AU AJ4 017 EFFECTIVE USE OF EARTHQUAKE DATA(U) NATIONAL RESEARCH I/|COUNCIL WASHINGTON D C CO ITTEE ON SEISMOLOGY

N ASS[IF I ED T V MCEVILLY ET AL. 1983 N00014-82-G-0016

N 110S110 FIG 8101 I*D*.**m***InIIIIIIIIIIIIhlSlllfffflllonIIIIIII

V 0.0

MCROCOPY RESOLUTOW TEST CHARTSMOSII.JA OF STANDAM -A

Ife ti1 L)s c

4k,4

I ~~TT,~~

NATIONAL RESEARCH COUNCILCOMMISSION ON PHYSICAL SCIENCES, MATHEMATICS, AND RESOURCES

2101 Comntiutiom Avenue Mshinton, D.C. 20418

BOARD ON ER T (1) 33W0cmasn.Tu ON MlOLOGY

October 24, 1983

Defense Technical InformationCenter

Office of Naval ResearchBuilding 5, Cameron StationAlexandria, VA 22314

Dear Sir:

The report Effective Use of Earthquake Data, twocopies of which are enclosed, was prepared under".Grant No. N00014-83-G-0016 In partial support ofCommittee on Seismology.

Sincerely,

&lcutive Secretary

Enclosures

to 4f. ad A! #wdui

*M 4biun -~6U

"Ag'Y,

Effctiv use of* Earthquake Data

Panel on Data Problems in SeismologyComnaitte op ScihaiologyBoard on Earth SciencesCOMMiuulon 00 Phyuical Sci9nce, MMadOBmi aMW

Ntion Rtesearch Council

DTV

O CT0098

NAniONAL ACADEWMY 0T69

This document has been approvedi .r public releaise and bale;- its

Lc ~~~ ~ ~ 3fiuini siie

~~J_ _ _ _ I

SInoZcu The project that is the subject of this reportwas approved by the Governing Board 6* the NationalResearch Council, whose members are drawn from thecouncils of the National Academy of Sciences, the

* National Academy of Bagineering, and the Institute ofMedicine. Te members of the comittee responsible forthe report Were chosen for their special caspetences andwith regard for appropriate balance.

This report has been reviewed by a group other thenthe authors according to pKgoedbas approved by a MogmtReview Comittee consisting of members of the NationalAcademy of Sce nces, the National Academy of, SAgIpeer Ingeand the institute of medlWiis.

The National Reearch Co.dicil was established by theNational Academy of Science#& in 1916 to associate thebroad commuity of science and technology with theAcademy's purposes of furthering knowledge and otadv ising the federal governmnt The Council operates inacoconc with general policies deteramed bl theAcakdemy 12e4r the authority of Its ongressional chaterof 1463, which eetablim Wi-% ecdemy as a pelvate#nomprofit, self-wgewralme ersI"ip costift. tbeC-2-11l has becme the princi~al operating wgenc of boththe Wetiomal Academy of Sciences and he Natinal Amabeyof Umifteer lA I the conduct of their ervie to thegwerremtt thes public, and the scientific endengineering - mities. it is administered Jointly byboth Academies and the Institute of ftedicine. ToeNatina hades of Vauibeeriag and the institute ofMedicine were, establimbed In IS" and 3970, respectively.

under the charter of do notiemams atei ofsciaoses.

Soi usa" Memssco mil

~iag . C. 00414

jib

Acession For

DTIC TAB ElUnac~nounccd 0lJustifigatio

cmxKIy= on SmzmOour

A,,a 1 a or

--C ALN OMI~ SafodUniversity ________DDO"ResseaerPolytedbuio Institute

a. 5 , t. LuisUniversityL-"HSDOKAWA , ClifoniaInstitute of 2eobn*Uqy

rRAMYN . Lage manProduction flesearob CaqpmyPAULW. QM S=kmtAseoiats

PAUL . RIHA~eLiamt-Daberty Gelogical Observatory,

DhVI W. nM=,Lamt-Dabetty Geolog tol. Observatory,

FARM23. WOLAu.ap U.S. Geological Survey. Meal* Park

Liaison Memers

LawN L. Dual",. U.S. nuclear magulatory CasiamWILLIAM J. SNW* Ait fteae Ofic cc iai.tti S hmeobMICEAL A. C3usa. Nationl Omn and At"Oherio

I dinistration

jao a. vuemK1 U.S. Geologic"l Survey3351 A. PL.~ NAtiewa Aerossttse md IN

j 1aym 3. 4 National Sises Vmiati1"ja AI. UMO1 .s.tam Orsuft Om Utsmert

p"el. am*nu , "acy Oa, bat asa ftwlm.t.7Mm ~ el 20016"1ona Ommsnie Agsmei

~~-4NdW'r A

At/

CARL F. RON=, fspiif Advanced Research Projects Agencymomh Knoi=, 6.epaztment of Energy

306313 1. slay, *at 'Aezomattiag and SpaceAdministrtiona

I. TIMUAMI, Natiohnal Science foeudation

Statf

306313 1. NE, JR.wY 3. uAME

I4

4.'

PANEL ON M!TA P0LM IN S3I8UMiOS

V1UGILBORT, UniVrsity Of California, San Diego,Chairman (until February 1963)

sml~TO s. ALUIAUZR Pennsylvania state University,ChaltMin (after Veobrary 1*93)

MILO. Me Din. University of Texas, AustinlAYNOU P. BULAND, U.S. Geological Survey. DenverRICESID?. LwaSS. Nasesaohustts institute of TechnologyKNIIBT L. LAMM, western Geogbysical Company of AmericaWILLIAN ff. Re LO33, U.S. Geological Survey. Menlo parkKAM C. MNLLY. University of California# Santa CruzMYVID W. SINIMCU. Lamont-Daberty Geological observatory.

PalisadesM. MIai TOSOS. Massachusetts institute of Technology

Liaison Numbers

1.5CM 1. SA=a.. 0.S. N10lea9 w*galatOy CommissionWILIM J. REST Alt ftre* Off ice of ScientifiAc UssearchVM~wZv w. SO= NatOUal Archives Md Mooed Service*.a. eNUSM. U06. G661001ioa Surveim a.e VUCMO U.S. Geologica SuveyPAM a.e vz . 90S. ArWmy C ar t oSMeera s

R3 . MMER vation.) bcomice so oMmsseiAiaistration

Lz . 2.u, .*. "tics aSew Me fntie3M. Nr. LMIS, Ofto al " mdeimobel

hdo"Itroug9,10ft a MAI # Go* asem 0

J ~~~ ~ ~ ~ no NOSMMO 011"o "" tse

WILLIAM a. MRAUI3U National Oceanic and AtmosphericAdminiatzatiou

K. !NIRNIALAIP National Sciwam Foundation

Staff

wy a. BAMeo

vLr

"k, IL

BOARD ON RAMS SCIENCES

4

WILLIAM R. DICKIINSE. University of Acisons, ChaltanSAWUEL S. ADANAr Adams and AssocLat e BoulderLLOYD S. CWUFF, Woodacd-Clyde Consultanter San franciucoWALTER R. ECEKUULM, Soh o Petroleum Corporation. HoustonMICHEL T. ALSOM, The Halbouty Center, HoustonWILLIAM W. BAY, University of ColoradoNELVIN J. HILL, Gulf Oil Corporation, HoustonCARUL AM HOGES, U.S. Geological Survey, Menlo ParkWILLIMA C. LTWE, Sandia National Laboratorles, ALbuquequeCURRIES J. NNE, Oklahoma Geological Survey, NormaV. DANA 10MET. University of MinnesotaJACK X. OL V. Cornell UniversitySTBPMN C. lMOER, University of WashingtonJ. WILLIAM SC30, University of California, Los AngelesB-AN M, U.S. Geological Survey, Reston

Liason Members

LEON L. BERATMAK U.A. NIlea Rgulatory CissieaNMIn MIN, National Science Foundation (until Sopteaer

30, 1962)PHILIP Co , U.S. GeologLoal Surveyur I In 9 1 V Defense a ing AgencyPaur I. rzwMEa. U.S. Army Cotpa of ngineersanwu a. iauuw. U.S. Geological euveyJm i. OB, Natio nal Soise l000iation 4freen October

1. 1962)JOlN 0. 1coK. Offes ot nawa feesatcaLIM M, i.t. geeapleal Surve6onm A. N&MVA, U.. "pwurmat d * r1U 11- . IM=* Natiemal s""* 9600itmMMS Lo. P=0K U.S. @seuvoesw uve

vii

* I i I _ . a . , *

"m ama Na.ntiona &ingmmui ms Space &dinistzationa0.amau W3181&! £.Gem 0.Z101 institute

KIII il N. maim, Nmrylua G4o0ogcal SnuTloPFAUK J. .S. Dupatasut of nwgy

Rx Officio

CIaXWCS a. win. Califogula institute of TeabmologyJCM C. CMIL. IALVeraty of 4Ca1focnia, Sauta Datbeaa

. awUS, sa0t *"a acio

I4,

AlAA

-*

CaMISSIOI on HYzCAL SCIZEcuSMATHIMATICSt AND ROUhh

UNFA.M PRMMNU*, national Research Council, ChairmanULEAM R. SLOWU Harvard Medil schoolWILLIAM WWR, Princeton University

EBNIAURD F. BURE, Massachusetts Institute of TechnologyENBE CHERNOM', Massachusetts Institute of Technology

MIL1D S. DNESSWIAUS, Massachusetts Institute ofTechnology

WALTER R. 3CUlIMAU, Shic Petroleum CopanyJOSSIM L. F1133, Office of the Governor, Comonvealth of

VirginiaJAMS C. nLUCK1R, University of PittsburghWILLIAM A. FOliSLR California institute of TechnologyGNI RU AU Brookhaven National LaboratoryEUBED A. FRIEW., Science Applications, Inc.EUBED D. GOiMS80, Scripps Institution of OceanographyCHARLS L. 3016XR, JR.* Pennsylvania State UniversityKoUmD B. KNMSOF. Stanford University

CHAX63S J. HAUMI*, Oklahoma Geological SurveyWAUAR H. MUIK, University of California, San DiegoGOIU 3. PAIS, X4ego feaaOh CenterR305 B. 8S , University of ColoradoBOUiD 3. sIlMUm, JR., 3.!. du Pont de Nemours a

Comawye Inc.JIN D. 313391.33. Harvard School of Public Health.Au S. yOma, JR., Carnegie Institution of Washington

gAmEN G. KASMPR xwutive Director

la

A*l ++ -

1* + r +

PRZFRCX

Seismic data from earthquakes, especially those recordeddigitally, have been accumulating rapidly in recentyears. At present, diverse types of high-quality dataare being generated under a wide range of program

supported by agencies such as the U.S. Geological Survey(USGS), the National Science Foundation (381), theDepartments of Defense and Energy (DOD, DOB), the U.S.Nuclear Regulatory Commission (NN)p and other state andfederal government and private Institutions. The recentreport of the National Research Council's Committee onseismology, Seisnmranhic Netorks Probles and Outlookfor the 1990o (national Academy Pres, Washington, D.C.,1983) provides Insight into cur rent data-acquisitionaspects of earthquake seismology and difficulties con-fronting those operations and gives recomndations onways to help eliminate problems in data collection.

It is clear that with the development of high dynamic.range, broadband digital seismic systems, the distinc-tion between strong-motion recording and sensitive high-gain seismic recording is disappearing. This means thatearthquake engineers and seismologists soo am share aceo seismic data base for their respective apoloo-

Stions. Therefore, the discussion in this report shouldbe taken to apply to all types of earthquake ground-motion data.

