Marine Technology Society Journal

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
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 UniversityCollege Station Counselor: Robert Randall Texas A&M UniversityGalveston 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
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
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
Crosstalk
MTS Journal Readers Comments
General Issue
61
Development of a Real-time Regional Ocean Forecast System with Application to a Domain off the U.S. East Coast
Laurence C. Breaker, Desiraju B.Rao, John G.W. Kelley, Ilya Rivin, Bhavani Balasubramaniyan
5
A Method for Rapid Hull Form Development and Resistance Estimation of Catamarans
V. Anantha Subramanian, Patrick Joy
80
The Legal Status of Autonomous Underwater Vehicles
Commentary by Stephanie Showalter
12
Satellite Data Assimilation for Improvement of Naval Undersea Capability
BACK COVER: Examples of ocean forecast products from NOAAs East Coast Regional Ocean Forecast System (see pages 61-79), image produced by the Environmental Modeling Center at NOAAs National Centers for Environmental Prediction.
84
Book Reviews
Peter C. Chu, Michael D. Perry, Eric L. Gottshall, David S. Cwalina
24
Use of Expert Systems for Optimum Maintenance of Marine Power Plants
K.D.H. Bob-Manuel
MTS Journal design and layout: Michele A. Danoff, Graphics By Design
30
The Science and Technology of Nonexplosive Severance Techniques
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. 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.
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
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. Please send all correspondence to: The Marine Technology Society 5565 Sterrett Place, Suite 108 Columbia, Maryland 21044 (410) 884-5330 Tel. (410) 884-9060 FAX E-mail: mtspubs@erols.com mtsdir@erols.com mtsmbrship@erols.com Online: www.mtsociety.org 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.
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
Phil Nuytten
Nuytco Research, LTD.
Bruce H. Robison
Monterey Bay Aquarium Research Institute
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 Geobases 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.
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
Marine Technology Society Journal
Readers 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, Ren 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 Authors 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 youd like to submit to Crosstalk, please limit it to 250 words or less and send it to: mtspubs@erols.com. 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
Readers 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
PAPER
A Method for Rapid Hull Form Development and Resistance Estimation of Catamarans
AUTHORS
V. Anantha Subramanian Patrick Joy Department of Ocean Engineering, Indian Institute of Technology
ABSTRACT
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 Michells 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
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 owners 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 . 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 + k) 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 Michells 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
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, Q (u, w) =
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
n +1 m +1 i =1 j = 1
Bi,j Ni,k (u) Mj,l (w)
Bi,js 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
Estimation of Total Calm Water Resistance

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)
The coefficient of wave resistance for catamarans may be expressed as, CW = RW/(AWU2) (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 + k) CF + CW (3)
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,
/2
The viscous interference factor () depends upon the speed of the vessel, slenderness ratio, and the hull spacing ratio. For a monohull, =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 + k) 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
RW = (g4/U6)
(P2 + Q2) sec5 d

0
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.
P and Q are the Michell wave functions defined by,

