AD-A128 032 SEISMOGRAPH4IC NETWORKS: POWLEUS AND OUTLOOK FOR THE 1/11980S(U) NATIONAL RESEARCN COUNCIL VASHINGTONCOiITTEE ON SEISMOLOGY T V IIKVILLY IT AL. 1983

IN(I ASS] F lTD FIG 5/2 NL

IhhlhhhhEND

Lmn~LMD I3

u.ROOP RESLUO TES CAU

1111125 IIfl L.2 00- .26

MICROCOPY RESOLU TION TEST CHARTNATIONtAL 9LJEAU OF STOAftS -1963 A

ADA !28032

SisD~nmorpi SNetworks:

Problems nd tpuloo for.th180

Seismographic Networks:Problems and Outlook for the 1980s

Report of the Workshop on Seismographic NetworksCommittee on SeismologyCommission on Physical Sciences, Mathematics, and

ResourcesNational Research Council

NATIONAL ACADEMY PRESSWashington. D.C. 1983

NOTICE The project that is the subject of this reportwas approved by the Governing Board of the NationalResearch Council, whose members are drawn from thecouncils of the National Academy of Sciences, theNational Academy of Engineering, and the Institute ofMedicine. The members of the committee responsible forthe report were chosen for their special competences andwith regard for appropriate balance.

This report has been reviewed by a group other thanthe authors according to procedures approved by a ReportReview Committee consisting of members of the NationalAcademy of Sciences, the National Academy of Engineering,and the Institute of Medicine.

The National Research Council was established by theNational Academy of Sciences in 1916 to associate thebroad community of science and technology with theAcademy's purposes of furthering knowledge and ofadvising the federal government. The Council operates inaccordance with general policies determined by theAcademy under the authority of its congressional charterof 1863, which establishes the Academy as a private,nonprofit, self-governing membership corporation. TheCouncil has become the principal operating agency of boththe National Academy of Sciences and the National Academyof Engineering in the conduct of their services to thegovernment, the public, and the scientific andengineering communities. It is administered jointly byboth Academies and the Institute of Medicine. TheNational Academy of Engineering and the Institute ofMedicine were established in 1964 and 1970, respectively,under the charter of the National Academy of Sciences.

DAccession ForOOJV WTIS GRA&I

1 DTIC TAB

Unannouncad 0JSJustific 1o . ._[!

Av lab f m utfa'1

Ge ogcl1 ie es Board BNat 1 R a r Council Distribui.ion/

2 n ._ u Availability CodesWash t , D.C 041 Iand/or

Dist Special

COMMITTEE ON SEISMOLOGY

THOMAS V. McEVILLY, University of California, Berkeley,Chairman

C. ALLIN CORNELL, Stanford UniversityRICARDO DOBRY, Rensselaer Polytechnic InstituteROBERT B. HERRMANN, St. Louis UniversityHIROO KANAMORI, California Institute of TechnologyFRANKLYN K. LEVIN, Exxon Production Research CompanyPAUL W. POMEROY, Rondout AssociatesPAUL G. RICHARDS, Lamont-Doherty Geological Observatory,

PalisadesDAVID W. SIMPSON, Lamont-Doherty Geological Observatory,

PalisadesROBERT B. SMITH, University of UtahROBERT E. WALLACE, U.S. Geological Survey, Menlo Park

Liaison Members

LEON L. BERATAN, U.S. Nuclear Regulatory CommissionWILLIAM J. BEST, Air Force Office of Scientific ResearchMICHAEL A. CHINNERY, National Oceanic and Atmospheric

AdministrationJOHN R. FILSON, U.S. Geological SurveyEDWARD A. FLINN, National Aeronautics and Space

AdministrationJOHN G. HEACOCK, Office of Naval ResearchLEONARD E. JOHNSON, National Science FoundationGEORGE A. KOLSTAD, U.S. Department of EnergyPAUL F. KRUMPE, Agency for International DevelopmentJAMES F. LANDER, National Oceanic and Atmospheric

AdministrationJAMES M. McDONALD, Office of Naval ResearchUGO MORELLI, Federal Emergency Management Agency

iii

CARL F. ROMNEY, Defense Advanced Research Projects AgencyEDWARD SCHREIBER, U.S. Department of EnergyJOSEPH W. SIRY, National Aeronautics and Space

AdministrationK. THIRUMALAI, National Science Foundation

Staff

JOSEPH W. BERG, JR.ROY E. HANSON

iv

WORKSHOP ON SEISMOGRAPHIC NETWORKS: PARTICIPANTS

THOMAS V. McEVILLY, ChairmanDepartment of Geology and GeophysicsUniversity of CaliforniaBerkeley, CA

C. ALLIN CORNELLDepartment of Civil EngineeringStanford UniversityStanford, CA

PAUL C. JENNINGSDivision of Engineering and Applied ScienceCalifornia Institute of TechnologyPasadena, CA

HIROO KANAMORISeismological Laboratory 252-21California Institute of TechnologyPasadena, CA

KAREN C. MCNALLYEarth Science BoardUniversity of CaliforniaSanta Cruz, CA

PETER H. MOLNARDepartment of Earth and Planetary Sciences 54-712Massachusetts Institute of TechnologyCambridge, MA

PAUL G. RICHARDSLamont-Doherty Geological ObservatoryPalisades, NY

v

ROBERT B. SMITHDepartment of Geology and GeophysicsUniversity of UtahSalt Lake City, UT

Ex Officio Member

KEIITI AKIDepartment of Earth and Planetary Sciences 54-526Massachusetts Institute of TechnologyCambridge, MA

Liaison Members

LEON L. BERATANChief, Earth Sciences BranchDivision of Health, Siting and Waste ManagementU.S. Nuclear Regulatory CommissionWashington, DC

WILLIAM J. BESTDirectorate of Physical and Geophysical SciencesAir Force Office of Scientific ResearchBolling Air Force BaseWashington, DC

JOHN R. FILSONChief, Office of Earthquake StudiesU.S. Geological Survey

Reston, VA

LEONARD E . JOHNSONProgram Director for Seismology and Deep Earth StructureDivision of Earth SciencesNational Science FoundationWashington, DC

JAMES F. LANDERDeputy Director, National Geophysics and

Solar-Terrestrial Data CenterEnvironmental Data and Information ServiceNational Oceanic and Atmospheric AdministrationBoulder, CO

vi

Staff

JOSEPH W. BERG, JR.Executive Secretary

Committee on SeismologyCommission on Physical Sciences, Mathematics, and

ResourcesNational Research CouncilWashington, DC

ROY E. HANSONSenior Staff OfficerCommittee on SeismologyCommission on Physical Sciences, Mathematics, and

ResourcesNational Research CouncilWashington, DC

Guests

SHELTON S. ALEXANDERDepartment of GeosciencesGeophysics ProgramPennsylvania State UniversityUniversity Park, PA

THOMAS C. BACHEProgram ManagerGeophysical Sciences DivisionDefense Advanced Research Projects AgencyArlington, VA

JOHN N. DAVIESGeophysical InstituteUniversity of AlaskaFairbanks, AK

WILLIAM L. ELLSWORTHU.S. Geological SurveyMenlo Park, CA

E. ROBERT ENGDAHLBranch of Global SeismologyU.S. Geological SurveyDenver, CO

vii

WALTER W. HAYSOffice of Earthquake StudiesU.S. Geological SurveyReston, VA

ROBERT B. HERRMANNSt. Louis UniversitySt. Louis, MO

JEFFREY K. KIMBALLU.S. Nuclear RegulatoryWashington, DC

ANDREW J. MURPHYU.S. Nuclear RegulatoryWashington, DC

PAUL W. POMEROYRondout AssociatesStone Ridge, NY

ROGER M. STEWARTOffice of Earthquake Studies

U.S. Geological SurveyReston, VA

viii

GEOLOGICAL SCIENCES BOARD

WILLIAM R. DICKINSON, University of Arizona, ChairmanSAMUEL S. ADAMS, Adams and Associates, Boulder, ColoradoLLOYD S. CLUFF, Woodward-Clyde Consultants, San Francisco,

CaliforniaWALTER R. ECKELMANN, Exxon Corporation, New YorkMICHEL T. HALBOUTY, The Halbouty Center, Houston, TexasWILLIAM W. HAY, University of ColoradoMELVIN, J. HILL, Gulf Oil Corporation, Houston, TexasCARROLL ANN HODGES, U.S. Geological Survey, Menlo Park,

CaliforniaWILLIAM C. LUTH, Sandia National Laboratories,

Albuquerque, New MexicoCHARLES J. MANKIN, Oklahoma Geological Survey, NormanV. RAMA MURTHY, University of Minnesota, MinneapolisJACK E. OLIVER, Cornell University, Ithaca, New YorkSTEPHEN C. PORTER, University of Washington, SeattleJ. WILLIAM SCHOPF, University of California,

Los Angelesai-AN ZEN, U.S. Geological Survey, Reston, Virginia

Liaison Members

LEON L. BERATAN, U.S. Nuclear Regulatory CommissionROBIN BRETT, National Science Foundation,

Washington, D.C. (until September 30, 1982)PHILIP COHEN, U.S. Geological Survey, Reston, VirginiaKENNETH DAUGHERTY, Defense Mapping Agency,

Washington, D.C.PAUL R. FISHER, U.S. Department of the ArmyBRUCE B. HANSHAW, U.S. Geological Survey, Reston,

Virginia

LX

JAMES F. HAYS, National Science Foundation,Washington, D.C. (from October 1, 1982)

JOHN G. HEACOCK, Office of Naval ResearchLINN HOOVER, U.S. Geological Survey, Reston, VirginiaGEORGE A. KOLSTAD, U.S. Department of EnergyJOHN F. LANCE, National Science Foundation,

Washington, D.C.DALLAS L. PECK, U.S. Geological Survey, Reston,

VirginiaMARK SETTLE, National Aeronautics and Space

Administration, Washington, D.C.A. G. UNKLESBAY, American Geological Institute,

Falls Church, VirginiaKENNETH N. WEAVER, Maryland Geological SurveyFRANK J. WOBBER, U.S. Department of Energy

Ex Officio

CLARENCE R. ALLEN, California Institute of Technology,Pasadena

JOHN C. CROWELL, University of California, Santa Barbara

Staff

JOSEPH W. BERG, JR., Executive SecretaryWILLIAM E. BENSON, Senior Staff OfficerROY E. HANSON, Senior Staff Officer

X

COMMISSION ON PHYSICAL SCIENCES,MATHEMATICS, AND RESOURCES

HERBERT FRIEDMAN, National Research Council, CochairmanROBERT M. WHITE, University Corporation for Atmospheric

Research, Cochairmaan

STANLEY I. AUERBACH, Oak Ridge National Laboratory

ELKAN R. BLOUT, Harvard Medical SchoolWILLIAM BROWDER, Princeton University

BERNARD F. BURKE, Massachusetts Institute of TechnologyHERMAN CHERNOFF, Massachusetts Institute of Technology

WALTER R. ECKELMANN, Exxon CorporationJOSEPH L. FISHER, Office of the Governor, Commonwealth of

Virginia

JAMES C. FLETCHER, University of Pittsburgh

WILLIAM A. FOWLER, California Institute of Technology

GERHART FRIEDLANDER, Brookhaven National LaboratoryEDWARD A. FRIEMAN, Science Applications, Inc.EDWARD D. GOLDBERG, Scripps Institution of Oceanography

KONRAD B. KRAUSKOPF, Stanford University

CHARLES J. MANKIN, Oklahoma Geological Survey

WALTER H. MUNK, University of California, San DiegoNORTON NELSON, New York University Medical Center

DANIEL A. OKUN, University of North CarolinaGEORGE E. PAKE, Xerox Research CenterCHARLES K. REED, National Research Council

HOWARD E. SIMMONS, JR., E.I. du Pont de Nemours and

Company, Inc.

