in
association with most sections.
In his references Stachiw generously includes the significant works of others that
have contributed to acrylic research and its
structural use. Examples include Auguste
Piccard’s work with conical viewports in
some of the first deep submersibles. He also
refers to the significant but little known
works of von Mises, whose graphs for cylindrical pressure hulls have been and are still
used but seldom credited, possibly because
they were developed prior to WWI for submarine hulls.
Among his accomplishments, Stachiw
completed the impossible dream of the early
pioneer in deep submergence, Auguste
Piccard. At the end of his brilliant career,
Piccard fantasized about a submersible with
a transparent hull made up of 12 spherical
segments to provide a panoramic view of
the undersea. This would permit better research than could be done with the limited
vistas from the small viewports that he had
developed. After artist Bruce Beasely dem-
onstrated to the acrylic manufacturers how
to cast very thick, large pieces of acrylic without flaws in 1969, Stachiw upgraded the 12
piece Nemo hulls into 2 hemispheres for the
two Johnson Sea Links, thus establishing the
standard of spherical pressure hull fabrication used currently by virtually all acrylic
deep submersibles. Until such time as there
is a major technical breakthrough in transparent structural materials, this handbook
will be the bible for the designing, fabricating and maintenance of acrylic pressure vessels and structures for human use.
The book focuses on several aspects of
acrylic window usage. The main topics covered in the book are the design, fabrication,
quality assurance, installation, service inspection, and maintenance of pressure resistant
windows in service. Major attention is devoted
to design procedure. Simple guidelines are presented that facilitate the conversion of previously published test data into maximum working pressures by application of conversion factors. For non-standard window configurations, a set of design stresses for service in different ambient temperatures is suggested.
An interesting feature of the section on
the design of different window configurations
is the figures depicting the mechanism and
propagation of fracture under over-pressurization or pressure cycling, and the locations
on the window where cracks generated by
different loading conditions originate. Distinction is made between critical and noncritical crack locations and magnitudes.
Spring 2004
Volume 38, Number 1
85
Although traditionally acrylic components of aquaria do not fall into the category
of pressure resistant windows, their design
is very challenging, particularly for submerged tunnels and domes. Aquaria have
come a long way from being simple flat windows set in concrete. Underwater observation chambers accessible by tunnels fabricated from plane and thermoformed acrylic
plates now complement the sheer expanse
of acrylic walls. Guidelines for the design,
fabrication, and installation of acrylic panels are presented. A specification for procurement of acrylic castings concludes this very
informative section.
An interesting feature of the section on
fabrication is the presentation of techniques
for the detection of residual stresses, the result
of incomplete annealing. The included tables
cover the whole gamut of thermal treatments
available for reduction of residual stresses introduced into the acrylic component by the
casting and machining procedures.
Of great help to the operator of any vessel equipped with acrylic windows is the
description of environmental conditions that
cause acrylic to deteriorate. Particularly useful is an extensive listing of chemical compounds whose contact with the acrylic may
be detrimental to the optical and structural
performance of the windows. The effect of
submersion in water and weathering on the
deterioration of acrylic is also noted and their
effect quantified.
One of the features of a pressure vessel
for human occupancy that generally is not
covered in other publications is the techniques for the illumination of the hyperbaric
chamber’s interior for the benefit of occupants undergoing treatment. Designs of
lights for illumination of the interior through
acrylic light pipes or windows are shown and
their salient features discussed.
No book on acrylic would be complete
without a description of bonding processes.
What makes the description of bonding procedures in this book unusual is that it addresses not only procedures for achieving
good bonds but also how to repair bad ones.
The Section on optical performance of
acrylic serves as a good introduction to optical effects one can expect from different
86
