thepuertoricotsunamiwarningandmitigationprogramsiaiacad09.univali.br/ics2004/arquivos/332_mercado.pdf ·...

The Puerto Rico Tsunami Warning and Mitigation Program

INTRODUCTION

LANDER

JANSMA MATTIOLI

SCHWAB

GRINDLAY

BRINK SMITH

MERCADO MCCANN

MCCANN

MERCADO MCCANN GRINDLAY

MERCADO LYNETT LIU MERCADO

(2002) reported that since 1498 there havebeen 91 “tsunamis” in the Caribbean region. Among them,Lander . has verified that twenty-seven are true, verifiabletsunamis and nine additional wave events are considered to bevery likely true tsunamis. In fact, in the Caribbean Sea regionwe have all of the potential tsunami-generating sources:submarine earthquakes, subaerial or submarine landslides, andunderwater explosions. According to and(2003), Puerto Rico and the northern Virgin Islands sit within abroad zone of deformation between two larger plates: the NorthAmerican to the north and the Caribbean to the south. Thisseparate microplate moves approximately 15 mm/yr relative toNorth America and 3 mm/yr relative to the Caribbean. Offshorefaults along which major earthquakes can occur have beenidentified in the boundary zone of the North American andCaribbean plates. The landslide-tsunami threat for the islandcomes from the existence of the Puerto Rico Trench just north ofthe island, along which there is evidence of a large slump, with a

potential volume of approximately 900-1500 km ( etal., 1993; , 1998) and many other smaller ones (ten

and , 2003).It is not strange then that the northeastern corner of the

Caribbean, the region encompassing the islands of Hispaniola,Puerto Rico, and the United States and British Virgin Islands,has experienced three destructive tsunamis in the last 136 years:1867, 1918, and 1946. All three tsunamis affected the island ofPuerto Rico, with the one in 1918 doing great damage along thewest coast of the island ( and , 1998). This“forgotten hazard” has led state and federal agencies to sponsora series of workshops and projects with the purpose ofevaluating the magnitude of this threat. The University ofPuerto Rico Sea Grant College Program and the Puerto RicoState Emergency Management Agency sponsored a CaribbeanTsunami Workshop in 1997 and a series of studies (and , 1997; , 1998; , 1998;

, 2001; and , 2002; , 2002)that set the stage for the Puerto Rico Tsunami Warning andMitigation Program (PRTWMP). The two years long PRTWMPwas funded by the USA Federal Emergency ManagementAgency and the University of Puerto Rico. It consisted of sixtasks: 1) preparation of tsunami flood maps for the wholeisland; 2) education about this "forgotten hazard" in theCaribbean (including a video, tsunami drills, workshops and theinstallation of tsunami warning signs along beaches); 3) local(Puerto Rico) and regional (Caribbean) seismic wave formanalysis for rapid determination of earthquake sourceparameters; 4) development of a tsunami warning andadvisories protocol for Puerto Rico and the Caribbean region; 5)preparation of a PC-Windows based Atlantic and CaribbeanHistorical Tsunami Database; and 6) participation in themeetings of the USA National Tsunami Hazard MitigationProgram.

It is the purpose of this presentation to summarize the mostimportant results from each task and the impact it has had inraising the awareness about this latent coastal hazard.

For the preparation of the tsunami flood maps we first had tocharacterize the system of local faults. For this, data recorded bythe Puerto Rico Seismic Network (PRSN) was used. Thisincluded obtaining the basic parameters that control the tsunamitectonic generation: quake moment (Mo), fault dimensions andrupture geometry. Significant progress in the understanding ofthe geologic and tectonic environment of the Puerto Rico regionhas been made during the past years. Details of the seismogeniczones are being identified and the concept of a block for PuertoRico and the Virgin islands has been established. However,there remain a number of fundamental questions. The two mostimportant ones are the relationsh

et al.

et al

3

PROGRAM TASKS

Task 1: Tsunami Flood Maps

the geologic structures and the stress regime in relation to theip between the seismicity and