HATTEN S. YODER, JR., Carnegie Institution of Washington

RAPHAEL G. KASPER, Executive Director

xi

PREFACE

Seismographic networks collect the data fundamental tothe science of seismology, providing recordings of groundmotion from natural earthquakes and other seismic sourcesranging in size from the great earthquakes to the smallestdetectable microearthquakes. This ground motion revealsthe passage of seismic waves through the earth, and thewaves may be recorded very close to the source or atgreat distances many minutes or hours after the event,having traversed paths that penetrate the entire earth.The data tell us much about the earth's interiorstructure and dynamics in addition to the nature ofearthquakes. They allow identification of the type ofsource, explosion or earthquake, and they provide detailsof the seismic source process. These observations areimportant not only to the science of seismology but also,directly or indirectly, to society. For example, theearth's magnetic field--the basis for navigation, forgeophysical exploration, and for a geological timescale--is generated mainly in the core, the structure ofwhich is determined seismologically; the differentiationof explosions and earthquakes by seismic means is basicto monitoring a nuclear test ban treaty; seismic meansare used to study inhomogeneities in the earth's mantle,which can lead to the discovery of mineral resources; themitigation of seismic hazards for general constructionand for critical facilities requires knowledge of boththe locations of expected earthquakes and the nature ofstrong ground motion; and the search for methods ofearthquake prediction relies heavily upon the existenceof networks of closely spaced seismographic stations.

Seismology is a young and vigorous science, and it isbeing called upon continually to address new problems.The laboratory of seismology is the earth, its data baseconstantly changing and its time scale set by geodynamic

xiii

~~1i6o~PAW S LAMWMO 71 lJ

processes. The seismographic networks are the basicscientific tools, analogous to major telescopes inastronomy or particle accelerator facilities in physics,that provide the continuing data base for the science.It is very important, for this reason, to keep U.S.-supported seismographic networks in the best operatingcondition, to provide networks with the latest technology,and to improve constantly the management and data basesof the networks.

These needs, unfortunately, have not always been met.The importance of observational data from seismographicnetworks has not been recognized consistently by decisionmakers allocating funds among competing programs. Thevarious governmental agencies responsible for networkoperations encounter many difficulties in obtainingadequate funding for the maintenance, upgrading, and theresearch associated with these important nationalfacilities. Even though the amounts of money needed aremodest, crises in support funding seem to occur regularly,as short-term objectives change within the agencies.

Unlike the otherwise analogous telescope oraccelerator facilities, seismographic networks are madeup of large numbers of individual, relatively smallinstallations, necessarily distributed widely withrespect to the features studied, and thus globally formany investigations.

The broad perception of networks as facilities isconsequently lacking, contributing to their vulnerabilityin times of financial stress. In addition, the ongoingprocess of upgrading involves simultaneous acquisition ofnew equipment for all stations in the network, includingforeign installations in the global networks, and theconsequences of this peculiarity in facility maintenanceare not readily accepted by funding agencies. Given theextremely rapid rate of technological advance in theareas of data acquisition and processing, maintainingstate-of-the-art capabilities in the seismographicnetworks is a difficult task indeed.

This report is the result of a workshop convened atthe request of several governmental agencies to reviewthe status and associated problems of and the outlook forseismographic networks. Recommendations have been madeto help solve the problems and to assure a viableobservational capability for the future. If this isaccomplished, the time and effort of the manycontributing scientists will have produced a majorcontribution to the nation.

Thomas V. McEvillyChairman

ACKNOWLEDGMENTS

This study was performed by the participants of the

Workshop on Seismographic Networks under the aegis of the

Committee on Seismology of the National Research Council'sCommission on Physical Sciences, Mathematics, andResources. The work of the Committee on Seismology is

supported by the National Oceanic and Atmospheric

Administration, U.S. Geological Survey, Air Force Office

of Scientific Research, Nuclear Regulatory Commission,

Office of Naval Research, Department of Energy, FederalEmergency Management Agency, Defense Advanced ResearchProjects Agency, and the National Science Foundation.The Committee on Seismology expresses its appreciation

for the interest and support of these agencies.Forty-five operators of seismographic networks completedquestionnaires for the participants of the workshop, andthese provided many insights into operations and problems.

xv

CONTENT S

1 EXECUTIVE SUMMARY 1

2 INTRODUCTION 4

3 GLOBAL NETWORKS 7

4 REGIONAL NETWORKS 17

5 NATIONAL NETWORK 27

REFERENCES 30

APPENDIXES

A GLOBAL NETWORK DATA 31

B REGIONAL NETWORKS: QUESTIONNAIRES,RESPONDENTS, SUMMARY 45

C SUMMARY AND MAJOR RECOMMENDATIONSOF U.S. EARTHQUAKE OBSERVATORIES:RECOMMENDATIONS FOR A NEW NATIONALNETWORK 59

D GLOSSARY 62

xvii

NaGPAW &AI6MO? 11

1

EXECUTIVE SUMMARY

An earthquake produces seismic waves, which radiate fromits focus, traveling around and through the earth, withsize and persistence proportional to the dimensions ofthe source. Only a very few of the thousands of earth-quakes cataloged annually affect mankind directly. Mostare perceptible only to seismographs--the scientific eyesand ears of the seismologist. From the seismograms,which display ground motion associated with the passingwaves, comes our knowledge of the global distribution ofearthquakes, of the internal structure of the earth, andof the earthquake source process. Interpreting therecorded seismogram requires sophisticated analysisprocedures. Recent advances in analytical methods andinstrumentation have increased dramatically theinformation to be gained from seismograms, but acquisitionof adequate seismological data requires wide coverage byseismographs, globally, nationally, and regionally.Instruments must be maintained and upgraded regularlywith the latest technology. Effective management iscrucial for operations and data handling. All of theseneeds require adequate financial support over longperiods of time.

Seismographic networks provide data essential toprograms such as the mitigation of earthquake hazards,the definition of geological structure on the margins andwithin tectonic plates, the safe siting of dams, powerplants, and other critical facilities, and theinvestigation of dynamic processes of the earth.Operating a typical seismographic network is not overlyexpensive, but it does require dedication of time andtalent by seismologists who run the stations. In manycases the major rewards are in providing data to helpsolve problems of national and global significance.

2

The large number of questions on seismographic networksbrought in recent months to the Committee on Seismologyis strong evidence that there are critical problems withnetwork operations. At the Workshop on SeismographicNetworks, prompted by these questions, participantsconsidered global, regional, and national networkscollectively as an integrated system and also as entitieswith specific problems. This report discusses eachcomponent of the system in terms of rationale andproblems, giving recommendations for solutions. A briefstatement follows of major problems and major recommenda-tions for the global, regional, and national networks.

Global Networks. Global networks are expected to providefor the scientific community a data base that continuesindefinitely. Unfortunately, managing agencies find itdifficult to recognize this long-term scientificimportance. The service function of the networks, i.e.,providing data for other users, must be considered infunding decisions by the managing agency. Globalnetworks require continuing financial support at anadequate level. It is recommended (1) that considerationbe given to transferring management responsibility forthe global network from its present organizational baseto another location within the U.S. Geological Survey oreven to another agency, if such a change seems clearlyadvantageous; (2) that stable funding for global networksbe sought from normal budgetary requests from within theU.S. Geological Survey, from the Defense AdvancedResearch Projects Agency, and from other agencies thatuse data from the networks; (3) that access to digitaldata and use of those data be improved while networkscontinue to meet fully the demand for and the globalcoverage provided by analog (i.e., visible) data at thepresent time; and (4) that procedures be established andfunding be provided for the orderly and continuinginteragency transfer of the most recent instrumentationand technology.

Regional Networks. Regional network operations are besetwith problems falling into three categories: functionaldefinition, funding difficulties, and operationalproblems. Functional definition is the planned lifetimeof a network, and a realistic estimate of it needs to beprovided. Funding difficulties are of two types: a lackof stability on a year-to-year basis, and the vulner-ability of research funding being decreased to maintain

3

network operations in times of fiscal stress when researchfunding is mixed with basic operational costs of thenetwork. Operational problems are seen in a lack ofcoordination among networks, the need for a morestandardized data base management system, and a growingobsolescence of network equipment. These problems areinterrelated and difficult to order in importance.

Recommendations are (1) that networks of planned,limited lifetime be reviewed every three to five yearswith respect to objectives and performance; (2) that theprovision of data fundamental to research on seismotec-tonic processes and earthquake occurrence in the regionbe acknowledged by funding agencies as the main purposeof regional networks; (3) that an adequate data set fromall regional networks be archived; (4) that data formatsbe standardized; and (5) that operations of networks becoordinated.

National Network. The concept of a national networklacks general acceptance and widespread support by theU.S. seismological community, within which there is atpresent little coordination of network operations. Theconcept is sound, and support will grow with formulationof a suitable plan for implementation.

Working Group on Seismic Networks. It is recommendedthat a Working Group on Seismic Networks be set up underthe Committee on Seismology to provide continuity anduniformity in consideration of the various policy mattersarising in network seismology. This group will providethe review functions recommended throughout this reportfor global, regional, and national networks. It shouldevaluate continually the health and status of regionalnetworks, and advise on the development of a nationalnetwork.

The contributions to the earth sciences from seismicnetworks of all types have been substantial in the pasttwo decades. We have entered the 1980s with majoradvances in data acquisition, management, and processingtechniques now available to seismology. The challenge isto build effectively on the present structure ofnetworks, creating a new capability for addressing thenext level of difficulty in the exciting problems ofgeoscience.

2

INTRODUCTION

The Workshop on Seismographic Networks was convened toconsider the problems confronting network seismology andto provide scientific, technical, and management guidanceto federal agencies, primarily regarding the operation ofglobal, national, and regional seismographic networks.It was prompted by a number of related questions broughtin recent months to the Committee on Seismology. Thisreport is the result of that workshop.,-'

The National Research Council report, U.S. EarthquakeObservatories (NRC 1980), recommended establishing anintegrated U.S. Seismograph System (USSS), the core ofwhich would be a new national network of modern digitalseismological observatories. The report called also fora guiding working group on the USSS to be established.Many of the problems discussed and several recommendationsin this report are similar to those in the 1980 study.Current constraints in federal funding, the potentiallydisastrous budget cuts nearly imposed in earthquakestudies by several agencies in the fall of 1981, and thepromise of continued stringencies all impart a sense ofurgency to the need for clear position statements by theseismological community on U.S. seismographic networks.

The global, national, and regional systems ofseismographic stations, spanning the earth much likemeteorological observatories, provide the fundamentaldata base for scientists to investigate the earth atdifferent scales, addressing problems of earthquakehazards and prediction, safer sitings for criticalfacilities, and the identification of underground nuclearexplosions, in addition to fundamental questions on thephysical and chemical composition and geologicalstructure and dynamics of the entire earth. Majoradvances in the earth sciences have come directly from

4

5

these data. Agencies with responsibilities for maintain-ing subsets of this worldwide seismographic system haveasked the Committee on Seismology for guidance inallocating their fixed or decreasing financial resourcesamong competing scientific efforts.

As the primary source of seismological data, networkshave been and continue to be essential to the scientifichealth of seismology. The committee perceives a range ofserious problems threatening this data base. The globalnetwork is insufficiently funded. All networks sufferfrom rising operational costs. Questions must beaddressed on appropriate operational lifetimes. Muchinstrumentation is obsolete. Modernization of dataacquisition and archiving methods is needed, and existingnew approaches to data base management should beintroduc-d to provide a creative environment for research.It is generally acknowledged that digital systems mustultimately replace analog, but at what cost? All ofthese efforts call for increased financial support at atime when the funding climate is inhospitable. Theseismological data base demands stable continuity ofsupport for operating networks that is independent ofvariations in research support.

An important aspect of this report is the considera-tion, perhaps for the first time in such depth, of theinterplay among the network systems of differing scalesand purposes--global, national, and regional--and theirdefinition as a major scientific resource for acquisitionof important data. In this unification the nationalnetwork becomes the linkage by which regional networksare integrated with the global seismographic systems.Thus, that portion of the global network located withinthe United States can be viewed as a subset of thenational network, which in turn is a subset of stationsfrom regional networks. The rationale for such astructure is to facilitate exchange of data, methods ofanalysis, and scientific results, by enlisting theinvolvement of present network operators in the system.

In order to review in depth the present state ofdifficulties facing the networks, the committee solicitedassessments and opinions from a range of networkoperators, users, and supporting agencies. A compre-hensive questionnaire was formulated and distributed tooperators of regional networks. Forty-five completedquestionnaires were returned by federal and stateagencies and universities, providing an unprecedented andsubstantial overview of regional seismographic networks

6

operated for a variety of purposes by U.S. scientists.Appendix B gives the questionnaire, a list of the

respondents, and a summary of the resulting information.Even before the workshop began, the need for a

continuing working group that would address practicalproblems of seismic networks from a perspective differentfrom that of any one agency was recognized. TheCommittee on Seismology plans to establish a workinggroup on seismic networks within the National ResearchCouncil, for if it does not, the committee's agenda foryears to come will be dominated by network questions.The working group will be asked for policy recommenda-tions on the interrelationships among global, national,and regional networks. Other, more specific, questionsfor the working group are given throughout this report.