window configurations. Particular emphasis is given to the optical effects generated
by spherical sector windows.
The book concludes with a very informative description of the ANSI ASME/
PVHO-1 Safety Standard covering acrylic
components of pressure vessels. In the discussion one is introduced to the origins of
the Standard as well as its objectives and the
major technical areas it addresses. Since the
author was one of the co-authors of the Standard, his opinions expressed in this Section
provide a valuable insight into some of the
fundamental concepts, such as design life,
service life, minimum safety margin, and
others promulgated by the Standard.
It was hard in this short review to present
an account that does real justice to the contents and presentation of this large volume.
Between the covers of this book the author
attempted to cover all facets of acrylic material, procurement, design, fabrication, installation, testing, and in service inspection. In
this endeavor he was quite successful, presenting the reader with both the empirical
design criteria and the test data from which
they were derived.
The Marine Technology Society chose
the right publication to sponsor.
BOOK REVIEW
The Autobiography of a Yankee Mariner:
Christopher Prince and the American Revolution
Michael J. Crawford (Editor)
Brassey’s, Inc., 2002
223 pp. plus appendices $26.95
Reviewed by Jonathan Michael Prince
UNOLS Office, Moss Landing Marine
Laboratories
“
I went aloft, and I saw in a moment
something which I knew was the topgallant
sails of a vessel. We bore away and made all
the sail we could. The wind was light, it did
not exceed a six knot breeze. I kept aloft until
I saw the topsails of the ship, and she was
seen off deck. At twelve o’clock, meridian,
we was alongside of her. After we boarded
her we found by her bills a cargo of £50,000
sterling. She was called the Lovely Lass. In
one hour we put a Prize Master on board
her by the name of Thompson, and sent her
to New Bedford, as the safest port she could
go to.”
This sort of matter of fact, first person
recount of events during the American Revolutionary War by Captain Thomas Prince
tells the story of New England mariners
during this important chapter of our nation’s
history. This is not the story of a famous
national hero, but the story of one of the
many Americans who struggled to find their
place in the conflict, contributed to the cause
and survived to tell the story. That is what
makes this book somewhat unique. Captain
Prince’s “brief sketch of my life,” according to
editor Michael J. Crawford, is one of only a
dozen full autobiographical narratives by
men that served on government warships
or privateers during the war.
Many of us have listened to the stories
and adventures of relatives and friends and
have thought that they (or we) should write
down their experiences for future generations
to experience. Captain Prince did not have
children of his own, but his nieces and nephews must have listened to his stories and
asked questions which most likely motivated
him to write his memoirs. His writings were
passed down through his brother’s family for
a number of generations before being transcribed and eventually donated to the Naval
Historical Foundation and subsequently the
Library of Congress. The original manuscript
has been lost and the story itself might have
remained in the archives, except for the work
of naval historian, Michael J. Crawford,
Ph.D., the head of the Early History Branch
of the Naval Historical Center. Dr. Crawford
was drawn to Prince’s autobiography and decided it should be published not just because
it was a first hand account of some important military and naval actions of the War of
Independence, but also because it reveals the
impacts of the war on an ordinary American. Not only does he describe the events,
but also how he felt, why he made some difficult choices and how these events affected
the remainder of his life.
Crawford gives us the text of Prince’s
transcribed manuscript word for word, with
just a few modifications. The manuscript
consisted of around sixty thousand words,
largely unbroken by chapter divisions or
paragraph breaks. The Editor broke the work
into chapters, provided headings and introduced additional paragraphing along with
removing redundant and crossed out words,
modernizing some spelling and punctuation
and inserting explanations when needed to
make the autobiography more readable.
In addition he has made his own contribution to the story by providing factual introductions to the book and to each chap-
ter, which allows the reader to follow Prince’s
narrative in the historical context. Through
the liberal use of footnotes, Crawford provides corroborating evidence for some parts