Journal of Coastal Research SI 39 1582 - 1586 ICS 2004 (Proceedings) Brazil ISSN 0749-0208

A. Mercado-Irizarry†; C. von Hillebrandt‡ and V. Huerfano‡

†Dept. of Marine Sciences,University of Puerto Rico, Mayaguez,00681-9013, Puerto [email protected]

MERCADO-IRIZARRY, A.; VON HILLEBRANDT, C. and HUERFANO, V., 2006. The Puerto Rico tsunamiwarning and mitigation program. Journal of Coastal Research, SI 39 (Proccendigs of the 8th International CoastalSymposium), 1582 - 1586. Itajaí, SC, Brazil, ISSN 0749-0208.

Arecent compendium of the history of tsunamis in the Caribbean reports that since 1498 there have been 91 tsunamisin the Caribbean region. In fact, in the Caribbean Sea region we have all of the potential tsunami-generating sources:submarine earthquakes, subaerial or submarine landslides, and underwater explosions. Past and recent studiessupport the interpretation that Puerto Rico and the northern Virgin Islands sit within a broad zone of deformationbetween two larger plates: the North American to the north and the Caribbean to the south. Offshore faults alongwhich major earthquakes can occur have been identified in the boundary zone of the NorthAmerican and Caribbeanplates. The landslide-tsunami threat for the island comes from the existence of the Puerto Rico Trench just north ofthe island, along which there is evidence of a large slump, with a potential volume of approximately 900-1500. ThePuerto Rico Tsunami Warning and Mitigation Program was funded by the USA Federal Emergency ManagementAgency and the University of Puerto Rico with the purpose of addressing the tsunami threat and concern in theisland. It consisted of six tasks: 1) preparation of tsunami flood maps for the island; 2) education about this"forgotten hazard"; 3) local (Puerto Rico) and regional (Caribbean) seismic wave form analysis for rapiddetermination of earthquake source parameters; 4) development of tsunami warning and advisories protocols; 5)preparation of a PC-Windows basedAtlantic and Caribbean Historical Tsunami Database; and 6) participation in themeetings of the USANational Tsunami Hazard Mitigation Program.

ADDITIONAL INDEX WORDS: Natural hazards, Caribbean.

ABSTRACT

‡ Puerto Rico Seismic Network, Department ofGeology, University of Puerto Rico, Mayaguez, [email protected]@midas.uprm.edu

Journal of Coastal Research Special Issue 39, 2006,

tectonic processes in the Local Zone of Puerto Rico (LZPR).The northeastern Caribbean region, encompassing

Hispaniola on the west and the Leeward Islands on the east(which defines our LZPR), was divided into 12 subregionsaccording to major tectonic features, seismic distribution andstress pattern: 19° Faults,Anegada Passage, Eastern DominicanRepublic,

Earthquake location programs like PR-HYPO, the one usedin the PRSN, need a realistic crustal model. and

(1994) proposed a flat homogeneous P-wave modelfor Puerto Rico and from Wadatti plots the Vp/Vs ratio (P waveto S wave velocity ratio) was determined. The complete PRSNseismic catalogue was recompiled with this model so that allsolutions are based on the same methodology. A criteria used todetermine the data quality is the network stability. Time history

analysis indicates that from 1992 to present the data obtained bythe PRSN remains mostly stable as evidenced by the number ofreadings and location errors. During the time period fromJanuary 1987 to June 2002, 4422 microseismic records from thePRSN catalogue with quality A, B and C (local scale) wereselected to determine the seismic potential in the LZPR. Qualityfactors are based on the number of readings, rms, hypocentralerrors and azimuthal coverage (gap). These factors allows us toidentify reliable solutions inside a specific region.

Based on the offshore seismic distribution of shallowearthquakes (depth less that 50 km) and mapping of known faultsystems, the 12 sub-regions were proposed, each one defined byspecific stress parameters. Each region was previouslyanalyzed by classic seismological methods like the Gutenbergand Richter b-value, Wells and Coppersmith relationships,composite stress patterns and the hypo-centering algorithm( , 2003). The Puerto Rico Trench (PRT) wassubdivided into eastern and western sections due to intrinsicdifferences in seismicity and tectonics. Finally wecharacterized 238 spatially linear potential faults, each onegenerating their hypothetical earthquake. These sources wereconsidered as local (relative to Puerto Rico) and it was foundthat the linear faults generate bigger tsunamis compared withthe 102 spatially nonlinear or segmented faults which were alsoconsidered ( , 2003).