The working group, which will report to the Committeeon Seismology, is to consist of at least five individualswith overlapping terms of about five years. Networkproblems are not quickly solved. It is thereforeimportant for a majority of members of the working group

to serve long enough to show some results from changes innetwork policy. We anticipate that the working group

will often need to seek the advice of specialists,particularly in recommending technological improvements.

The report considers global networks in Chapter 3 and

regional networks in Chapter 4, reviewing the nature anduse of each, identifying the problems peculiar to each,and recommending various approaches toward solutions tothe problems. Chapter 5 reviews the concept of a newnational network and its place in the present networkstructure. By virtue of the variety of issues, therecommendations range from specific actions toacknowledgments of remaining outstanding difficultiesrequiring further attention. Background information forthe workshop is presented in Appendixes A through C.

3

GLOBAL NETWORKS

REVIEW

The history of global seismographic networks* is closelytied to the U.S. national need for improved capability indetecting and identifying underground nuclear explosions.The World-Wide Standardized Seismograph Network (WWSSN)was established in the early 1960s as a part of ProjectVela Uniform, a program of fundamental and appliedresearch in seismology managed by the Defense AdvancedResearch Projects Agency (DARPA). Since then DARPA hasbeen responsible for virtually all advances in globalseismographic networks, but has declined to commit fundsfor routine WWSSN operations. Key elements of the WWSSNare standardized three-component long- and short-periodseismographs with uniform calibration, and the means fordistributing the seismograms to the earthquake researchcommunity. The WWSSN today comprises 110 stationsoperating in 54 countries; its role is to produce thedata needed for fundamental research in seismology. Theresponsibility for installing and managing the WWSSN wasassigned by DARPA to the U.S. Coast and Geodetic Survey(USCGS). The network was essentially completed by 1963.The WWSSN serves as a worldwide organization base towhich network improvements and modernization can beapplied. The network was partially funded by theNational Science Foundation (NSF) between 1968 and 1978.In 1973 the WWSSN and other elements of the USCGSearthquake program, apart from the data services, weretransferred to the U.S. Geological Survey (USGS) and,

*See Appendix A for further information on the global

networks.

7

8

since 1978, the network has been funded entirely by theUSGS under the National Earthquake Hazards ReductionProgram (NEHRP). Since the inception of the WWSSN, morethan 5 million original seismograms have been microfilmedand 60 million high-quality film copies have been suppliedto research workers by the Environmental Data Service ofthe National Oceanic and Atmospheric Administration(NOAA). Despite the superiority of the more recently

available digital seismic data for many purposes, theanalog seismograms from the WWSSN remain the foundationfor much fundamental seismological research, not only inthe United States but around the world. WWSSN encounteredits first difficulty when attempts were made to transferits financial support from the Department of Defense (DOD)to NOAA in the late 1960s, a transfer that had beendiscussed and coordinated over a period of years. Opposi-tion centered on the U.S. support of the foreign stationsof the WWSSN, and the NSF assumed the responsibility forpartial support of the foreign stations of the WWSSN in1968.

In the late 1960s, DARPA sponsored the development andinstallation of 10 high-gain long-period (HGLP) seis-mographs. The HGLP seismographs were superior to theWWSSN seismographs in the detection of long-period earthmotion because of their installation in special airtighttanks to protect them from temperature and pressurechanges. The HGLP seismographs were also the firstglobally deployed seismographs to be equipped withdigital recorders. The USGS was assigned responsibilityfor the HGLP network and managed it as a complementarypart of the WWSSN. Later, five of the HGLP systems weremodified: short-period seismometers were added, and theoriginal digital recorders were replaced by more advanced,computer-controlled versions. The modified HGLP seis-mographs, now called Abbreviated Seismic ResearchObservatory (ASRO) systems, are still in operation.

The HGLP seismographs, especially the horizontal com-ponents, are affected by earth tilt caused by atmosphericloading of the earth's surface by wind and variations inbarometric pressure. The resulting ground noise isattenuated rapidly with depth. A joint effort by privateindustry and the government led to the successfuldevelopment of a broadband low-noise force-balanceaccelerometer that could be installed in a small-diameterborehole. Operated at a depth of 100 m, the boreholeseismometer is virtually unaffected by wind noise at thesurface. In 1973, the USGS, with DARPA funding, began

9

the development and global deployment of 13 SeismicResearch Observatories (SRO) that combined the newborehole seismometer with an advanced analog and digitalrecording system.

The availability of high-quality digital data producedby the SRO stimulated utilization of digital data by manyresearch organizations. New analytical techniques andsoftware were developed, opening exciting new directionsof research such as the determinations of sourceparameters for all large earthquakes in a given timeinterval. In the late 1970s, again with funding fromDARPA, the USGS developed a digital recorder that couldbe attached to existing WWSSN systems. Seventeen suchrecorders are currently being installed at WWSSN stations(termed DWWSSN stations). NSF has funded theirinstallation at six foreign stations, and the USGS isfunding installation at the remaining sites.

ASRO, SRO, and DWWSSN stations have been termedcollectively the Global Digital Seismograph Network(GDSN) and observatories are located in Figure 1. TheGDSN and the WWSSN are complementary networks andtogether have been termed the Global Seismograph Network(GSN). Operation of the GSN, with stations in more than60 countries of the world, is a notable example ofsuccessful international scientific cooperation and dataexchange. Hundreds of seismographic stations areoperated by other countries that go into a total globalseismographic network effort, and data from these areavailable to U.S. seismologists. There are manyseismologists, both in the United States and in foreign

countries, who do not have access to computer facilitiesand can work only with analog data, who prefer to workwith analog data, or who need the denser global coverage

of stations provided by the WWSSN. In its relativelyshort life, the WWSSN has generated an historical data

base that is important as a baseline for testing newhypotheses; continuing data from many of the samestations are essential to eliminate the possible effects

of station location on recordings of seismic events. TheGSN would continue an infrastructure that stimulates andsupports international cooperation in seismology. It nowprovides the USGS with convenient and ready access toseismological organizations in more than 50 countries.In many countries, the GSN stations represent theprincipal or only national facilities for support ofin-country seismological programs. Termination of someWWSSN stations may have serious and long-lasting

10

II

Ike

11

political as well as scientific repercussions. Eventhough numerous seismographic stations exist in manyother countries, their data are not as readily availableto the researcher as are GSN data because of theestablished data distribution system.

While the SRO systems were being installed, theUniversity of California at San Diego developed and beganinstalling digitally recording gravimeters that aredesigned to record very-long-period vertical-componentseismic data, including earth tides and free oscillations.Seventeen are currently in operation, and the network iscalled the International Deployment of Accelerometers(IDA). The digital data provided by the IDA network anddistributed by NOAA are used by many research groups forvery-long-period studies of the earth's structure and theearthquake source mechanism. Implementation of IDA hasbeen funded by the National Science Foundation and aprivate source, but, as manager of GSN, the USGS has beensuggested by NSF as the agency that can best providelong-term future support for the network.

The role of DARPA is limited to the development anddemonstration of new technology; hence DARPA support forthe operation of the GDSN ended in FY 1979 even thoughmuch DARPA-supported research today makes substantial useof GDSN data. Following recommendations set forth in theNRC reports entitled Trends and Opportunities inSeismology (NRC, 1977b) and Global Earthquake Monitoring(NRC, 1977a), the research community strongly urged thatthe USGS provide long-term operational support of theGDSN beginning in 1980. Funding for the continuedoperation of the network by the USGS was planned in thereport entitled Earthquake Prediction and HazardMitigation: Options for USGS and NSF Programs (NSF,1976), which outlines the structure of the NationalEarthquake Hazards Reduction Program (NEHRP), beginningin FY 1980 under all budget options in the report. Therecommendations were accepted, and, beginning in FY 1980,the USGS has provided both management and funding of theGDSN.

Network support operations are managed by the USGSAlbuquerque Seismological Laboratory (ASL). ASL furnishesoperating supplies routinely and replacement parts orcomponents on call and repairs defective equipmentreturned from the stations. Contract maintenancetechnicians operating out of ASL install new equipmentand service operating stations. Optimally, two or threemaintenance teams are overseas at any given time and can

12

respond to provide assistance to stations. Duringroutine service visits, the maintenance personnel performtests, calibrations, and any software or hardwaremodifications that are required. In operating the WWSSNstations, participating organizations contribute 2 to 3times more than the annual costs to the U.S. government.

The most recent (to 1982) development has been theinstallation of a five-station telemetered network bySandia National Laboratories as part of a RegionalSeismic Test Network (RSTN). The stations, which areunmanned, acquire signals from borehole seismometers(improved versions of the SRO sensors) and transmit ninechannels of data continuously in three data bands:short-period, mid-period, and long-period. A geosynch-ronous satellite provides two-way communication betweenthe network control station, located in Albuquerque, andeach RSTN station. Current plans are to merge RSTN data(events only in the case of short- and mid-period signals)with GDSN data on the network day tapes for generaldistribution. RSTN data are also received and processedin the new DARPA Center for Seismic Studies (CSS) inArlington, Virginia.

The establishment of the WWSSN, and its subsequentdevelopment, was certainly one of the outstanding U.S.accomplishments in international support of science. Ithas been the major factor in developing seismology into aquantitative and precise science that provides much ofour basic knowledge of the earth's structure, of theactive processes that deform it, and of the hazards andrisks of earthquakes. One can well argue that therevolution in the earth sciences that followed therecognition of plate tectonics derived from two mainsubfields, i.e., marine geology and seismology, and thatthe contributions from seismology derive largely from ananalysis of WWSSN data. We have every reason to believethat advances in the coming decades will flow similarlyfrom analyses of data from today's global networks.

ASSESSMENT OF PROBLEMS

Global networks have been plagued throughout theirexistence by the lack of sufficient long-term fundingfrom a host organization within which the operation isviewed as a significant element and thus receivesenthusiastic support. The funding problems of the WWSSNdid not cease with its transfer to the new Branch of

13

Global Seismology of the USGS in 1973. It now had to

compete for its funding with other elements within theUSGS of the large and underfunded NEHRP, and with theincreasing support needed for the GDSN. Global networks,in the competition between USGS service- and mission-

oriented programs, have been subject to funding pressuresthat annually threaten their stability and continuity.Within the broad USGS seismological research program, a

fairly small fraction of scientists works with data from

global networks, so that internal advocacy, whilededicated, represents a minority. Consequently, the

maintenance of global networks, an important service to

the seismological community, lacks universal support

within the Survey. Despite this situation, USGS

continues to support the GSN. In contrast, a large part

of university-based seismological research on the

earthquake process, earth structure, tectonics, and

earthquake prediction relies heavily on the global

networks. Thus many non-USGS seismologists depend upon

that agency for service in providing high-quality globaldata while most USGS seismologists do not need such data

for their research. This dichotomy produces competition

between programs with strong Survey interest and those

that provide services largely for non-Survey personnel.

Nevertheless, USGS administrators have allocated funds

for the maintenance of the WWSSN and to allow for the

steady upgrading of the worldwide system by the addition

of a digital capability, with support from DARPA and NSF.A serious problem exists in that global network funds

have remained fixed for the past three years. This year,once again, the seismological community and the USGS havebeen faced with options to meet the budget ceilings.*These include reducing WWSSN, reducing GDSN, and obtainingmore funds, the latter never succeeding despite strong

efforts. The community has always reaffirmed its strong

backing for retaining the capabilities represented by the

WWSSN essentially as it now exists. This annualoccurrence makes it clear that some means must be found

to ensure the continuity of the WWSSN and its systematicupgrading into the digital era.

The GSN should be insulated from the vagaries offunding of the NEHRP. Support for the GSN must be thoughtof in terms of decades and in terms of its international

impact and in terms of the health of seismology. The

*See Appendix A for details of global network options.

14

USGS must continue to meet its long-term commitment tothe GSN, even if the NEHRP should be terminated. Giventhe time scale of processes controlling global seismicity,it is essential to establish and maintain a data basethat is uniform for many decades.

