TECTONICS, VOL. 19, NO. 6, PAGES 1021-1037DECEMBER 2000
Neotectonicsof Puerto Rico and the Virgin Islands,
northeastern Caribbean, from GPS geodesy
Pamela
E. Jansma
•'2,GlenS.Mattioli
• AlbertoLopez
• Charles
DeMets
3
TimothyH. Dixon4,PaulManns,andEricCalais
6
Abstract. The boundary between the North American and
Caribbeanplates is characterizedprimarily by left-lateral motion along predominantlyeast-weststriking faults. Seismicity
and marine geophysical survey data are consistent with at
least two, and possibly three, microplates in the diffuse
boundaryzone in the northeasternCaribbean:(1) the Gonave,
(2) the Hispaniola,and (3) the PuertoRico-northernVirgin Islands (PRVI). We discuss results fi'omGPS geodetic measurementsacquiredsince 1994 to test the microplatehypothesis, define PRVI translationand rotation within the boundary
Puerto Rico trench and offshore faults north of Puerto Rico.
Differences in GPS-derived velocities from Hispaniola and
PRVI yield east-westextensionacrossthe N-S trendingMona
rift of a few millimetersper year when estimatedelastic strain
accumulationeffectsalongthe north Hispaniola deformedbelt
and the Septentrionalfault zone are considered. The opening
rate impliesan ageof the Mona rift of 2-3 million years,agreeing with marinegeophysicaldatathat supporta young age for
the structure.
zone, and constrain PRVI neotectonics. GPS-derived veloci-
ties are analyzed with respect to both North American and
Caribbean plate referenceframes. Integrated displacements
acrossPRVI are limited to a few millimetersper year, consistent with a rigid PRVI and permitting calculation of an average velocity for PRVI. The motionsof PRVI relative to North
America and the Caribbean are 16.9+1.1 mm/yr toward
N68øE+3ø (lo)and 2.4ñ1.4 mm/yrtoward S79øWñ26ø (lo),
respectively.In contrastwith somerecentmodels,ongoingrotation of PRVI about a nearby (< 25ø distant) vertical axis is
not supported by the geodetic data. In addition, we argue
against eastwardtectonic escapeof PRVI and favor a simple,
progressive increase in velocity across the plate boundary
zone, requiringthat the summedmagnitudeof strike-slip fault
slip rates will equal the total plate motion rate between the
Caribbean and North
America.
GPS data are consistent
with
components
of left-lateralstrike-slip faulting along the Muertos trough south of Puerto Rico and shortening acrossthe
Puerto Rico trench. Comparisonof GPS velocities for PRVI
with respect to North America with total North AmericaCaribbean relative motion suggestsup to 85% of North
American-Caribbeanplate motion is accommodatedby the
tDepartment
of Geology,
University
of Puerto
Rico,Mayagtiez.
2Nowat Department
of Geosciences,
University
of Arkansas,
Fayetteville.
3Department
of Geology
andGeophysics,
University
of Wisconsin,
Madison.
4Rosenstiel
School
of MarineandAtmospheric
Sciences,
University
of
1.
Introduction
Tectonic models for the northern Caribbean [e.g., Byrne et
al., 1985; Mann et al., 1995] propose active microplates
withinthe plateboundaryzoneon the basisof geologicand
earthquake
evidence,
but no previousstudieshaveattempted
to usegeodeticdatato isolatemicroplate
motionin the region.
In thispaper,we applyGPSgeodesy
to testfor thepresence
of
an independentlymovingPuertoRico-northern
Virgin Islands(PRVI) microplate
andto defineits translationand rotation rateswith respectto the larger,adjacentNorth American
and Caribbean plates.
The boundarybetweenthe North Americanand Caribbean
platesis characterized
primarilyby left-lateralmotionalong
predominantly
east-west
strikingfaults(Figure1). In thewest
the structureis relatively simple,consistingof the Swanand
Oriente transformfaults, which define the E-W trending Cay-
mantroughandboundthe short(-100 km), N-S trendingmidCaymanspreadingcenter. In contrast,the easternhalf of the
boundaryin Hispaniola,
PuertoRico andthe Virgin Islandsis
a complexdeformation
zone -250 km wide, whosenorthern
and southernlimits are definedby the PuertoRico trenchand
the Muertostrough,respectively.Threeproposedmicroplates
lie within this diffuseboundaryzone(Figure 1). From west to
east,thereare (1) the Gonave[Mannet al., 1995], (2) the His-
paniola[Byrneet al., 1985],and(3) the PuertoRico-northern
Virgin Islands(PRVI) [Massonand Scanlon,1991]. Sucha
microplate
modelassumes
thatnearlyall of the deformation
associated with North
Miami, Miami, Florida.
America-Caribbean
motion is concen-
tratedalongthe faultsthat boundthe threerigid blocks: the
Oriente,Septentrional,
Enriquillo-PlantainGarden,andAnegada faults,the Muertostroughand North Hispaniola de-
'•Institute
forGeophysics,
University
ofTexas
atAustin.
6Geosciences
Azur,Centre
National
deRecherches
Scientifique,
Valbonne, France.
formedbelt, andthe Mona rift faults northwest of Puerto Rico
Copyright2000 by the AmericanGeophysicalUnion.
(Figure 1). Bothrotationabouta nearbyverticalaxis [Masson
and Scanlon, 1991] and tectonicescapeto the east [Jany et
al., 1987]havebeenproposed
for PRVI. A rigoroustestof the
Papernumber 1999TC001170.
0278-7407/00/1999TC001170512.00
1021
1022
JANSMA ET AL.: NEOTECTONICS OF PUERTO RICO AND VIRGIN ISLANDS
a)
25øN
ß
ß
.
;"X, 1}.."X,•<.,
,•iiti!%....:,
'
. :.:
...• -' u
- . ...-...
20øN
North
America
.ß
.o,,,:,
o. ,,,,, ,.
,
!;' ,
?.,,
.•...'..I
15øN
r.,•'ii..'
-"..'•:,'-•!
--''-'½.;:..
-j!;.?.:•
'•"
"""•:•••••f
'"'"'
'•"•""••"
''"'"•'';'•'•,-'--•-•'•4•'
.•ß
Caribbean
-,,.,;..½..,?.:?•,,.•--.,-'---•,,---•••:•••:..,....,....,•i•i,'*:•:•'.'•!t:..,.:
..
.
'
".:.:.*.' ':;" :'""-:-:•.?-5:gj;,'•!
•'?-•j•i:'
.. .:
10øN
ß - :;.
II
5øN
ß
.- .' ':',
't•?•'.'.•l•
..... '
.,'
-
: -,,
. .., .
65øW
60øW
ß -.'
II Cø,,øs
:. .
90øW
b)
AVES
'•.'"
ß'
,'..
- ' •.: .... '*" .............
•:•.'.Ei
-• '• :,"•*'•
85øW
80øW
75øW
70øW
25øN
.•.--.•,
•..-.-X•...,
_
....• '•'. ,.• ,.•;•..,••••.•D
20øN
OF
,'*
.•,5
15øN
90øW
North
America
85ow
••........... ..... •
.... ' •
B
Caribbean
.- -':' ..,.':
ß
.
HR..- -'" "'
j'BR;
. : ..'..
80ow
75ow
70ow
•
PRT
•
SB
MT
•
'.:•
•2'
[•
•g
65ow
60ow
Figure 1. (a) Map of Caribbeanplate and regionalseis•nicity.Epicentersare for earthquakesabove depths of
60 km with magnitudes> 3.5 from January1, 1967, until April 28, 1999 (USGS). The four GPS sites used to
constrainCaribbeanreferenceframe are plotted as stars: SANA, San Andres Island; ROJO, Cabo Rojo. Dominican Republic; CRO1, St. Croix, US Virgin Islands;AVES, Aves Island. Original Caribbean-North America Plate Experiment (CANAPE) GPS sites are ROJO and CRO1 plus the five squares:CAPO, Capotillo.
Dominican Republic; FRAN, Cabo FrancesViejo, Dominican Republic; GTMO, GuantanamoBay. Cuba:
ISAB, Isabela,Puerto Rico; TURK-Grand Turk, Turks and Caicos. (b) Map of northern Caribbean plate
boundaryshowingmicroplatesand structures.AP, Anegadapassage;BP, Bahamasplatform; BR, Beata ridge:
CT, Caymantrough spreadingcenter; EPGF, Enriquillo-Plantain Garden fault; GP, Gonave platelet: HP.
Hispaniola platelet; HR, Hess rise; LAT, LesserAntilles trench; MR, Mona rift. MT, Muertos trough; NHDB.
north Hispaniola deformedbelt; OF, Oriente fault; PRT, Puerto Rico trench; PRVI, Puerto Rico-Virgin Islandsblock; SB, Sombrerobasin; SITF, Swan Islandstransformfault; SF, Septentrionalfault; SPRSF. south
Puerto Rico slope fault; WF, Walton fault.
microplate hypothesis for the northeastern Caribbean, however, has not been conducted to date.
Global Positioning System(GPS) geodesy is a powerful
tool that can be usedto assessthe presenceof blocks and microplates and their independent motion [e.g., Dixon et al.,
2000] but, until recently,was limited in its applicationto
plate boundarydeformation
in the northernCaribbeanby insufficient data and lack of a well-defined
Caribbean reference
frame. Theselimitationshavebeen steadily reducedby acquisition of GPS data at continuous
stations installed
in the
JANSMA
ET AL.: NEOTECTONICS
OF PUERTO RICO AND VIRGIN
ISLANDS
1023
northernCaribbeanbetween1995 and 1997, throughyearly
campaigns
since 1994, which capitalizedon initial occupationsat a few sitesin 1986[Dixonet al., 1991, 1998],andby
the development
of a Caribbeanplate reference
framefromG PS
geodeticmeasurements
in the Caribbeaninterior [DeMets et
al., 2000]. Thesenewdataallowusto testtherigidityof a po-
19.4ñ1.2 mm/yr and yields a revised azimuth toward
N79øEñ3ø (1o), in agreementwith our earlier estimate within
tential PRVI block and to constrain the kinematics of the
tional undocumented
northeastern
cornerof the Caribbean
plate. We analyzeGPS-
[Leroy and Mauffret, 1996]. Recent GPS-derived velocities
derived velocitieswith respectto both North Americanand
Caribbeanplate reference
framesto assess(1) rigidity of the
relative
one standard error in rate and two standard
errors in azimuth.