L/2 0
FIGURE 2
Isometric and Profile view of a generated catamaran demihull form (L = 30 m)
P + iQ = f
L/2T
b(x,z) e(iwox + koz) dz dx
b(x,z) is the local beam of the demihull. 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) sec2 uo = ko sin wo = ko cos
Spring 2004
Volume 38, Number 1
FIGURE 3
Body plan views of the generated demihull forms (L = 30 m)
TABLE 1
Hull form parameters (L = 30 m, = 90 m3) Parameter b (m) D (m) T (m) bWL (m) CB CP AW (m ) L/b b/T L/
1/3 2
Form 1 3.20 3.00 1.78 3.00 0.57 0.60 130.40 9.37 1.79 6.69
Form 2 3.60 3.00 1.86 3.28 0.49 0.52 131.10 8.33 1.93 6.69
Form 3 3.90 3.00 1.70 3.56 0.49 0.67 125.20 7.69 2.29 6.69
FIGURE 4
TABLE 2
1/3 2
Form 4 3.20 3.00 1.70 2.70 0.48 0.67 155.50 12.50 1.88 8.92
Form 5 3.40 3.20 1.65 2.92 0.47 0.65 152.40 11.76 2.06 8.92
Form 6 3.60 3.20 1.60 3.12 0.46 0.62 151.10 11.11 2.25 8.92
FIGURE 5
TABLE 3
1/3 2
Form 7 3.20 3.00 1.50 2.77 0.45 0.94 173.00 15.62 2.13 11.15
Form 8 3.60 3.00 1.40 3.06 0.43 0.92 169.20 13.89 2.57 11.15
Form 9 4.00 3.00 1.25 3.34 0.43 0.90 166.53 12.50 3.20 11.15
FIGURE 6
Coefficient of total resistance (s/L = 0.15)
Spring 2004
Volume 38, Number 1
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 Suchitran. 1999. .R. 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
Peter C. Chu Michael D. Perry Naval Ocean Analysis and Prediction Laboratory, Department of Oceanography, Naval Postgraduate School Monterey, CA Eric L. Gottshall Space and Naval Warfare System Command San Diego, CA David S. Cwalina Naval Undersea Warfare Center Newport, RI
ABSTRACT
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 Navys 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.
INTRODUCTION
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. Navys 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.
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.
12
2. Navys 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 1/2 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 1/2 in the open ocean increasing to 1/4 in coastal seas and increasing again to 1/8 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 POMs 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),
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
+
N j ij j=1
i = Gi ,
(1)
where i are the weights, is the signal to noise ratio, ij is the autocorrelation between locations i and j, and 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 days 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
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.
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.Navys 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).
14
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) input , or between the two sets of output weapon preset data (running using GDEM and MODAS) output ,
(r,t) = (r,t) 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 M and G are the variables (either input or output) using GDEM and MODAS, respectively. We may take the probability histograms of M and G to show the difference between the statistical characteristics.
4.2. Root Mean Square Difference

GDEM and MODAS have different grid spacing: 1/2 x 1/2 in GDEM and 1/12 x 1/12 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 files name and the latitude and longitude of the profile. FIGURE 4
GDEM and MODAS data points
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.
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.
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
16
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-
FIGURE 7
Comparison between GDEM and MODAS temperature histograms at (a) the surface, (b) 100 m depth, and (c) 2000 m depth.
18
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 users standpoint, this means that both sets of data predict about the same 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.
20
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 convey 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 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.
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
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 perFIGURE 13
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
RMSD for shallow depth band coverage: left panel for ASUW scenario and right panel for ASW scenario.
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.
22
FIGURE 15
RMSD for deep depth band coverage: left panel for ASUW scenario and right panel for ASW scenario.
Acknowledgments
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
K.D.H. Bob-Manuel Rivers State University of Science and Technology Port Harcourt, Nigeria
ABSTRACT
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 authors experience in the management of ferries, there is an optimum amount of maintenance effort for any given condition.
INTRODUCTION
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 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. 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.
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.
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
Breakdown/Repair-only Maintenance Policy

A breakdown/replacement policy considers a component that has suffered a major
24
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:
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
(1) where
> 0 determines the shape of the

distribution.
> 0 is the characteristic life of the

component.
t0
determines the origin of the distribution time factor.
(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
25
TABLE 1
Analysis of lubricating engine oil from 3 (three) tugboats using Spectrometric Oil Analysis Procedure (SOAP) (Chugbo, 1999) PARAMETERS Density @15 C(kg/l) Water Content (% vol.) Kinematic Viscosity @ 40C (cst) Kinematic Viscosity @ 100C (cst) Total acid No. (Mg KOH/gr) Sulphated Ash % Wt * Values are off specification 2.0 16.0 Too dark to observe colour change 1.73 1.09 0.46 2.56 0.43 14.4 33.49* 13.21* 8.95* Nil 182.0 0.40 150.23* 0.15 125.23* 68.45* LUB-OIL SPEC. 0.916 TUGBOAT A 0.9253* TUGBOAT B 0.9026* TUGBOAT C 0.8887*
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 engineercomputer interaction i.e. the diagnosis of the failed component
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 FIGURE 2
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.
Constituent and total maintenance cost for breakdown and planned maintenance policies (Shields et al., 1995)
26
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.
Spare Part and Store Management

Expenditure on spare parts for marine power plants accounts for a substantial amount of the vessels 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-
FIGURE 6
Ultrasonic signature before and after cylinder liner replacement for a medium-speed diesel engine (Haller and Kelleher, 1999)
28
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.
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 maintenancea qualitative approach. I.Mar.E. London: Marine Media Management Ltd.
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.
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
Mark J. Kaiser Allan G. Pulsipher Center for Energy Studies, Louisiana State University Robert C. Byrd Twachtman Synder & Byrd, Inc.
ABSTRACT
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.
INTRODUCTION
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
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.
30
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
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
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 cuttera small explosive blast that utilizes less than 5 pounds of explosiveand 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 cutstorch cutting and arc gougingare 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
32
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 Councils 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
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).
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
FIGURE 2
Schematic of a Mechanical Cutting System (Courtesy of Hydrodynamic Cutting Systems)
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
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