The major short-term problem thus rests with fundingfor the global system. Because of recent increases inthe costs of supplies for the WWSSN and of maintenance

required for the GDSN, it is becoming increasinglydifficult for the USGS with inadequate program resourcesto maintain the present and the projected level ofoperation of these networks. On the other hand, diverseusers in the seismological community require both analogand digital data with no interruption of continuity orcoverage of either type of data. Eliminating either theWWSSN or the GDSN should not be considered an acceptablesolution to current funding problems. (Absorbing a FY1983 $500,000 shortfall in this manner would represent a35 to 40 percent reduction in the data collected.)*

Seismologists do not consider the problem as one ofanalog versus digital data. Until very recently, analogseismograms constituted the fundamental data base ofglobal seismology. Although analog records are morelimited in dynamic range and more difficult to use forquantitative analysis than digital records, they appealto the trained eyes of seismologists and other earthscientists more directly. Very often subtle changes inthe waveform on analog records are used for the determina-tion of the depth of earthquakes and for information onthe fine structure of the earth's interior. Initialhints that led to important discoveries have often beenfound in such features of analog seismograms. Analogrecords represent many more stations than digital records,and they can be utilized without sophisticated hardware

and software.*Digital seismograms, on the other hand, are far

superior to their analog counterparts when we know whatparameters are to be extracted from them. A large numberof data can be processed in a relatively short time toobtain accurate results. Over the past few years, thenumber of high-quality normal mode data and sourcemechanism solutions has increased by more than an order

*See Appendix A for budgetary information.**See Appendix A for a discussion of the utility of

analog data.

15

of magnitude, which brought about several major break-

throughs in seismology. The basic fact of data superior-ity assures the ongoing conversion to digital acquisitionof seismic data.

The importance of the global data base to the seis-

mological community is such that the overwhelmingconsensus of the workshop participants was that nostations should be eliminated with the possible exceptionof redundant WWSSN analog and GDSN digital stations. Itis anticipated, however, that digital stations willeventually supplant the analog WWSSN stations at selectedsites of the global network and that analog (microfiche)seismograms will be generated from the digital recordings.The remaining analog WWSSN stations should continue tooperate until it can be demonstrated to the scientificcommunity that they are no longer needed as part of theglobal data base.

A third type of problem lies in the slow pace ofutilization of the digital data. Because of the diversityof users and their in-house capabilities, there aredifficulties in the transfer of technology in going fromanalog to digital recording. This includes instrumenta-tion changes, software development for users, anddissemination of information to users. Moreover, theflow of digital data to users is not yet adequate.Tutorial data packages on selected special events areneeded, to allow researchers to gain experience with thenew data at reasonable cost and to learn how to use themfor their individual problems.

High-technology, high-cost facilities transferred from

one government agency to another create a fiscal problemfor the receiving agency. We here refer to such transfersas that of the WWSSN from DOD to NOAA to the USGS; thatof the SROs from DOD to the USGS; the pressures fortransfer of IDA from NSF to USGS; and the possibility ofa transfer of RSTN from DOE to USGS. The problem is notthat the receiving agency is surprised by an unexpectedrequest--in the above cases, transfers have been plannedover a period of years. Rather, the problem is that basefunding is almost never transferred along with theresponsibility for the facility operation. The receivingagency thus gets the funding problem as part of thetransfer. It should be emphasized that these transfersare usually regarded as the best alternatives to thegovernment and to the science. The option of dismantlinga valuable resource is not an acceptable alternative.

16

RECOMMENDATIONS

It is essential to maintain a global data base that isuniform for years or even decades. This requiresestablishing a supportive home environment for GSN.Relevant U.S. government agencies must realize that along-term commitment is implicit in the very existence ofGSN, independent of the mission-oriented programs of thehost organization. We suggest that a solution to thelong-term problem may be to transfer responsibility forthe global networks from its present NEHRP base toanother location within the Survey or even to anotheragency so the GSN does not compete directly for fundswithin a largely unrelated program.

In regard to the options open to the USGS over theshort term, we recommend seeking funds to support theglobal networks at the required level from normalbudgetary requests, from within the USGS, and from otheragencies, such as DARPA, that use data from the networks.Should this fail, we recommend that efforts be directedtoward maintaining the WWSSN at the present level at theexpense of a slowdown in the completion of the GDSNprogram and of a data loss that may result from a reducedGDSN maintenance program. Consequences of this approachshould be reviewed yearly. At the same time the USGSshould seek assistance from other agencies in providingoperational support during this difficult interim period.In particular, DARPA cannot ignore the pertinent role ofthe GDSN data base in DARPA objectives, and the seriousdegradation threatened by inadequate funding.

The problem of digital data usage must be addressed

while the demand for analog data continues to be metfully. Additional efforts are needed to familiarizeseismological researchers with the use of digital dataand associated software for analysis. Digital data willultimately replace the analog WWSSN. Methods for betteraccess to the data need to be developed, and new dataproducts at reasonable cost to the user should begenerated and provided by agencies responsible for datadistribution to promote wider use of the data.

The interagency transfer of network facilities is aremaining outstanding issue. The associated problems canbe largely resolved if adequate funding is transferredalong with the operational responsibility, as was done,for example, in the recent transfer of the strong-motion

network from NSF to USGS.

4

REGIONAL NETWORKS

REVIEW

Regional networks are those that have been installed to

study seismological problems on a geographic scale of 100to 1000 km. Such networks have been established widelyin recent years, and today there are approximately50 regional networks in the United States, each consistingof tens to hundreds of individual seismographic stations.These networks are supported by a number of differentgroups (e.g., USGS, U.S. Nuclear Regulatory Commission(USNRC), DOD, DOE, and other federal, state, and localagencies) for a number of different reasons.

The role of regional networks is, in general, todelineate the time and space distribution of earthquakeson a fine enough scale to contribute to our scientificunderstanding of earthquake occurrence and relatedtectonic processes and to provide important baseline datafor engineering investigations (e.g., earthquakeprediction and hazard assessment). Figure 2 shows theearthquake occurrences in California for 1980. U.S.regional networks date from the 1887 installation of theUniversity of California seismographic stations.

The first telemetered network in the U.S was that of

the USGS in Hawaii. Developed during the mid-1950s, thenetwork had four original stations around the summit ofKilauea Volcano with the information telemetered to theHawaiian Volcano Observatory. By July 1958 this localnetwork had expanded to about 15 km across.

An early U.S. regional network of seismographicstations that was connected by FM telemetry to a centralsite was installed by the University of California atBerkeley in 1960 to monitor and study seismicity in

central California. Regional networks with increased

17

18

0.

0 0

0 0, ..

00

o,.

0 00 200

KI LOMETERS o IO

- 9

FIGURE 2 California seismicity for 1980 from data of

three regional networks - Caltech, Univ. of Nevada, and

USGS.

station density were installed by the USGS in the late

1960s to study in greater detail the San Andreas Fault inCalifornia. Within a very short time, well-definedspatial patterns of earthquakes were delineated, providinga clearer fine-scale picture of the spatial distributionof earthquakes. Given this initial success, throughoutthe next decade, regional networks were extended inCalifornia and established in other states with ongoingseismicity.

19

Only since 1970 or later have the major seismologicalproblem areas of the United States been consistentlymonitored at high sensitivity by regional networks. Theincrease in data flow resulting from this large number ofseismographic stations has, in part, been managed by theintroduction of automatic detection and digital dataprocessing. The acquired data have been used forengineering planning, disaster mitigation, and fundamentalscientific investigations. In many cases the results ofa network installation and operation transcend theoriginal purpose of its installation. Some relativelysmall networks have contributed important scientificresults on topics ranging from plate tectonics to themechanics of crustal deformation and from inducedseismicity to the prediction of volcanic eruptions.

The proliferation of regional networks, usually fundedinitially by various mission-oriented agencies for veryspecific purposes, has resulted in some problems.Coordination of the establishment, distribution, andshutdown of local and regional networks has proved to bea difficult task. An increasingly acute problem for theUSGS is its role in continuing the operation of networksthat have lost support from their original fundingagency, and the impact upon other USGS programs iffinancial support is provided.

The prime reason for this workshop was a perceivedneed to assess the future of regional networks in termsof federal agency funding constraints in late 1981. Theparticular trigger was the threatened budget cuts in the

fall of 1981 that could have decimated this vital database. Important decisions are being made now on the1983-1984 network funds. This report marks the firsttime that the activities and problems of regional networkshave been examined inclusively. The basic background

data were obtained from a thorough questionnaire used bythe committee to obtain data and opinions from theresponsible operators of some 45 regional networks. The

collected information helped in the assessment of theproblems and the formation of the recommendations thatfollow.

ASSESSMENT OF PROBLEMS

Given that there are more than 1600 U.S. seismographic

stations grouped functionally and operationally into50-odd independently managed networks, with funding

20

coming from about 10 different federal agencies, an equalnumber of states, some cities, utilities, and even privateuniversities, it is not surprising that there are problemsin regional network seismology. Mixed with this basicdiversity of purpose and support are the particular goalsof the operators, usually established research scientists

with their own perception of purpose for the network theymanage.

The committee, in considering agency questions andoperators' comments, has defined a number of problems

facing the regional networks. They fall into threeseparable categories: functional definition, fundingdifficulties, and operational problems. A functional

definition, that is, a clear statement of network goalsand a realistic estimate of its planned lifetime, must

always be provided. Funding difficulties are of twotypes: a lack of stability on a year-to-year basis and

the vulnerability of research funding being decreased tokeep networks operating in times of fiscal stress whenresearch funding is mixed with basic operational costs ofthe network. Operational problems are seen in a lack of

coordination among networks, the need for a morestandardized data base management system, and a growing

obsolescence of network equipment. These problems areinterrelated and difficult to order in importance. Arobust funding environment would go far toward solvingmost of them. Unfortunately, such is not the presentsituation. We consider these problems now in turn.

A significant part of the overall problems of regionalnetworks is the difficulty of designating a realisticdesign lifetime of a particular network. It is a well-recognized fact that some networks are operated toprovide a specific data base and that when that data basehas been collected the network is expendable to thesupporting agency.

There are a number of support agencies that aremission-oriented, serving either a regulatory or aservice function for the government as contrasted to aresearch function. In addition, there are agencies thatrepresent a gradation between functions. The USNRC andthe U.S. Army Corps of Engineers might be taken asexamples of the former, whereas the USGS is an example ofan agency in the middle ground, serving both a serviceand a research function.

The particular mission of the support agency has adefinite controlling influence on the design lifetime ofa network. While it may seem appropriate to operate a

21

small regional network in the immediate vicinity of aproposed critical facility to obtain seismological datafor design purposes, any such window of data must berecognized as a very short-time sample of naturalphenomena that occur infrequently. Furthermore, thefacility itself may alter the seismological character-istics of the site, as in the impoundment of a largereservoir.

Some regional networks such as those in Californiasupported by the USGS or those in the Northeast andSoutheast supported by the USNRC are intended to provideseismological data over a very broad region--notspecifically for a particular site. Such regionalnetworks may require a minimum of 15 to 20 years toobtain a representative sampling of seismicity.

Clearly, an initial understanding of design lifetimeis important, as well as a regular evaluation of theinitial plan in light of the results of operating thenetwork.

Perhaps the most significant problem perceived bynetworks is the lack of a rational approach to stablenetwork funding. Many operators attribute this situationto a lack of realization at high levels in supportagencies that many of the applied and basic researchproblems addressed by regional networks require long andcontinuous data bases. Operators see vagaries in fundingas a result of this situation.

Separation of the research activities from theoperation of regional networks appeared as a common themein the non-USGS operators' responses to the questionnaireand leads to our identifying the following problem: forfunding purposes the operation of a network (all stepsthrough bulletin preparation) is often lumped togetherwith research (the scientific analyses of the networkdata). In times of financial stringencies, we thus seeelimination of research funds, since operational costsare fixed and subject to inflation. When the scientistin charge can cut only research funds in combinedbudgets, he assumes more and more the role of atechnician providing services. Research and networkoperation are usually not evaluated by differentstandards, but they should be. For the former thestandard is scientific merit, whereas for the latter itshould be stability, quality, and service.