Although the site at Cabo Rojo in the Dominican Republic is
south of the southernmoststrike-slip fault in the northern Caribbean plate boundary zone, the possibility exists of addi-
to North
deformation
America
in the offshore farther south
from the interior
of the Caribbean
plate at San Andres Island in the west and Aves Island in the
east,however,are similarto the velocity at Cabo Rojo within
error in both rate and azimuth [DeMets et aL, 2000], supportNorth American-Caribbean
convergenceacrossboth the east- ing the assumptionthat the southernDominican Republic is
ern PuertoRico trenchandMuertostrough,and (5) boundary part of the Caribbeanplate.
parallelextensionand the openingof the Mona rift.
The NUVEL-IA
velocity azimuth for Caribbean-North
America relative motion is consistent with nearly pure leftlateral strike-slip displacementalong plate boundarystruc2. Relative Motion Between the Caribbean
tures in the northeasternCaribbean, although minor converand North American
Plates
gence is allowed within the uncertainty.However, the velocity azimuthsobtainedfrom earthquakeslip vectors [Deng and
The relativemotionof the Caribbeanwith respectto North Sykes, 1995] and GPS data [DeMets et al., 2000] are consisAmericahasbeencontroversial,
rangingfrom11+3 mm/yrto
tent with some transpression east of Cuba. Evidence for a
the eastas predictedfi'omNUVEL-1A (NorthwesternUnivercomponent of convergence across the eastern Puerto Rico
sity Velocity Model IA) [DeMets et al., 1990] to 37ñ5
trench includes deep focus earthquakesdefining dipping
mm/yearto the northeastas derivedfromearthquakeslip vec- zones analogous to Wadati-Benioff zones, shallow thrust
tors[Sykeset al., 1982;Dengand Sykes,1995].GPS geodetic earthquakes,folds and thrust faultsaffectingyoung sedimendata from the Caribbean-NorthAmericanPlate Experiment tary rocks, and elevated topography [Sykes et aL, 1982;
(CANAPE) constrainmotionof the Caribbeanplatewith re- Frankel, 1982; Mann and Burke, 1984; Byrne et al., 1985;
spectto North Americaat Cabo Rojo in the southernDominiStein et al., 1988; Mann et al., 1990; Rosencrantz and Mann,
can Republic as 20.6ñ1.2 mm/yr toward N89øEñ3
ø (1o)
1991; Calais and Mercier de Lepinay, 1993; Deng and Sykes,
[Dixonet al., 1998]. Incorporation
of additionalunpublished 1995]. The convergenceimplied by these features,however,
datafromour 1998observations
decreases
the rateslightlyto may recordlocal block tectonicswithin the broad,deforming
PRVI block,(2) ongoingrotationof PRVI abouta nearby(<
25ø) verticalaxis,(3) eastwardtectonicescapeof PRVI, (4)
20øN
•
19øN
.
ß
'•
'•:2---7-••N•
RSF' _ '•/?._..
•
•••. •
ß'./' ' '• _ •"
ß. e,,••
•• •
' ß
ß
?
ß ß
'SPRSF
' :" •.?•;•':':.
ß
• '•....'•'
. .. . '. ß...•
ß
.
....•
18øN
......
::::::::::::::::::::::::::::::::::::::::::::::::
HispaniOla
17øN
69øW
68øW
67øW
66øW
65øW
64øW
Figure 2. Focal mechanisms
for depth < 35 km for easternHispaniola, Puerto Rico, and Virgin Islands.
Sourcesare the HarvardCentroidMomentTensor(CMT) catalogue,the PuertoRico SeismicNetwork, De/?g
andSykes[1995],andMolnar andSykes[1969]. DotsareUSGSepicenters
for earthquakes
abovedepthsof
60 km withmagnitudes
> 3.5 fromJanuary1, 1967,untilApril 28, 1999(USGS). AP, Anegada
passage;MR.
Monarift; MAR, Main ridge; MT, Muertostrough;NPRSF, northPuertoRico slopefault; PRT; Puerto
Ricotrench;SF, Septentrional
fault; SPRSF,southPuertoRicoslopefault; YR, Yumarift.
1024
JANSMA ET AL.: NEOTECTONICS
OF PUERTO RICO AND VIRGIN
ISLANDS
plateboundary
ratherthanrelativemotionof the rigid Carib-
and Pennington, 1990; Masson and Scanlon, 1991; Deng
and Sykes,1995]. EastwardalongPRVI, ongoingsubduction
Hispaniola and PRVI are underthrust in the north and south
of Caribbeanlithosphereis less clear. Indeed, Malfait and
by NorthAmerican
and Caribbean
lithosphere,
respectively, Dinkelman [1972] and Burke et al. [1978] argued that sub[Dolanet al., 1998] andthus likely behaveindependently ductionendedin the Oligocene. Side-scansonarimagesshow
fromthemajorplates.
that the accretionaryprism narrowseastward and disappears
near 65øW, southwestof St. Croix [Mauffret and Jany, 1990;
Masson and Scanlon, 1991]. A minimumof 40 km of under3. Puerto Rico-Northern Virgin Islands Block thrusting is proposedat 68.5øW, whereasno underthrusting
(PRVI)
is thought to have occurredat 65øW [Ladd et al., 1977], implying an eastwarddecreasein convergenceacrossthe MuerSeismicitywithin and aroundPuertoRico andthe Virgin
tos trough [Masson and Scanlon, 1991].
Islandsaverageshundredsof earthquakes
per year(Figures1
To the west, PRVI is separatedfrom Hispaniola by the
and2). Although mostare small(< 4.0), severallargeevents Mona passage(Figure 2). Studiesof seafloorstructure[Larue
have occurredduring historic times, including the 1916,
and Ryan, 1990; 1998; van Gestel et al., 1998] suggestE-W
1918, and 1943Mona passage
earthquakes
(Ms = 7.2, 7.3, and
extensionoccursin this region with the formationduring the
7.5, respectively),
the 1867 Anegadaearthquake
(Ms = 7.3),
last few million years of the N-S trending Mona rift offshore
the 1787 PuertoRico trenchearthquake(M = 7.57), and the
northwesternPuerto Rico. A significant earthquakein 1918
1670SanGermanearthquake
(M - 6.5?) [Pachecoand Sykes, (Ms = 7.3) in this areawas accompaniedby a 4-6 m tsunami in
1992]. With mosteventsconcentratedoffshore,currentseiswestern Puerto Rico. Rupture along four segmentsof a N-S
micitymimicsthe patternof the large,historicevents(Figure trending normal fault in the Mona rift were used in numerical
2), leadingseveralworkersto proposea rigid PRVI in the
simulations of tsunami run-ups in western Puerto Rico
northeastern
comerof the Caribbean[Byrneet al., 1985; Mas[Mercado and McCann, 1998]. The modelyielded resultsin
son and Scanlon, 1991].
reasonableagreementwith observedaccountsof the tsunami,
The northside of PRVI is boundedby the east-weststriksupporting the normal motion mechanismand thus the asing PuertoRico trench,which lies-•100 km offshore,reachesa
sumption of active extension in the Mona rift. Focal mecha-
bean plate with respectto North America. Indeed, eastern
waterdepth> 8 km,andcoincideswith the largestnegative
free-airgravityanomaly
onEarth(Figure1). Despiterecognitionof this featurein the 1950s[Officeret al., 1957]the nature of deformation in the Puerto Rico trench remains contro-
versial. Obliqueconvergence
acrossthe PuertoRico trenchis
supportedby earthquakeslip vectors,a diffusezone of south
dipping earthquakesbelow the island of Puerto Rico, and
low-anglethrustsimagedin seismicprofilesof the western
PuertoRicotrench[Sykes
et al., 1982;McCann,1985;Deng
andSykes,1995;LarueandRyan,1998]. In contrast,
Speed
and Larue [1991] andMassonand Scanlon[1991] inferred
extension across the trench offshore northwestern Puerto Rico
from evidence of subsidencesince the Pliocene of the northern
insularshelfof PuertoRicoimpliedby Mioceneshallowwater
limestones
in waterdepthsof 5000 m [Moussaet al., 1987].
Althoughour previousGPS-derivedvelocities[Dixonet al.,
nisms of some more recent events offshore northwestern Puerto
Rico are consistentwith normalfaulting along NNE striking
planes (Figure 2). The nature of the putative boundary between the PRVI and Hispaniolamicroplatessouthof the Mona
rift is unclear. The N-S orientedYuma rift has been mapped
southwest of the Mona rift (Figure 2) [dany et al., 1987;
Grindlay et al., 1997], but the kinematics of the structureare
not well constrained.Side-scanimagessupportcontinuity of
the Muertos trough from southeasternHispaniola to southwesternPuerto Rico [dany et al., 1987].
EasternPRVI is boundedby the ENE trending Anegada
passage(Figure 2), which connectsthe NeogeneVirgin Island
and Whiting basins in the southwest(Figures 1 and 3)with
the Sombrero basin in the northeast and defines a zone of
probable transtension along which displacement from the
easternend of the Muertos trough is transferredto the Puerto
Rico trench [danyet al., 1987; Masson and Scanlon, 1991].