Based on the 340 potential faults determined under themethodology described above, a simulation was made of eachone of them using the Japanese non-linear shallow watertsunami TIME model. Three nested grids were used, startingwith the outer grid with a cell size of 27 arc seconds, followed bythe intermediate grid with cell size of 9 arc seconds, and theinner grid with a resolution of 3 arc seconds. The region coveredby Figure 1 shows the extent of the lower resolution grid, whilethe two rectangles show the location of the intermediate andhigher resolution grids. It is in the higher resolution grid that themodel is run in its non-linear mode, allowing for thecomputation of coastal flooding. The induced sea bottomdeformation was determined for each one of the potential faultsusing the and (1971bathymetry, recently acquired SHOALS bathymetry was used

Leeward Islands, Mona Passage (identified asMcCann's faults), Mona Canyon, Muertos Trough, NorthPlatform, North Puerto Rico Trench, Puerto Rico West toSoutheast, Septentrional and Sombrero (see Figure 1). Theseismicity associated with faults in the LZPR are currentlylocated using Geiger's method ( and , 1995), thatis a linearization of the travel time equation in a first orderTaylor series that relates the difference between the observedand theoretical travel times to unknown adjustments in thehypocentral coordinates through the partial derivates of traveltime with respect to unknowns. There are some technicsavailable to improve the quake locations with the goal ofidentifying structural details, such as the location of activeplanes, into well defined focus. But in general, all methodsshow a diffuse picture of seismicity. A very importantobservation is that the spatial distribution of smallerearthquakes delineates areas in which larger earthquakes arelikely to occur ( and , 2000), but this requires avery reliable earthquake location method to delineate faultsystems. Based in the hypothesis that any quake that occurs isdue to the fracture along a specific fault, and that multiple andsuccessive quakes delineate the fault plane, (2003)proposed a complete set of faults using two methods. The first isbased on the composite focal mechanism analysis while thesecond is based on geometrical analysis of hypocentersstratified by magnitude.

LAY WALLACE

KAFKA LEVIN

HUERFANO

HUERFANO

BATAILLE

HUERFANO

HUERFANO

MANSINHA SMYLIE ) method. For nearshore


Figure 1. Location map of LZPR (see text).


1583

for deeper shelf waters National Ocean Survey data was used,and for deep water we used the latest and(1997) bathymetry. For topography use was made of the USGSDigital Elevation Model data for the island, with a resolution of30 meters. The tsunami model and the data have been shown toprovide good estimates of the observed runup due to the 1918Puerto Rico tsunami ( and , 1998). Thisreference can also be consulted for further details about themodeling aspects.

Adopting a standard hazard mitigation approach ofpreparing for the worst, the maximum flooding, irrespective oftime of occurrence, at each computational cell for eachscenario, was saved. Next, the maximum of the maximums(MOM) was obtained at each computational cell by goingthrough each one of the 340 runs and picking out the maximumamong all the runs. The maps prepared are based on the MOM.As in the simulation of the 1918 tsunami, a constant Manningcoefficient of 0.025 was utilized.