of the story and corrections to others. As he
notes in the introduction, a story written
from memory will often include events that
are authentic, but perhaps not completely
accurate when it comes to remembering
when in the sequence of events it occurred
or with regards to details such as dates or
names. The Editor also provides us with
numerous maps and illustrations that enhance the authenticity of Prince’s story. These
are indexed just after the table of contents.
A final contribution by the Editor are appendices that include Captain Prince’s obituary, a narrative by Ethan Allen of his captivity on one of the vessels in which Captain
Prince sailed, other historical references, and
a useful glossary of sailing terms.
During the course of Prince’s story we
are presented with the pre-Revolutionary
War life of a New England youth with a
strong desire to follow the sea like many of
his relatives. Succeeding in that vocation, he
moves from being a Grand Banks fisherman
to being an accomplished merchant seaman
in the employ of his uncle by the time the
war begins. He soon finds himself caught
up in the war as a crewmember of the sloop
Polly, which was commissioned to take the
members of loyalist families from Boston to
Halifax. In Halifax the ship and crew become prisoners of the British and are made
part of a fleet of guard ships on the St.
Lawrence River. From this vantage point he
Spring 2004
Volume 38, Number 1
87
becomes personally involved in the war and
witnesses the American invasion of Canada,
the imprisonment of Ethan Allen, and after
his release from captivity, the retreat of the
American Army down Lake Champlain and
the Hudson River.
The American Army had been decimated by smallpox and, during the retreat,
Captain Prince contracted the disease, but
was fortunate enough to be in New York
City when the effects hit him. He received
better than normal care, survived and recovered in time to participate in attempts to
prepare for the defense of New York. He
describes the efforts to obstruct passage up
the Hudson River by sinking several vessels.
He then goes on to serve in vessels of the
Connecticut Navy and finally in several privateers sailing from New London. There are
several exciting descriptions of action at sea
and narrow escapes from danger.
Captain Prince concludes his story with
descriptions of his postwar life as a merchant
vessel master and his growing religious convictions—a typical Yankee mariner.
I recommend this book for anyone who
has an interest in maritime history, history
of the American Revolution or historical biographies in general. It is a good and sometimes exciting story that puts this important
part of our history into the perspective of an
ordinary American involved in extraordinary
times.
For me, this story brings to life the experiences of my ancestors. Captain Prince and
I share the same emigrant ancestor; I am a
direct descendant of his great, great grandfather. I have seen the records of many of these
ancestors that show date of birth, date of
death and perhaps a short sentence about
serving on board one or more vessels, but no
details, no insight into what their lives were
like. This volume provides that richness of
detail for me and for many others whose
ancestors made their living as seafarers in the
early years of our country. The Autobiography of a Yankee Mariner brings history to life
and at the same time, through the efforts of
the editor, keeps the history accurate.
88
Marine Technology Society Publications Listing
The following Marine Technology Society
publications are available for purchase.
Prices for 2004 are listed below. Members
are granted a discount of 10% off the purchase order.
You can purchase our publications by:
1. Calling MTS at 410-884-5330 with
your publication(s) order and credit
card number. Please use reference
number SM38:1 when placing
your order.
2. Log on to our website at
www.mtsociety.org, click on icon
publications, sidebar publications list.
3. Circle the items you would like to
order, fill out the form below and
mail to MTS.
JOURNALS
Human-generated Ocean Sound and
the Effects on Marine Life ........................... $20
Ocean Observing Systems ................................ $20
Science, Technology and Management in the
National Marine Sanctuary Program ............ $20
Ocean Energy—an Overview of the State of
the Art ........................................................ $20
Marine Archaeology and Technology—
A New Direction in Deep Sea Exploration ... $20
Technology in Marine Biology ......................... $20
Ocean Mapping—A Focus of Shallow
Water Environment .................................... $20
Oceanographic Research Vessels ....................... $20