Digital data are opening exciting new ares ofresearch and applications that until now have not beenpossible even with the best analog data. To realisethese potential scientific braktromighs In seimalegyfully, these data must be ffectively dssmited to awide user gneity concernad with beof baui* reseaeand applications of seismic data. This report Isaddressed primarily to the problems of data management In

ZII

A I

I i=slo~became O55hocalsiu Iin the present data-handling and -distribution practices constitute the

greatest barrier to effective use of earthquake data of

fte ane onDataProlem InSeismology wasestablished by the Comittee on Seismlogy to reviewseismic data problems and make recommendat ions for theorganization, selection, storage, and distribution of thedata. It was specified that the study should includeboth existing and anticipated analog and digital datafrom global, national, regional, and local seismicnetworks and strong-mtion data. further, the panelwaasked to consider procedures and policies to ensure theavailability, timely distribution# and analysis of thesedata. (Mhe charge to the panel Is given in Appendix A.)

The results Of the Panel's findings presented In thisreport fcus On seismological data from earthquakes. ftegeneral discussion and conclusions should be relevant forseveral years, but details of the report reflect thecurrent situation in a rapidly changing era of datacollection and distributioni of advancing capabilitiesand availability of casputersi of varying data storagecapacities Of present systemer of increasing need forcomuters for solving casPlex analytical problemig ofIncreasing availability and use of digital as osprdwith analog datal and of perpetual uncertainties in otherfactors, schb an the yearly funding levels of federalagencies.

MWe data problems in seiology are of such keyImportance for achieving Potential scientific advancesand so changeable with time that continued vigilance will

* be needed to ensure that new developments in technologyare ispleMented in a timely manner, enabling United

* States scientists and engineers to stay at the forefrontof modern seismology.

te suppot of the Committee on SeiinOLog by thefollowing federal agencies Is appreciated: nationalOceanic and Atmospheric Adminiatrationp Nationa scienceFoundation, Division of Barth Sciences (AM) j U.S.Geological Survey a Office of Naval "eeahla AitrforceOffice of Scientific Research a Nationa scienceFoundation, Division of civil and anirouamtalUngineerims (M) r U.S. Nuclear regulatory CasmissiouiDefese Advanced Reseach projects Agency U.S.Departmient of saewly and federal amergemoy MsaaementAgency .

Shelton Be Alexander, MaimePanel on no"a problems In Seismology

Xii

3..szecutiv* Sumary 7

2. 1ntodticu 7

3. Current Statu,/ Froblea 10

(4.*3oi~dtin 22

5. Reforrs~cea and Xiblogrqahy 34

Appendix As Marge to the panel ca Data PrablesIn selawmg 36

Appendix A: Botigates Of Present and MuitreDigital Deta guatitiesj 37

Appendixc C: IM! yosolutio. ca IntermaticMalData 2uhhanve~ 41.

Appendix Ds befense Mvenod ]%search ProjectsIqeoy's Center tat Selumic Stuidiesj 43

A"Ond$.x as hit 1occ oWealol ApsIUlomCentects seiftotasl Activities) 46

Appendix Fs Proposed Puttiosa bqMiraM~ts forthe Ustimual Caster tog Ietamuoal,

Ste~l..)47

Appendix Gg AbbgeyIstims AM fttmia.1oM so5

xiii

-ago

After consider ing a wide variety of data problems inseismology, the Panel has identified as the primacychallenge in the imdiate future the develcgmmnt of acoordinated national effoct in the oolletie *torage,and dissemination of digital ear tbuh k data to aeethat our soat advanced tahology Is used effectively Inseismological research and engineering mplicatiems.Indeed many of the moat inpctent and challenging eels-mological studies of today equie both digital data fromstate-of-the-art instruments and ceqpuer facilitiescapable of analyzing large data sets and modelng theprocesses that explain the observations. Providingseismologists* with easy (remote) access to dosired eventinfamation and waveform data (lncluding real-timrecordings from global or national noetcks) at a NatimalCenter foe Seismological Studies should Increasescientific productivity significantly without aggreciablyincreasing overall data management and diesemiatcosts.

We foresee major advancs in seinology resulting fromthe increasing use of digital data. Hwny of tiesadvances will slmply follow from ref Isnmets in doingbetter what has already been done. e.g.. teprwdvelocity modeA, mce accurate hcoentec losations, aoecomplete earthquake catUlogs, and mace ,eftfient sesarigand sorting of data baes. ftere are many area ofresearch, bowever, where digital data are opening

*In this report -sismologst- refers Interchmgeby toboth scientists ad engineers and eoimollr includesthe full rage of "Lose oed * Ne lg studles8including strong motion.

A

.* *'

2

possibilities that have until now been beyond thelimitations of the beat analog datal among these aredetailed structure and heterogeneity of the larth'einterior from high-resolution .igenfrequency spectra offree oncillationar freqeny-epndn attenuation (Q)structure of the Earth,1 detailed dynamic models of theearthquake source from broadband waveform studies,real-time or near-real-time analysis of unclipped,broadband records from very large earthquakes recordedboth teleseismically and in the near fieldi and theroutine determination of moment- tensor representations ofearthquake sources.

If no steps are taken to address the problemsidentified In Chapter 3, the situation is likely todevelop that only the operatocs of seismograph stationswill have ready access to the data they record. Effectivedata dissemination to the seismological user cossuiitycurrently exists only foe the World Wide StandardisedSeismograph Hetwork (U63) anal" fiMM data, and thatservice, Is threatened by escalating costs. There is isoI comprehensive directory of infrme -tien that describeswhat data sets exist, who has them. ad how data can beobtained, Thaee Is no cm obeesisve national earthquakecatalog, and additional Phase data are not generallyavailable at all. Information Is fragmente and most beobtained from many different organizations. Distributionof digital data Is limited because of Current policiesand procedures-for dissemination, ombhined withsignificant direct or Indirect costs to users.,

These problem, coupled with the Defense AdvancedResearch Projects Agency's (mwA) recent implemmntationof a modern computational ospility for seismologicaldata storage, retrieval, and dissemination to theirresearch contractors, have, led the Panel to onclude tMatestablishing a MMUA L uals iMli~ data M be A

atiomal center ter gseoela tue it hoamdesirabl and-fible at this t*. A Utional Centerthat meets the famttafta reqmemte elaberated inAppendix V would overcome key daamnaeet and-dissemination problem and e~mesignificantly theavailability of high-qoality defta sets sad theireffective use by the entire seismological, inusuity.

Effective use of this national data base will teouirethe followingst

(a Upgrading the present daaMH 0ems system toprovide adequately for user needs.

dew-

lol,

3

(b) Develomnt of a omptehnsive ditectogy ofavailable data, data sourcea, 0palityl and foggets,together with a computer ised search capabilityimplemented at a National center aid regional. oenters.

(c) Development of a caupressivo catalog ofnational and global hypocenters, including associatedphase data for at least the larger events, together witha computerizes earch capability Implemented at aNational Center.

(d) Standardisation of data formats for the purposeOf data exchange.

(*) Ustablisbasent of a policy for data retention andarch iviag.

Those subjects are discussed in the text andappropriate recommeandat ions are given.

The Principal reosmndations of the Panel are amfollows&

global, national, selecte 14L aod stranon-tionOahg~kSWO Ms, Newmpebilities afforded br thiscenter would Include real-timne or near-real-tine accessto global and national network recordifts and rumoteterminal access to a wll-oeganixed seiamolegia" database. suggestions for the funding and management of anNOBS are given In this isport.

2. In calendar year lP)3 the me Committee onseismology should Initiate stae to coene a meeting ofrepresentatives of fsfding amnie, use gros, and theuniversity research community to discuss implimntationand funding of a National Center for SaimmozolmiaStudies using MARPA'S Canter for Seismic studie. (COO) asi ts prototype.

Other Important recomemdatioas; aire as follawas

s ar fe ibed i n kh tot alteenative teubtaigmefar6 4aeminting sisg data abuld be onsidered byUMA to kep awe Soots at a aftim.

4. Alterntive types of global Digital beimegrgh,botmock NW) digital data subsets Shuld he nudeavailable rem"WInul to "an In SPtina to the stndr

Ir -"Ift

g. _

4

5. Th e Ugional Data Cenbts shou continue toprovide Global Digital Seismic Uetimk data to visitingresearchers.

G. Data f ems regional and local networks shouldt continue to be olleotedt anaysed, mnd arcbved by

individual etwrk operators and sode available in astandard format to other users directly or via theNational Center or other appropriate regional datacenters.

7. Strong-motion digital data should be archived anddistributed by the National Center.

9. At least a subset of the Mir lcroe TechnicalApplication Center's Global Surface System (005) data forthe larger earthquakes should be made available forgener al use.

*9. Selected subsets of special data collectionscurrently at teledyne-Geotech's BoSismi" Data AnalysisCenter (SDC) in Alexandria, Virginia, should be keptindefinitely as part of an active, accessible nationalseismological data basesm these subsets should includedigital even t tapes of all ab 5.3 and greater.tealeseismic event reordings and all ab 3. and abovetregional aid local event rcordings.

10. A centralimed data directory should be developedand maintained as part of the national seimmUloical dataI base accessible to sers via computer terminalsr andglobal hypocenter catalogs should be augmented with acouprehensive national catalog derived from local andI regional networks and associated phase data be providedfor the larger events. These catalogs should beavailable from a national facility that provides co~utersearch capability to sers.

11. Rasearch granting and contrating agencies shouldplan allocatio of maroisaly 10 percent of awardedmonies for manageament and dissemination of data forstudies involving the acquisition of new data or analysisof existing data.

12. The national Research Council's Committee onSeismology should organise a wokso for the Purpose ofestablishing stanard data exchange forest* anidstandarisation oil event catalogs. For the presentr theInternatiOnal Association of Seism1logy and Physics 09the North's Interior C ZASIE) staid for internationalexchange of digital data should be used.

*gIs" Item need Imdiate actio

"n-

421,

13. Use Natioala Nesearb CMoUmei a Cmittse ONSeimal@W shoul asem the lest role ia eselibeoverall policy for the long-tern retm~e ofseismological data ourrently beig .Llses It to

esetal that Esuttag avemiesoser .esmatms --the university reseerok oinitg actively partilete in*stahblishing overall policy to ese that inasfel ftta

14. Seosus at the rapid tekMi"166 Obft iSoa0pUttisal and dftt&-hWAmlmg emodlities ANational nseak ftil's Comitftes bee3OWshot"d eview tdo states 09 dafta PWbLs La mirniegon a year-to-year bes"s to ernie efwetie ~s ofearthqoake data.'

-. 71dumb AM

2

Ueimmolmg, in cam m n with most of the physical sciences,has entered the 1930. with Increasing challenges andapportuaities presented by advances ian the technology forgather lag and analysing data. 2hese developmnts haveresulted in .apiqd increase of the amunt of data, ashift io data collectiona from primarily analog to digitalroortdiag, ad the growiag applicationi of advancedcemste teaalogy. n they provide opotunities toimems sipefiatly the scientific returns fromseismi dat@ all types.

Vones a dslqats pervade all aspects of seismologsi we qpinmat is strmgt iona stadioss explotationstivitteer etimestal sad swmar reflection profiling#

glebsi. regiemal, md Lo"a metuack mad array studiesimad lIng.-sOsl global studies at structure, attenuation,med Swnson ese smumisa. 3mbk of these areasbeAte iSm see ad spesif ie data peablems. Wei reportis uessed primariy a data peablemat o the global

aehs 04wil ma regismal metuerkofan strong-mmtlm Wes 106lM. Problms ssciated with seismic

POWUM e ma,. alised data-gathering activitiesLa Me ~ meo Setq metwoef iamed above havebarn, desue Me1 : - prablam of eatqu data

inqt6401e bass asee@.Us past of Oni rWImmeses is the usmmt of

donae apsedall di15 data, Is MUMin felt in manyfisids. Offsmam is Pertlsmi. Ubs imeresas In theSmOMtd of mb~m data Is a cmesmsre at the rapid

sqmOwl Nom "as dea oil doe ee - IOilM I I" M -0, ft the world md the

amn 4w o d st segiitift methods. TableS1 is 8490" a 96ves m iadiestlam Of the amt of

ONG ~ ~ b PaIMal 6gleA61.q aM&m reglsmal metmeks and

-- *.-4

* '

!

the amount of storage capacity required to save thesedata.