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 1990s 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.
FIGURE 7
Diamond Wire Cutting System Deployed from a ROV for use in Pipeline or Small Member Cutting (Courtesy of CUT USA, Inc.)
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.
4.2. 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 tools 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.
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.
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.
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
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
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 phasesexploration, development, production, and decommissioningand 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-
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
38
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.
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.
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 todays 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. 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
Dan Ohno 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
ABSTRACT
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.
INTRODUCTION
he mission of deep submergence research vehicles built to datesuch as SHINKAI 2000, SHINKAI 6500, Alvin, Nautile, and othershas 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-
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. Mans 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.
40
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 mankinds well-being today and to ensure prosperity of the future as international collaboration advances ocean exploitation technologies.
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
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 1020 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.
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 JAMSTECs
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
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. 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.
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)
Integrated deep sea life environmental studies
Development studies of deep sea survey. direct-connecting social and economic activities
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.
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
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
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
42
TABLE 2
Results of conceptual designs of submergence vehicles Principal Items of Submersibles Max. operating depth (1) Dimensions (LBH) (2) Weight in air (3) Cruising speed (4) Navigation range (surface) (5) Bottom operating time (6) Crew (7) Pressure hull (8) Power source Max. operating depth (1) Dimensions (LBH) (2) Weight in air (3) Cruising speed (4) Navigation range (surface) (5) Bottom operating time (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 (3h21h) (2) Cross over ocean ridges (3) Reserved inner utility space by 2 spheres (4) Capable of AUV / ROV operation (Launch, recovery & supervising) (5) Large capacity of payload (6) Wide range of vision for TV camera & view ports (1) Extended operating time on the 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 (Launch, recovery & supervising) (5) Large capacity of payload (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 (Launch, recovery & supervising) (6) Very large capacity of payload (1) Capable of ocean-going /no need of support ship (2) Long term & large area (3) Extended operating time on the sea floor by using FC-AIP system (4) Comfortable large space habitable for a long term operation (5) Capable of AUV/ROV operation (Launch, recovery & supervising) (6) Very wide range of vision with large acrylic spherical view port 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)
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 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 Li-ion secondary batteries
(2)
(3)
Max. operating depth (1) Dimensions (LBH) (2) Weight in air (3) Cruising speed (4) Navigation range (surface) (5) Bottom operating time (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 Li-ion secondary batteries
(1) ~(5); Do. (6) Research on the dangerous points /narrow courses by AUV/ROV
(4)
Max. operating depth (1) Dimensions (LBH) (2) Weight in air (3) Cruising speed (4) Navigation range (surface) (5) Bottom operating time (6) Crew (7) Pressure hull (8) Power source Max. operating depth (1) Dimensions (LBH) (2) Weight in air (3) Cruising speed (4) Navigation range (surface) (5) Bottom operating time (6) Crew (7) Pressure hull (8) Power source
(5)
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
Spring 2004
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
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.
48
TABLE 3
Design conditions of AIP study Items Deep submersible research vehicle 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) 0.4 kg/kWh (target value) 0.45 kg/kWh (target value) 0.55 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
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 Outer pressure vessel Inner pressure vessel Outer pressure vessel Inner pressure vessel
Buoyancy adjustment tank
Material
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
It has been fifteen years since Shinkai 6500 was launched in 1989, providing the worlds 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.
References
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 Directory1990-91. 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 EarthOcean Sciences New Direction. The 30th Anniversary Commemorative Issue, JAMSTEC. 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.
Morr, B. 2001. Autonomous Submarines Broadening Industrys 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.
Spring 2004
Volume 38, Number 1
51
PAPER
Temperature and Salinity Variability in the Mississippi Bight