Technology as a Driving Force in the Changing Roles
of Aquariums in the New Millennium ......... $20
Submarine Telecoms Cable Installation
Technologies ............................................... $20
Deep Ocean Frontiers ...................................... $20
A Formula for Bycatch Reduction .................... $16
Marine Science and Technology in the Asia Region,
Part 2 ......................................................... $16
Marine Science and Technology in the Asia Region,
Part 1 ......................................................... $16
Major U.S Oceanographic Research Programs:
Impacts, Legacies and the Future ................. $16
Marine Animal Telemetry Tags: What We Learn and
How We Learn It ........................................ $16
Scientific Sampling Systems for Underwater
Vehicles ...................................................... $16
U.S. Naval Operational Oceanography ............ $16
Innovation and Partnerships for Marine Science and
Technology in the 21st Century .................. $16
Marine Science and Technology in Russia ........ $16
Public-Private Partnerships For Marine Science &
Technology (1995) ...................................... $16
Oceanographic Ships (1994–95) ...................... $16
Military Assets for Environmental Research
(1993–94) .................................................. $16
Oceanic and Atmospheric Nowcasting and
Forecasting (1992) ...................................... $12
Education and Training in Ocean Engineering
(1992) ........................................................ $12
Global Change, Part II (1991–92) ................... $10
Global Change, Part 1 (1991) .......................... $10
M A R I N E E D U C AT I O N
Education and Training Programs In Oceanography
and Related Fields (1995) ....................... $6
Operational Effectiveness of Unmanned Underwater
Systems .................................................. $99
State of Technology Report—Ocean and Coastal
Engineering (2001) ................................ $7 dom
............................................................. $9For.
State of Technology Report—Marine Policy and
Education (2002) ................................... $7 dom
............................................................. $9 For.
PROCEEDINGS
Oceans 2003 MTS/IEEE (CDROM) .............. $50
Artificial Reef Conference ................................ $25
Oceans 2000 MTS/IEEE CD-ROM ............... $50
Oceans 1999 MTS/IEEE ‘99 Paper Copy ........ $80
Oceans 1999 MTS/IEEE CDROM ................. $40
Ocean Community Conference ‘98 ............... $100
Underwater Intervention 2002 ........................ $50
Underwater Intervention 2000 ........................ $50
Underwater Intervention ‘99 ........................... $50
500 Years of Ocean Exploration
(Oceans ‘97) ............................................. $130
The Coastal Ocean—Prospects for
The 21st Century (Oceans ‘96) ................. $145
Challenges of Our Changing Global
Environment (Oceans ‘95) ........................ $145
MTS ‘92 ....................................................... $140
ROV ‘92 ....................................................... $105
MTS ‘91 ....................................................... $130
ROV ‘91 ......................................................... $95
1991–1992 Review of Developments in Marine
Living Resources, Engineering and
Technology ................................................. $15
ROV ‘90 ......................................................... $90
The Global Ocean (Oceans ‘89) ...................... $75
ROV ‘89 ......................................................... $65
Partnership of Marine Interest (Oceans ‘88) ..... $65
The Oceans–An International Workplace
(Oceans ‘87) ............................................... $65
Organotin Symposium (Oceans ‘87, Vol. 4) ..... $10
ROV ‘87 ......................................................... $40
Technology Update–An International
Perspective (ROV ‘85) ................................. $35
Ocean Engineering and the Environment
(Oceans ‘85) ............................................... $45
Ocean Data: Sensor-to-User (1985) ................. $33
Arctic Engineering for the 21st Century (1984) ..... $20
Marine Salvage: Proceedings of the 3rd
International Symposium (1984) ................. $21
M T S P U B L I C AT I O N S O R D E R F O R M — R E F : S M 3 8 : 1
S E N D A P P L I C AT I O N T O :
Please print
_____________________________________________________________________________________________
Mr./Ms./Dr.
First name
Last name
_____________________________________________________________________________________________
Address
_____________________________________________________________________________________________
City