The inland limits of the flood was overlaid on top oftopographic maps and recently acquired IKONOS 1-mresolution color photos and made available through papercopies, CD's, and the Internet

Together with the preparation of the tsunami flood maps, theeducational phase of the projct is the most important one.Caribbean tsunamis (with the exception of the 1755 Lisbonearthquake) have all been local events. As such there is notmuch a tsunami warning system can do, and by having citizenswith knowledge about this hazard is the only way we can insurethat the risk is minimized. For this purpose workshops were

given throughout the island, in which representatives of variousgovernment agencies participated. This was accompanied bythe distribution of a Spanish language tsunami video to allschools, government agencies, and interested citizens.Newspaper, radio and TV shows were contacted in order toreach the widest audince possible. Two tsunami drills werecarried out at two public schools. Tsunami warning signs wereprepared to be placed at all coastal areas which, according to theflood maps, are potentially vulnerable. For these, theinternationally accepted figure and message for tsunamis wasused. Finally, a WEB page was developed

With the goal of establishing a Caribbean Tsunami WarningSystem for Puerto Rico, local and regional seismic wave formanalysis algorithms were adapted to the Caribbean region. At-risk regions need real-time determination of earthquake sourceinformation to assess the nature of the hazard in order tooptimize emergency response. In 1998 FEMA and theUniversity of Puerto Rico funded a project to upgrade theinstrumentation of the Puerto Rico Seismic Network (PRSN) ofthe Department of Geology of the University of Puerto Rico atMayaguez. As part of this project, 9 digital telemetered broad-band stations were installed on the main island of Puerto Ricoand offshore islands. The digital waveforms are recorded andsaved at the facilities of the PRSN. Also, since 1998 the PRSNoperates the data center for the Middle-America SeismographCthe routine computation of earthquake source parameters were

SANDWELL SMITH

MERCADO MCCANN

Task 2: Education

Task 3: Local and Regional Seismic WaveformAnalysis

(http://poseidon.uprm.edu).

(http://poseidon.uprm.edu).

onsortium (MIDAS). Currently available methodologies for

Mercado et al.-Irizarry

Figure 2. Location of earthquake and tsunami historical data, based on CHTDB.


1584

modified and adopted for use in the Caribbean region based onregionally and locally-recorded seismic waveforms.

An analysis of digital broadband waveforms from the PuertoRico Seismic Network (PRSN) for local analysis and theIncorporated Research Institutions for Seismology (IRIS)Digital Management Information Center for regional analysis -is carried out to provide earthquake-faulting information innear-realtime. If possible, PRSN data will be supplementedwith realtime broadband waveforms recorded by GlobalSeismograph Network (GSN) stations in the Caribbean and datasubmitted to the MIDAS Data Center.

For the regional analysis, the techniques developed byAmmon, Randall and Owens ( ., 1995) wereimplemented to perform a regional moment-tensor inversionusing all three components (vertical, transverse, and radial)available from the digital seismograph stations. An analogousmethodology, now based on . (2001) wasadapted for local analysis. In both methods, an inversionscheme is applied to identify the focal mechanism that bestreproduces the seismic waveforms observed at regional andlocal distances (less than 12 degrees). For regional analysistheoretical waveforms are computed by applying a reflection-matrix technique that uses a prescribed crustal-velocity model.For local analysis theoretical waveforms are computed byapplying the (1991) method that uses a prescribedcrustal-velocity model. Prior to inversion, the observed recordsare deconvolved to remove the instrument response.

The application of these moment-tensor inversion methodsin a semi-automatic manner should provide source parametersin near-realtime for earthquakes in Puerto Rico and theCaribbean.

We have also adapted to the Caribbean region the“Earlybird” system developed and used in the West Coast andAlaska Tsunami Warning Center. This package provides rapidearthquake location and magnitude and will provide themoment-tensor solution.As part of this the PRSN has acquired adisplay system, composed of 9 PC monitors under one CPU,allowing the acquisition and display of waveforms, earthquakelocation on a local and regional basis and iterative waveformanalysis, all on separate displays.

These activities are in harmony with the long term plans ofNOAA to establish a Caribbean Tsunami Warning Center atMayaguez.