AUTHORS
Sergey Vinogradov Nadya Vinogradova Vladimir Kamenkovich Dmitri Nechaev Department of Marine Science, University of Southern Mississippi, Stennis Space Center
ABSTRACT
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
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
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 setsprimary and secondaryso 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
52
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
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-
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
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 February 1999 May 1999 November 1999 January/February 2000 August/September 2000 Total: Total number Duration of Number of time- Number of Number of of stations survey (days) series stations spatial groups transect groups 245 219 312 385 178 1339 13 10 14 15 15 67 7 2 6 4 2 21 4 6 6 5 3 24 5 5 4 16 7 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
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 6-7 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 100500 m layer, and two curves bulged down from the right side to the left. All T-S groups tend to 8 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 4-5 C. The variability analysis shows that temperature variations at the surface layer 0.58 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 10 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
54
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 50480 m portions of the Gulf transects where the temperature monotonically decreases with depth from 30C to 7C 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 FIGURE 2
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 C 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 2.56.0 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 C (January-February 2000 survey, spatial scale of about 23.8 km, 1.0 m depth).
(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. (2a)
(2b)
(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 5090 m depths in the vicinity of the shelf break reveal significant temperature drops in the near-bottom layer (Fig.3c). The upper layer (065 m) at station A with a total depth of 81 m is strongly homogeneous (25.5C, 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,
(3b)
(3c)
56
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 (510 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 23 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
(4c)
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.
58
FIGURE 4
Time-series stations located near passes and straits, continued. (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
Time-series stations located near passes and straits, continued. (4f)
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. 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.
(4g)
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. 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
OMB Contribution No. 206
Spring 2004
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 todays 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
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.
62
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 1990s was primarily responsible for the closure of over 50% of Americas 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
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
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 atmosphereocean 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 1980s, 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-
64
TABLE 1
User Requirements for Ocean Forecasts Activity Temperature Search & Rescue Oil Spill Models Estuarine Forecast Models Ecosystem Models Mariculture Marine Weather Forecasting Commercial Fishing Commercial Shipping Recreational Boating Salvage & Mining Oil & Gas5 Fisheries Mgmt. & Research Ship Routing Military Applications Coastal Zone Management Marine Science Community
1 4 2
Forecast Variable Salinity Currents X X X X X X X X X X X X X X X X X X X X X X X X