State
Zip
Country
_____________________________________________________________________________________________
Telephone
FAX
Marine Technology Society
Please call MTS at 410-884-5330 for a
MTS Membership Application or see
following pages for an Application.
_____________________________________________________________________________________________
Total E-mail
P A Y M E N T : Make checks payable to Marine Technology Society (U.S. funds only)
Credit card: ❏ Amex ❏ Mastercard ❏ Visa ❏ Diners Club
FAX (410) 884-9060
_____________________________________________________________________________________________
EMail: [email protected]
Card Number
Exp. Date
Signature
Spring 2004
Volume 38, Number 1
89
UPCOMING MTS JOURNAL ISSUES CALL FOR PAPERS
Summer 2004
Innovations in Ocean Research Infrastructure to
Advance High Priority Science
Alexandra Isern—National Science Foundation
The Marine
Technology Society
is accepting papers,
Fall 2004
Environmental Threats from Submerged Vessels
Lisa Symons—NOAA
technical notes, and
commentaries for
Winter 2004/2005
U.S. Commission on Ocean Policy: Implications
and Opportunities
Terry Schaff—U.S. Commission on Ocean Policy,
and Dan Walker—National Research Council and
Ocean Studies Board
Spring 2005
Acoustic Tracking of Marine Fishes: Implications for
the Design of Marine Protected Areas
James Lindholm—Pfleger Institute of Environmental
Research
Summer 2005
General Issue
MTS Journal for the
following issue:
Summer 2005
General Issue
If you have material that
you would like considered,
please forward it to:
MTS Journal Managing Editor
ADVERTISING RATES
SPACE
INSERTION RATE (US$)
1x
2x
4x
Width
Depth
Full Page
525
475
435
7"
10"
SPACE DIMENSIONS
Vertical
350
320
300
3½”
10"
Half Page
Horizontal
350
320
300
7"
4½”
One-Quarter
205
185
160
3½”
4½”
One-Eighth
135
115
105
3½”
2"
90
publication in the
Or contact:
[email protected]
to request information or
to submit your paper online.
You can also visit
our homepage at
www.mtsociety.org
Marine Technology Society Membership Application
and MTS Journal Subscription Request
■ Yes, I want to become a member of the Marine Technology Society.
■ I wish to purchase a subscription to the print version of the MTS Journal. (MTS Journal online is a free
member benefit not available to non-members.)
■ I’m not ready to join at the present time, but I would like to receive news and information from MTS.
Please keep my name on your mailing list. (Please fill in member information section and return to MTS.)
MEMBER INFORMATION (Please fill in the address where MTS publications and correspondence should be sent.)
Name: _______________________________________________Company (if applicable): _____________________________
Title:________________________________________________Email:____________________________________________
Phone:______________________________________________ FAX:_____________________________________________
Address:________________________________________________________________________________________________
_______________________________________________________________________________________________________
MEMBERSHIP CATEGORIES
■ MEMBER: $75 Any person with active professional interest in the marine field or a closely related field. Includes
the online version of the Journal.
■ ASSOCIATE: $75 Any person with an interest in the
marine field and supporting MTS objectives. Includes the
online version of the Journal.
■ STUDENT: $25 Must be a full-time undergraduate or graduate student. Includes the online version of the Journal.
■ PATRON: Any person supporting MTS objectives by contributing $100 or more annually. (Contribution above $75
dues is tax deductible.) Includes the online version of the
Journal.
■ EMERITUS: $40 Any person who is retired from active
professional interest in the marine field. Includes the
online version of the Journal.
■ LIFE: $1000 One-time payment. Includes both the online
version of the Journal and the paper copy of the Journal.
■ CORPORATE Fortune 500: $2000 Any corporation
endorsing the objectives, policies and activities of MTS.
May appoint 16 representatives to the Society, who have
the same rights and privileges as Members. Includes the
online version of the Journal for each representative. The
main contact receives a subscription to the paper copy of
the Journal.
■ CORPORATE Non-Fortune 500: $1000 Any corporation
endorsing the objectives, policies and activities of MTS.
May appoint 11 representatives to the Society, who
have the same rights and privileges as Members.
Includes the online version of the Journal for each representative. The main contact receives a subscription to