A related but specialized difficulty is seen inmaintaining a balance of support for operations andnetwork development between funding agency and funded

22

operator when there is a joint responsibility for thatnetwork.

Lack of coordination appears as a generic problem that

emerges in one fashion or another in the majority of theproblems that confront the regional seismographicnetworks. Effects of a lack of coordination show up inthe areas of data exchange, network boundaries, softwarecompatability, data archives, and duplication of effortsin a number of technical developments, including bothsoftware and hardware. The net result has been to reducethe effectiveness of the regional network operations as awhole.

A class of problems involves the lack of standardizedmethods for data base management. This problem involvesa number of elements, including data management, datacenters, software portability, and manpower usage.Network operators customarily face two conflicting

objectives. On the one hand, they contribute to anarchive of data that can be used long-term, together withother geophysical data (perhaps including other seismicnetworks) for an overall synthesis. On the other hand,the scientist-operator must undertake (usually on a moreshort-term basis) scientific research objectives specificto the immediate field area, and drawing heavily on thedata set.

There exists a need to achieve a standard set of

regional seismographic network data that can be easilyaccessed by general users for both service and researchpurposes. We anticipate that the archive that is suppliedby the aggregate of networks will constitute a majorscientific resource for understanding seismicity patternson a regional and national scale.

With respect to present scientific and technologiccapabilities, regional networks suffer from obsoleteequipment. The existing regional networks of seismicstations, including more than 1600 stations, almostuniversally employ short-period, vertical seismometers.Most of the signals from these instruments are trans-mitted by narrow bandwidth, low-dynamic range FMtelemetry that was first employed more than 20 yearsago. The seismograms obtained from these networks haveproven to be an economical and effective scientific toolfor solving problems of earthquake location and, to alimited degree, for defining their source mechanisms.However, these data are inadequate for analysis by manypowerful seismological methods developed during the pastdecade, and the data consequently do not provide

23

critically needed new information on earthquake sourceproperties and the structure of the earth's crust. Theproblem is particularly acute in the regional networkslocated outside of California, where network stationspacing is large, and in areas where there are fewhigh-quality instruments of other types operating withinthe short-period vertical networks. The lack of higher-quality data not only affects the basic researchcapabilities of the networks, but also may compromisetheir potential for answering mission-oriented questionsposed by funding agencies. Within California, forexample, it may be argued that present coverage, withhundreds of obsolete short-period stations, could bereplaced to economic and scientific advantage by fewermodern instruments.

RECOMMENDATIONS

There are no easy answers to the problems summarizedabove. It is probably not possible to assign rationallya priori a design lifetime to any but the most local,site-specific networks. On the other hand, tiaitialguidelines must be set by funding agencies when a limitedduration is planned. Problems of regional seismicitywill not be "solved' in a three- or a five-year period.Rather, it will normally require lifetimes of decades toobtain a representative sampling of the seismicity.Nevertheless, networks of planned, limited lifetimeshould be reviewed on a three- to five-year basis so thatthe state of knowledge gained by operation can be weighedagainst the network purpose.

There exists a disturbing range of quality amongregional network operations. Different networks currentlyoperate to quite different standards. While someoperators can produce on demand a bulletin updated towithin a few days prior to a request for information,others have essentially no bulletin. Some produceseismicity maps with accurate information on magnitudesand focal mechanisms, but others have been operated withignorance of instrument gain or polarities. The periodicreview should involve the supporting agency or agenciesand address a network's intended purpose and performance,considering all specific mission-oriented goals. Anetwork should also be reviewed with respect to alter-native approaches that implement new technology toachieve equal or better performance at comparable or

24

lower cost. We believe also that all regional networks,regardless of size or funding source, should be reviewedby the same review panel.

There is, of course, no way to guarantee stable funding

of regional networks. However, an irrational slashing ofnetworks can be alleviated by an increased understandingof the broader role of regional networks by the fundingagencies and by increased coordination of networkoperations at the national and local levels. Further-more, the above review procedure can result in a firmcommitment to support (or terminate) until the next three-to five- year review. Agencies have the responsibilityto decide when a network has satisfied their needs and toarrange for an orderly closure. Since networks ofteninvolve other users (who probably are not sharing insupport), these Oshadow" users should be informed earlyof decisions to terminate. Exhibit 3 in Appendix B givesexamples of the wide variety of such users for fournetworks. The mission-oriented support agencies shouldalso recognize and carefully evaluate their role andresponsibility to support a national network independentof their current needs.

We recommend that the network funding agencies admitto a realization that for most operators the main purposeof a regional network is to provide data fundamental toresearch on seismotectonic processes and earthquakeoccurrence in the region. This simple acknowledgmentwill allow open and rational discussion of the separatecosts of routine operation versus research. A good faitheffort must be made to remove the serious vulnerabilityof research support in the packaged funding practice. Itwill go far in eliminating different opinions betweenoperators and funding agencies on true operational costs.

We believe that it is time to emphasize the archivalfunction of regional networks. Minimum desired archiveddata include earthquake summaries, phase lists, anddigital seismograms at some minimu magnitude cutoff.Network specifications, including station parameters andresponse characteristics, are necessary. This functioncan be used as the major distinguishing featureseparating what we call a network from a portable array.

Of the 45 U.S. regional networks responding to thequestionnaire, there are 11 that operate dedicatedcomputer-based recording centers. Many of these centerswere established with direct support from the USGS.These regional centers provide the beginning of a skeletonframework for a recommended system of U.S. Regional

25

Network Data Centers that should operate with standardizeddata formats that provide easily transportable data and

that are well documented. A specific effort should bemade to provide easy access via telephone terminal orequivalent to the developing Regional Network Data

Centers.Coordination is not easy, but the above initiatives

will in effect force a semblance of coordination andcooperation onto the regional network scene. We con-templated a recommendation for a regional coordinator whowould work toward a greater integration of operations in

a given area. It seems best, however, to hope than anevolution toward Regional Network Data Centers willprovide a natural focus for coordination among operators,users, and support agencies.

Seismologists throughout the United States are workingon a variety of theoretical and operational software--ata great investment of manpower and money. Through an

effective coordination among networks, duplication can beminimized and significant developments can be madeavailable to the rest of the community on a timely basis.

Throughout the history of seismology in the United

States, technical, theoretical, and observationalproblems have been tackled by seismologists. With therapid expansion in seismological efforts and withadvances in modern electronics, computation capabilities,numerical modeling, etc., it is cost effective to useexperts in these fields to help solve seismologicalproblems, and to disseminate the solutions, as a means ofallowing seismologists time to pursue their own research.

There exists within the broad seismological communityconsiderable expertise and experience in data collectionsystems other than those in common network usage. Suchsystems employ broadband sensors and wide dynamic range

digital telemetry and recording. The opportunitiespresented by this technology are in many respects com-parable to those apparent in 1960 when specificationswere set for the worldwide standard seismograph network

first proposed in 1959 by the Berkner panel. Severalgroups are proposing new generations of instruments, forvarious applications. It is important that input comesalso from the regional networks on standardization and

the type of recording and data collection systems to beused.

It is recommended that a standing working group be

established by the Committee on Seismology (the samegroup that represents the seismological community on the

26

problem of a national network--a problem that is closelyinterrelated) to evaluate continually the overall healthand status of regional networks. This working groupwould provide all the review functions recommended forregional networks in this section. It would work closelywith the network operators and with agencies that fundthe network and provide advice on related matters. Mostsuch agencies were represented at this workshop, andtheir representatives indicated a need for ongoing reviewof the type recommended here.

5

NATIONAL NETWORKS

REVIEW

The rationale for and a proper design of a state-of-the-art national digital seismic network were the subjectof a comprehensive report prepared by the Panel onNational, Regional and Local Seismograph Networks of theCommittee on Seismology entitled U.S. EarthquakeObservatories: Recommendations for a New NationalNetwork (NRC, 1980). The major conclusions and overallrecommendations from that report are included as AppendixC. Many of the recommendations set out in this reportparallel closely those in the 1980 report. This chapterconsiders those 1980 recommendations in the context ofglobal and regional networks, and in light of the currentfunding picture.

U.S. seismographic networks were considered on anational basis, and it was recommended that a concept beadopted of an integrated U.S. Seismograph System (USSS)so that the effect of damaging earthquakes on long-termnational economic and security matters could be properlyassessed and ameliorated. This application representsonly a part of the overall role of the USSS, which wouldbe a basic research tool in a variety of seismologicalinvestigations such as the detection and location ofseismic events, studies of earth structure and processes,and site evaluations for critical facilities. The USSSis perceived to be a national and integrated systemconsisting of

1. a national digital seismograph network2. regional networks3. data archiving and dissemination functions4. management function including (a) a working group

27

28

guiding the development of the USSS and (b) a workinggroup on instrumentation and data handling

5. the National Earthquake Information Service (NEIS)

A national digital seismograph network would consistof broadband, high-dynamic-range, three-component seis-mographs at a few tens of sites in the United States. Itwould serve as a stable interface between global seis-mographic networks and regional networks. It wouldprovide a high-quality long-term data resource requiredfor scientific and engineering purposes, giving uniformnational coverage for earthquakes of magnitude 3.5 andgreater.

The availability of high-quality, three-componentbroadband, high-dynamic- range digital data will allowapplication of sophisticated, but proven, analyticalmethods to extract new information on earthquake sourceparameters, and on path properties and geometries thatheretofore were not available in the lower-quality data.

At present, elements of a national digital networkexist, but they are not integrated. A mix of stationscould be specified as selected units from

1. national digital WWSSN2. elements of regional networks3. RSTN4. national SRO/ASRO5. U.S. Telemetered Seismic Network of 1980, which

serves the NEIS early location responsibilities by drawingon selected short- and long-period data generated byregional networks

6. National Tsunami Warning Network

Clearly, coordination and some standardization of selectedelements from these networks could constitute a nationaldigital network.

ASSESSMENT OF PROBLEMS

To date there has been no commitment to or funding forthe USSS concept. In the past, the seismologicalcommunity has been slow to recognize the value of acoordinated and integrated national approach to somescientific as well as practical problems. Operators ofthe more mature regional networks stress that some oftheir problems could be alleviated by coordination andintegration on a national scale. It is likely that costs

29

would be significantly cut with an integrated system.Because of the low priority given to USSS by most of theseismological community, and because the funds recommendedfor this purpose do not seem to be obtainable in the nearfuture, the federal government has not yet seen fit toendorse the total concept in the form of a commitment byany agency. It is important for the community to perceivethat major elements of a national digital network existin fact, and that evolution of USSS can be controlled ina manner designed to satisfy national needs andpriorities.

Many of the present regional networks, however, areproducts of parochial applied research objectives. Thereis currently little coordination at a national scale.Instrumentation, data handling functions, data processing,data exchange, and data archiving are thus not standard-ized. This state of affairs is not conducive to thedevelopment of long-term national integration of stationsinto USSS.

RECOMMENDATIONS

The committee endorses the adoption of the concept of anintegrated U.S. Seismograph System.

It is recommended that the Working Group on SeismicNetworks to be established by the Committee on Seismologybe charged also with guiding the development of the newintegrated USSS and a move toward a national digitalseismograph network. The nature of the working group isdiscussed in the introduction to this report. Thisworking group should be the same group that representsthe seismological community on the overall problems ofseismograph networks, augmented with specialists fromtime to time as needed for problems specific to a nationalnetwork. Members and specialists would include networkdata users, instrument specialists, computer specialists,academic network operators, an operator of an USNRC-funded network, and a USGS network operator. All membersshould be qualified, experienced seismologists, and,calling upon specialists as needed, they will recommendstandards and options in the national network contextrelated to the following topics:

" instrumentation* network optimization* data processing and products• data management and archiving* data dissemination and centers for regional data

REFERENCE S

National Research Council (1960) Specifications for aWorld-Wide Network of Standardized Seismographs,Committee on Siesmological Stations. Washington,D.C.: National Academy of Sciences.

National Research Council (1977a) Global EarthquakeMonitoring: Its Uses, Potentials, and SupportRequirements, Panel on Seismograph Networks, Committeeon Siesmology. Washington, D.C.: National Academy ofSciences.

National Research Council (1977b) Trends andOpportunities in Seismology, Committee on Seismology,Washington, D.C.: National Academy of Sciences.

National Research Council (1980) U.S. EarthquakeObservatories: Recommendations for a New NationalNetwork, Panel on National, Regional and LocalSeismograph Networks, Committee on Seismology.Washington, D.C.: National Academy of Sciences.