3
Water Level X X X
Depth EWC1 surface EWC EWC EWC
Forecast Characteristics Area of Frequency Interest ECD2 ECD WA3 WA WA ECD ECD ECD ECD ECD WA ECD ECD Littoral Zone ECD WA hourly hourly 4/day 1/day3 1/day 4/day 4/day 4/day 2/day 4/day 4/day Variable 4/day 4/day 1/day Variable
Forecast Period 0-48 hrs7 0-48 hrs 0-72 hrs UA4 UA 0-72 hrs 0-72 hrs 0-72 hrs 0-48 hrs 0-72 hrs 0-72 hrs Retrospective 0-72 hrs 0-72 hrs UA Variable
X X X X X X X X X X X X X X X X
surface EWC
EWC surface EWC
EWC EWC
X X X X
surface EWC EWC EWC
X X X
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 Centers 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 models 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/ 16x 1/16in 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 UKMOs 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 (Systme Ocanique de Prvision Rgionale en Atlantique Nord Est) is used by the French (Giraud et al., 1997) as part of their ongoing SOAP (Systme Oprationel dAnalyse et de Prvision) program. The system is based on a 1/10 quasigeostrophic model of the Northeast Atlantic from 24N54N, 35W 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
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 NOAAs 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 48N, and from the U.S. East Coast out to 50W (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
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 earths 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 proFIGURE 1
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. Navys 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-
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.
66
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: NCEPs 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 Navys 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 50W, 90 Svs exit the domain between 37 and 40N 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 SSTs will be described later.) This completes the hindcast cycle and provides new initial conditions for starting the forecast cycle each day. FIGURE 2
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
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 10C 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.
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 Buzzards 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)
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
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 Buzzards Bay, Massachusetts. The comparison covered a period of 80 days from OctoberDecember, 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 NOAAs 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 (36.9N, 75.7W) 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 models 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 JuneDecember, 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 2-3C. Off the mouth of the Chesapeake Bay, the improvement is even more striking. In this case the improvement is closer to 5C. At most buoys throughout the model domain improvements of ~1C 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 38N, 73W 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
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 38N, 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 41N, 63.5W, 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)
70
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): SSHAs 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 SSHAs are calculated from a three-year mean surface elevation field. Optimal interpolation is used to interpolate the SSHAs 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 SSHAs 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 39.5N, 65W 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 39N, 65W, 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 37.3N, 52.1W 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.
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 37.3N, 52.1W. 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 days 24-hour forecast as the initial condition for the next days forecast. We also show a forecast using the present 24-hour hindcast cycle, as discussed in section (4d), to generate initial conditions. Data from NOSs 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
72
versus the conditions from the previous days 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 models 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- 42N, and from the coast out to 70W) 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
NCEPs Marine Prediction Center (MPC, which is now called the Ocean Prediction Center), and NOAAs 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 5C 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-
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 USGSs 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, SSTs from satellite retrievals and from in situ reports are being assimilated and their
74
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 XBTs and ARGO type floats. But, unfortunately, data from these sources are sparse. Often less than 10 XBTs 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 XBTs 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 36N and 74W. 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
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.
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.
76
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.
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 NOAAs 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.
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.
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 MDs 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.
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. 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. Hay, M. and P.Jumars. 1999. Collaborating ocean ecologists assess achievements, prepare for challenges. EOS, Transactions of the American Geophysical Union. 80:77,81. Hinrichsen, D. 1998. Coastal Waters of the World: Trends, Threats, and Strategies. Washington, D.C.: Island Press. 275 pp. 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. 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.
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. 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. Leaman, K.D., R.L.Molinari, and P.S.Vertes. 1987. Structure and variability of the Florida Current at 27N: April 1982 - July 1984. Jour Phys Oceanogr. 17:565-583. McWilliams, J.C., 1996: Modelling the oceanic general circulation. Ann Rev Fluid Mech. 28:215-248. Miller, J.L., M.A. Goodberlet and J.B. Zaitzeff. 1998. Airborne salinity mapper makes debut in coastal zone. EOS. 79:173. 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.
78
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 youre 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. 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. 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.
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. Stommel, H. 1958. The Gulf Stream: A Physical and Dynamical Description. Berkeley: University of California Press and London: Cambridge University Press. 202 pp. 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
UVs, Autonomous Underwater Vehicles, are the cutting edge of technology used to explore the worlds 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 AUVs 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.
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 lets assume for a moment that AUVs are vessels. One class clearly would have to adhere to the COLREG provisionsthe 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).
80
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 operators 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 mammalAUV 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
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 governments 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-
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 takingthose 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
82
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 Captains 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 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 Orgerons 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 NASAs 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 readers 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
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 Valleys 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 doesnt 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 millionthe largest maritime salvage award in U.S. history. The crew received $1,752,642, what remained after paying interest, costs, and Keystones 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.
84
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 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 Piccards 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 chambers 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
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.
86
BOOK REVIEW
The Autobiography of a Yankee Mariner: Christopher Prince and the American Revolution
Michael J. Crawford (Editor) Brasseys, 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 oclock, 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 nations 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 Princes 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 brothers 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 Princes 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 Princes 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 Princes 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 Princes story. These are indexed just after the table of contents. A final contribution by the Editor are appendices that include Captain Princes 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 Princes 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
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 convictionsa 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.
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 (199495) ...................... $16 Military Assets for Environmental Research (199394) .................................................. $16 Oceanic and Atmospheric Nowcasting and Forecasting (1992) ...................................... $12 Education and Training in Ocean Engineering (1992) ........................................................ $12 Global Change, Part II (199192) ................... $10 Global Change, Part 1 (1991) .......................... $10 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 Underwater Intervention 98 ........................... $50 500 Years of Ocean Exploration (Oceans 97) ............................................. $130 Underwater Intervention 97 ........................... $50 The Coastal OceanProspects for The 21st Century (Oceans 96) ................. $145 Underwater Intervention 96 ........................... $50 Challenges of Our Changing Global Environment (Oceans 95) ........................ $145 Underwater Intervention 95 ........................... $75 Underwater Intervention 94 ........................... $95 Underwater Intervention 93 ........................... $95 MTS 92 ....................................................... $140 ROV 92 ....................................................... $105 MTS 91 ....................................................... $130 ROV 91 ......................................................... $95 19911992 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 OceansAn International Workplace (Oceans 87) ............................................... $65 Organotin Symposium (Oceans 87, Vol. 4) ..... $10 ROV 87 ......................................................... $40 Technology UpdateAn 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
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 Energyan 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 MappingA 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
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 ReportOcean and Coastal Engineering (2001) ................................ $7 dom ............................................................. $9For. State of Technology ReportMarine Policy and Education (2002) ................................... $7 dom ............................................................. $9 For.
PROCEEDINGS
Oceans 2003 MTS/IEEE (CDROM) .............. $50 Oceans 2002 MTS/IEEE (CDROM) .............. $50 Oceans 2001 MTS/IEEE (CDROM) .............. $50
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
Please print
S E N D A P P L I C AT I O N T O :
First name Last name
_____________________________________________________________________________________________
Mr./Ms./Dr. Address
_____________________________________________________________________________________________ _____________________________________________________________________________________________
City Telephone Total E-mail State FAX Zip Country
Marine Technology Society 5565 Sterrett Place, Suite 108 Columbia, Maryland 21044 Please call MTS at 410-884-5330 for a MTS Membership Application or see following pages for an Application. FAX (410) 884-9060 EMail: mtsmbrship@erols.com
_____________________________________________________________________________________________ _____________________________________________________________________________________________ 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 _____________________________________________________________________________________________
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 IsernNational Science Foundation Fall 2004 Environmental Threats from Submerged Vessels Lisa SymonsNOAA Winter 2004/2005 U.S. Commission on Ocean Policy: Implications and Opportunities Terry SchaffU.S. Commission on Ocean Policy, and Dan WalkerNational Research Council and Ocean Studies Board Spring 2005 Acoustic Tracking of Marine Fishes: Implications for the Design of Marine Protected Areas James LindholmPfleger Institute of Environmental Research Summer 2005 General Issue
The Marine Technology Society