the paper copy of the Journal.
■ BUSINESS: $550 Any business firm whose gross annual
income is less than $4 million endorsing the objectives,
policies and activities of MTS. May appoint six representatives to the Society, who have the same rights and
privileges as Members. Includes the online version of
the Journal for each representative. The main contact
receives a subscription to the paper copy of the
Journal.
■ INSTITUTION: $550 Any library, government unit or
other qualified nonprofit organization endorsing the
objectives, policies and activities of MTS. May appoint
six representatives to the Society, who have the same
rights and privileges as Members. Includes the online
version of the Journal for each representative. The main
contact receives a subscription to the paper copy of
the Journal.
MTS JOURNAL PRINT VERSION SUBSCRIPTION
Domestic Subscription
■ Member Rate: Journal Subscription........................$ 25
■ Non-Member Rate: Journal Subscription................. $120
Foreign Subscription
■ Member Rate: Journal Subscription.......................$ 50
■ Non-Member Rate: Journal Subscription.................$135
Spring 2004
Please
side91
Volume
38,complete
Number other
1
Marine Technology Society Membership Application (continued)
EDUCATIONAL INFORMATION
Please fill out the following information about yourself:
Check your highest level of education:
n
■ High School Diploma
■
n Associate (2 yr.) Degree
Check all that apply:
n
■ B.S.
■
n B.E.
n
■ YES
Do you have a P.E. license?
■
n Four Year Degree
■
n B.A.
■
n M.S.
■
n Graduate Degree
■
n M.A.
■
n M.E.
■
n Ph.D.
■
n Doctorate
■
n Sc.D.
■
n NO
BUSINESS/PROFESSIONAL INFORMATION
Name of current employer: _________________________________________________________________________________
Your employer’s primary line of business at your location: ________________________________________________________
If you don’t work for an employer, please identify your business: ___________________________________________________
If military, rank: _________________________________________________________________________________________
Your principal job
function/responsibilities:
__Engineering Management
__Science Management
__Sales
__Marketing
__Administration
__Policy Making, Regulatory
__Public Affairs
__Engineering Design
__Mechanical Engineering
__Software Engineering
__Education/Teaching
__Legal
__Consulting
__Retired
__Other (please specify)
_______________________
Your job title:
Check areas of interest:
__President/CEO/COO
__Owner/Partner
__VP, Senior Manager
__Project Manager,
Engineering
__Project Manager, Other
__Corporate VP, Engineering
__Engineering Director
__Chief/Senior Engineer
__Chief/Senior Scientist
__Project Manager
__Engineer
__Operations VP
__Scientist
_______________________
__Autonomous Underwater
Vehicles
__Dynamic Positioning
__Manned Underwater
Vehicles
__Ocean Energy
__Oceanographic
Instrumentation
__Remote Sensing
__Remotely Operated
Vehicles
__Underwater Imaging
__Marine Geodesy
__Marine Living Resources
__Mineral Resources
__Ocean Pollution
__Oceanographic Ships
__Physical Oceanography &
Meteorology
__Seafloor Engineering
__Buoy Technology
__Cables & Connectors
__Marine Archaeology
__Diving
__Marine Materials
__Moorings
__Offshore Structures
__Ropes & Tension Members
__Coastal Zone Management
__Marine Education
__Marine Law & Policy
__Marine Recreation
__Merchant Marine
__Marine Security
__Ocean Economic Potential
_______________________
Optional Information:
n
■ Male
■
n Female
What is your age?
■
n Under 30
■
n 30-40
■
n 41-50
■
n 51-60
■
n Over 60
MEMBERSHIP AND JOURNAL PAYMENT
Payment Method:
■
n Check Enclosed
■
n Master Card
Make checks payable to the Marine Technology Society (U.S. funds only)
■
n Visa
■
n Diners Club
■
n Am Ex
Card #: __________________________________________________________ Expiration Date: ________________________
Signature: ________________________________________________________ Date: _________________________________
TOTAL PAYMENT:
Membership:
$_____________
Journal:
$_____________
TOTAL:
$_____________
Four easy ways to join!
Mail:
Fax:
Online:
Phone:
Send application with check or credit card info to:
Marine Technology Society / 5565 Sterrett Place, Suite 108 / Columbia, MD 21044
Fax application to: 410-884-9060 (credit card payments only)
Apply online at www.mtsociety.org
Contact us at: 410-884-5330
Notes
Notes
Notes
Notes
Marine Technology Society Member Organizations
C O R P O R AT E M E M B E R S
Alstom Power Conversion, Inc.
Houston, Texas
C-Mar America, Inc.
Houston, Texas
Compass Publications
Arlington, Virginia
Cortland Cable Company
Cortland, New York
Dynacon, Inc.
Bryan, Texas
ExxonMobil Upstream Research Company
Houston, Texas
FMC SOFEC Floating Inc.