On a global basis the number of oountries and organiza-tios currently engaged in digital seismic monitoring isalready ipresive, and others are moving rapidly towarddigital recording. A joint Whh/UNW publication,Directory of Wmrld Digital seismic Statiois (1982),describes many of the existing operations. Severaladditional countries are installing digital systems, andin others they are planned or initiated.

One of the major advances in observational seismologyin the predigital era was the developmoent and installationof the World-Wide Standardised Seismograph Network (UVOSE)in the early 19O. The network was installed with DARPAsupport under the VEA UNIFOf program for fundamentalstudies, including those related to the detection anddiscrimination of nuclear explosion. This networksucceeded in its primary goal of upgrading observationalseismology. For example, modern theories of platetectonics are rooted in global patterns in seismicity andearthquake fcal mechanisms, the observation of thesesystematic patterns can be traced directly to USSEdata. Global earthquake catalogs, used In a variety ofapplications from earthquake prediction to tectonicstudies and seismic risk analysis, have relied heavily onVMS data. Our knowledge of the internal structure ofthe Barth has also been advanced significantly by studiesusing WAS data. Because of the widespread use of thesedata and the operation of stations by many universitygroups, the IRN has also played an important role inthe education and training of seismologists, both in theUnited States and internationally.

The prima y reaons for the outstanding success of theMM5 are inherent in its titles It Is ffcikw]lproviding global covrgel it uses a standscljjd set ofinstrumentr providing uniform responses at all stationsand It is a coordinated AuL#e with a smmesfulmechaniam for data collection and distribution thrumghthe National Geophysical Data Center. Our oballenge aswe enter the era of digital seimolegy is to repeat thesesuc sss of the mum, while taking advantage of theenhanced quality and rfeolutioa of digital data..

fte Senor used i the UN iON rp atd staft-of-the-art Nsioetty of that time toia 1960) e Meinstruaitation selected fllwed eareful emaide seUof the available te boly teted mees at tedata. The vidom of the boices mode to regected to the

34-

* 06__

continued viability of these stations today. omee,because of the limitations imposed by the photographicrecording medium, the 3156 observations are limited inboth frequency and dynamic cange. The M38 data havebeen mst useful In those studies that require measure-ments of tie and amplitude of particular phases, e~g.,hypocenter locations magnitude determinations fault-plane solutions, and velocity structure. They have beenused considerably less in detailed waveform and spectralanalysis because of the necessity that they first beconverted to digital form. Some of the long-period.records have been successfully digitized by hand (atedious and time-cnsuming process, with limited spectralresolution) for me-detailed waveform and spectralanalyses, the short-period data have seen more limiteduse in waveormc and spectral studies, because of thedifficulties in digitization from the compressed timescale of the photographic records.

Recent developments In senor technology (especiallyfeedback seissometers), coupled with stable, broadbandamplifiers and digital recording and playback methods,now provide the technological capability to record andanalyse, ground motion with high fidelity over a broadfrequency and amlitude range. in a frequency handspanning many orders of magnitude, from periods ofthousaids of seconds to frequencies of hundreds of hertz,there are no longer any teBhalml limitations indetecting and recording ground motion ranging inamplitude from beekprousid noise at the quieteet sites toI the strong actions 1produced In the neaw field by largeearthiquakes. hs discussed elsewhiere, in this report, themaimr problem now lie In determfiin what data will beI callected, how nuah will be archived, and how data willbe distributed to noses (ie., dota-IgeI policiesand procedures) and funding of dmemmeagst ad-distribution activities. Mile faul ofamtag of theenew technologies has mot yet bees takes In emy areas OfSOAGODegy, the emergime ilasaoado divitalrecording, especially in taleseesmic and now-fieldstudie, aready points to the selastiti. awataes,gained from the mee of high-eseleutem data [mefft

&M~ fa~ha wm~mmvA4 m (JOAS),

At~~- xwo 1178o -l

9

In this report we deal only peripherally withquestions of instrumentation fo primary data collection(seismograph systems). These instrumentation issues arediscussed at length in the network reports listed aboveand (except for regional networks, which are discussedbriefly below) excellent prototype seismographic systemwith broadband digital recording are already available.We also have not emphasized the computer hardware to beused in accessing the data base, nor have we consideredin detail the computer facilities required for extensivecomputational studies using the data. In these areas,technology is developing rapidly in response to demandsfrom a wide variety of scientific and nonscientificusers, and the data volumes anticipated in seismology donot appear to present any significant tohnologicalproblems that will not be satisfied by the advancingstate of the art. Rere, as with seismographic iustru-mentation, our problems lie not in enouraging thedevelopmeat of new tecnologies but In ensuring thatmodern facilities are sof available to the seilmologicalcoummity and used wisely. This will require adequatefunding and good data-.mngemt: policies and procedures.

To realise fully the potential scientific teturns inseismology requires that high-quality data not only begathered but also effectively disseminated to a wide usercommnity concerned with both basic ressaxh andapplications of seismic data. This report is addressed,primarily to the problem of data aagemet in seimologybecause Of 41hUOrtoInUs in tke Irset.dt-wdii-distribution nraoticos.

aim

3

CORmN= SIAM AUD PWALS

It is convenient to character ize earthquake, data sourcesby function into near-fileld Or Strongmtio recordingsand far-fileld or network recrdings. Network data ay befurther categorized by areal extent into local, regional.national, and global scale*. ideally, each data sourceis described by a station catalog listing the positionand opeg sting characteristics of each receiver an afunction of time. Siahb data source produces raw datarecorded in Analog and/or digital format. Routinely.phase arrival ties and amplitudes or event durations arederived from raw network data, and evest hypoostrallocations, origia time, and magitudes are -derived fromthese phase data.

In general, the success of efforts to provide stationcatalogs, event catalogs, event-associated phase data#and ray data to other users reflects both the miaturity ofa particular data source and the national percept ion overthe past several decades of its social relevance. thustthe distribution of strong-miotion data Is relatively welldeveloped because of Its relevance to lifesavingearthquake-resistant building design and because the datavolum is smeall. Similarly, global network datadistribution is better developed than that for regionalor local networks because of Its applicability to a widerange of seismological and tectonic scientific proble1msand to practical issues such as mmitring nuclearexplos ions and asessmuats oil seismic risk over largemrea. the umaller netwomrks tend to be of oresqecislised interest and Uinited in lifetime, resultingin little or no distribution of data to secondary users.

Ida

The distribution of anlgrecordings is better

developed than that for digital data fcc severalreasons. The latter is based on now technology and havebeen loe readily available for perusall many data usershave limited familiarity with the use of digital data andcommonly lack facilities for exploiting digital analysistechniques in their research. The following discussionfocuses primarily an the status and problem with globaland national network data and strong-motion data.

DATA SOUMCBS

The Global Seismic Network (CAN) was developed by theDefense Advanced Research Projects Agency (RPA) in aseries of successive projects, and it is currentlyoperated by the USGS. The oldest coaumlan1ent, the USD.had an intended size of 125 stations. About 105 stationswer* installed in the early 1960s, and 96 are stilloperating. These stations feature analog recording ofthree-component long-period r 15-sec peak period) andthree-coqiionent short-period ( I-sec peak period)instruments. In the late 1960s and early 1970a, the 05Mwas augmented with 10 digitally recorded high-gainlong-period (HmPL) instruments featuring three-cqmonentlong-period data sampled at a rate of one sample persecond (ups). During the mid-1970s 12 Seismic ResearchObservatories (SFs) were added. Sorehole, force-balanceinstruments in the Sans feature digital recording ofthree-component long-period, data sampled at 1 sps andvertical component shoct-period data at 20 ape. in thelate 1970s, a short-period vertical component sensor wasadded to five of the u9Lfs. which were upgraded to anSMO-compatible recording system and dubbed AbbreviatedSeismic Research Observatories (AMM. The other IlL?stations were closed. In the early 1930ap 14 of theWWSSN stations wers upgraded to digital recording andrenaged Digital World Wide Standardised SeismographNetwork (MURY stations. The USUIs featurethree -,. P : mes!t long- and intermediate- periLod channelssampld at 1 spe and 1l spe, respectively,and a vertioal omonest sbort-period channel samled at20 ups. The S~lsp ArS.a and UUUUfs ate colistivelyreferred to as the Global Disital Solimgrab Network(OWIN). Anl M gutlss featurie odatls~s recedingOf long-peioid wftels a"i field "aent-uri"*eriag torecord signals from short- and intermediate-period

ado1

12

channels. The GDSN is complemented by the internationalDeployment of Accelecometers (IDA) operated by theUniversity of California at San Diego. The 17force-balance gravimters in this global network supplydigitally recorded, vertical-component information atperiods longer than 60 sec, sampled every 10 sec.

Several nay networks are being added to the GM~i. Theinstruments feature borehole force-balance operation,real-time satellite telemetry, and three components ineach of three period bands all continuously digitallyrecorded at rates of 1, 4, and 40 spa for long-,intermediate-, and short-period bands, respectively. Afive-station North American network of such stations,known as the Regional Seismic Test Network (lUST)operated by Sandia National Laboratories is already inoperation. A subset of these continuous data (withlaboratory event-triggered short- and intermediate-periodchannels) is currently being included on the network-daytapes (NlDTs). Another similar network, known as theGlobal Telemetered Seismic Network (GTSN), is now beingplanned by the USGS. it will consist of 17 stations inthe southern portion of the western hemisphere. These 17new stations will represent a significant increase in theGDSN data volume that will require substantial increasesin the level of effort required for effective archivingand distribution. In turn, costs to users will besignificantly increased if the smallest unit of data thatcan be obtained continues to be a network day (26 hoursof data for all stations). Appendix 2 describes newglobal data collection activities of the Air ForceTechnical Applications Center (ArTAC) utilising similarinstrumentation.

There is currently no national seismic network.Dowever, by telemetering about 60 short-period verticalsignals from existing observatories in the continentalUnited States and Alaska to Golden, Colorado, the USGShas created what is de facto a rudimentary nationalnetwork. The data produced are digitally recorded at 20spa for use by the National Barctquake Informationservice (MB) of the USS. Because these are narrow-band and loo-dynomia-range data, no effort has been madeto distribute them to secoadary users. However, thePanel believes there is a possibility of establishingselected bigb-dymmic-range, broadbemd regional networkstations that together with existing ON Stations(including ) would form a national network of highquality.

4

13

There are numerous regional and local networks through-out the United States. They are typically operated byuniversity personnel and funded by either the USGS or theNuclear Regulatory Commission (NBC). In aggregater theycomprise about 1,600 stations, of which about half aredigitally recorded (event windows) at rates of 50-250ape. For the most part, the instrumentation is vertical-component only, narrow band, and low dynamic range,although there are an increasing number of broadband,high-dynamic-range, three-component stations. Thesenetworks are used mainly to measure arrival time, signaldurations, and amplitudes of microearthquakes forearthquake-hazard-reduction research. The chief productsof these networks usually consist of an event catalog andassociated seismicity maps.

Although some of the larger networks share camonhardware and software, there are many data formatsextant. There is little data exchange at present andthen only on an informal basis. There is no comprehensivecatalog of stations and events for these networks. Thuspotential users cannot easily determine whetherappropriate data even exist for their problem.