Both right-lateral [Nemec, 1980; Matthews and Holcombe,
1976;danyet al., 1987] and left-lateralstrike-slipcomponents
1998] indicated motion along the Puerto Rico trench as
largelyleft-lateralstrikeslip,we were not ableto differentiate
withinerrorbetweenminorconvergence
or divergence
across
thestructure.ThePuertoRicotrenchis traditionallyassumed of motionhave beenproposedfor the Anegadapassagefault
to accommodate
muchof the current highly oblique North
[Hessand Maxwell, 1953; Donnelly, 1964]. Dextral faulting
American-Caribbeanrelative plate motion. Additional
implieseastwardmotion of PRVI fasterthan that of the Caribmapped offshore faults between the north coast of Puerto Rico
and the Puerto Rico trench, such as the south Puerto Rico
slope fault (SPRSF),however,also maytake up somedisplacement[Grindlayet al., 1997].
The Muertostrough,an east-west
strikingbathymetric
featureof > 5 kmdepth,defines
thesouthern
limitof PRVI (Figure
1). A northdippingzoneof earthquakes
to a depthof 100 km
and an accretionary
prismalong the lower slope south of
southeasternHispaniola and southwesternPuerto Rico are
consistentwith overridingof Caribbeanlithosphereby
southwestern
PRVI alongthe Muertostrough[Laddet al.,
1977; Byrneet al., 1985; Larue and Ryan, 1990; McCann
be'an
interior
andsupports
tectonic
escape
of PRVIwithinthe
plate boundaryzone. Shallow seismicityis localizedalong
the edgesof the Anegadapassage(Figure 2), although most
large historic events occurrednorth of the Virgin Islands
[Murphy and McCann, 1979; Frankel et al., 1980; McCann,
1985] and no earthquakeswere recordedin the easternAnegada Passagebetween63.5ø and 64.5øW by a local seismic
networkdeployedin the Virgin Islandsduring the late 1970s
[Frankelet al., 1980]. Whetherthe lack of seismicityreflects
long recurrenceintervalsor no displacement
alongthe eastern
portion of the Anegadafault is unknown.Eventswell north of
the Anegadapassageand the northernVirgin Islandsrecorda
JANSMA ET AL.: NEOTECTONICS OF PUERTO RICO AND VIRGIN ISLANDS
1025
20 N
NORTH
'---:
-
Septentrional
Fault.•.•
5o•"'
PRW block
'no•hern DR)
•::•ANEG
PUNT
IOMA
BOCA
PUR3
ISAB
NOVA
•!•1•._•
AB NOVA
_ SanJuan .....
GORD
-
----_..___
18N
Hispaniolaplatelet
MuertosTrough
'•ß campaign
continuous
sites
sites
•
CARIBBEAN
ROJO (southernDR- Caribbean)
68 W
NUVEL-1A
PLATE
66 W
64 W
Figure 3. CurrentmixedmodeGPS geodeticnetworkin the northeastern
Caribbean.GNPRFZ, greatnorthern
PuertoRico fault zone; GSPRFZ, great southernPuerto Rico fault zone; CJF, postulated Cordillera-Joyuda
faults;LV, Lajas Valley (mediumgray shadedrectanglein southwesternPuerto Rico). Major offshorestructuresalsoare shown. Light gray shadedregions are zonesof inferredextension. WB, Whiting basin: VIB.
Virgin Islandsbasin. Arrows are GPS-derivedvelocities relative to North America and the NUVEL-IA velocity, derivedfrom globalplatemotions,of the Caribbeanwith respectto North America. For scale.length of
NUVEL-lA arrow corresponds
to a rateof I 1+3 mm/yr. Circlesare 95% confidencelimit. Error ellipse is not
shownfor NUVEL-lA for clarity.
combinationof reverseand sinistral motion along roughly EW striking faults, which is consistentwith ENE directedrelative motion of the Caribbeanwith respectto North America.
The structuralgrain of the Anegadapassageis dominatedby
NE/SW or E-W striking faults, which on seismic reflection
profiles,clearlyhave a significantnormalcomponent[danyet
al., 1987; Holcombe et al., 1989; dany et al., 1987; Masson
and Scanlon, 1991] requiring extensionacross the structure
[Speedand Larue, 1991]. Throws along the major faults decreasesouthwestward to the Muertos trough from> 4 km to
zeroover a distanceof 50 km, implying rapidly decreasingextensionwestward [Masson and Scanlon, 1991].
Some neotectonic models for PRVI advocaterigid body,
counterclockwise
rotation
of PRVI
about a vertical
axis lo-
catedeither in southeastPuerto Rico or immediatelyoffshore
to the northeastin responseto left-lateral motion along the
North American-Caribbeanplate boundary[e.g.,Masson and
Scanlon,1991;Huerfano, 1995]. Thesemodelspredictextension to the NW (Puerto Rico trench offshore northwestern
Puerto Rico) and SE (Anegadapassage)and shorteningto the
NE (Puerto Rico trench offshorenortheasternPuerto Rico)
and SW (Muertos trough). The opening of the Mona riff is
problematicin this interpretation.
A rotationmodel is supportedby paleomagneticdata from
Cretaceousand Eocene rocks exposedon Puerto Rico, which
record > 45 ø and 24.5 ø of counterclockwise
rotation
since the
Eoceneand Miocene, respectively[Fink and Harrison, 1972;
Van Fossen et al., 1989; Reid et al., 1991]. Becausethe paleomagneticsignatureof Pliocene (4.5 m.y.)carbonatesfrom
northernPuerto Rico was not statisticallydifferent from that of
the present,all the rotation observedin Miocene strata was
assumedby Reid et al. [1991] to have occurredbetween I 1
and 4.5 m.y. with no rotation occurringduring the last few
million years. This viewpoint is supported by Larue and
Ryan [1998] and van Gestel et al. [1998], who argue, on the
basisof evidencefrom seismicreflectionprofiles,that PRVI rotation stoppeda few million years ago and that the current
phaseof deformationis contractionalacrossthe entire E-W
1026
JANSMA
ET AL.: NEOTECTONICS
OF PUERTO
trending Puerto Rico trench, including the western segment
offshorenorthwesternPuerto Rico. In their models,opening
of the Mona rift occurs in responseto eastward motion of a
rigid PRVI during the last few million years. Collision with
the BahamaBank effectively pins Hispaniolaand precludesits
eastwarddisplacementwithin the plate boundary zone, yielding extensionbetween PRVI and Hispaniola. Mauffret and
Jany [1990] also favor recenteastwardtectonicescapeof PRVI
on the basis of fault patterns within the Anegada passage,
which are compatiblewith dextral transtension [Jany et al.,
1987].
RICO
AND
VIRGIN
ISLANDS
all have chokering antennaeand recordat 30 s rate to 5ø elevation. GEOL, FAJA, and PUR3 use Trimble 4000SSi receivers, while the CRO1 site is equipped with an AOA Turbo
Rogue receiver.
Data from continuous sites PUR3, GEOL, and CRO1
(Figure 4) and campaign sites ISAB, PARG, MIRA, ARC1,
ZSUA, and GORD (Figure 5) are consideredin this paper. For
the remainingsites,time seriesare not yet of sufficientduration to provide robust velocity solutions and they are not
used in our analysis.
5. Data Analysis
4. GPS Data Acquisition
GPS measurements
were first collected
in the northeastern
Caribbeanin 1986 at sevenlocations(Figure 1): Grand Turk
(TURK), Turks and Caicos; Guantanamo(GTMO), Cuba; Cabo
Rojo (ROJO), Capotillo (CAPO), and Cabo Frances Viejo
(FRAN) in the Dominican Republic; St. Croix (STCX), U.S.
Virgin Islands;and Isabela (ISAB), Puerto Rico [Dixon et al.,
1991]. These sites were reoccupiedin 1994, and the network
was densifted to include an additional
15 sites in the Domini-
can Republic, Puerto Rico, and the Virgin Islands as part of
CANAPE. Measurementswere collectedeach year since 1994
on subsetsof the network. In 1995, a permanentInternational
GPS Service for Geodynamics (IGS)station was established
in St. Croix (CRO1), and a vectortie to the original 1986 site,
STCX, was established[Dixon et al., 1998].
The GPS network in Puerto Rico and the Virgin Islands
(Figure 3) consists of the original 1994 CANAPE locations
(ISAB, PARG, and GORD) plus campaign sites MIRA
(Miradero-Mayagtiez), ZSUA (San Juan), MONA (Mona Island), DSCH (Desecheo Island), ADJU (Adjuntas), ARC1
(Arecibo), CCM5 (Ponce), FAJA (Fajardo), LAJ1, LAJ2, and
LAJ3 (Lajas Valley), and ANEG (Anegada, British Virgin Islands) and continuoussites GEOL in Mayagiiez and FAJA in
Fajardooperatedby the Departmentof Geology,University of
Puerto Rico, and PUR3 in Aguadilla maintainedby the U.S.
Coast
GPS geodetic data were processedusing the GPS inferred
positioning system, orbit analysis, and simulation software
package (GIPSY/OASIS II)developed, distributed, and supported by the NASA Jet Propulsion Laboratory (JPL)
[Lichten, 1990]. Analysiswas performedeither at the University of Wisconsinor Universityof PuertoRico, Mayagiiez,using identical versionsof GIPSY (version 2.5, update8a)and
processingscripts. Receiverindependentexchange(RINEX)
format datawere processed
with preciseorbit and clock products from JPL [Zumberge et al., 1997]. A nonfiducial pointpositioning strategywas adoptedfor all station days, following Dixon et al. [1998]. Free-network solutions were transformed
into
the
international
terrestrial
reference
frame
(ITRF96) [Sillard et al., 1998] (Table 1) [e.g., Blewitt et al.,
1992: Heftin et al., 1992]. Errors shown for daily site positions are scaled lo errors. Velocity estimatesuse a weighted
least squaresfit to daily site positions. Uncertainty on derived
velocities
reflects an estimate of both
white
and time-
correlatednoise,usingthe methodologyof Mao et al. [1999].