National Research Council (in preparation) Data Problemsin Seismology, Panel on Data Problems in Seismology,Committee on Seismology.

National Science Foundation (1976) Earthquake Predictionand Hazard Mitigation: Options for USGS and NSFPrograms., NSF Publication 76-49. Washington, D.C.

Oliver, J., and L. Murphy (1971) WWSSN: Seismology'sglobal network at observing stations, Science174:254-261.

Panel on Seismic Improvement, L.V. Berkner, Chairman(1959) The Need for Fundamental Research inSeismology, panel established by President Eisenhower'sSpecial Assistant for Science and Technology,Departi ent of State, Washington, D.C. 214 pp.

30

APPENDIX A

GLOBAL NETWORK DATA

A-i Map showing distribution of WWSSN stations.

A-2 Map showing distribution of GDSN, IDA, andRSTN stations.

A-3 WWSSN data services.

A-4 Objectives and funding options in EarthquakePrediction and Hazard Mitigation: Options forUSGS and NSF Programs (NSF 1976), under theGlobal Seismology Sub-Element of theFundamental Earthquake Studies Element.

A-5 Funding history for Global Seismology Branch,USGS. Total funds include, in addition todirect program funds, funding received fromother agencies and program elements. FY 1983numbers are projected.

A-6 Summary of global network options and possibleconsequences.

A-7 Statement in support of analog WWSSN data.

31

32

;"'i ( .' i.--2

, : .i t J i , I t <

If-- " ---act)

i7 ix

C, C

00 0 ..

i U' -- x--: 8 " i 4.-4

it~~4

A *isC

~-'~ ~ b5~@o~, S

Vj .0E4

33

*E--

K' 04

p. 41

*7 0 ,-

34

EXHIBIT A-3

WWSSN data services.

An essential part of the WWSSN from its inception was themicrofilming of the original records and the provision ofhigh-quality film copies. Since its beginning in 1961,more than 5 million original records have been copied and60 million copies supplied to users. Currently, thereare several hundred requests per year for seismogrammicrofilm. The network had an intended size of 125stations and still operates with about 110 stations.Originally, the seismograms were filmed on speciallydesigned 70-mm panoramic cameras at 8X reduction. In1978 filming was changed to put 24 images (4 days ofnormal operation) on a single 105-mm microfiche at 32Xreduction. Standing orders of the whole network havebeen provided to eight institutions (Lamont-DohertyGeophysical Observatory; Institute of GeologicalSciences, Edinburgh, U.K.; University of Tokyo;California Institute of Technology; MassachusettsInstitute of Technology; USGS/Menlo Park; USGS/Golden;NGSDC/CIRES), and substantial parts of the network havebeen supplied to five institutions (University ofTexas/Galveston; Cornell University; University of Otago(New England); Los Alamos National Laboratory;USGS/Albuquerque).

The network is augmented by copies of the visiblerecords from the HGLP (1), ASRO (4), and SRO (12)networks, from the Canadian network on 35-mm film since1966, from the People's Republic of China 17-stationnational network on 70-mm since 1980, and for large-magnitude or seismologically important earthquakes fromseveral hundred additional stations including those ofthe USSR under the International Data Exchange.

The system is operating primarily with contract laborand with about 8 weeks being required for the cycle fromreceipt of original records to supplying copies tousers. Fifty percent of the network data is generallyavailable for distribution within 8 months after therecording interval. In general the archival film copy ismade at NOAA expense with the cost of copy being borne bythe user. Present costs to users are $0.80 per fiche,but this will undoubtedly increase as contract costs rise.

35

EXHIBIT A-4

Objectives and funding options in Earthquake Predictionand Hazard Mitigation: Options for USGS and NSF Programs(NSF, 1976), under the Global Seismology Sub-Element of

the Fundamental Earthquake Studies Element.

Global Seismology--Collect and disseminate seismologicaldata from around the world.

1. Operate the World-wide Standardized SeismographNetwork (WWSSN) and reestablish a maintenance program forthe stations that lapsed several years ago.

2. Operate the data acquisition and processing

capability of the National Earthquake InformationService, including use of satellite telecommunications,issuance of new seismicity maps, and routine computationof the parameters of the earthquake mechanism.

3. Upgrade about half of the WWSSN and establish thecapability to produce integrated tapes of digital seismicdata.

4. Acquire and operate a ten-station array oftransportable broadband seismographs for global seismicstudies.

5. Operate an integrated digital network consistingof high-gain long-period stations, Seismological ResearchObservatories, and the upgraded WWSSN stations called forin activity 3, and produce integrated tapes of digitalseismic data.

6. Acquire, install, and operate 10 ocean-bottom

seismographs.

Present and Proposed Funding OptionsElement: 1. Fundamental Earthquake Studies

Option A will allow a stable, minimally sufficient,

operation of the WWSSN and operation of the dataacquisition and processing capability of the NationalEarthquake Information Service (NEIS) in FY 1978-1980, avery limited start in upgrading a few of the WWSSNstations in FY 1979, and the incorporation of the

36

FY 76 FY 77 FY 1978 FY 1979 FY 1980

Sub Element Act. Req. A B C A B C A B C

a. The Earth-quakeProcess NSF 1.1 1.6 2.3 2.6 3.4 2.6 3.0 3.6 3.0 3.3 3.8

b. The Impli-cations ofPlateTectonicsfor Earth-quake HazardReductionNSF 1.5 1.9 2.4 2.7 4.1 2.7 3.0 4.0 3.0 3.3 3.9

c. GlobalSeismologyUSGS 1.9 1.7 2.3 2.6 3.2 2.5 3.1 3.6 3.4 3.6 4.0

TOTAL 4.5 5.2 7.0 7.9 10.7 7.8 9.1 11.2 9.4 10.2 11.7

(Amounts are in millions of dollars)

existing high-gain long-period stations and Seismic

Research Observatories into an expanded WWSSN in FY 1980.Option B will allow a partial reestablishment of the

maintenance program that lapsed several years ago and the

upgrading of about half of the WWSSN stations to produce

integrated types of digital seismic data by the end of FY1980.

Option C allows the acquisition and operation of a10-station array of broadband seismographs for globalseismic studies.

37

C>C

0O 0'C,

en en CC- co -

- 4 0' 0

U, cnc co

LA G - .1.)

0 Lt- rCO4

A CE e

41(3 - u4'0

1.1 uJ4 rn.J G

oo 4

0 a0

' 04 41 , C-1

' ~ ~ i 0' > 4j041

E-40.4

0' _

AHG

to 4

38

EXHIBIT A-6

Summary of global network optionsand possible consequences.

The following options are those considered by the USGS inits deliberations for the FY 1983 budget, and offered tothe Committee on Seismology for its comments andrecommendations on behalf of the scientific community.They are an excellent example of the options that havebeen considered in recent years and that have given riseto this workshop:

1. Terminate all USGS support for the WWSSN. It maybe assumed with certainty that the network will cease tofunction if the USGS withdraws support. Most of theforeign stations are located in developing or under-developed countries that do not have the hard currenciesneeded to purchase operating supplies even though theannual costs of supplies are small. Even more important,all of the stations, whether they have adequate fundingor not, depend upon the USGS Albuquerque SeismologicalLaboratory as the only source of replacement parts andcomponents, which are no longer manufactured by privateindustry. Thus a decision to withdraw support would meanthe demise of the WWSSN within 1 or 2 years. Of course,the data exchange, which is the purpose of the entireprogram, would end abruptly.

2. Seek support for the WWSSN from the stationsand/or foreign governments. Based on past experience, weare not optimistic that such an appeal would produceresults of substance. Many WWSSN stations continue tooperate only because of traditional obligations to globalseismic data exchange engendered by the network. However,it is doubtful that this goodwill will extend to sup-porting with their internal funds what in many cases isconsidered obsolete equipment in comparison with other,more modern stations serving their national needs. Inthis competition for funds, WWSSN stations are bound todeteriorate and finally terminate operations. Anotherconcern is the apparent incongruity between the con-.derable funds being expended on the Digital WWSSNupgrade and other new programs and a plea on our part forseveral thousand dollars in support funds from a hostcountry. Such a plea might also send out unintentionalsignals that the WWSSN has lost or is losing itsimportance to the worldwide seismological community.

39

3. Terminate photographic supplies to all U.S. WWSSNstations. Termination of supplies to 21 U.S. stationswould result in an annual saving of about $70,000.Although international obligations will be met, ournational needs will not, since many U.S. stations will beforced to terminate operations.

4. Reduce trace spacing from 10 mm to 5 mm on allWWSSN long-period components. The reduction inrequirements for photographic paper would result in anannual saving of about $75,000. Present fiche format isarranged to line up all six components for each day inone column for the ease of users. There would be a slightinconvenience to users in that in the new recording pitchlong-period data would appear only in the first and thirdcolumn. The degradation of data will in some instancesbe unacceptable to users.

5. Replace WWSSN photographic paper recording byrectilinear recording on heat sensitive paper. Thischange in the WWSSN recording medium would result in anannual saving of about $335,000. Operationally, thechange would result in less losses due to developingerrors and light intensity problems and in more uniformmicrofiche images. The total one-time cost forprocurement and installation of heated pen assemblieswould be about $1,400,000.

6. Eliminate WWSSN short-period horizontal recordings.The reduction in requirements for photographic paper wouldresult in an annual saving of about $100,000. Loss ofshort-period horizontal data would seriously curtailcurrent studies of the earth's anisotropic properties andof the regional discrimination problem.

7. Terminate all USGS support for the GDSN. A capitalinvestment of more than $10,000,000 would be lost. Anopportunity to establish a resource of great potentialwill be irrevocably lost to seismology for many years inthe foreseeable future. Principal advantages of GDSNdata are the wide dynamic range, bandwidth, resolution,and the ease and speed with which large amounts of datacan be processed. Without these data it will not bepossible quickly to test and verify recent and futureadvances in theoretical seismology by comparing syntheticwaveforms and spectra against large volumes of high-quality digital data. The routine use of many digitalprocessing techniques, until recently impeded by the lackof resolution in analog recording, will not be possiblewithout digitally recorded data.

40

8. Support a Global Seismograph Network (GSN) thatcombines the GDSN and WWSSN stations. Given this choice,three options are open to the USGS: (a) reduced supportfor the WWSSN in favor of the digital stations, (b)reduced support of the digital stations in favor of theanalog stations, and (c) redirection of funds from otherelements of the Earthquake Hazards Reduction Programwithin the USGS to the support of the global networks.

41

EXHIBIT A-7

Statement in support of analog WWSSN data.

The WWSSN has been by far the most productive general-purpose network of seismograph stations ever operated.The instruments now consist of moving coil pendulumscoupled with recording galvanometers. Free periods are 1and 0.75 s, respectively, for the short-period instru-ments and 15 and 100 s for the long-period. However, thecost of photographic paper to record three components ofshort-period and three components of long-period motionat some 100 stations amounts to about $300,000 annually.This expense can reasonably be questioned since the under-lying technology here is about 25 years old. Thus we seepressures for the option of reducing the analog WWSSNsystem.

Let us examine the underlying WWSSN issues in threeparts, the first two being amenable to scientificdiscussion and the last, more nebulous.

Analog or Digital Recording?

Digital recording will supplant analog. The presentadvantages of analog have largely to do with merits thatcan be maintained as the transfer to digital recording isaccomplished. Thus digital now offers nothing like filmchip distribution nor the archiving that is so easilydone with analog. This last point is very serious. Wehave the example of LASA digital data--of which verylittle remain. In the year 2000 or beyond, let us imaginethat we wish to look back on the previous decades andapply new theories as we study some of the seismicactivity preceding a great Alaska earthquake that mighthave occurred in, say, 1990. The archive will beessential.

We should not underestimate the present advantages ofwell-written paper records. As one looks over the sheet,there is an enormous amount of detail in a compressedformat contributing to a sense of where one is, withrespect to noise levels, across a quite wide frequencyband (for WWSSN long periods) and a substantial dynamicrange. The trained eye can absorb this informationrapidly. Although accepting the merits of properlywritten digital equivalents (absence of overlapping

42

traces, lack of human error in digitizing, suitabilityfor data analysis), there is a considerable loss in goingfrom analog to digital facsimilies of LP WWSSN at onesample per second.