Irreplaceable data from many li ited-duration stationdeployments have been gathered during the past 35 yearsand now are in danger of being lost forever unlessimmediate action is taken. Millions of dollars have beenspent on collecting these special research data sets fromnetworks such as the national network of teorarystations, which comprises the Long Range Seimic Plonitor-ing (LiU) network and permanent observatories fornuclear test ban monitoring that were in place in thelate 1960a and the early 1970a! the KW network, theSeismic Data Collection System (SDCS)l and from special-purpose research arrays such as the Large AperatureSeismic Array (LKA) and the Alaskan Long Period Array(ALPA). To the present, T! e -. esh's Seismic DataAnalysis Center in Alexandria, Virginia, has archivedthese analog and digital ata, but the operation of thatfacility is scheduled to terminate on October 1, 1963.Larae subswt from these dat should be achivedindetinitely became of their valm tor gurmt mdfuture seimlUwic studies. To do this effectively willrequire a policy for data retention for all earthquakeseismic data, since none exists at peesnt, and theassignment of organizational resee sibility for thepreparation the data subsets selected for archivLng.

IWO.

-S. /

14

The National Strong-Votion program has two ooiponents(a) the National Science Foundation is responsible forthe research program, and (b) the USG operates andmanages the strong-motion netwocks, including datahandling. An extensive body of strong-motion data hasbeen collected under this program and organised into adata base that includes both analog and digital waveforms.A strong-motion instrumentation program Is operated bythe state of California, and seleoted data from thisprogram should be included in a national data base. Samearrangement should be made for getting stong-motion dataroutinely on a global basis. The details of thesestrong-motion data collection systems are smmried inthe reports entitled Strong-Notion ButbAmM InstrumentArrays (1978) and U.S. Stona-Motion SartbmakeInstr mentation (1981).

DATA AIK VIiG AND DISTPZUITIOM

Since its beginning in 1961, more than S million originalWMW analog reoords have been copied and 60 millioncopies supplied to users. Currently, there are severalhundred reuets por year for selamgram miorofiche.Originally, the seimograme vere filmed on speciallydesigned 70-rn panoramic cameras at ft reduction. in1978 filming was chmaged to put 4 images (4 days ofnormal operatio.) on a single 1OS-a microfiche at 32zreduction. Standing ocders of the Wo3le metwork bavebeen purchased by eight institutions [(Lmt-DohertyGeological ObmeevatoYi Institute of Geological Sciences,dinburghr .9.t University of 26kyal California Institute

of T hnologyi Mlssumetts Imstitute of ftaologyjUUM/Menlo Parko I0 /Golden; Uatioal Gopbyoalo DataCenter (NW)), ad substantial parts of the netwock baebeen supplied to five institations (Uiversity of !Ta /Galveston; Cornell Univerityl University of Otago (MewSealand); Los AlalOm NUtiemal Uadetoaty M/

Te m. metork data are ampsmiked oapies o thevisible (malog) reowde I rm de AM (S) md B (12)notworks, f i the Co im neawo t em 3s,,5-m lla *twiolp6, and frm the Pe"e BMabl of MU Malmoc1979. Large-ngitude or eismlgieelly isportentearthquak fraom several Immred additimal stations,including those of the Union of Soviet booialiatRepublics, are provided under the International Data

A A

Irchnge(I=) agreement. Also, selected historicalanalog seinograms are being filmed under a jointUsGSMA Proj ect-

This analog data distribution system is operatingprimarily with contract labor and with about I weeksbeing required for the cycle from receipt of originalrecords to supplying of copies to users. Fifty percentof the network data Is generally available for distribum-tion within 8 months after the recording interval. tearchival film copy is made at OMt expense, with the costof each additional copy being borne by the users, which Isa direct reflection of the policy of the U.S. Departmentof Commerce. Present costs to users have Increasedsignificantly for some services# which poses a barrier tocontinued wide use of these analog data.

MRAL has cmiled station catalogs of ME stations andhistorical stations (both U.S. and foreign) on hardcopy.The USGS maintains, In a computer data base, a currentstation catalog of U.S. and foreign stations thatcontribute derived phase data to the =IlS, but it is notcur rently available to secondary data users. 2he E=1Scollects phase arrivals and amplitute from some 2,000contributing stations by telex and airmail letter. Some500,000 of these data per year are culled# associated#and used to locate 5,000 to 6,000 earthquakes. Uarthquakebulletins (hardcopy) are prepared and distributed by theUSGS. Waent lists are also merged into a NM catalogfor distribution on- both hardcop and computer-compatiblemedia. In addition, a catalog of historical U.S.earthquakes Is being compiled by the U3GS. Asoiatedphase data are distributed only on hardcopy listingsanedeven that distribution Is limited. Both of theselimitations constitute a major shortcoming of the preentdat-management proce~idures.

Since 1976, OMM data have been collated into EMlf irst In a raw binary format by WA and since 1900 bythe UNG in a binary format with Aimerican NationalStandard Code for Information Interchange (MCI) headers(Including conideable information about stationoperating characteristics). lAW station tapes containingabout 14 days at data end oerator le"s covering the ametime period awe ollected ftrm ach statim in theby the Albwqme Seiamol~gioal Laberaliory (MM) of theUSS Station-Obannel headers inoor,tm the logInformation ar* generated for each Lag. FIMalli allheaders and station tqw data foe suf day e incluedIn OWN, each stme.hne et appearing in sqss

I"

A&3

as a file on the tape. Te Ion are made availabl, tousers approximately 12 weeks after the data are initiallyrecorded. This operation in also plagued by rapid costincreases due to increases in the cost of personnel a"dmter ials and because of the rapid growth in the size of

the uto th e GM is ready in iminent

hardware. To date, all casts of generating the Met havebeen borne by the UMG as a service to the usercomunity. Mhe mWO are currently archived by the UMGand slang with XOR data tapes are archived by MWc fordistribution of copies to users.

There are a number of difficulties with the gndatacollection, archiving, and distribution. Although theanalog data handling and distribution is smooth andeffective, this operation is necessarily labor-intensive,resulting in hIgh costs that have chronically threatenedits continued existence. Of the catalog data, only theevent catal" my be omputer soehe at preent, andeach user mast sapy hin own searcob software or requesta searchl by ainc, whicb necessarily involves some delayin receiving the desired event information. Te digitaldata distribution problems is the mat severe of all.tese data are available only in Increments of m day

(o9 gkday tapes) at 61" per tape. oupled withaternative Seems of acquiriLog mef this ha represented

a significant harrier for users ceosally gaiversityresearchers), resulting in low demad at UMC.

To Illustrate the nature of thi barriers a typicaltuniversity study of the nature of earthquake sourcemechanim that character ise a seismic area of interest

reaonalywould involve at least 20 erthquakes,resulting In a cast of 82,000 for 20 metwork-my tapes.

The researcher must thea preprocess these 20 fall tapesto extsact the event windo in Lteresto thus substantialadditional costs are likely to he incurred before anydata analysis is dome. in practise sers have acquired="S data inmally, seta at so costs. f rom DAMA

an otome"the ue of daoa. the tES recentlyestablished three regiama2 seamts that have acquired

cmpaca eo" medi cfat eels UmieverIthis watoonly perlalevidte cWuergo" pSoel beuse o n

costs are incurred sad the numer of Ses that must be

17

obtained and preprocessed by the user remains the siim.Thse cost-induced barriers vill continue to ensureexcessively high per-user cost of aquiring and utilisingGM data, unless alternative moa of distribution areiplemnted.

These distribution problem rais merious questions ofbow effective distribution at affordable costs can beachieved and which organization should ultimately beresponsible for distributing GM. data. This wholesituation will be further exacerbated by the impendingaddition to the GM of several v networks representinga new generation of technology, as discussed earlier.Very soon a network day viii not fit on a single 10 bpi9-track magnetic tape.

The national strong-motion progran is funded andmanaged by the USGS from Menlo Park# California. Brent-recording catalogs have been ocMiled and made availableto users via on-sits and remote interactive access to arelational data-base manager. Original analog filmrecordings fro the western hemisphere are archived inMenlo Park, California* and derived digitized wavefocmare available from AMC. bouever, because of a constantstate of flux in the UM ooauting envirocmnst and ashortage of data-bandling resources, it has beendifficult to maintain accessibility to the strong-motiondata base to met the aeeds of the engineering caamintyor even to keep it current. To ensure its viability inthe future, this invaluable and already hLghly organiseddata resource needs to be housed and mawaged in same morestable. central, national earthquake data archiving anddistribution facility where it can be maintained in areadily accessible form for users.

A PAWWN "T"ZOML C m OS S3 GCAL 59UNS

An item of imdiate ipsortamce, with long-term splUca-tions for the future of seismology, ad specifically ftoovercoming many of the data problems discussed above, isthe current develoment by DaM of the Center forSiLic Studies (CU) # recently establihe in loeslyneVirginia. Appendix D Ddesribes this faLlity mawefully. Own copleted in 19*4, the Center Vwil includestate-of-the-art nsofutational and aamenmet

capabilities designed fto seiLugical applatims andresearch, a mall resident researc staff, and provisLonfor visiting scientists waoking on & researchprograms in test ban treaty verification.

/,,

To support MRA objectives, data ar* being collectedfrom a variety of sources through several types of datalinks. These range from RMU data received via asatellite North station located at the Center toalphanumeric data received via the world meteorologicalOrganizationls (1010) Global Telecomnicat ions systiem.GOOU and other data are received on tape. At varioustimes special experiments will be conducted as part ofthe United Nations Comittee an Disarmament Group ofScientific Expects activities and will include nationaland international data exchange.

As currently conceived, this facility Is developingmany of the capabilities desired for general seismologioaldata management and distribution. The seismic data basewill consist of basic event information (such as stationphase and amplitude readings, hypocentcal locations,magnitude, and other souce characteristics) a"d recordedsignals, consisting of event waveforms, som available asthe events ocur, but others, such as GOM day tapes.delayed by weeks to months. A variety of analysissoftware Is also being developed and implemented toprovide routine user computational services.

In addition, remote user access to the Center is beingdeveloped by MDfl in the form of relatively inexpensive($30#000 to $50,000 each) Deote Seismic Terminal IMT0).work stations. These terminals will be capable of@toring and displaying a significant amount of eventInformation ane weform data, powerful local processingfunctions can also be carried out on the PAT*. DesigningInterfaces with many other remote user system should bereasonably straightforward as well.

The Center will be accessible vid the AISM. Tymeet.dial-up, the 710s, and, in a fewcose, dedicated lines.Through them semas, DAMr researchers will have accessto both the data aid computing resources. These resourceswill Include two Digital Wr~ipment Corporation (MC) VAX11/760 computers and three EEC PIS 11/44a limned by ahigh-speed Iesm) local netuork. Analysis My becarried cut *a var osc interactive graphics tonrmlals.The Cnet has bees designed to be am flexible mdaccessible to researchers as possible.

A small nmber of researchers hove afteadybeu usialthe CB ompeting and data-handia resouce am a trialbasis to help to evate it and to Peim i IUq vn meets--tio the system As development continue. eg imaiag infisa 1@0ar 19IM, the Cetert expects to OVPt NW*extensive use of the facilities by UM1-aisoe

roeearcers. Within at muta OWew s Se OeWSsmmlresponsibility tat the Ceter is plinod to be pooled toso other ogmisatiom. a* Castor VmM an bejointly funded by WAWA* the apertiag orgaetio. aniother user aefiiSS.

The panel views the develapmmst .1 One CMasssignif icant oPportusity to benefIt amieftely sum therapid avances in osaputor tookesog and ase-a.magemt that have profoundly lusromsed digital datastorage, retrieval, and analysis oqp&litee Aig Mepast decade. Other Mn-pn~e effects isno hresthave led to maimr eeiamlogima anumoe, mmer tyexaples are the ErI and fLila chip data distibutiont iathe 1960s and the new ONglobal digital motmrk withstandard network-day tooe data distribution. N~eriewith this advanoed system at the Center will beexceedingly Important to the deopUment of intermediate-and lag-tern capabilities needed by the seismlogicalcomeuity for research and applications in the digital

Sm"=m or iwAmM

'Rh. abome discussion of the current status of sias-graphic network data handling includes problem ofoperations and data management. Specifically theproblems that now seriously limit or threaten to lisitthe effective use of earthquake seismic data are thefollowings

1. A ocqarebensiveo directory of information onearthquake data sources does not exist in hardcopy orcomputer-aoessible form. 1bus, for potential ws itommnly requires a tiecnaig Ambeious effort lostto deterame mhat relevant dta"5 exist, Ame theymre, aroived, amd how they b e topic"ednd In uAtfP cc o formats. 2%14 situation is exacerbated by rapidincreases in data volumel. changing station distribution.and a stoody transition to digital Cdeim.