6.
GPS-Derived
Velocities
for
Puerto
Rico
and the Virgin Islands
6.1. Microplate Rigidity, Translation, and Rotation
Guard.
GPS campaigndata were collected using two differentreceiver/antennaecombinations. All observations from May
1994 to June 1996 were made using Trimble 4000 SSe 9channel,dual-frequency,code phasereceiversequipped with
Trimble SST antennaewith ground planes. All observations
afterJune 1996, with the exceptionof the 1998 campaignsat
Before solving for the motion of sites in Puerto Rico and
the Virgin Islands relative to the North American and Caribbean plates, we first use the velocities of GPS sites in Puerto
Rico, expressedrelative to ITRF96 (Table 1), to addresstwo
questions:Does the microplatedeforminternally,and do GPS
velocities fi'omPuerto Rico constrain all componentsof the
SANA and AVES, were obtained with Trimble 4000 SSi 12-
motion of the PRVI microplate? The dispersionof geodetic
velocities about the predictions of an angular velocity that
bestfits thosevelocitiescan be usedto assessthe rigidity of a
plate [Argus and Gordon, 1996; Dixon et al., 1996] or block
channel,dual-frequency,code phasereceiversequipped with
Trimble Dorn-Margolin type choke ring antennae. Data were
collectedand archived at a 30 s epoch using a 10ø elevation
mask. A minimumof 8 hoursof GPS dataper UTC observation
day was collected during all campaignoccupations. The majority of siteshad significantly moredata as a result of 16-24
hoursof observations
duringeachUTC day. Exceptfor a brief
period in October 1995, both anti-spoofing (A/S)and selective availability (S/A) were on during data acquisition. The
University of Puerto Rico, Mayagilez (UPRM) (GEOL and
FAJA), National Oceanographicand AtmosphericAdministration (NOAA) (PUR3), and IGS (CRO1) continuous stations
[Dixon et al., 2000], albeit only approximatelywhen sites are
located in the zone of elastic
strain accumulation
of an active
fault. Using the method described by Ward [1990], we inverted horizontal velocities from ARC1, GEOL, ISAB, MIRA,
PARG, PUR3, and ZSUA to find the weighted least squares
bestfit angularvelocitiesfor the sevensites. The averagerate
misfitto the 14 horizontal velocity componentsis 1.2 mrn/yr,
with only two velocity componentsmisfitby > 2 mm/yrand
onemisfit at a level exceedingone standarderror. The data are
JANSMAET AL.' NEOTECTONICS
OF PUERTORICOAND VIRGINISLANDS
60
,
i
i
PUR3
i
102'7
60
i
PUR3
Rate = 12.8 mm/yr
Rate =8.5
mm/yr
• 40
IB 40
ß
ß
'.'.,.: ' .';.b:
•
20
•
0
..ß • iIIl•e"
.,* e,.• I.:;..•.•,'•
o•20
•o
'"
WRMS
,
60
60
GEOL
t
,
•
GEOL Rate =8.2 mm/yr
Rate = 12.0 mm/yr
40
ß
20
ß
ß
'e"I
0
ßß
"".. '
60
40
. ß
ß •.
ß
••
,
20,
ß
2ß
&e
ß
,
.
•40
.•20
Io
o
RMS
ß ee
&..,•,i._. ß
• 60CR01 Rate= •.0 mm/yr
._•••
•
ß
,'
vvF•
••m•-' '"'"'""'
'•?"-
•o
WRMS = 5.0 mm
CRO1
Rate=IS.5
mm/yr
ß .•=.,
ß .'
.•20
ß
ß
ß
• 40
ee•I ß
ß
"ßß
ß'"fi"";a•
WRMS
=7.3
mm •'
= 5.0 mm
ß
.
' •: • .•,.4•k•..,2
...- .•,.ß _,ß .'/4,.
..:....-,.
.ß. '•'.
4 6 mm
".
1936
' '',1' '
'1937
1938
1939
80
60
PUR3 Rate = 1.6 mm/yr
40
ß.. ß .... "<:_.. ..:
ß* .*'*t.•.
•,. *
20
0
-20
ß
-40:
ß•1..
ßß
V•RMS= :14.0mm
80
(50
40
GEOL Rate = 5.2 mm/yr
ß
ß
.*
'1,.
. .
.. .•,
.•,
"I_.
= 12.7
mm
20
'd":;
0
-20
ß
ß
WRMS
80
60
qO
CRO1
Rate = 3.5 mm/yr
ß
ß
ß .'
,ßß*
ß * ...%** .* ß ..
..*•, g
ß -,,-.-' ;.'-• ..' .."-,'• ß ,.,:'_a-.'.,
ß
i*
-20-
-40:
WRMS
1•96
1•97
1•98
= 13.9 mm
1•99
Figure4. GPSstation
coordinate
timeseries
for conti•uous
stations
PUR3,GEOL,andCRO1. Dailypoint
positions
arein international
terrestrial
reference
frame
(ITRF96).Formalsolutionerrorsarenotshownfor
clarity. Note the correlatednoiseamongthe three sites.
1028
JANSMA ET AL.' NEeTECTONICS OF PUERTO RICO AND VIRGIN ISLANDS
a)
b)
ISAB
WR'MS
=' 4'
.7 ................
mm
E 60
ARC1
30
•
•
,
i
,
,
,
WRMS= 5.0mm
'½
E 20
• 50
• 40
= 30
•
• 20
.
•10
0
;
'-' 60
I
WRMS
"
0
Rate = 11 4 mm/yr
I
,
10
,'
i
WRMS
= 7.9 mm
i
= 6.8 mm
i
Rate = 11.9 mm/yr
I
i
i
i
ß
• 40
• 20
•
ß
•e ee
•0
•
-20
Rate
=9.6
,m•/y,
r
WRMS
'•,.•o
I
Rate= 6.6mm/yr
•WRMS=12.7mm
= 16.3 mm
ß
•
ß ee
'•o
o•,•
ß
•-30
ß
+.
Rate= -7.1rn,rn/•r
,
1994 ' )9'95'
d)
MIRA
I
,
,
,
i
lB 60
ß
1999
PARG
l
WRMS
=5.9
mm
20.
19•8
1997
c)
,
Rate = -I5.9 mm/yr
-60
,
i
I
,
,
,
WRMS
= 6.1 mm
•
= 6.6 mm
I
,
,
,
I
,
•
,
I
,
,
,
i
,
E 50
•ø•
•
40
= 30
=o •
•
20
'• 10
ß
-20
20-
Ra,
te= 10.5
I mm,/yr
WRMS = 6.8 mm
ß
0
,-, 60
WRMS
40
Rate
=11.1
,mr•/y,•
*" 20
0
,•
,•
,-, 30-
I
I
,
0
, 3.½
l R,
ate=
mm,/yr
- 10.0
,m/y,
r
,.,ß , I , , , I • , , I , Rate
, , t, , ,rn
WRM$
WRMS = 18.0 mm
= 11.0 mm
ß
'-'
Rate =-2.3 mm/yr.
--,
19'97
19'98
19•9
ß
0
Rate=-3.6 mm/yr•
,
1994' '1•95 '1996 1997 1998 1½99
Figure 5. GPS stationcoordinate
time seriesfor PRVI campaignsites: ISAB, MIRA, PARG, ARCI. ZSUA.
andGORD. Pointpositions
are in ITRF96. Solutionerrorsarescaled1o formalerrorsfromGPS inferredpositioningsystem,orbit analysis,and simulationsoftwarepackage(GIPSY/OASIS II).
JANSMA ET AL.' NEOTECTONICS OF PUERTO RICO AND VIRGIN ISLANDS
e)
• 50
•
GORD
ZSUA
i
WRMS
•
•
i
i
•
•
•
i
= 6.2 mm
•' 60
• 50
40
•
'• 30
•
20
•
10
40
•
10
•,
60
Rate = 12.5 mm/yr
WRMS
Rate = 13.6 mm/yr
0
= 6.6 mm
: ':
,
WRMS
I
,
,
•
I
,
•
,
t
,
,
,
t
,
'
,
I
= 5.1 mm
• 40
,.• 20
•
40
•= 30
20
0
•
1029
0
0
Rate = 8.8 mm/yr
• 30
WRMS
: •: : I : : , I • , , I , , , I , , , I
0
= 17.5 mm
E -30
ß
•
0
• -60
;;" -90
-30-
1•96
1•97
Rate
='-5.3
mm,/yr.
1•398
1999
i
1994
.
.
1995
,
•
,
,
,
1996
i
,
,
1997
,
i
1998
i
1999
Figure 5. (continued)
thusfit within their estimateduncertaintiesof 1-3 mm/yr,providing an approximateboundon the level of internal deformation of the PRVI microplateand suggestingthat the block is
rigid within our velocity uncertainties.