Global Coverage

There are now about 100 more stations in the WWSSN than

in the GDSN. Many seismologists claim that without thatextent of global coverage, we would go back into the darkages of seismology. This point of view applies not onlyto those studying tectonics and regional problems, butalso to those studying the upper 700 km of the mantle.The many structures proposed differ from one anotherbecause of the earth's lateral variability. The problemsof inverting seismic data from too few stations have ahistory as long as that of the science itself.

For many scientific purposes, the main pressure shouldbe on expanding coverage rather than improving technology.Thus the many special problems associated with Alaska,and comparisons with the coverage by standard stations inCanada (about 34), indicate about six to eight moreAlaskan stations are necessary. Work is required toimprove South American stations. They sometimes do notsubmit data for days on which South American earthquakesoccur. Of course, island stations in the world's oceansare critical.

Work Habits of Analog Users, As Compared to Those ofDigital Users

This is the nonscientific part of the issue, but it must

be addressed because opinions are strongly held.In many cases, seismology is not applied as a *stand

alone" science. Rather, it has links to materialsscience, structural geology, gravity, heat flow, geo-magnetism, and remote sensing. Those scientists who wishto proceed on a broad front with data from variousdisciplines will typically not now make what theyperceive to be a heavy investment in digital hardware(tape drives, plotting devices, etc.), nor take the timeto master a digital facility if offered. This remarkobviously does not apply to those close to discrimination/verification problems, or to those in exploration/prospecting. In those fields, a digital revolution hasalready been accomplished and is appropriate.

43

For the broader view of seismology, consider a coupleof examples. A recent review of geophysical and geologicevidence strongly suggests the Makran region of Pakistanand Iran is now an active subduction zone and probablyhas been during most of the Cenozoic. Oceanic portionsof the Arabian plate currently subduct northward towardEurasia with a relative motion of about 5 cm/year. TheMakran region consists of a nearly complete trench-arcsystem; however, some of its tectonic features aresomewhat atypical. For example, an abundant supply ofsediments seems to lead to a shallow dip for the sub-ducted Arabian plate, and it does not permit a bathymetrictrench to develop; the accretionary prism is very wide,and a large part of it is subaerially exposed rather thanbeing submarine; only moderate seismicity occurs in theshallow-dipping thrust zone; at subcrustal depths thedipping seismic zone has a weak and sporadic expressionto depths of only 80 km and is not documented at largerdepths; the volcanic arc is poorly developed with largespacing (about 100 km) between its major Quaternaryvolcanic centers; the trench-volcano gap measures 500 +100 km, more than twice the width of a typical trench-volcano gap.

Despite these peculiarities, geologic, geophysical,and plate tectonic data suggest an active plate boundarywith ongoing subduction beneath the Makran region. It isa rapidly accreting continental margin, large portions ofwhich are still underlain by a mobile oceanic basement.

This work in Asian tectonics, which made major uses ofWWSSN data, would not in practice have reached the sameinsights if only the sparser GDSN data had been available.To say whether that is important or not is a valuejudgment.

Consider another example: In the proceedings volumeof the last Ewing Symposium--Earthquake Prediction, AnInternational Review (American Geophysical Union, 1981),edited by David W. Simpson and Paul G. Richards--aboutone quarter of the papers used data coming in large partfrom the WWSSN. Furthermore, except in a couple ofcases, the future research anticipated by these paperswould still use such a data base.

It is healthy that some seismologists are driven todiversity in the direction of other earth sciences, andsome to diversity in the direction of signal-processingand information theory. The present merits of film-chipdistribution are recognized by all seismologists, andsomething like this widespread distribution system for

44

visible records should be maintained even with a digitaldata base. This is stated in Global Earthquake Monitoring(NRC, 1977a, pages 31, 42) and in the report of the Panelon Data Problems in Seismology (NRC, in preparation).

APPENDIX B

REGIONAL NETWORKS: QUESTIONNAIRES, RESPONDENTS, SUMMARY

DATA BASE

In order to conduct informed discussion on the problemsfacing regional network seismology in the United States,a reasonably complete data base was required. Ourapproach was to find as many networks as possible,through a concentrated investigative effort, and tosurvey the operators for factual data on their networksand for their opinions on a number of questions.

Fifty-three more-or-less separate networks were found,distributed throughout the United States, includingHawaii and Alaska, along with four relatively permanentnetworks operated in foreign countries by U.S. inves-tigators. Two of the networks included were recently

shut down, and others soon may be, due to fundinglimitations. Network size ranged from 2 to 315 stations,with about half of them in the 10 to 40 station-sizerange. Exhibit B-1 of this appendix gives the listing ofthe networks found, along with the approximate number ofstations (some uncertainty creeps into this number due tojoint operations, network overlaps, etc.), and anindication of those networks for which we received aquestionnaire response. The 47 operating U.S. regionalnetworks involve 1631 separate stations, according to ourfigures.

The questionnaire used is reproduced as Exhibit B-2.It was formulated with the intent that it be relativelyeasy to complete (only 2 to 3 weeks were available beforethe workshop), and that it allow alternative ways ofproviding the basic network data and operators' opinions.Forty-five responses, in varying detail, were received.We view this near-unanimous response as a unifiedrecognition of a serious need to address the problems ofregional networks. The following section sumarizes thedata and opinions supplied by the 45 respondents.

45

QUESTIONNAIRE RESPONSE SUMMARY

Sections I, V

All but one respondent favored holding the workshop andthought that useful results were possible. Concern wasexpressed over the lack of wider representation fromoperators and over the overwhelming complexity of theproblems.

'Crucial questions" posed by respondents covered suchmatters as rising telemetry costs, the mixing ofoperational and research costs, need for stable policyand funding, precariousness of the northeast UnitedStates network support, wide variability of data quality,instrumentation improvements, coordination of hardwareand software development, the unique role of Alaska inU.S. seismology, future of global and national networks,an assessment of the value of current network practices,and the growing usage of seismologists as technicians.

Section II

If we exclude the Washington and Caltech cooperative USGSnetworks (WC) and the 10 other USGS networks (G10), theremaining 33 regional networks (33) are operated by theorganizations with the following average number ofpersonnel and total budgets devoted to seismologicalstudies:

Senior research personnel 3.9Graduate students 5.6Technical staff 3.0Total annual seismology budget $236,000

46

47

The $60,000 average total yearly support per senior

research scientist holds for all non-USGS institutions.

Section III

Here again we must separate the population according tothe W , G10, and 33 groupings, with the following results:

WC GIO 33Annual data acquisition

plus processingcosts per station $4,000 $5,000 $6,000

Ratio of acquisitioncosts to processingcosts 1.4 1.2 1.0

" All but three networks utilize telemetry.* Eleven networks operate on-line to dedicated

computer systems.0 For the (33) networks, operators estimate

that an average 35 percent of their institutions'seismological programs (in terms of both cost andscientific output) are supported by the network.

0 Most of the networks are felt to be ofindefinite lifetime, althoughseveral of the special-purpose networks were initiallyinstalled with definite 3- to 7-year expected durations.

* The question on nonpaying users brought awide response, with almost all networks involvedroutinely in providing data in some form to a variety ofindividual, industrial, and governmental users. ExhibitB-3 illustrates the diversity of users.

Section IV

A 60 percent majority opposed the idea of reducing their

network size for fewer higher-quality stations.The national digital network was favored by 33 and

opposed by 5 respondents, with a lack of strong feelingeither way.

Questions 23, 24, and 25 revealed a division of opinionbetween the Gl0 and 33 populations:

48

G10 33Yes No Yes No

Instrumentation adequate? 5 4 10 16Need to standardize? 4 5 19 6Separate operations and

research funding? 2 7 19 5

A 70 percent majority felt that the present mix of fundingagencies is satisfactory. 'Major problems facing networkoperations" were perceived as follows:

No. Responses Problems18 Funding, in general, at stable and

continuing level5 Vagaries of agency policy toward

regional networks5 Rising telemetry costs4 Funding to upgrade equipment1 Generally poor data quality1 How to cut operational costs

49

EXHIBIT B-1

Compilation of regional networks surveyed for workshop.

U.S. Regional Seismic Networks

Questionnaire Northeastern Seismographic Networks Approximate No.Returned (and operators) of Stations

Yes 1. New England Network--Boston College 38

Yes 2. New York Network--LDGO 38No 3. MIT Network--MIT 9

Yes 4. Penn. State Network--Penn. State 16No 5. Delaware Net--Delaware Geological Survey 3

Yes 6. SUNY at Stony Brook 2Subtotal 106

Southeastern Seismographic Networks

Yes 7. Virginia Network-VPI 31Yes 8. So-thern Appalachian Regional Network--

Tennessee Earthquake Information Center(Memphis State) 18

Yes 9. South Carolina Seismic Program--Universityof South Carolina and USGS Charleston Net 38

Yes 10. Central Georgia Net (Wallace Dam)--GeorgiaTech.

Yes 11. Northern Alabama Seismic Network--AlabamaGeological Survey and Georgia Tech. 12

Yes 12. Georgia Tech. Networks--Georgia Tech. 15Subtotal 19

Central United States Networks

Yes 13. New Madrid Network--St. Louis University 70

Yes 14. Kansas Network--Kansas Geological Survey 19No 15. Oklahoma Seismographic Network--Oklahoma

Geological Survey 10Yes 16. Ohio, Indiana, Michigan Network--University

of Michigan 14Yes 17. Central Minnesota Seismic Array--University

of Minnesota 6Yes 18. Nebraska Nemaha Ridge Seismic Net--Nebraska

Geological Survey 4Yes 19. Memphis Area Regional Seismographic Network--

Tennessee Earthquake Information Center(Memphis State) 13

Subtotal 136

Intermountain Seisamoramhic Networks

No 20. Montana Network--Montana Bureau of Mines 7Yes Yellowstone Network--USGS (discontinued

October 1961)No 21. Southeastern Idaho Network--DO and Bureau

of Reclamation 6Yes 22. Southern Intermountain Net--University of

Utah e1

50

Questionnaire Approximate No.Returned Intermountain Seismoqraphic Networks of Stations

Yes 23. Northern New Mexico Net--Los Alamos 24Yes 24. West Texas Network--University of Texas,

El Paso 6Yes Albuquerque Network--USGS (discontinued

October 1981)Yes 25. Socorro Network--USGS 12Yes 26. San Juan Basin Network--USGS 5Yes 27. Southern Nevada Network--USGS 53Yes 28. Nevada Network--University of Nevada, Reno 40No 29. Southeast Utah Network--Woodward-Clyde 23

Subtotal 259

Alaskan Networks

Yes 30. Southern Alaska Network--USGS 53Yes 31. Shumagin Islands--Lamont-Doherty

Geological Observatory 35Yes 32. Adak Network--CIRES 15Yes 33. Unilaska, Dutch Harbor Array--Lamont

Doherty Geological Observatory (one halfwill close summer 1982) 7

Yes 34. Central Alaska Network--University of Alaska 12Yes 35. Western Alaska Network (Seward Peninsula)

(will close summer 1982) 18Yes 36. Kodiak, Cook Inlet Network (will close summer

1982) 33Subtotal 173

West Coast Network;

Yes 37. Washington Network--University of Washington 132Yes 38. Oregon Network--State University of Oregon 6Yes 39. Cascade Network--USGS 48Yes 40. Berkeley Network--University of California,

Berkeley 18Yes 41. Central and Northern California Network--USGS 315Yes 42. Southern California Network--Caltech 34 ) com-

binedNo 43. Southern California Network--USGS 156 ) networkYes 44. Los Angeles Basin Network--University

of Southern California 43Yes 45. ANZA Network--University of California,

San Diego 17No 46. California Department of Water Resources 20Yes 47. Hawaiian Volcano Observatory Network--USGS 49

Subtotal 838

Selected Other Networks

Yes Caribbean Network--Lamont-DohertyGeological Observatory 25

Yes Resnor Network--University ofCalifornia, San Diego 14

Yes Nurek Reservation, USSR--LDGO 6Yes Toktogal Reservation, USSR--LDGO 6

45 Responses

mast Coast 225 stationsCntral United States 134Intermun~atain 259Alaskia 173West coast 836

Iotal 1631 stations In United States

~-ZL

52

EXHIBIT B-2

Operator's questionnaire: local/regional seismographic network.