2. onee One mot emist a oeremsive stalog ofglobal amd national hypoomatets with inseelated phasedata. ame the limi1ed data that emit narmlydistributed In haidmOW! fors, Kireft oompteaeS tosu" data by ose is severely limited or Iftmssible at

Ioi

20

3. Uboi nUag an distribution of Global Seismictegk fstaleg 4et is well cgs by the UC personnel,ut current D"WImenat Co CMoe polik requires that

eapylag and diatribmtim of aeimgras be charged to theWece oe the data# and these costs have already increasedso greatly that the ocotiamation of the present dataservices at arffdable "at& to users Is seriouslytbreaamed.

4. ftept got the am netwock-day tapes, there areno standard foemats foe digital data echange. Becauseof the variety of formats used fot recording and archivLngdata amg the past and current seimic ata-collectionoperatiom, direct excbange of data is exceedingly cumber-son and difficult. there needs to be a standardisationof national and International data-exchange formats forglobal, regional, and local networks.

5. Because the operation of smaller regional and localnetworks Is focused on seismic problems of specialisedinterest, there Is little or no distribution of data tosecoday users and there Is no standard exchange format.because of limited funding, these network data comonlyare not archived in a readily accessible form fordistribution. fhere is also no readLly available centraldirectory of information on the events reoorded or evenon the station locations and instrumentation that make upthese networks.

G. The extensive data bos of high-quality strong-notion "to is not adequately available to users as itwould be if the strong-motion data were part of anational seimmlogical data base boused in a stable.national eartbquake data atchiving and distributionfacility.

7. Decase of Depattment of Cemeroe cost-recovetypolicies and a low initial demand, network-day tapes ofthe GM have been priced by the IWMC at l100 per tapecopy. Decause the network-day tape is the allestLiccemant of data currently available, the user typioal ymust obtain (and pCeprocess) many network-day tapes justto obtain the event data of Interestr these event datausually oupS only a small fraction of the tapes thatmust be acuired te hig coats have resuLted Inno k-ay tpe being obtained by users, with lessconveseno but essentLall at no cost, on user-wApliedtap" from DNW the am, and at four regioaLcenters. Orgamisational responblity for distiutingthe amSE data is fragmented and unsttled, and It needsto be determined j, under the condition that standa"

-7i

21

types of digital data subsets must be readily availablej~t aUmeste tow anosteedta ta

S. now di-bota nto-day fotatin e.ety en9estabihede wilsoo el for a sgnicyfcat irentino

al lrhue s isic data, teh andld Thgaise pwillt

prgentdaraae thes data s o lont ter m arehuvieg

thie roblem ish vomedae of d ate r ehgh-pravitydseia data toases acqui th mus otAM. taLIA thSDCS. nd otheor network orray pe ain fteps

9. yer t aed retly arpcye aot deaeayre-teob'sl i artak Analysi Center ih Aeanraised Virgiia o

ptet fite slcheed to ber cloedar attebiningofhiscarolear 1to4 immedaead te fctor i theed igqtoasueiathsilal data baesacuiedbyte iM, 1ASerAVAOWSEC. ather eeoen by araoean of the Ceteoesm

3t5des tat aeof-the-at atave anayi To sisidatamicba mAnaement fCilty, i represent siiima.l

Thisydevloed nd scevalued t e fcli tyi t beinntune fsalver 19 4 sommoteatet astm i ded ogi-t io that will halbe doomthe mam sngtWspu

daibt or itsagcntfaind rpetouatpresentaeuyit fatbem isd annopprtniy The proble isa this

auly eveoen vautd h facilityprery otfgs o to ertune of uses" oulderv as t prostomtye ofgaie

Ifis-ate intionl ceter for t s is thiat stuhis.

Such a facility, properly smaged and adequately fundedecould owrome most of the Current problems that thepanel has Identified and greatly enhance the ose atavailable earthquae seismic data by both 0.A. andinternatial ecientislts.

I q9f

4

sucuuna& zau,

If no steps ae taken to address the problem identifiedin Chapter 3 the situation is likely to develop that onlythe operators of eismograpb stations vil have readyaccees to the data they reoocd. Zffective data dLsseaina-tion to the seisnological user coammity currently existsonly got the oldwide Standardised Seismograph network(1Wsn) analog (film) data, and that service is threatenedby escalating costs. fTere is no ocmprebensive directoryof information that descrLbes what data sets exist, whohas the., and how data can be obtained. isting globaleathquake catalogs are difficult to coeputer search, andthere is no aqarehensive national earthquake catlog.Also, additional phase data are not generally availableat all. Wat information exists is fragmented and mustbe obtained from many different organizations. Distribu-tion of digital data Is limited because of fragmentedresponsibility, current polioies and prcceduces, anddirect and indirect costs to users.

'Ehese pableam, coupled with the Defese AdvancedResearch Projects Agenay's (MM&'s) recent implemntatLonof a deu comqutational capablity for seismologicaldata storage, retrieval, and dissemiation to users (theDm~ Center for Seismic Studies) and for data analysishave led te poel to ocanouie that establishiog anational eaemoagical digital data base an a nationalseimemi data faoliLty Is boW1 desirable and feasible atthis tim. oefcee, -the t -- t .- "l >of this Imel is tat a otianaI Cewto for em lMal

lta4 be "Wailed thot will ,OrCoft0 t"J Uff deta-

!EM t W sM tat we hv identified sombsnathe avalab ity a i efMtLie us of hiLO-m"L t datasots by the entire smloM2Uica gce nit,

22

23

aftective use of this national seismological facilityvill require the followings

(a) Upgrading of the present data-management systemsto provide adequately for user needs.

(b) Development of a coprehensive directory ofavailable data, data sources, data quality, and format#together with a search capability implemented at theNational Center and at regional facilities, accessiblealso by remote terminals.

(c) Development of a comprehensive catalog of nationaland global hypocenters including associated phase datafor at least the larger events, together with a camputersearch capability Impleented at the National Center.

(d) Standardization of data formats for the purposeof national and international data exchange.

(e) 3stablishment of a policy for long-term dataretention and archiving.

To meet these goals, the panel's conclusions andrecomendations are as follows

1. DAMPA's Center for Seismic Studies (CS) hasseveral objectives of direct relevance to this panel'sobjectives. The first is to develop the Capability touse digital data from global stations and networkseffectively. The second is to support U.S. initiativesin the United Nations Comaitteoe on Disarameent (U),Group of Scientif ic xperts (03) V to develop improveddata-exchange provisions for future test ban treaties.The third objective is to provide a test facility fordeveloping now tools for seismic data analysis and toprovide access to organised data bases to support DARPAresearch.

The panel vies the development of this Center as asignificant op;ortunity to benefit immdiately from therapid advances in camputer technology that have profoundlyIncreased digital data storage, retrieval, and analysiscapabilities during the past decade. Rxperienoe withthis system will be exceedingly Important to thedevelopment of intermediate- and ong-tem capabilitiesneeded by the seismological oammnity. A discussed inChapter 3, other DAMP-SpaLoaced effocts in te past have,led to major seimlogical advenmes, and this ame clearlywill also.

DRM S Caabr "I" 3be hacosierd oottmsfo t ued lmat of antoa

24

Center for Seismological Studies. Its operation shouldbe structured to provide data and services needed by the

* seismological research community in general and by several

interested federal agencies. Operational costs of anational center are estimated to be bptween $2 millionand $3 million per year. The required level of fundingshould be provided by the participating user organiza-tionsi federal agencies that can be shown to have a needare the Natonal Science Foundation (NSF), DAPA, the U.S.Geological Survey (USGS), the Department of Energy (DO),the Air Force Technical Application Center (lINAC), theAgency for International Development (AID), the NationalpOceanic and Atmospheric Administration (MA), theFederal Emergency Management Agency (1A), and theNational Aeronautics and Space Administration (RASA).

Therefore, the highest Prioritv recomeandation of thepanel is that a National Center for Selsoolaical Studiesbe established to assure the effective use of global,national, and selected regional and strong-motionjearthquake data. The National Center should meet atleast all the functional requirements given in Appendix Fof this report.

The min goal of the National Center for SeismologicalStudies should be to encourage the continuing applicationof state-of-the-art equipment and techniques to currentproblm In seismology. A ocmponent of in-house researchand facilities for visitors will be essential to ensurethe continued effectiveness of the Centaer one of itsmajor functions will be to ensure the availability andeasy distribution of data to external users. stablishingformats and a data-base structure for the archived dataand developing standardised software for the routinemanipulation of both waveform and parameter data withinthe Center will de facto solve many of the problem ofstandardisation that currently inhibit the easy exchangeof data. NMny of the current activities of federalagencies responsible for earthquake monitoring andresearch can be coordinated through the Center withsignificant ispcveet in quality, efficiency, andresponse, time for major seilic events. Same or all ofthe functions of the National Eartquake InformationService (OIS, WIS) cam be inoporated within theCenter. The cloe oordinatios of the activities of theCenter with the ongoing work of MEPA in testing andimplenenting nm technologies will help to ensure thatmodern facilities wre mintaime4, e ftCMer winl

provide anatural focus for the analysis and distribution

I _________________________________________

25

of data, either via direct access or through a distribu-tion facility such as that nov operated by the NGDS.

2. The operation of the national Center for Seismo-logical Studies should be the designated responsibilityof a lead organization, such as a private nonprofitcorporation like that of the national Center forAtmospheric Research, the National Science Foundation, oranother federal agency. However, the policies andprocedures for the National Center for SeismologicalStudies should be established by the participating userorganizations through an advisory board. The funding foroperation of this facility should be planned and formallycommitted on a continuing basis by a consortium offunding agencies to minimize future support problem.Figure 1 shows how the proposed Center could be organizedto meet the desired objectives discussed in this report.

Using the Center for Seismic Studies (CBS) as aprototype or a nucleus to develop a national datafacility will require immediate action. The C88 willsoon be operational (at the end of fiscal year 1984) as amodel facility designed for use in comprehensivemonitoring of a test ban on underground nuclearexplosions. As such, It has a limited operational lifefor these requirements, and then it will be transferredto some other (currently undetermined) organization orshut down. The CB represents a rare opportunity toadvance our seismological research capabilities takingfull advantage of the state-of-the-art computing and database management technology that has been implemented.

* The panel recommends that in calendar 19$3 the NationalResearch Council's Committee on Seismoloay initiate stepsto convene a meting of representatives from fundingagencies. other user organizations. and the universityommunitv to discuss the Possible implementation andfundina of a National Center for Seismological Studiesusing the CBS as its prototype.

3. The present distribution system for analog datashould be maintained and supported. The technology foracquiring and analyzing data is changing rapidly, butmany productive and innovative seismologists in theUnited States and in other countries will continue to useanalog data for at least another decade because newtechnology will not be available to them and, perhaps,because of lack of training in the use of digitaltechnology. Too, some seismographic stations recordinganalog data fill important gaps in the global networkcoverage. ]ven though the trend is to replace analog by

V

US

* *6.4U ____

a* U-.

* UIL U 4 Uo Uo 4'

£ a4 £a I

U 3~88I5 I.'U lull 0

mu '.4

4 0 a- Sa U

U I0liii 5' 1 1

4 0.

'4'

U U~2111111

45I I U

U4~4

'4'0

I~ 45I

111111 "4

iimm

26

'~

27

digital recording in the long tern, provisions must bemade for handling both types of data.