is accepting papers, technical notes, and commentaries for publication in the 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 Marine Technology Society 5565 Sterrett Place, Suite 108 Columbia, Maryland 21044 Or contact: mtspubs@erols.com to request information or to submit your paper online. You can also visit our homepage at
ADVERTISING RATES
SPACE INSERTION RATE (US$) 1x Full Page Half Page Vertical Horizontal One-Quarter One-Eighth 350 350 205 135 320 320 185 115 300 300 160 105 3 7" 3 3 10" 4 4 2" 525 2x 475 4x 435 SPACE DIMENSIONS Width 7" Depth 10"
www.mtsociety.org
90
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.) Im 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.
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.
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.
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.
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.
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.
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.
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 Number 1 Volume 38,complete other side 91
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 Associate (2 yr.) Degree n Check all that apply: n B.S. B.E. n n YES B.A. n NO n Four Year Degree n M.S. n M.A. n Graduate Degree n M.E. n Ph.D. n Doctorate n Sc.D. n
Do you have a P.E. license?
BUSINESS/PROFESSIONAL INFORMATION
Name of current employer: _________________________________________________________________________________ Your employers primary line of business at your location: ________________________________________________________ If you dont 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) _______________________ Optional Information: n Male Female n What is your age? Under 30 n 30-40 n 41-50 n Your job title: __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 __Other (please specify) _______________________ Check areas of interest: __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 __Other (please specify) _______________________ 51-60 n Over 60 n
MEMBERSHIP AND JOURNAL PAYMENT

Payment Method: Check Enclosed n Master Card n Make checks payable to the Marine Technology Society (U.S. funds only) Visa n Diners Club n Am Ex n
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. San Diego, California SeaCon Brantner and Associates, Inc. El Cajon, California Sippican, Inc. Marion, Massachusetts Sonsub, Inc. Houston, Texas SonTek/YSI, Inc. San Diego, California 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 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. San Diego, California 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
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 San Diego, California DTC International, Inc. Houston, Texas Falmat, Inc. San Marcos, California Gilman Corporation Gilman, Connecticut Impulse Enterprise San Diego, California InterOcean Systems, Inc. San Diego, California 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.
Marine Technology Society

5565 Sterrett Place, Suite 108 Columbia, Maryland 21044
Postage for periodicals is paid at Columbia, MD, and additional mailing offices.

Marine Technology Society Journal

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Marine Technology Society Journal

Загружено:

Авторское право:

Доступные форматы

General Issue

Marine Technology Society Officers

PROFESSIONAL DIVISIONS & COMMITTEES

Volume 38, Number 1, Spring 2004

Peter C. Chu, Michael D. Perry, Eric L. Gottshall, David S. Cwalina

MTS Journal design and layout: Michele A. Danoff, Graphics By Design

Mark J. Kaiser, Allan G. Pulsipher, Robert C. Byrd

The Marine Technology Society