I. The Workshop:I. Do you favor its being held? (Y) (N)

If no, why?

2. Are you comfortable with the proposed workshop agenda

and plan? (Y) (N)If not, suggest modifications:

II. Your Institution:3. Personnel involved in all seismological studies:

Number

a. Senior Research Scientists:b. Graduate Students:

c. Technical Staff:

4. Approximate total annual budget: $

III. Your Network: (Include station map and complete this section foreach clearly separate network.)

5. Number of Stations and Types:a. SP (Z only)b. SP (3-component)c. Broadband/LPd. (Other)e. Total No. Data Channels

6. Data Transmission:a. No. stations telemetered

b. Approximate total line miles

c. Percent radio telemetered

d. No. stations recorded locallye. Percent also telemetered

7. Data Acquisition/Recording: Sta. Cha.a. No. stations & channels on-line to computerb. No. stations & channels recorded analog mag tapec. No. stations & channels recorded digital mag tape

d. No. stations & channels event recorded onlye. No. stations & channels analog recorded only

8. Data Acquisition Costs (yearly):Full-time staff

a. Telemetry costsb. Out-station maintenance

c. Central station maintenance

9. Data Processing and Analysis Costs (yearly):

a. Dedicated full-time equivalent staffb. Personnel costsc. Supplies and expenses costs

10. Purpose of network as originally installed:

53

11. Is purpose still same? (Y) (N)If no, how changed?

12. Funding sources/annual support:

13. Was there initially a general agreement on the expected duration of

operation? (Y) (N)

14. If yes on 13:

a. How many years?

b. Basis for this lifetime:

15. If no on 13:

a. Do you have a proposed duration to accomplish present purpose?

b. If so, what?

c. What fraction of the net, if any, should remain indefinitely?

d. Explain b. and/or c.:

16. Dissemination of Results:

a. Do you generate a regular bulletin? (Y) (N)

b. Do you generate seismicity maps? (M) (N)(from network--not historical)If yes, send recent sample.

c. Have research papers been published in openliterature using network? (Y) (N)If yes, send list of papers 1979 to present.

d. What is your opinion as to the major scientific or engineeringresult(s) of your network operation and what do you hope to learnfrom its continued operation? (Use additional sheets asnecessary.)

17. Are there other (non-paying) users or agencies relying on yournetwork data? (Y) (N)

If yes, describe their usage and try to estimate the cost of thisservice.

54

18. What percentage of your institution's seismology program (excludeteaching) does this network support?

a. In terms of total cost:

b. In terms of scientific output:

IV. Some of your opinions:

19. Can you see any merit in reducing your network size for fewer higher-quality stations? (Y) (N)Explain (Y) or (N):

20. A national digital network:

a. DO you see value in it? (Y) (N)Comments:

b. Would you assign your best upgraded station(s) to such anetwork? (Y) (N)Explain no:

21. What do you see as the major problem facing your network operation?

22. Your ideas on a practical solution to this problem:

23. In general, do you think your instrumentation is adequate? (Y) (N)If no, explains

24. Do you see any need for an effort to standardize networkinstrumentation throughout the country? Explain.

25. Do you feel network operational and analysis costs should be fundedseparately from basic research? (Y) (N) Why?

26. Do you think that there are too many (or too few) agenciesresponsible for local/regional network operations and/or national andglobal networks?

55

27. Where do you think policy-making responsibility should rest forsetting directions and priorities for new developments in networkseismology?

V. The Workshop (again):

28. Do you think it can produce a useful analysis and recomaendations?(Y) (N)rf no, why?

29. Finally, please list your 2 or 3 *crucial questionso you would like

the committee to be certain to address:

a.

b.

C.

Many thanks for your time.

56

EXHIBIT B-3

Example users of regional network data.

Earthquake data, as recorded by a regional seismicnetwork, are usually made available to the public in theform of network bulletins, catalogs, and maps. Thesedata, together with the original seismograms, are used,usually by the collecting institution, as the basis forfundamental research on the nature and distribution ofregional earthquakes. Results of such research arepublished in scientific journals and reported at scien-tific meetings. This research is usually of a quitedirected nature, if funded by the agency or institutionsponsoring the network and is more fundamental if fundedby the National Science Foundation or similar agencies.

It is not often realized the extent to which some ofthese data are used by other than the sponsoring agencyand the scientists of the collecting institution. It isnot possible to list all of the users of all of thenetwork data. It is instructive, though, to discuss afew specific cases where such information is reported.

As a first case let us consider the well-known190-station Caltech network (as currently augmented by anumber of USGS stations). Network operators routinelycooperate with each other and with the National EarthquakeInformation Service (NEIS). In the case of the combinedSouthern California Network, data from the followingnetworks are used in routine locations:

USGS Southern CaliforniaUSGS Nevada Network (selected stations)USGS Walker Pass NetworkCaltech Network (telemetered stations only)California Department of Water ResourcesUniversity of California, BerkeleyUniversity of Southern CaliforniaUniversity of Nevada

Also, telemetered signals are sent out from the Caltech

net to the following groups:

California Department of Water Resources (and then toNEIS, Golden)University of California, San DiegoUniversity of California, Riverside

57

California State University, FullertonNaval Weapons Center, China LakePasadena City College

Groups on Caltech's routine emergency call list (called34 times last year) include:

California Office of Emergency ServicesCalifornia Department of Water ResourcesCalifornia Division of Mines and GeologyU.S. Bureau of ReclamationU.S. Army Corps of EngineersU.S. Geological Survey, Seismic Engineering BranchFederal Emergency Management AgencyFederal Power CommissionNational Earthquake Information Service (USGS)American Red CrossLos Angeles Department of Water and PowerPacific Tsunami Warning NetworkLos Angeles County Emergency Information Bureau

Companies and agencies that financially support theSeismological Laboratory as members of the CaltechEarthquake Research Affiliates, because of their interestin southern California earthquake problems, include:

Atchison, Topeka, and Santa Fe RailroadC. F. Braun and CompanyDames and Moore, Inc.Factory Mutual EngineeringERTEC, Inc.International Business Machines, Inc.Kinemetrics, Inc.ExxonLos Angeles Department of Water and PowerLockheed California CompanyMetropolitan Water District of Southern CaliforniaPacific Telephone and Telegraph CompanySan Diego Gas and Electric CompanySouthern California Edison CompanySouthern California Gas Company

Southern Pacific CompanyStandard Oil Company of CaliforniaUnion Pacific Railroad CompanyUnion Oil Company of CaliforniaWoodward-Clyde ConsultantsBank of America

4

58

Bulletins are sent monthly to 250 interested parties--particularly to engineering and geotechnical consultingfirms.

The 81-station Utah network, in the IntermountainRegion, lists their chief, nonpaying users as follows:

Utah Geological and Mineral Survey (for geologicalhazard studies)

Mining, utilities, construction, petroleum exploration,and miscellaneous engineering consulting companies(for specific site-evaluation studies)

State and county government offices (for planning andemergency services)

USGS/NEIS (for regional earthquake recording)U.S. Bureau of Reclamation (dam safety)U.S. Forest Service (geological hazard studies)U.S. Soil Conservation Service (dam safety)U.S. Army (geological hazards)

Lamont-Doherty Geological Observatory, in the East,reports that their network operators spend approximately10 percent of their time in providing data to engineersand private consultants who require information abouthazards, plus the news media and private citizens whoregularly contact them for information on earthquakehazards with the desire to be educated about earthquakesin general.

Virginia Polytechnic Institute and State Universitylists:

Virginia State Office of Emergency and Energy Services(for planning)

Virginia State Fire Marshal's OfficeThe Virginia Chamber of Commerce (for industrial

siting studies)

These few examples demonstrate that the users ofregional seismic data are many and varied, includingother networks and the NRIS; a plethora of federal,state, and local agencies for planning and emergencyservices! major utility companies; engineering andgeotechnical consulting firms; petroleum explorationcompanies; and many other public and private groupsinterested in seismicity and earthquake hazards.

APPENDIX C

SUMMARY AND MAJOR RECOMMENDATIONS OFU.S. EARTHQUAKE OBSERVATORIES:

RECOMMENDATIONS FOR A NEW NATIONAL NETWORK

The report on U.S. Earthquake Observatories is the firstattempt by the seismological community to rationalize andoptimize the distribution of earthquake observatoriesacross the United States. The main aim is to increasesignificantly our knowledge of earthquakes and theearth's dynamics by providing access to scientificallymore valuable data. Other objectives are to provide amore efficient and cost-effective system of recording anddistributing earthquake data and to make as uniform aspossible the recording of earthquakes in all states.

Many problems of major national importance related toearthquakes remain to be solved. Earthquake predictionand the amelioration of earthquake hazards, for example,require uniform, continuous, and standardized earthquakerecords over the entire country using modern computer-coupled instrumentation. We cannot anticipate all thescientific gains that will accrue from sharply improvingthe national capability to observe, measure, and studyearthquakes, but we can be reasonably sure of manysuccesses. Among the research goals are the quantitativestudy of sources of earthquakes above magnitude 3.0 up tothe greatest earthquakes in the entire United States, acapability never before possible; more reliable under-standing and prediction of strong ground shaking; the

precise definition of fine structure in the earth's crustand deep interior using high-resolution techniques; andthe close mapping of regional tectonics, related togeological hazards, and location of natural resources.In particular, there is a need to monitor and analyzequickly short-term stress variations in active faultzones for earthquake prediction purposes, a high nationalpriority.

59

t:

60

0

(44

Aj

Aj

C44

0

-4

40

44 x

61

Two recent developments make the present an appropriatetime to move ahead by redesigning and consolidating theuncoordinated mixture of local, regional, and nationalearthquake observatories, many of which are obsolescent.The first development is the new technology based ondigital sampling of signals. The second is the decisiveadvance in theoretical seismology, including powerfulcomputational ability, that has created a need forhigh-quality observations of seismic waves, with widedynamic ranges in both frequency and amplitude.

For the attainment of research and applied goals, dataanalysis, archiving, and retrieval capabilities in theUnited States need streamlining, partly centralizing, sothat digital tapes, seismograms, and the derived seis-micity data from all stations are available in a shorttime to all users.

In order to bring together these components, thecentral recommendation of the Panel is that the guidingconcept be established of a rationalized and integratedseismograph system consisting of regional seismographnetworks run for crucial regional research and monitoringpurposes in tandem with a carefully designed, butsparser, nationwide network of technologically advancedobservatories. Such a national system must be thought ofnot only in terms of instrumentation but equally in termsof data storage, computer processing, and recordavailability. Exhibit C-1 shows the suggested locationsof seismograph stations for the proposed National DigitalSeismograph Network (NDSN).

In order to take advantage of recent technological andtheoretical advances, the concept of an integrated UnitedStates Seismograph System (USSS) should be adopted in theUnited States so that enhanced information on earthquakesources, seismic hazards, ground motions, and earthstructure is available.

w mm m m m m

APPENDIX D

GLOSSARY

ANZA Local seismicity array located near Anza,

California

ASL Albuquerque Seismological LaboratoryASRO Abbreviated Seismic Research ObservatoryCIRES Cooperative Institute for Research in

Environmental SciencesCSS Center for Seismic StudiesDARPA Defense Advanced Research Projects AgencyDOD Department of DefenseDOE Department of EnergyDWWSSN Digital World-Wide Standardized Seismograph

NetworkGDSN Global Digital Seismograph NetworkGSN Global Seismograph NetworkHGLP High-Gain Long-PeriodIDA International Deployment of AccelerometersLASA Large Aperture Seismic ArrayLDGO Lamont-Doherty Geological ObservatoryMIT Massachusetts Institute of TechnologyNDSN National Digital Seismograph NetworkNEHRP National Earthquake Hazards Reduction ProgramNEIS National Earthquake Information ServiceNGDC National Geophysical Data CenterNOAA National Oceanic and Atmospheric AdministrationNRC National Research CouncilNSF National Science FoundationRSTN Regional Seismic Test NetworkSRO Seismic Research ObservatoriesSUNY State University of New YorkUSCGS U.S. Coast and Geodetic SurveyUSGS U.S. Geological SurveyUSNRC U.S. Nuclear Regulatory CommissionUSSS U.S. Seismograph SystemVPI Virginia Polytechnic Institute and State

UniversityWWSSN World-Wide Standardized Seismograph Network

62