The National Geophysical Data Center (OGDC), of MORA,has had the primary responsibility of handling analogdata from the Worldwide Standardized Seismograph Network(WSSN) from its inception responsibilities include themicrofilming of original record& and the provision ofhigh-quality film copies to users. These services havebeen and will continue to be valuable to the seismologicalcommunity for the foreseeable future. The policy of the

Z Department of Commerce is that the costs of reproductionand distribution are recovered from the users. Ratherthan considering curtailment of this activity, HOAshould recognize the importance of this national dataresource to the seismological oeunity and continue toprovide funding for manpower, equipment, and storagefacilities adequate for this analog data preparation anddistribution, Funding organisations should providesufficient monies through their grants or contracts toresearchers for copying and distribution of the analogrecords. The panel recpmends that the analo dataarchiving and distribution facility be mainte- .andthat the costs of operation be Wed as dearibed above.Alternative tchnigues for arcbiving and diseminatineanalog data should be considered by M C to kee costs tousers at a minimum.

4. The Albuquerque Seismological Laboratory (AL) ofthe UG has the primary responsibility for assemblingthe digital data from the evolving GWN. This is beingdone as described in Chapter 3. The ML monitoa globaland national digital data collection, collates the dataand generates network-day tapes (M8s). The data ateavailable through the NOC in increments of one day (26hours) for $100 per tape md UEOCIhA sbould makeeffats toeard reducing this price. To avoid this highcost, users have acquired data informally at essentiallyno cost (except tapes) from DAMPA and USS. The UNS hasestablished thre regional contrs-at the CaliforniaInstitute og Tohnology, Saint Louis Uiversity, andBarvard nivority-all of which have acquired cmpletests of recent S. Visitors cm use regional centerfacilities to copy Ues.

All of these current modes of distribution of digitaldata ate either too costly or overIT cmbrsome to users,because the minimum increment of data available is omNM consasting of 26 bours of data from all the stationsin the USS. Thus a large nufter of tapes may be

. . s -

28V involved in a single research Project and the Usert commonly must engage in a significant preprocessing

cooputational effort just to get the desired eWant timewindow fog analysis. This situation will worsen

considerably In the near future as all of a network daywill not fit on a single l500-bpi 9-track tape. It in

the panel's understanding that at present almost anetwork-mth of GfhE data resides in active memory orwas storage on the system used to generate the inM atthe AOL. Thus network tapes of any time increment couldbe generated with minimal perturbation to the presentprocedures and no change in format. With modest upgradingof the aging conputer systems now used, several networkmonths could reside In the active archive and thus beavailable for supplying data increments other than 26-hournetwork days. Other simple types of subsets could easilybe generated, such as only selected stations or long9continuous Intervals of lonig-period data needed for freeoscillation studies. Mhe Danel roameads thatalternative types of GMI divital data incorements be maderoutinely available to users in addition to the standardnetwork-ah twoa.

S. The problem of which organization is responsiblefor distributing the GMI data should be settled withoutdelay. A new generation of technology is coming on line,and adequate planning is required to assure effective useof the resulting data. The current and future earthquakedigital data are and will be computr-intensive. Userorganizations should help with the planning and shouldshare In the data-handling costs. An with the analogdata, adequate support of data distribution to usersshould be provided ini the contracts and grants made toresearchers.

fthe 9a0el reoAen that GRO data be UMde availableat a National Caster Lor Seismoloical, ftudies and thatIthe ftiimal. Data Centes ontinu to Dimvide Gfun datato Visiting researcher. it is reooguized that ProvidingMe at the regional centers Is only an immediate,partial ansemr to Pesent distribution problems. but Itdoes provide relatively easy acces to the data W omreseacchers ad has the advuitage that new omae cn gainInitial familiarization with the data f rom r"eoalcenter personnel. 3,entually these centers should aliomhave available other standard types of GM3 soe thatnight be geneated, ams as evn tapes.

6. Data from regional as well as global end nationalnetworks must be considered in developing a national,

low,

29

seismological data base. Regioanal networks have almostexclusively used large numbers of inexpensive., narrow-band, low-dynaic-range stations for gathering arrivaltime, duration, and amplitude data for accurate locationof hypocenters and local magnitude determinations. Ifear thauake catalogs produced byregional and localnetwork& should be included -in -the national seiinoloicaldata base. Procedures should be established to monitorthe qualiy of the"e catalogs and provide at least minimmuniformity in location and smagnitude determinationprocedures.

In order to expand the use of regional" networks beyondthe simple cataloging of earthquakes, It Is oesentialthat wavefocm data of high dynamic ranoe and broadf reuenm ontent be satboared frtoo a subset of stationswithin each network (as discussed In q= u1Networks: Problems and _Outlo)ok for the UWO.. 1983).Such data are necessary to extend the understanding ofregional propagation and seismic hazard. Mhe mission-oriented agencies concerned with these problems sholdprovide the funding support necessary to develop andinstall Improved instrumentation. Naveform data fromsuch high-quality broadband stations should beincorporated into a national seismnological data base.

The 2ane recoinnds that daft froM Iftlial and localnetworks should contnue to be collecte. analvaui. andarchived, by indiviul networek opetors be ht Aafavailable In & standard datw-~nw ina fout tothNational Center as Dart of the natonl selemicUncadata base.

7. As discussed in Chapter 3, a high-quality digitaland analog strong-motion data base has been established.but its continued effective maintenance and distributionto users by the ==G and MOC Is problistioa. There-

S. Righ-quality data sets will he gemtated IffANrW's ne m netwook as desoribsi In Appeedim 3. butdata from It will act be Imade WNutinely avaloable.

t. The value of mt saimlogiia &ata doe stdreciate, with time. Large mts of mome here bees

expendI (millions og dolas) by DMis. DO MM@, 9004pMe, nd possibly other agencies, in obtaining seimic

ago___

30

data for special purposes. zasiples of such data arethose froM the UN network, LOSA, and the BOLP networkof stations. Th data from such operations can becOnSId*ed analogous to other valuable, unique worksl thestewards of such unique data should have a moralobligation to preserve it for future use of science andtechnology. A large collection of those unique seta ofdata are contained in Teledyme-Geoteh' la eismic DataAnalysis Center (MW) in Alexandria, Virginiai but It Isscheduled to terminate operations on October 1, 193.The Panel r2qe nds that selected subsets of thesesiecal data ollectLons be kept indefinitely as part ofan active. accessIbl national sismolocal data base.Specifically, with regard to the analog and digitalmagnetic tape data at S1DC, we recend that d italdata subse for al temimMi-c events of L 5. andgret~ or l egn g levents of gL5orr,..ter, be oMerated and vd. Tes data auld be

culled and digital events tapes made from the usaleanalog and digital magnetic tapes. The original record-ings could then be stored in data warteouses, given awayto any Interested organisation, or discarded. The digitalevent tapes should be put into a national semlsoLegicaldata base. he ane gecmends that the load oruanisR-tics fot the Ntial ,eter should have the re&mouMsbil-ity to archive thee data sumets as gat of ad atialageiloeLon] data base. fte costs should bedistrbutean g the cnsortaim of funding agencies listed earlierin this ahapter. apWrtomd according to the Current andproJected future use of the data.

10. Who has what data? That are the storage mediaand formats? Are services, facilities, or specialsoftwr available to deal with the datea? The seed for acentealiaed oaspta-eoessl. dLrectory of aminfomatio is real and Imdiate beasso ms, emists atpresent. Objoetics to a rUal sed servie providingthis kind of LAimaticu at* that the laft m em willnot be complete at aamuate md tbz- the ocgmmisatim addomeat burd m thn o met Ultimtely PrOvidthe Information to the estalised failtywl he I

aoeptc My large. Our view Is that these prl esamoaly gr worse with time, and only, by starting wm illwe bg a hmuae of providing this diretogr inaftationin later Yeats when the volmeo of data wil have iamreased

signifleatly. Intually global. natima. resiosel.and lava catalogs, should be Included.

Are .

31

A first step should be to select an agency or organiza-tiestatI egg~ coiate~ for Wevewawnand proviuingsuch services. One logical Imdiate choice offpears tobe N~aA/=UC. They already prowide many related dataservices and have recent experience with prototypedirectory development. Alteratively. this could be adesignated activity at. the national Center. In initiatingsuch a service, major potential contributors should beapproached to determine, their villingfess to par ticipateand to Indicate, just what Information about their datathey would be willing to furnish routinely. Informationshould not Just Indicate what data exist but should alsoindicaste Intended retetion time. If data are discardedOr becM unusable, this information should also be madeavailable to the data information center. As an incentivethe contributing organizations should be provided with!1muh11k11esju copies of the directory. Ibrecumsods that a oentraliaed data directarr be devpelopedend maintained as on integral Dart of the date base thatIs acceseible to users via roast omuter terminas.This activity should be considered as pert of the missionof 90h/R=D or the National Center in developing andmaintaining a national seisaslogical data base.

11. good, well-planned data management is an essentialelement In effective utilisatios of seismic deat and itis a keystone to future advances In earthquake msmolMg.The problems of data handling result from poor plamninesfrcagmented, uncoordinated responsibility foe differentoperations i and poor f inancial support. These problemsmust be rectified to derive desired societal benfitsfrom the advancing techmologiceal capabilities of thenation and to ensue the health of seismlogy as ascience. The recomendations given above address all the

major issues except the sustained funding Impliit In the

scones at law cost to &U.ta s.sts that -It teiparticular needs. collectively this will be a befItbecause the users cam deve acre effort to the analysisand interpretation, ad sew users will me he disbouragedby the effort required to obtain the -det they want.

12. Theta is a neod for stamiard, digital exchangeformas Itersatiomally, I"M ANS hasfeMelted stamdards

Opp,

32

fog data exchange. fte peael recompds the adoption ofthe IAMS standards for International exchange of dAIitalGata In"e Appendix C).* This currently incorporates the

few different standard foxmats may be acceptable, but nomore than a few should be adapted.

The data dexived from regional networks have beenPreviously described in Chapter 3. fte problem associ-ated with network operations axe discussed In plain-oravbic -Betw .k RU Polqso nOutlook fox- the 1980 a(1983) * There are about 50 regional networks operatingabout 1.600 seismograpbic stations in the United States.The Purpose of the networks is to provide data fundamentalto research On seimotectonic processes and earthquakeOccurrence In the region. There is need for coordinationof regional activities, because the seismic data have useto Other network operators as well as to private userssuch as utility companies. Standardimation of dataformats Or at least data-exchange formats will enhancethe Usefulness of the data. Te panel recomnds thatthe CAMMittee on Seismology take steps to convene aworkshop fog the purposes of establishing standard dataAnd data-exchange formats and standardization of datacatalogs.

13. TMe nnel, recoMmens tht the ational LesearchCouncills COsMIte on Atismame sOuld sIM the Leadorole inl establishing vrl os o h loa-tmretention of dama McurretU bein collted. Nowever,the consortium of supporting organizatins, shouldparticipate in the development of the policy that is I

adqae. this Policy should Include saving indef initelythe NW?., network event tapes, triggered wavefoce datafrom a national network, near-f ield (strcag-metion) data.historical ad current special data sets, and specialdata sets from regional networks.

with the advanoes in data stort&e capaity anddecreasiftg coats Per unlt VOLam of storage, It likely Istechnically feasible to keep iedeflitie3y a larg trac-tion of the data collected to"e Appendix a for ourreatA anticipated data volin) * 200thste data-base-

management costs could be maepalylairge if all thedata at* kept imactive (i aely accessible) massstorage. te guidelimes that the peal suggests forwavetorm data retention ar* as follmws a(a) Keepindefinitely as sek of the recorded event data aspoesibleo but use a 6torage-retrieval archivin architec-tare that moves less-fregmently used data into Odesper.O

Z.