Once a felt earthquake has been characterized in terms ofmagnitude, location and faulting mechanism and itstsunamigenic potential has been determined, this informationneeds to reach the affected communities as fast as possible.Warning protocols need to be developed at the municipal, state,and federal levels, and also within the rest of the Caribbeancommunity. The key agencies in the Pacific states for warningmessages and emergency response coordination are theNational Oceanographic and Atmospheric Administration(Pacific Tsunami Warning Center, West Coast and AlaskaTsunami Warning Center and the National Weather ServiceField Office, NWS), the USGS regional seismic networks, andstate and local emergency response agencies. A similar setupwill be established in Puerto Rico and the US Virgin Islands, andwill include the Puerto Rico Seismic Network (PRSN) of theUniversity of Puerto Rico, NWS, and the Puerto Rico StateEmergency Response Agency. This will also involve theparticipation of Puerto Rico's Warning CoordinationMeteorologist - WCM - in the regular Tsunami WarningCoordination Meteorologists Training Workshops held on theWest coast; the Pacific states WCM's have now includedtsunamis as part of their coastal hazard communitypreparedness activities. The WCM's have been designated asthe point of contact between the States and NOAA.

For this purpose, a “Tsunami Protocol Meeting” was held atthe Mayagüez Campus of the University Puerto Rico during the15-16th of January 2003. The objective of this meeting was toset the groundwork for a tsunami warning system capable of

rapid dissemination of information throughout the Caribbeanon potentially tsunamigenic events in Puerto Rico and theVirgin Islands.

A comprehensive earthquake and tsunami database,Historical Tsunami Database for the Atlantic has beendeveloped containing the written history of earthquakes andtsunamis that have occurred in theAtlantic Ocean, including theCaribbean Sea. The database includes all of the earthquakeinformation available, while the tsunami information is basedon the comprehensive compilation by (2002).

This database contains a PC-based Graphical User Interfacefor ease of visualization, and it was developed under contract atthe Novosibirsk Tsunami Laboratory, Russia. The wholepackage is contained on CD, but it can also be run on-line at

Figure 2 shows a plot of thetsunami runup and earthquake data available for the Caribbeanregion. A bar is shown at each location where runups have beenmeasured and the bar height is proportional to the runupelevation. The same information can also be listed as a table forprinting.

At the national level there is a National Tsunami HazardMitigation Program (NTHMP), which presently includes thestates of Washington, Alaska, Oregon, California, and Hawaii.This program is sponsored by FEMA, NOAA, USGS, and stateco-sharing. During the PRTWMP we were able to attendvarious meetings and workshops sponsored by the NTHMP andwe hope to be able to become a member in the near future.

Based on the amount of feedback obtained from relevantgovernment agencies (for example, Education Department,Puerto Rico Planning Board), private industries, civicorganizations, and the public in general, we can conclude thatwe have reached our most important goal of raising thegovernment and public awareness of this forgotten hazard.

This project was supported by the USA Federal EmergencyManagement Agency (through Commonwealth of Puerto RicoGovernor's Authorized Representative of the Office ofManagement and Budget) and the University of Puerto Rico.The Sea Grant Program of the University of Puerto Rico and thePuerto Rico State Emergency Management Agency alsocontributed to the success of this program. The authors wouldlike to acknowledge the significant contribution of Mr. HarryJustiniano (System Manager at the Physical OceanographyLaboratory) in all phases of this program. Other participants inthe program were Dr. Carlos Mendoza (seismologist),

M., 1981. A simple method to calculate Green'sFunctions for elastic layered media. Bull. Seism. Soc. Am.,71, 959-971.

N., 1998. Volume and density approximations ofmaterial involved in a debris avalanche on the south slope ofthe Puerto Rico Trench. Technical Report. Sea GrantCollege Program, University of Puerto Rico, 13 p.

V., 2003, Susceptibility of Puerto Rico Due toLocal Tsunami Effects. Mayaguez, Puerto Rico: Universityof Puerto Rico, Ph.D. thesis, 108p.

RANDALL

ZAHRADNIK

BOUCHON

LANDER

BOUCHON,

GRINDLAY,

HUERFANO,

et al

et al

et al.

Task 4: Warning Messages and Protocol

Task 5: Development of a Caribbean HistoricalTsunami Database (CHTDB)

Task 6: Participation in the USA NationalTsunami Warning and Mitigation Program

CONCLUSIONS

ACKNOWLEDGMENTS

LITERATURE CITED

Dr.Havidan Rodriguez (social scientist) and Ms. Maritza Pagan(AdministrativeAssistant).

http://omzg.sscc. ru/tsulab/.