The GPS site velocitiesfrom the PRVI microplateconstrain
only the componentsof the angularvelocity that predict the
uncertainties in the linear velocity predicted by the best fitting angularvelocity achieve minimumvalues of 4-0.65 mm/yr
and 4-2.8ø in western Puerto Rico, but increase to 4-4 mm/yr
and 4-5.5ø in easternPuerto Rico. Velocity uncertainties thus
increaserapidly with angular distance l?omwestern Puerto
Rico, as expected,given that the available data imposefew or
rate and direction of motion in western Puerto Rico, where all
no constraints
but one of the PRVI GPS sitesare located.For example,the 1o
Along the Muertos trough and Anegada fault, which separate
on motion
outside
of western
Table 1. Velocities of PRVI Sites and CRO1 in ITRF96 a
VelocityNorth
(ITRF96),
mm/yr
Velocity East
(ITRF96),
mm/yr
Latitude,øN
Longitude,øE
ARC1
11.9 + 2.5
6.6 +3.3
18.35
293.25
CRO1
13.5 + 1.5
9.0+
1.8
17.76
295.42
GEOL
12.0+ 3.1
8.2+
1.7
18.21
292.86
GORD
13.6 + 2.5
11.3 + 3.3
18.43
295.56
ISAB
11.4 + 0.9
9.6 + 1.3
18.47
292.95
MIRA
10.5 +2.5
3.6 + 3.3
18.23
292.87
PARG
11.1 + 1.1
10.0+2.1
17.97
292.96
PUR3
12.8 +2.8
8.5 + 1.6
18.46
292.93
ZSUA
12.5 +2.4
8.8 + 2.2
18.43
294.01
a Uncertainties are 1 standarderror.
Puerto
Rico.
1030
JANSMA ET AL.' NEOTECTONICS
OF PUERTO RICO AND VIRGIN
ISLANDS
a)
North American plate (fixed)
20 N
18N
ROJO
PARG
CRO1
10 mm/yr
ß
'"
Caribbean plate
74 W
72 W
70 W
68 W
b)
66 W
64 W
62 W
60 W
lO
PUR3 ZSUA
ARC
I•GE"-:'
OL
............
MIRA
............
,.............
'""'":'"'""':••
ISAB
.........
i................
i.............PARG
ß ..:.:.:.....
j
......
j
=========================
.....
::::.::'•:::
......
=================================
I
5
I
10
15
Velocity east (mm/yr)
Figure 6. (a) Velocitiesrelativeto fixedNorth Americadefinedaccordingto DeMets and Dixon [1999].
Confidence
ellipsesaretwo-dimensional
95%.(b) Velocities
of PRVI sitesrelativeto North Americaplotted
at a common
origin.Arrowis average
PRVI velocityat MIRA derivedfromPuertoRico-NorthAmericaangular velocities and their covariances.
the PRVI microplatefromthe Caribbeanplate, the lo uncertainties of severalmillimetersper year exceedgeologically
basedestimatesof slip ratesalongthosefaults.
Because the present velocities impose meaningful constraintson PRVI motion only in westernPuerto Rico, we restrict the subsequentkinematicanalysisas follows: predictions of the motion of the PRVI block relative to both the
6.2 Velocities Relative to the North American
To describe the velocities
Plate
of GPS sites in the northeastern
Caribbeanrelativeto the North Americanplate (Figure 6 and
Table 2), we usethe angularvelocity that describesmotion of
the North American plate relative to ITRF96 to predict the
North America ITRF96 velocity at each of our GPS sites. We
NorthAmericanandCaribbean
platesaremadeonlyfor a point
then subtractthe predicted plate velocity at each site from the
site velocity and sum the covariancethat describesthe uncerRico(18.3øN,67.0øW).We thenmakethereasonable
assump- tainties in both. The North Americanplate angular velocity
tion thatthe blocktranslates
rigidlyat the velocityof western we useis derivedfrom the velocitiesof 16 continuously operPuerto Rico, supported by the fact that GPS site ZSUA in
ating GPS stations in the stable interior of the North AmerieasternPuerto Rico has a velocity similar to that for sites in
can plate [DeMets and Dixon, 1999].
at the geographiccenter of the GPS sites in western Puerto
western Puerto Rico.
Well-constrained
velocities from sites
GPS-derived
velocities
are similar
for all sites in Puerto
in easternPuertoRicoareneededto testmorerigorouslyfor Rico at the 95% confidencelimit (Figure 6). The meanvelocrotationof the PRVI block and to refinethe interpretations ity for motionof PRVI relativeto NorthAmerica,computedfor
presented below.
westernPuertoRico, is 16.9+1.1 mm/yrtoward N68E+3 (1o).
JANSMA
ET AL.: NEOTECTONICS
OF PUERTO
Table 2. Velocitiesof PRVI SitesandCRO1 With Respect
to Stable North America a
North
ARC1
6.9 + 2.6
ISLANDS
1031
whether GORD is part of PRVI and intrablock deformationis
severalmillimeters per year betweeneasternand western PRVI
or whetherPRVI encompasses
only PuertoRico and someportion of the easternVirgin Islands and a tectonic break exists
America
between
VelocityNorth,
mm/yr
RICO AND VIRGIN
VelocityEast,
mm/yr
13.5 + 3.4
CRO1
7.7 + 1.6
15.7 + 2.0
GEOL
7.1 + 3.1
15.1 + 1.9
GORD
7.8 + 2.6
18.2 + 3.4
ISAB
6.5 + 1.1
16.6 + 1.5
MIRA
5.6 + 2.6
10.4 + 3.4
PARG
6.2 + 1.2
16.8 + 2.8
PUR3
7.9 + 2.9
15.5 + 1.8
ZSUA
7.2 + 2.4
15.7 + 2.4
GORD
and eastern Puerto
Rico.
If the motion
at
GORD does prove to be the sameas that at CRO1 over the
long term (decadescale), the implication is that the eastern
Virgin Islandsare completelycoupledto the Caribbeanplate
and that the displacementalong the Anegada passagedecreases to zero between
eastern Puerto
Rico
and the eastern
British Virgin Islands. The entire North American-Caribbean
relative plate motion then must be accommodatednorth of the
eastern Virgin Islands either by increased slip along the
Puerto Rico trench or displacementalong unmappedoffshore
faults. A zone of increasedseismicitydoes occur along the
Sombrerotrend north of the northeasternVirgin Islands
(Figure 2), where an earthquakeswarm occurredin the late
1970s [Frankel et al., 1980].
6.3 Velocities Relative to the Caribbean
Plate
a Uncertainties
are 1 standard
error.Velocitiesanderrorsarebased
on values in Table
1.
The velocities
We hereafteruse this velocity to describethe motion of PRVI
relative to North America. At the same location, a new Carib-
bean-NorthAmerica angularvelocity derived principally from
GPS sites in North America and the Caribbean plate [DeMets
et al., 2000] predicts that motion of the Caribbeanplate relative to North America is 19.2+1.3 mm/yr toward N69E+3 (lo)
(Figure 6b). The velocity of PRVI with respectto North America is thus2 mm/yr slower than that of the Caribbeanplate at
this location. Motions relative to the Caribbeanplate interior
of sites in PRVI
relative to the Caribbean
plate are shownin Figure 7 and Table 3. Data usedto define
the velocity of the Caribbean plate relative to ITRF96 come
fromROJO in the southernDominican Republic, SANA on
San Andres Island in the western Caribbean, AVES on Aves
Island in the eastern Caribbean, and CRO1 on St. Croix
(Figure 1). Our results do not change significantly if the
CRO1 velocity is eliminatedfromthe angular velocity solution. Detailsof the intraplatesitesare given by DeMets et al.
[2000]andarenotrepeated
here. The methodologyfor gener-
atinga velocityrelativeto the Caribbean
plateis analogous
to
that describedabove for deriving velocities relative to the
are discussed in section 6.3.
Comparisonof Caribbean-NorthAmerica motion in western
Puerto Rico to the meanvelocity of sites in western Puerto
Rico suggeststhat a minimumof 85% of the total motion betweenthe North Americanand Caribbeanplatesat the longitudeof westernPuerto Rico is accommodated
by faults north
of Puerto Rico with the remainder
of the island.
accommodated
to the south
We note that the azimuth of PRVI
motion rela-
tive to North America (and that of rigid Caribbeanwith respectto North America) is parallel to the trend of earthquake
slip vectorsfrom the northeasternCaribbean[Deng and Sykes,
1995] and to the SPRSF [Grindlay et al., 1997]. SPRSF is
predominantlystrike slip, whereasthe Main ridge is a compressional feature oriented perpendicular to the PRVI-North
America relative motion vector [Muszala et al., 1999].
The GPS-derivedvelocity with respectto North America at
GORD in the British Virgin Islands of 19.5+2 mm/yrtoward
N75E+5ø (lo) closelyapproximatesthe predictedrelative motion betweenthe rigid Caribbeanand North Americanplates
at GORD, 19.4+1.3 mm/yrtoward N68E+3ø (lo), and differs
fromthe velocity of westernPuertoRico, 16.9+1.1mm/yrtoward N68øE+3ø (lo). With only two epochsof observations
at GORD, the difference in the rates relative to North America
Table 3. Velocities of PRVI Sitesand CRO1 With Respect
to Stable Caribbean a
Caribbean
VelocityNorth,
mm/yr
VelocityEast,
mm/yr
ARC1
-0.1ñ2.7
-4.3ñ3.5
CRO1
0.4ñ2.0
-2.4ñ2.2
GEOL
0.2ñ3.2
2.8ñ2.1
GORD
0.5ñ2.8
0.3ñ3.5
ISAB
-0.4ñ1.4
-1.2ñ1.8
MIRA
-1.3ñ2.7
-7.4ñ3.5
PARG
-0.8ñ1.5
-1.1ñ2.4
PUR3
1.0ñ3.0
-2.3ñ2.0
ZSUA
0.1ñ2.6
-2.1ñ2.5
of GORD and PRVI is barelyabovethe 2 mm/yrlevel of the
noise associatedwith GPS geodeticdata. More data in the
Virgin Islands area will be required before we can assess
a Uncertainties are 1 standard error. Velocities and errors are based
on values in Table 1.
1032
JANSMA ET AL.' NEOTECTONICS
OF PUERTO RICO AND VIRGIN
ISLANDS
a)
"..• ,'" NorthAmerican
plate
(predicted)
NUVEL-1A
20 N
18N
ROJC
[
10 mm/yr
....