33

less-0ostly storage where It is still acessible but withsm delay. (b) Aply a uniqueness criterion to therecore4d events that are opinedered for deletion, thatis, how coma Is the event, and how many opportunitiesexist to obtain equivalent data if needed. Exasplecandidates for deletion on that basis Include microearth-quakes in areas of high seLmaicity or repeated quarryblasts at the same location. (c) Update the olderarchived data sets to conform to current storage andretrieval formts, that is, older data sets should be asreadily accessible as the newer data.

14. Finally, because of the rapid technologicalchanges in couputational and data-handling capabilities,the fanel reomemends that the Comittee on Seiamoloqreview the Stau of data Problems In SeIsmOLOa n aYeAto-yarC basis to ensure effective utili!ation ofearthuuake seimic data.

-- m. a mm mm im i =- =--

Directory of World Diacital Seismic tationa 1982. WorldData Center A for Solid Daith Geophysics, Report83-32. R. Gan". and C. R. Butt. Boulder, Colorado.

439 pp.Global Bartbausko Nonitorings Its Uses, Potentials, and

Suvmort iMourement%. 1977. Panel on SeimographNetworks, Comittee on Seismology, Assembly ofMathematical and Physical Sciences, National ResearchCouncil. National Academy of Sciences, Washington,D.C. 7S pp.

Soisinoraybic Networks. Problem and Outlook for the1960.o. 1963. Apecot of the Workshop on SeismographicNetworks, Comittee on Seismology, Commiss ion onPhysical Sciences, Mathematics, and Resources,National Research Council. National Academy Press,Washington, D.C. 62 pp.

StraoAH-otion Siartbuake Instrumnent Arrays. 1978.Proceedings of the International Workshop onStrong-Ntio Earthquake Instrment Arrays, May 2-5,1976, Uonolulu, Eaael. Sponsored by the NationalScience Foundation and United Nations Educational,Scientific and Cultural Organization. W. D. Ivanted. California Institute of Technology, Pasadena,Calif. 103 pp.

MeS. Narthuake Obsertorie Rmgcindations Mo a ew-bMAMmLhJIIM. 1960.0 Panel an National, Regional,and local Boismograph Networks, Camittee onSeismology Assembly of Mathematical and PhysicalSciences, National RearcMh Council. National AcaemPrews, Wasington, D.C. 122 pp.

34

S -O

U.S gt=N-t&L zatbieke ntrwientation. 1.981.Proeedngsof he .S.National WorkshoP On

1981.*Oia Z&tqk Intuen~~n April 12-14#198, ant BrbaaCalilognit Iarthquak

NngineeK lug bseatak instituate Universities Councilfor z aske ani wring Rsarebe and at the

II California Institute of 'fechnology.

Il

APPMWIX A:cl ToTMPAIEL On DA PROST lW iU SZN OLOGY

Seismic data, and especially digital seismic data, havebeen increasing rapidly in recent years. Furthermore,data sources are diverse, and data are being generatedunder a wide range of program supported by USGS, NSF,DOD, UOA, MC, and other government and privateinstitutions. In order to obtin maxim=m scientific

benefit from seismic data, their collection, storage,distribution, and analysis require careful planning. ThePanel on Data Problems in Seismoloyg, of the Cosmittee onSeismology, should review the seismic data problem andmake reamnendations for organization, selection,storage, and distribution of the data. Study shouldinclude both existing and anticipated analog and digitaldata from global, regional, and local seismic networksand strong-motion instruments. Seismic profiles, grounddefoemation, and other complementary data sets shouldalso be oamidered. The panl should recommend proceduresand policies to ensure the availability, timely distribu- ,tion, and analysis of these data.

j

34

Clio

--W

2he following oharacterises the magnitud, of the presentand future digital eismic data Waveform. problem forglobal anid regional seismic networks.

te RegioalWoland tst edvanoed (2M) stations arerepesetatve f te sot ftecedanddigital-data

station produces three s9hoetprodhannels sampled at

40 Va, three medivio-band anvnels somled at 4 sps, andthre lon-pe fo cannl toamled aer perS meow.Al

samples are 16 bit, encoded in gaire-rwaged (pedo-

samling rate Is -reduced to 20 ase it Is also possible

to have broadband Instumets, sampled, at a rate.approprilate for short-period data, from which all bandscan be omutationally extracted. But the PAW represets

agood basis on which to ohaeacter ize the data problemfor global metvorks*

Regional or Iowa networks are eTvotae Oriented,toward short patio"s a"d high-frequnmy data. Oftent onlya single Vertical comoent is of Itetest. soots"rated for digital ata are typically In the W0. s"arange* In what follow ne sms that a sile, staticsin a regional de Issal betwek will preduse 100 samlesof 16-it defta ac se"WO

the numer of advmcnsd digital stations avaiolabl*today for Loal~svI in a global. network io an the orderof 30. It is not uareI . -l to asaMs that aft OM tobe as much s 10. The 6asis for this ea~mtio Is areview of IMMU data. A subset of som, 60-100 MS

37

sttin ha roue the v::t majority of all OWSN

arrival times and film data. It seems reasonable toassme hatthe digital network may grm o that sz u

probably no larger.The numer of seismometors in regional and local

networks is very large already, although most are not yetdigital. In the United states alone there ace som* 1,600stations In sufh networks (Stemo-granbic NetworkstProblem* and Outlook for the 1980s, 1983), with perhapshalf of those collected as analog data but processed andhandled in digital form. This processing is done byvarious universities and government agencies, Which actas regional network analysis and processing centers.Based on this we will asme that there might be as manyas 2,000 digital stations at some future time.

Table 3.1 summarizes the magnitude of the digitalseismic wavefocm data problem for global and regionalnetworks. First, It is clear that in terms of bulk ofdata the regional networks represent the largest potentialproblem. Dren after selection of Intervals for rietention.the amount of data is probably beyond what is practicalto retain for long periods of time. Also, the nature, ofsuch networks and their data may not justify saving largeamounts. In the case of global networks, the totalamount of data may be impractical to save using currenttechnology and reasonable funding exetations. But byselecting short-period event windows to save permnently,the quantity of waveform; data is reduced to a feasiblelevel for archiving.

Figure 3.1 presents additional Information concerningseismic data accmulation~ and storage media. the figurecan he used to gain an aggreciation of the ais ofdigital seismic data bases. it shows graphically hwtheeise of the data base Increases with the average smyplerate, the numer of stations, and the time of opeation.For eusap cosidet the global network Statioparameters used is the Tabl 8.1 that tesulted In 10seleated tapes per year. Mtwi ?igme &.1 we se thatthis osereads to en averages *WMg rates of abot 7

GO e-get "at uWio& Is O"Bet.The sawt nubrs of statis.o station aaterls-

tics, tap storage caps.ity, data seLoot"* polLcie, andother related factors wre not iqertant, for tOiSdisussa prewided they are r easomahl. there a twomain Points to be noted$ fIirst* that dvgtteL glob"netwrks wll rd. date qmUttes that are V=MhWl

Ago! oi

tfig

Ad a

B:it

.Alp

.,. I -

323* *:ii a a. Mu:

hiU' Ii

40

Ubrary of SITS*

Library of02 .!P *

1010

101 4Large Disk (600 M~ytes)

Enc~dpedla-0, Sandard MagneticEncycloedia ~Taps (623 BPI)

Inup~suu

10.1 1 10 100

AVERAGE SAMPLE RATE

PIMM 3.1 Cooprative exaeles of data volumes.J

too large for total retention by Currently availablemoo, but judiicious selection should reduce the data Ito

amagablequantities, while oaving almost all data likelYto be of significant scientific valuep seconde thepasestla qtantity of digital data from regional networksis at least an crder Of Weina tuf. Is""er j

IASPEI.1 EA ILDDT

intligsrn oinfa mnsi rcgie ~r a ri eino predicted

earhqukes tad the moin ont pet ablee aryofsonmtin 1963on byt the nmnal~a

egiastuk songieas inthe deloing unweri, e.caywhteh~ fUe

IAPENeon that the I-SE mont pvojer bt unin 1%3 bymu ptica oinvting

poksor frm abroad, seeking national or iniernational Ranub to achiev, this.

lAS~l.13. STORAGE OF MISOMICAL DATA

Riemonzing the importance of reaming old ueamcgraphs. b"amhi. ualublishedlredns cdock correction and calibratona r ao and other uhmeslogic infiiston.

Notng a mth rd t thse ay be los de o bof pew or holites. orfor other rmasns.

Emmcuruqes all stationa "d1*Mt msI MMMt mostorag conditions in order to pmasave these invaluable dim, If neceary msingfinancial and technical help from national or intaioaal sourus.

14. REGINAL SEMINARS

IASPEI.RerogmNi:In the useftulmn Of vloei uutm whic loe been do wed so

specifc simogcltop"cs such n tdit held by CEREWOAS on i menroau- lo1 inLim in November. 1978, and- that hod by de JIME on eusghneing a S3ml inJapan in April 19310.

Noting that these seminars hav particular vahu lor sclnien as devoincountries

ROMA"s t0 encourage na ol d inoauneulani101 be"h 00eutssalrrgoaseminars and sympoeia in and/or fo the developing comenas.

1g0Un I#0ipQS Cblacatc-. no* 152, no6emS ll1.

41

4

42

AP,=Dlz C

15S. DIRECTORIES OFIDIGITAL STATIONS

IASPEI.Noduig the continuing devdopnment of new digitalsiunagph station by many

Urve' that dircte of digial recodin stations be updatd at kamnualy, andthat these directories be nvde avail"l in computer acossible brunt.

16. DIGITAL RECORDING FORMAT

lASM,Neon h variety of des fornmats currmgWy in use for digta deta.Rwftvmwu&ds tlat t &e glba VWgi sesmograph networ day Wae format beadopted es do is" leaded for iunternational daea exchange ald that dewa imin this

formtob made availble for maitrary (umn defined) ~ ant he Irindows,Further, avcoamwen& that one or moe demonstado data tape be devaloped tohep UWev

17. DIGITAL ANALYSIS

lA94ELRecowizb,, that digiAl wR1 ar analysis is a maile proeu..(I'ge that softwaen for stniple type of analsi be made avat"bl to asimnologis

who amt siutivey iexperlienced igit O somOwY

18. DIGITAL DATA EXCHANGE

Rucnwamradae that dWWta saismograuns in an ienhoalyaccepted format beindudes in the Inlernaulonal Dat Exchang dat US.

19. REUOUJTION OlFTHAI,=

IASPE,foade~gt enase Of the 191 Assembly.

naelvilif t, Ia we*k aiii t wast inved in ppmltExprmn its thabksto to Uiftwrsltof W&Wen sabbIn, t Dartmsat of

eophyssm and t 10Wa orgaukin OMntese for the fbel Wlt" w"t wt madeavailale and for all the preporalt wh0 contue to is aselent Autubly.

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ AA

4 -k

YM611 sawn

DAMP is supporting Gsvelcpmt of a now facility Calledthe Center for Seismic Studies for Aisic there are two

neon of WA-sppoc ted rese to iopcoue the U.S.

second is to develop the oability tomttelater-

objetive topces woloom o feted seem sorwe*CMr fg n aaing UMV# fromemo digital data ttita iasdelt

capablity eed tqe.pli md d igmeia Leisytmvi ruledata foe braemVoc&rsarc, objectives. A

advaceddat cter epent tati aed an a networkomibicfoutershasmbo Stehlesan it bein deeopedat

the entr. oumse ~da aItte a oomiio mfromsowragig ro dga~el ath ttea td ia sa aite

Regioal Seismic Te Pst mAwkto p danlataOourafmueia O HOV teunsml at. -t,e tNM. Data

soorces~~~ w-, umI, -"'ma

syte wi cesmlwi dat9tmoeso atWrselctbl crtraa tnp tta o a

Tbe QS.I fo 09"45$ &It t ef~

ostabisbo tobo b'et5P',aSm"l~

ree~dO atM. 4o it'ttiftsaiiatl5, it7, MI

I. 44developed tator of international data exchange aNd

Of the operational prototype to support the researchanInternational daacenter functions is schedule for lt1904.