1585

HUERFANO, BATAILLE

JANSMA, MATTIOLI,

KAFKA, LEVIN,

LANDER, ; WHITESIDE, LOCKRIDGE,

LAY, WALLACE,

LYNETT, LIU,

MASINHA, SMYLIE,

MCCANN, MERCADO,

MCCANN,

MERCADO, MCCANN,

MERCADO,

MCCANN,

RANDALL, AMMON, OWENS,

SANDWELL, SMITH,

SCHWAB, W. C.; DANFORTH, W. W.; SCANLON, KMASSON,

TEN BRINK SMITH,

ZAHRANDNK, J.; JANSKY, PAPATSIMPA,

V. and , K., 1994, Crustal structure andstress regime near Puerto Rico: preliminary locations ofearthquakes recorded near Puerto Rico, Puerto RicoSeismic Network, Seismic Bulletin, January-December,Univ. of Puerto Rico, Mayaguez, Geology Department, 9 p.

P. and G., 2003. Diffuse extension acrossan active faulting within the Puerto Rico and northernVirgin Islands microplate: GPS geodetic results from 1994-2002. Abstract. Seismological Soc. of America AnnualMeeting, San Juan, P.R.

A. and S., 2000, Does the spatial distributionof smaller earthquakes delineate areas where largerearthquakes are likely to ocurrs?, 724-738.

J. F. L. S. and P.A., 2002.A brief history of tsunamis in the Caribbean.

, 20(2), 57-94.T. and T. C., 1995. Modern Global

Seismology. California:Academic Press, 521 p.P. and P. L.-F., 2002. “A Numerical Study of

Submarine Landslide Generated Waves and Runup,”., 458, 2885-2910.

L. and D. E., 1971. The displacementfields of inclined faults. ., 61(5),1433- 1440.

W. and A., 1997. Analysis of thetsunami hazard of northwestern Puerto Rico. EarthquakeAdvisory Board, Puerto Rico State Civil Defense. EarthScientific Consultants, 38 p.

W., 1998. Tsunami hazard of western Puerto Ricofrom local sources: characteristics of tsunamigenic faults.

Technical Report. Sea Grant College Program, UPR, 88 p.A. and W., 1998. Numerical simulation

of the 1918 Puerto Rico tsunami. , 18,57-76.

A., 2001. Determination of the tsunami hazard forwestern Puerto Rico from local sources. Technical Report.Sea Grant College Program. University of Puerto Rico, 154p.

A., 2002. Investigation of the potential tsunamihazard on the north coast of Puerto Rico due to submarinelandslides along the Puerto Rico Trench. Technical Report,Puerto Rico State Emergency Management Agency and SeaGrant College Program, Uinversity of Puerto Rico, 245 p.

E., C. and T., 1995, Momenttensor estimation using regional seismograms from aTibetan Plateau portable network deployment,

D. T. and W. H. F. (1997). Marine gravityanomaly from Geosat and ERS-1 satellite altimetry.

, 102, 10039-10050.. and

D., 1991. A giant submarine slope failure on thenorthern insular slope of Puerto Rico. , 96,237-246.

, U. and S., 2003. High-resolutionbathymetric map of the Puerto Rico Trench: Implications forearthquake and tsunami hazards. Abstract. SeismologicalSoc. ofAmericaAnnual Meeting, San Juan, P.R.

J. and K., 2001. Focalmechanisms of weak earthquakes from amplitude spectraand polarities. ., 158, 647-665.

Bull. Seismol. Soc.

Science ofTsunami Hazards

Proc.Royal Society of London A

Bull. Seismol. Soc. Am

J. Natural Hazards

Geophys.

J.Geophys. Res.

Marine Geology

Pure and Applied Geophys

Res.Lett., 22, 1665-1668.

Mercado-Irizarry et al.


1586

thepuertoricotsunamiwarningandmitigationprogramsiaiacad09.univali.br/ics2004/arquivos/332_mercado.pdf ·...

Documents