--.•/Caribbean plate(fixed)
74 W
72 W
70 W
68 W
66 W
64 W
b)
62 W
'
60 W
2.5
PUR3
I
7.5
I
5
I
2.5
- 2.5
Velocity west (mm/yr)
Figure 7. (a) Velocities relative to Caribbeanreferenceframe. Confidenceellipses are 95%. Arrow of predictedNorth America motion is that predictedusingthe Caribbeanreferenceframe of DeMets et al. [2000]. (b)
Velocitiesof PRVI sitesrelativeto Caribbeanplottedat a commonorigin. Arrow is averagePRVI velocity at
MIRA derived from PuertoRico-Caribbeanangularvelocitiesand their covariances.
North American plate The mean velocity of sites in Puerto
Rico relative to the Caribbean,computedfor western Puerto
Rico, is 2.4+1.4 mm/yr toward S79øW+26ø (lo). Individual
site velocities(Figure 7) rangefrom 1.4+0.8 mm/yr toward the
SW at PARG to 2.5+1.5 mm/yr toward the NW at PUR3
(Figure 7). The velocity of GORD relative to the Caribbean
plate is negligibleat 0.5+4 mm/yr.
To assesswhetherthe present GPS velocities fi'omsites in
Puerto Rico are consistent with significant motion of PRVI
relative to the Caribbeanplate, we tested for the existenceof a
separatemicroplateusing the F ratio testof Stein and Gordon
[1984]. This methodcomparesthe least squaresfits of models
that use two angular velocities and one angular velocity, respectively, to fit a set of kinematic observations. The test is
insensitive to systematic overestimatesor underestimatesof
velocity uncertainties,an important considerationwith GPS
velocities. Fitting the seven PRVI velocities and four Caribbean GPS velocities with a single angular velocity gives a
weighted least squaresmisfit of 13.0. Fitting the two sets of
velocities with separateangular velocities gives a summed
misfit of 9.8. The improvementin fit in the latter model,which
stemsfrom usingthreeadditionaladjustableparameters,
gives
F-- 1.7, whichis significantat only the 80% confidencelevel
for 3 versus16 (=22-6) degreesof freedom.The test for a separate PRVI microplatethus fails at the 95% confidencelevel.
Althoughthis resultis consistent
with no currentmotionof a
PRVI block relative to the Caribbean plate, it implies that
PRVI bounding faults are inactive. We believe this is unlikely given the existenceof significantseismicityalong all
the boundary structuresexceptthe easternAnegada passage
JANSMA ET AL.' NEOTECTONICS
OF PUERTO RICO AND VIRGIN
fault. The interpretation
we preferis thatthe GPS velocityuncertaintiesare still too largeto detectthe predicted2-3 mm/yr
of PRVI block motionrelative to the Caribbeanat high confidencelevels. Assumingthat the presentestimateof Caribbean
plate and PRVI block velocities is correct and the GPS veloc-
ity uncertaintiesdecreaseprogressively
throughtime,numerical experimentssuggestthat two to three additional years of
measurements
at the existing sites will be required beforethe
presentvelocitiesare sufficientlypreciseto detect2.5 mm/yrof
PRVI-CA
motion at the 95% confidence
level.
Velocitiesin the DominicanRepublicrelativeto the Caribbean are fasterand more SSW trending than that of western
PRVI (Figure 7a). This differencesuggeststhe existenceof
separatePRVI and Hispaniola microplatesand constrainsmotion along the bounding structures. Predicted deformation
acrossthe Muertos trough is predominantlyleft-lateral strike
slip at the longitude of Puerto Rico with up to 2.3 mm/yrof
convergence
permissiblewithin error. The Muertostrough is
characterizedby almost complete convergencenear Hispaniola.The kinematicsof the Mona rift canbe assessed
by examining the velocities of westernPuerto Rico and those in the
easternDominican Republicand is discussedbelow.
7.
Geological and Tectonic Implications
7.1
Intrablock
Deformation
of PRVI
The velocity estimatefor sites in Puerto Rico yields errorsthat are < 2.0 mm/yr, definingthe upper bound on permissible deformation
within
the island.
All
sites have velocities
that are equivalent at the 95% confidencelimit. The discrep-
ISLANDS
1033
ancy in the velocities betweenPuerto Rico sites and GORD
in Virgin Gordaat the easternextremeof PRVI is closeto the
resolution of the current GPS geodetic data, giving relative
motion between western(the island of Puerto Rico) and eastern (Virgin Gorda) PRVI at 3+3 mm/yr. Although the scatter
of the GPS velocitieson PRVI permitsseveralmillimetersper
yearof slip,the absence
of any obviousgeographicpatternin
the velocities leads us to argue against substantial organized
deformation
within
PRVI
in favor of deformation slower than
the severalmillimetersper year resolution of the presentG PS
velocities. We solved for a stronger upper bound on intraPRVI deformationby examining the evolution of baseline
length between the two continuous sites in western Puerto
Rico, GEOL and PUR3 (Figure 8). Since continuous measurementsbegan at both stations in June 1997, the baseline
length has remainedconstantwithin error at -0.5+0.3 mm/yr
(lo). Intrablock deformationthereforeis small, and most deformationis likely accommodated
along the bounding structures of the microplate.
These data constrainthe rigidity of PRVI and bear on the
issue of seismic hazard. Specifically, do faults exist within
PRVI that are capable of producing locally damaging earthquakes? Shallow microseismicitydoesoccur onshore(Figure
2), but the historic record is consistent with major events
confined to the offshoreregion [McCann and Pennington,
1990]. The highest levels of onshore seismicity are in the
southwestcornerof PuertoRico in the Lajas valley [Asencio,
1980], an E-W trendingNeogenefeature. Faultsof similar orientationare mappedoffshoresouthof the Mona rift (Figure 3)
[Meltzer, 1997]. The Lajas valley is between our sites at
GEOL and PARG, whose velocities agree at lo, arguing
against significant deformationacrossthe structure. The island of PuertoRico also is traversedby two major northwest-
15.0
Rate - -0.5+0.3 mm/yr: WRMS - 4.0 mm
10.0
5.0
0.0
-5.0
-10.0
-15.0
,
1997
I
,
I
1998
BaselinelengthbetweenGEOL and PUR3
1999
Figure 8. Evolution of residualbaselinelengthbetweencontinuoussitesPUR3 and GEOL fromJune 1997.
Rate of changeof the baselineis-0.5+0.3 mm/yr. Statederror is 1o formalerrorwithout any provision tbr
monumentnoise.Median baselinelengthis 28898.713m. Numberof datapointsis 663.
200O
1034
JANSMA ET AL.: NEOTECTONICSOF PUERTO RICO AND VIRGIN ISLANDS
southeaststriking fault zonesthat were active during the Eoceneand are coveredby little deformedNeogene strata(Figure
3): (1) the great northern Puerto Rico fault zone (GNPRFZ)
and (2) the great southern Puerto Rico fault zone (GSPRFZ).
We note that the baseline between the two continuous
sites
(Figure 8)crosses the GSPRFZ. Elastic strain effectsfrom
GSPRFZ on GEOL and PUR3 time seriesare unlikely. Simple
two-dimensional elastic strain models using distances from
GSPRFZ of 5 and25 km for GEOL and PUR3, respectively,a
baselinelength changebetween GEOL and PUR3 of 0.5+0.3
mm/yr, and assumedvertical fault orientation with locking
depths of 10-20 km suggestthe permissibleupper bound on
GSPRFZ motion is 1.5+2.0 mm/yr. Field evidence for postOligocene motion along either the GSPRFZ or GNPRFZ is
sparse,which further supportsa rigid PRVI sincethe Miocene.
7.2
Counterclockwise
Rotation
Counterclockwiserotation of a rigid PRVI about a vertical
axis in southeasternPuerto Rico predicts large lateral linear
velocity gradients across the island, extension acrossthe
westernPuerto Rico trench, and shorteningacrossthe Muer-
tos trough south of southwesternPuerto Rico. Velocitiesof
all sites on the island, including ZSUA near San Juanin the
east, however, are equivalent within lo. Furthermore,the
GPS-derived velocity for PRVI is consistent with convergenceacrossthe PuertoRico trenchand a significantcompo-
the northern
and southern
boundaries.
Thus the tectonic
es-
cape hypothesis is not compatible with the GPS geodetic
data.
This may have moregeneral implications for the kinematic
behaviorof microplatesand blocks in transfolTnplate boundary zones. Our measurements
imply a relatively simple,monotonic increasein velocity (i.e., the velocity gradienthas a constant sign) along a path from the interior of one bounding
plate(e.g., North America)acrossthe PRVI block into the interior of the other (e.g., Caribbean) plate. In contrast,the tectonic escape model, with a mixture of right- and left-lateral
strike-slip faults, predicts a velocity gradient that would
change sign along a perpendiculartraverse from the North
American to the Caribbean plate. The simple monotonic increasein velocity that we predict implies that the summed
magnitude of strike-slip fault slip rates will equal the total
plate motion rate between the Caribbeanand North America.
In the tectonic escapehypothesis,however,the summedmagnitude of slip rates along the sametraversewould be higher
thanthe total plate rate. Our preferredmodelthereforemay represent less total energy dissipation comparedto the tectonic
escapemodel. Perhaps this is commonlythe casefor smaller
tectonicblocks and microplates,where frictional edge effects,
which scale by fault length and slip rate, may be large comparedto basal driving forces,which scaleby plate area.