CGIPUTING AND DATA R3QUIRUUWE

To support these objectives, capability to required tocollect, process, and organize high-quality digital dataanalysis capability is required to support easy displayand manipulation of seisic wavefoma and nonwaveform(mps, focal solution) data. ft e nature of seismic datamakes the problem complex In that It Is recorded onvariouis media including paper records, film, and magnetictape in analog and digital format. Receipt Of datavaries from close to real time to several months andIsometimes years. Paxameter data muse be integrated withw~avor data, aind historical data must be combined withcurrent data to construct crglaised data bases forresearch. a pecotoatyp seismic data veter was designedaNod evslqd to provide these functions and to meetcurrent and eoving research aid dmaeaugo needs.

The required fuctiams ad the neod fwor epasion basebeen satisfied Useoiq the doselpwat ot a distributedcomawuter, spsa roiteatere. A nubr od A miii atacOwMnted by a 10=al *own%" metwork. -owm the mejort

subsytems of th e petotwes Cw11Atoatios Dase,

co"Ps~~ever-t opo M " sys/W amm *.amsA11/44 $"oa.a~ Sit"~ mai4m pqie a e.&e

operaum *ztmos U* s 000vasI am # oi

teoma a40-to "mwmu "new.E 4410 aad do

sv"f":WQ , Se49SA NO

~V

ways to oammiate with the Center for ov9nient a00esand data exchange. These imlude the A Wm . T r,dedicated limes dial-.p psb"Ilty, 1all, and the DmoteSeamic ve dmax. through the oaticas Interfaceresearchers access the Center's data and €ting

The Database amnagement 8Ubsystem receives, organises,and archives all the data at the Center. It is based onIngres, a relatioaal data base sufforted by UIX. Throughthe data bas, requests for data can be fogmalated inseismological tern and efficiently stored and retrieved.Data requests can take the form "get all the data from aparticular seismic regiong and the systen will providethe pareter and waveiorm data in an integrated fernindependent of its storage leatie an disk and tape.The digital date at the Custer Will Aalude all hiStoroal&and future Global Digital Seimie Natowk data onedensedinto archlve formt and refereneod by. event rather thantiN. The Internatimnal Data Collection EagerimantDatabase preqpre4 by Smm de for the GOap of SoiestificKzperts and other selected research data baes will alsobe in the Center data library. The parameter data baseincludes the Natiosal uthqusk I fatation 3rvie andInternatiaenal SeiAmlOgUSI Centre (Me) -oatlgs,parameters received f9m Canadian an OR artra, and datareceived over the Wocld N teorologioal Orgamisati esystem. The entire NI* idJe asimic Statin Neworkfilm library is also at the Ouster.

The heart of the system is the 4ei180e AnalysiSSubsystm, which has been desigled to peovide a suite of

alwoiti for testis large data base as well as thecapability for eay tegratis of ne idea in smicm "signal Processing. Zteraotive signal anlses Ispossible through the we of the gSahios terminals. OAeteminal dislays wavefem IWle the other om di&laynaps, foca ouItifms. ray-treing selstisns, me otheanalysis tqols.

The qp jbUis .ad tools deled ftr the Custefortuei Stooloo ar*el ra gspsttt of ane epupiltdesire& fjor the Ntioal Canter isessed is this imrt.

6I

i 2At'Lem

pK

AIR am

fhe United states Air Yom is developing a mjorlatovnamnt in th. semic segmut of the Atomic AweaDetotio n Syatn (aM).' Ts iampn vemt lnown as theGlobal Sufaesm Iwt t (655) will provide the am with afully Integrated digItal-tefnology-based system capableof detecting and locating dmlib events in the Unio ofSoviet Socialist Republics %tRS5) and the Pe"les smpubllcof China (PW) ahortly after thir occurrece. the systenwill be designed for equakmn to accept and process datafroI additional souros tat will eventually lead toWorldwide coveage and, wben cabined with additional"sytem .IWrOvemnt, will allow preot euplsio

Selected AM@ stations are being upgraded to providean umwm d digital data Collection -A tranesissioncapability (facility) to forward data from an ePadeset of nsOr arays.. Sieumb wavegom data Oollected atthese rete stat ill be transmited throug h -quality digital tramisIsion circuits uIng rro-detection and efrrort euua. tumissinpaths are preomiinte eV c ad of satellite liss,,with the delivety of a data sizauit to la Narth statioat the Control heeUsrta facilty.

A Central darters IDINlea Opeatca et(M0) Is being deeti ta pft,* a hbith e69"patas .tift It the ueIe-tole of " LA6 aei data.Oswt'e-Um-art sigmml ad a Amiatioalgaithnum will petrom the PrlIUt seiaMIC datascreening such that emeriesed seismologists can devotetime to evalmti ts at inte et using advuIutochipis. bitematioI will pr1oide a such beee eventdeoIstife empabIity With le smOW4 than reviou

stems. & dite achv will provide histatlea filesfot subsequen researc or dv~n eft4 its.

: : _ _. .. . . . :, :" • , ' , , . id i

, a -- •-. : : ' . , ' . .:: .

... L " ? ' : " , ' .. .. .,?. .'. . .'. . . , : , j ,' ,i!

2 "',n. . . . .. .. i : ' : ° . . .. .. .

Mbe National Center should ham, at least the fOllOWLngfunctional requirmente.

A. Maintaining am intonational data bass ootainingso, A oomprehemsiye directory of national and

international melanie data soces.o Buletine of hypoasater an" associated pbase

data (Intonation"l Seismological -osatre OWI) v INZS# andgogional and local rpuct))*

o Special desocipthom og eartupaalie (*.I.#swithsonian huletin of oect-livei Ophe-Imes. dining. andgeologic obseuvetimis a" tateasIty doat io*e.. 31).

3. Maintaining a digital an aalog wvotoom database ontalaings

oW data (nomusstday tape) and latest *egrentof oomtlwms data (669g.F M, 91W).

o1 Dnentus eptomet of Acelermeter# (IDA)network data.

0 selected high-vautty data fro& foreign stations.o Matiemal meteark date (ossistlag af selected

static"a Is reglonal atuotie ohat Soeaee d1Igiaroortding high dynmic range, and beed -banwith

* special ""et dota mete, (*S. 0 diial data,and/e date from 4igittuu amaeg - "am" from 4mcest

sto, m ti ing. &mS.uo o 'Id aIlo vngett

* m SUe chi"s orgaI&Oed boah bg sttIcS amdnetwok-day Sequencs.

o Anaog U28a reordlage tae Other historic dataGets IeSg. MB -~Pemem~t ObSuiatorIS).

47

a Near-f bied (sttromg-motion) analog and digital

C. Developing and mintaining an effectivedata-baes-mnagement systam to provide users with desiredseismological informlation and data in a timly Manner.

fteWs hould be versatile to meet the typs andcombinat ofo features s speCified b h aauesD- Providing User services at the =CSas f ollows I

0 Quick look and browse for waveform data.o Preprocessing of data (e.g.. rotation of

COsPOnents, filtering, spectra, roed sections.reduction to a ommon instrumnt resonse).

" A variety Of graphical display choices.o Unmited access by users to c=aters for

Intensive data manipulations and preliminary analyse ofdata.

3.Ixlemeting NeW technDoog and software toimprove the UCS' s capabilities to provide data servicesand products.

?- Providing periodic training mseis to educatenew users.

G. maintaining software neoessaq Se the foLLOiinga" mootin event locatio ndm bulletin, Iemesatlicn." Preprocessing of9 doS"*" selected analytical teabaigme." 2eal-tim data maipultiss gmr eM **etse

and location using a limited amS ag s6&ais.He Distributing data to VOSrs w OMMUa mea .tims.

o Network-day tooe (in stard GsmWt0 Nfetwork event tape (in standard SesMs)o hemote access via terminaso Analog wae~foas0 Agoe*" to reatona dMaa baese0 $PeciL event tape (in stamiad farmato Rnteemationsl data einbemge

z.contismia iWmiomeat og tasesmb seLIOD"is"s.in the a"e of the NUB PoSsibl "cook a visitinscientist program.

17, ismaatig SOe seet ata sets Vensaeat the *e"M data se"s ohm" "e qgsitjjjete t

calibration oe *e@Pame "mail *a *"e fepreprocessing so that ssimmlisa With lam*"eexpelft imeWith digital Peesesm M SWMt Withthe data.

49

9. Creating a comprehensive event data baseconsisting of all global network digital data for eventsof ab 5.5,p plus selected special data sets (e.g.1.mainshok and aftershock sequences) that also Includestrong-motion and regional netwk observations, andother ancillary information oheraotstising the sourceregion.

L. asperiments designed to provide mear-real-timcapaility to determine source characteristics, of largeearthquakes needed for purposes sufh as warnings (e.g.,tsunami),. damage assessment , and deployment of specialInstruentation In the epicentral ares.

K. Flexibility to incorporate now types of event andwavefoem data In the future and to accept special digitalevent data* such as dAIgitsed analog seiamgzsm.

N. Archiving of all daa with etrievel, architecture,structure to acoodate specific user mae andfrequency of use,

0. Developing "An distributing of experimental.msimmio analyis software packapee mmmally usinghigher-order languages such as VQAU. that are easy to

* operate an the mst commn opater ystems extant in theuser oommmnity*

P.* Wide dissintio, to tkA seiGImelagloa aeMMnityof Infoirmation about the NCSS ap iities ad theavailable data bases.

Q. Periodic review sessiomsp or special syxpoia atuatim)/aenation"l inkings to "iscss researchresults and capabilities and to Identify additional, usersee.

at4

ARMS Atomic anergy Detection systemASTAC kir Vooe veobsical AUpplicatiofl CenterAID Agency for Iatoraational DeelopmntALPA Alaskao, Lang Perioad ArrayASL Nobges a Seismological LabogatOf VASCII Mericat Nstloala standard Code for Information

ABDO Abbeited "eionic lesearch ObservatoryCBS center fot Sismic studies (DWWAIOwC Dilital SPIVMiet COr-POAtAORDM Dae hae"" ameatch Pojets- Agenc~yDOD Depar tment Of Defense)DOE Dspatt of 3.ervymum. Oiital Wwld Wift Steaiardised kiOgrqia

13£ federal smargenay I~ma einAt AgencygumN global Digital SeLamgqb U.tuOrkan - Grw of Scientifi 1.inperts

653 Global Selamoa MetwckGo5 Global surf me System

653 Global ftlg"%teed Selamnogaph NotUOrknow bli gains long PeriodshalS Imtsamational Asoslatidh of 80180101y end

Phystals of the Sethle InterimrISe&aisa frei4ymmt ofil leosttI= uetina Data vw**ssacmtamlSiin~in Centre

LAMA lals f.r twoeii Array

30" inatiomsl MreaftfAS sa ftwo AftAsetatleimos liemi cet ai sgelStudies

to

51

URIS National Sarthquake Information ServiceVGDC National Geophysical Data Center (EWA)MOM National Oceanic and Atioepheric AdainisttatiOfl

sw national Regulatory Camsissionam National science Voun4ation35! Remte Seimic TerminalRonU Regional Seismic Test NetworkSOAC Seismic Data Analysis CenteramC Seismic Data Collection Systemspa Saples per seond

Sao Sei Research ObservatoryTOO Technical Operations SubsystemUUIM United Nations Ccimittee on DisarmmsntU908 U.S. Gmeological SurveyODC World Data CenterWHO world Hteorological Organisation3653N Nocz1tif Standardized Seimgraph Network

Swent Datas Intormation such as station phase andampitude reaings, hypocentral locations.magitude, and other semcecharacteristics.

Waveform Data$ Consisting of event signatures, soosavailable c-line in real tMinS bdt Others,Such aS ODI network-day tapes. delayed bydays to weeks.

-4

mo ME ff p