Houston, Texas
Fugro Chance, Inc.
Lafayette, Louisiana
Fugro Geoservices, Inc.
Houston, Texas
Fugro-McClelland Marine Geosciences
Houston, Texas
Fugro Pelagos, Inc.
San Diego, CA
General Dynamics/ATS
Herndon, Virginia
Geospace Offshore Cables
Houston, Texas
Innerspace Corporation
Covina, California
J. P. Kenny, Inc.
Houston, Texas
JDR Cable Systems, Inc.
Houston, Texas
Klein Associations, Inc.
Salem, New Hampshire
Kongsberg Simrad, Inc.
Houston, Texas
Kvaerner Oilfield Products
Houston, Texas
Maritime Communication Services
Melbourne, Florida
Mitsui Engineering and Shipbuilding Co. Ltd.
Tokyo, Japan
Mohr Engineering & Testing
Houston, Texas
Nautronix, Inc.
Houston, Texas
Navatek, Ltd.
Honolulu, Hawaii
Neptune Sciences, Inc.
Slidell, Louisiana
Ocean Design, Inc.
Ormond Beach, Florida
Oceaneering International, Inc.
Houston, Texas
Oceaneering Technologies
Upper Marlboro, Maryland
Oil States Industries, Inc.
Arlington, Texas
Orincon Hawaii, Inc.
Kailua, Hawaii
Pegasus International, Inc.
Houston, Texas
Perry Slingsby Systems, Inc.
Jupiter, Florida
Phoenix International, Inc.
Landover, Maryland
Planning Systems, Inc.
Reston, Virginia
RD Instruments
San Diego, California
Reson, Inc.
Goleta, California
SBM-IMODCO, INC.
Houston, Texas
Schilling Robotics, LLC
Davis, California
Science Applications International Corp.
SeaCon Brantner and Associates, Inc.
El Cajon, California
Sippican, Inc.
Marion, Massachusetts
Sonsub, Inc.
Houston, Texas
SonTek/YSI, Inc.
South Bay Cable Corp.
Idyllwild, California
SubConn, Inc.
Burwell, Nebraska
Subsea Seven
Houston, Texas
Technip
Houston, Texas
Thales Geosolutions, Inc.
Houston, Texas
The Tsurumi-Seiki Co., Ltd.
Yokohama, Japan
Tyco Telecommunications (US) Inc.
Morristown, New Jersey
BUSINESS MEMBERS
4 Controlled Solutions
Houston, Texas
Aanderaa Instruments, Inc.
S. Attleboro, Massachusetts
Applied Subsea Technologies, Inc.
Providence, Rhode Island
Ashtead Technology, Inc.
Houston, Texas
Bennex Subsea, Houston
Houston, Texas
Bluewater Offshore Production Systems USA, Inc.
Houston, Texas
C.A. Richards and Associates
Houston, Texas
C & C Technologies, Inc.
Lafayette, Louisiana
Deep Marine Technology, Inc.
Houston, Texas
Deepsea Power and Light
DTC International, Inc.
Houston, Texas
Falmat, Inc.
San Marcos, California
Gilman Corporation
Gilman, Connecticut
Impulse Enterprise
InterOcean Systems, Inc.
Makai Ocean Engineering, Inc.
Kailua, Hawaii
Marine Desalination Systems, L.L.C.
Washington, DC
The Marine Technology Society gratefully acknowledges the critical support of the Corporate, Business, and Institutional members listed.
Member organizations have aided the Society substantially in attaining its objectives since its inception in 1963.
Matthews-Daniel Company
Houston, Texas
Natural Resources Canada
Dartmouth, Nova Scotia, Canada
Oceanic Imaging Consultants, Inc.
Honolulu, Hawaii
OceanWorks International
Houston, Texas
Prizm Advanced Communication Electronics, Inc.
Baltimore, Maryland
Pro Staff Engineering
Houston, Texas
Reel In, Inc.
College Station Texas
Remote Ocean Systems, Inc.
Saipem, Inc.
Houston, Texas
Sonardyne, Inc.
Houston, Texas
Sound Ocean Systems, Inc.
Redmond, Washington
Tension Member Technology
Huntington Beach, California
TSC Holdings Group, Inc.
Palm City, Florida
Videoray, LLC
Exton, Pennsylvania
Weatherguy.com, LP
Kailua, Hawaii
INSTITUTIONAL MEMBERS
British Embassy
Washington, DC
CEROS
Kailua-Kona, Hawaii
Consortium for Oceanographic Research and Education
Washington, DC
Harbor Branch Oceanographic Institution, Inc.
Fort Pierce, Florida
MBARI
Moss Landing, California
Mitretek Systems
Falls Church, Virginia
National Ocean Industries Association
Washington, D.C.
NOAA/PMEL
Seattle, Washington
Naval Facilities Engineering Service Center
Port Hueneme, California
Naval Meteorology and Oceanography Command
Stennis Space Center, Mississippi
Scripps Institution of Oceanography
La Jolla, California
Service Argos, Inc.
Largo, Maryland
SW Research Institute
San Antonio, Texas
U.S. Coast Guard
Washington, DC
U.S. Naval Academy
Annapolis, Maryland
University of British Columbia Library
BC, Canada
University of California Library
Berkeley, California
Postage for periodicals
is paid at Columbia, MD,
and additional mailing offices.

Vol. 38, No. 1 - Marine Technology Society

Transcription

Similar documents

Tree Top Ocean Views

Ocean Essence Pacific Sea Salt Scrub

Buku program SPP81 - Faculty of Mechanical

R. Lee Ermey served 11 years active duty with the US Marine Corps

Marine Core Service Overview

www.hotondo.com.au Major sponsor Hotondo Homes Batemans

MFCU Scholarship Application - Marine Federal Credit Union

ECOMARE - Oceano XXI

cancellation of business name

Print - Travel Advantage Network