7.4 Muertos Trough and Anegada Passage
nent of left lateral strike-slip faulting along the Muertos
troughsouthof SW PuertoRico. On the basisof both observationswe argue againstsignificantongoingrotation about a
proximalaxis and favor instead northeastwardtranslationof
PRVI relative to North America. The geodeticresults are supported by focal mechanisms
fromhistoric earthquakes,which
also yield convergenceacrossthe Puerto Rico trench offshore
Assumingthat the Puerto Rico block is not rotating significantlyrelativeto the Caribbeanplate and that the 2.4+1.4
mm/yr towardS79øW+26ø (lo) of predictedwestwardmotion
northwestern Puerto Rico [McCann and Pennington, 1990;
Deng and Sykes, 1995; Dolan and Wald, 1998]. Offshore
seismicprofileswere interpretedto documenta changefrom extension to shortening across the Puerto Rico trench in the
past few million years [Larue and Ryan, 1998] in agreement
with paleomagneticresults from Miocene rocks exposedon
ponent. Within error,the amountof convergence
permissible
Puerto Rico that are consistent with cessation of rotation -4.5
million years ago [Reid et al., 1991].
7.3 Tectonic Escape of PRVI
Jany et al. [1987] and Mauffret and Jany [1990] suggestedthat the PRVI block experiencestectonicescapeto the
east,squeezedlike a pumpkinseedbetweenconvergingCaribbeanandNorth Americanplates. If correct,this model bears
someanalogy to the eastwardmotion of the larger Caribbean
plate, which maybe driven, at least in part, by the westward
increasing convergencebetween North and South America
[Sykeset al., 1982; Dixon and Mao, 1997]. The tectonic escape model makeskinematic predictions that can be tested
with our new data. Specifically, it predictsleft-lateral strikeslip motion along the northernboundary of the PRVI block
and right-lateral motion along the southernboundary. Our
new GPS data,however,predictleft-lateralmotion along both
in central Puerto Rico can be used to describe motion of the
PRVI block south of Puerto Rico, displacement along the
Muertos trough has a significant left-lateral strike-slip comacross the Muertos trough is 2.3 mm/yr. Convergence is
greaterto the west offshoresoutheasternHispaniola, where
folds and reverse faults are mapped [Ladd et al., 1977] and
GPS velocities relative to the Caribbean are more southerly
trending(Figure 7). Southof southwesternPuerto Rico the
presenceof an accretionarycomplexadjacentto the Muertos
trough[Ladd and Watkins, 1978] and seismicityat depths>
100km [Byrneet al., 1985] favor convergencein this region.
Two moderateearthquakes
(M = 4.7 and 5) occurredin August
1993 near the eastern termination of the Muertos trough at
65øW where the accretionaryprism disappears[dany et al.,
1987; Masson and Scanlon, 1991]. Focal mechanismsgeneratedby the PuertoRico SeismicNetwork are consistentwith
left-lateral strike-slip motion along an E-W striking fault
[Huerfano, 1995], corroborating
preliminaryGPS-derivedvelocities(Figure 2). E-W strikingfaults along which predominantly left-lateral slip also is inferred are mappedoffshore
southwesternPuerto Rico [Meltzer, 1997]. Shorteningalong
the Muertos trough is greatestoffshoresouthwestPuerto Rico
and dies out to the easttoward the Anegada passage[Masson
and Scanlon, 1991].
Assumingthat motion along the Muertos trough is conservedand transferredto the western end of the Anegada Passage offshore southeasternPuerto Rico, then displacement
JANSMA
ET AL.: NEOTECTONICS
OF PUERTO RICO AND VIRGIN
ISLANDS
1035
along the westernAnegadapassage,the Virgin Island and
north Hispaniola deformedbelt, a submarinezone of folds and
thrusts, which was the locus of a series of large thrust earthPuerto Rico is sinistral transtension of the order of a few milquakes in 1946 and 1948 [Russo and Villasehor, 1995; Dolimetersperyear(Figure3). Evidence
to support
activeslipin lan and Wald, 1998].
the regionincludeshighseismicity[Frankelet al., 1980] and
Two-dimensional elastic modeling of strain accumulation
theoccurrence
in 1867of a large(7 < M < 7.75) tsunamogenic across Hispaniola, using GPS-derived velocities from three
earthquakealongthe northwall of the Virgin Islandsbasin sites in the Dominican Republic and one site in Grand Turk,
thatcaused
extensive
damagein St.CroixandSt.Thomas[Reid seismogenicdepths of 15-20 km, and imposed vertical fault
and Taber, 1920].
geometries,yields fault-parallelslip estimatesalong the north
The kinematicsof the easternAnegadapassage
canbe exam- Hispaniola deformedbelt and the Septentrional fault of 4+3
ined by comparingthe velocity of St. Croix (CRO1) with that
mm/yr and 8+3 mm/yr, respectively [Dixon et al., 1998].
of Virgin Gorda (GORD). Results from GORD must be conGeologicallyderivedHoloceneratesyield maximumslip estisideredas preliminarysincethereareonly two occupations, matesalong distinct segmentsof the Septentrional fault of
albeitseparated
by 5 years. As discussed
above,the velocity 13+4 mm/yr and 23+6 mm/yr [Mann et al., 1998].
at GORD is verysimilarto thatat CRO1, indicatingdisplaceAlthough the GPS velocities are preliminary becausethe
ment acrossand along the Anegada Passagein the eastern geodeticdata set is spatially limited, they do provide some
Whiting basins, and/or other structuresoffshoresoutheastern
BritishVirginIslandsis < 2 mm/yr. The datacurrentlydo not
distinguishbetweencomponents
of left-lateraland rightlateralstrikeslip if the existence
of strike-slipmotionis assumed.Janyet al. [1987] arguethat dextralmotionalongthe
Anegadapassagebeganin the Plioceneand is limited to 6-15
constraints on the kinematics of the Mona rift. The eastward
projectionsof both the northHispanioladeformedbelt and the
Septentrionalfault lie offshore northernPuerto Rico. If we as-
sumethat the displacementalong the north Hispaniola de-
km, implyinga displacement
rateof 1.5-3.7mm/yr,whichmay
formedbelt andthe Septentrional
fault is representative
of slip
offshorenorthernPuertoRico, the GPS-derivedvelocity for
not be resolvablewith GPS geodeticdata for severalmore
years.The lack of earthquakes
observedalong the Anegada
Puerto Rico with respectto North Americashould reflectthe
integratedslip alongoffshorestructuresand the displacement
passageduring deploymentof a seismic network in the north-
ods of seismicquiescenceor little motion betweenGORD and
CRO1. The possibility does exist that CRO1 and GORD ve-
associatedwith the Mona rift. The GPS-derivedaveragevelocity for Puerto Rico with respectto North America is
16.9+1.1 mm/yrtowardN68øE+3ø (lo). The integratedslip
along the north Hispaniola deformedbelt and the Septentrional fault estimatedfrom2-D elasticmodelingof strainac-
locitiesareaffectedby strainaccumulation
alonga locked,fast
slipping, easternAnegadaPassagefault. We believe this is
unlikely,however,becauseof the similarityof CROl's velocity with the otherrigid Caribbeansites (AVES, SANA, and
Puerto Rico moveseastwardrelative to central Hispaniola
southof the Septentrional
fault at an estimatedvelocityof 5+4
mm/yr. This rate is the upperboundon the openingrate of the
ern Virgin Islandsin the 1970s [Frankel et al., 1980] supports eitherlong recurrenceintervalswith interveningperi-
ROJO).
The westward motion of Puerto Rico relative to
GORD in the Caribbeanreference
framerequiresE-W extensionbetweeneasternPuertoRico andthe easternVirgin Islandsof a few millimetersper year. The locationof structures
alongwhich extensionnorthof the Anegadapassage
maybe
accommodated
are not constrainedby the GPS data reported
here.
7.5
Mona
Rift
The offshoredeformationbetweenHispaniola and Puerto
Rico reflectsrelativemotionsof easternHispaniolawith respectto westernPRVI. The openingrate for the Mona rift can
be estimatedfromcomparison
of the averageGPS-derivedvelocity for Puerto Rico with those of the northern Dominican
Republic. The calculatedextensiondependson the amountof
elastic strain accumulation on active faults in the Dominican
Republic and the errors associatedwith the GPS-derived velocities.
The two major E-W to WNW trendingfault zonestraversing the island of Hispaniola, the Septentrionalfault zone in
the northandtheEnriquillofaultzonein the south(Figure1),
havepresent-day
slipratesin excess
of severalmillimeters
per
year as estimatedfrom geologic evidence [Prentice et al.,
1993;Mann et al., 1998]. Somedeformation
also likely occursimmediately
offshore
to thenorthof Hispaniolaalongthe
cumulation is 12+4 mm/yr toward the ENE.
Thus western
N-S trendingMona rift offshore northwestern Puerto Rico, requiring that all the eastwardmotion is accommodatedacross
this structure. (A small northwardcomponentof motion of
PuertoRico relative to central Hispaniola also is possible,
which would impart minor left-lateral slip along the N-S
trendingboundingfaultsof the Mona rift.)
The total extension across the Mona rift constrained by
seismicprofiles is -6 km [van Gestel et al., 1998]. Using an
openingrate of 5 mm/yr estimatedfrom GPS velocitiesyields a
minimumage of the structure of 1.2 million years, confirming
the youth of the Mona rift and postdating the 4.5 million year
age for the end of PRV! counterclockwise rotation inferred
from paleomagneticdata [Reid et al., 1991]. The minimumand
maximumages of the Mona rift, within the errors of the G PS
data, are < 1 million and 6 million years,respectively.
The GPS velocitiesfrom the northeasternCaribbean permit
us to estimate preliminary displacementsalong the major
boundingstructuresof the PRVI block. We argue that at least
one, and possibly two, extensionalbelts occur between Hispaniola and the eastern Virgin Islands along which slip is
transferred
from the
southern
limit
of the
North
America-
Caribbean plate boundary zone to the northern portion. The
first and more significantis the Mona rift acrosswhich extension is capturedby our GPS data. The secondpotentialbelt is
the area between easternPuerto Rico and the Virgin Islands,
which may experienceminor extensionin responseto differential motion between easternPuerto Rico (ZSUA) and Virgin
1036
JANSMA ET AL.: NEOTECTONICS
OF PUERTO RICO AND VIRGIN
Gorda (GORD) that remains unresolved within current uncertainties. PRVI behavesrigidly betweenthesetwo extensional
zones. The N-S trendingMona rift accommodates
-•5 mm/yrof
east-westopening,which results in a reductionof slip along
the southernplate boundaryzone from -•8 mm/yr alongthe Enriquillo fault, estimatedfrom 2-D modeling of GPS data acquired between 1986 and 1995 in Hispaniola [Dixon et al.
1998], to-•3 mm/yr alongthe Muertos trough. East of Puerto
Rico an additional •-2 mm/yr must transferfrom the western
Anegada passagenorthwardto the easternPuerto Rico trench
to allow
for little
or no motion
between
GORD
and CRO1.
The increased seismicity and focal mechanismsnorth of the
northeasternVirgin Islands are consistentwith this interpretation (Figure 2).
8.
ISLANDS
easternBritish Virgin Islands (GORD) closely approximates
that of St. Croix (CRO 1) on the rigid Caribbeanplate. We infer
that PRVI is attachedto the Caribbeanat its easternedge. Motion of PRVI
relative to North America
is slower than that be-
tween the rigid Caribbean and North American plates, precluding eastward tectonic escapeof PRVI within the plate
boundary zone.
GPS velocities predict left-lateraltranspressionacrossthe
Puerto Rico trench and the Muertos trough south of Puerto
Rico, left-lateral transtensionacrossthe western Anegada pas-
sage,and east-westopeningacrossthe Mona riff. This deformation pattern is not compatiblewith rotation of a rigid PRVI
about a vertical
axis located in the southeast of the island of
Puerto Rico but is consistent
with northeastward
translation
of PRVI relative to North America and easternHispaniola.
Conclusions
We document the existence of a distinct PRVI
block within
the diffuseboundary zone between the North Americanand
Caribbeanplates fromGPS geodeticdata collectedin Puerto
Rico, the Virgin Islands,and easternHispaniola. At the 95%
confidencelevel, GPS velocitiesfor sites in PRVI are equivalent, with the exceptionof one site (GORD) in the eastern
British Virgin Islands. Intrablock displacementthereforeis
lessthana few millimetersper year and deformationassociated
with North American-Caribbean
relative plate motion is limited to the block-boundingstructures. The velocity of the
Acknowledgments.This work benefitedfrom the involvementof
numerouspeople over the years. We are particularly indebtedto J.
Camacho,L. Jaramillo,D. Martinez, H. Rodriguez,D. Lao, and M. Canabal of UPRM
for assistance in data collection.
We also thank the De-
partmentof Marine Science,UPRM, the AgriculturalExperimentStation
in Isabela,PueaoRico, the FAA Headquaaersin SanJuan,Pueao Rico,
andthe GuavaberryVacationHomesin Virgin Gordafor continuedaccess to their facilities. Work was supportedby NSF grants EAR9316215, EAR-9628553,
EAR-9807289,
EAR9806456,
and HRD-
9353549, NASA grantsNAG5-6031 and NCCW-0088, and several
NASA Solid Eaah and Natural Hazard grantsto T.H.D. Puerto Rico
EPSCoRand the College of Arts and Sciences,UPRM, also provided
funding. UTIG contribution1504.Geosciences
Azur contribution
318.
References
Argus, D. F., and R. G. Gordon, Tests of the
rigid-plate hypothesisand bounds on intraplate deformation using geodetic data
from very long baselineinterferometry,J.
Geophys.Res., 101, 13,555-13,572,1996.
Asencio, E., Western Puerto Rico seismicity,
U.S.Geol.Surv.OpenFile Rep. 80-192, 135
pp., 1980.
Blewitt, G., M. B. Heftin, F. H. Webb, U. J.
Lindqwister, and R. P. Malla, Global coordinateswith centimeteraccuracy in the International
Terrestrial
Reference
Frame
usingGPS, Geophys.Res. Lett., 19, 853856, 1992.
Burke, K., J. Fox, and A.M. Celal Sengot,
Bouyantocean floor and the evolutionof
the Caribbean,J. Geophys.Res., 83, 3949-
North America plate motion:Implications
for platerigidityandobliqueplate boundary
convergence,Geophys.Res. Lett., 27, 437440, 2000.
Deng, J., and L. R. Sykes, Determinationof
Euler pole for contemporary
relativemotion
of Caribbeanand North American plates
usingslipvectorsof interplateearthquakes,
Tectonics,14, 39-53, 1995.
Dixon, T. H., and A. Mao, A GPS estimateof
relative motion between North and South
America,Geophys.Res.Letr, 24, 535-538,
1997.
Plate boundaryzone, edited by J. Dolan,
and P. Mann, Spec. Pap. Geol. Soc. Am.
Spec.Paper, 326, 143-161,1998.
Dolan, J., H. Mullins, and D. Wald, Active
tectonics
of the noah-central
Caribbean:
obliquecollision,strainpaaitioning,and opposingsubducted
slabs,in Active strlke-slip
and collisional tectonics of the northern
CaribbeanPlate boundaryzone,editedby J.
Dolan, andP. Mann, Spec. Pap. Geol. Soc.
Am., 326, 1-61, 1998.
Donelly, T. W., Evolutionof easternGreater
Antillean island arc, Am. Assoc. Pet. Geol.
Dixon,T., G. Gonzales,
E. Katsigris,
and S. LiBull., 48, 680-696, 1964.
chten, First epoch geodeticmeasurements Ewing, M., A. G. Lonardi,andJ. I. Ewing, The
with the Global Positioning
Systemacross
sedimentsand topographyof the Pueao
the noahern Caribbean plate boundary
Rico Trenchand outerridge, Trans.Caribb.
3954, 1978.
zone, J. Geophys. Res., 96, 2397-2415,
Geol. Conf., 4, 325-334, 1965.
1991.
Byrne, D. B., G. Suarez, and W. R. McCann,
Muertos Trough subduction--Microplate Dixon,T. H., A. Mao,andS. Stein,Howrigidis Fink, L. K., Jr., and C. G. A. Harrison, Paleomagneticinvestigations
of selectedlava
tectonics in the noahern Caribbean?, Nathe stable interior of the Noah American
units on Puerto Rico, Trans. Caribb. Geol.
ture, 317, 420-421, 1985.
Plate?,Geophys.Res.Lett., 23, 3035-3038,
Conf, 6, 379, 1972.
1996.
Calais, E., and B. Mercier de Lepinay,
for
Semiquantitativemodeling of strain and Dixon,T. H., F. Farina,C. DeMets,P. Jansma, Frankel,A., A compositefocal mechanism
microearthquakesalong the northeastern
kinematics along the Caribbean/North
P. Mann, and E. Calais, Caribbean-North
border of the Caribbeanplate, Geophys.
Americastrike-slipplateboundaryzone, J.
American plate relative motionand strain
Res. Lett., 9, 511-514, 1982.
Geophys.Res.,98, 8293-8308, 1993.
partitioningacrossthe northernCaribbean
Conoily,J. R., and M. Ewing, Sedimentationin
plate boundaryzone from a decadeof GPS Frankel, A., W. R. McCann, and A. J. Murphy,
Observations from a seismic network in the
the Pueao Rico Trench, J. Sediment.Petobservations,d. Geophys. Res., 103,
Virgin Islandsregion: Tectonic structures
rol., 37, 44-59, 1967.
15,157-15,182, 1998.
and earthquakeswarms, d. Geophys.Res.,
DeMets, C., and T. Dixon, New kinematic
Dixon,T. H., M. Miller, F. Farina,H. Wang,
85, 2669-2678, 1980.
models for Pacific-Noah
America motion
and D. Johnson,
Present-day
motionof the
from 3 Ma to present, 1, Evidence for
Sierra Nevada block and some tectonic im-
steadymotion and biasesin the NUVEL-1A
model,Geophys.Res.Lett., 26, 1921-1924,
plications
for theBasinandRangeprovince,
1999.
DeMets,C., R. G. Gordon,D. F. Argus, and S.
Stein,Current plate motions,Geophys.J.
Int., 101, 425-478, 1990.
DeMets, C., P. Jansma,G. Mattioli, T. Dixon, F.
Farina, R. Bilham, E. Calais, and P. Mann,
GPS geodetic constraints on Caribbean-
Noah American Cordillera, Tectonics,19,
1-24, 2000.
Dolan, J. F., and D. J. Wald, The 1943-1953
north-centralCaribbeanearthquakes: active tectonic setting,seismic hazards and
implicationsfor Caribbean-NoahAmerican
platemotions,in Activestrike-slipand collisionaltectonicsof the northernCaribbean
Grindlay,N. R., P. Mann, and J. F. Dolan, Researchers
investigate
submarine
faultsnorth
of Pueao Rico, EOS Trans.AGU,
404, 1997.
78 (38),
Heftin, M., et al., Global geodesyusingGPS
without fiducial sites, Geophys. Res. Lett.,
19, 131-134, 1992.
Hess,H. H., and J. C. Maxwell, Caribbeanresearchproject,Geol.Soc.Am. Bull., 64, I6, 1953.
JANSMA ET AL.: NEOTECTONICS OF PUERTO RICO AND VIRGIN ISLANDS
Holcombe, T. L., C. G. Fisher, and F. A.
Bowles, Gravity-flow depositsfrom the St.
Croix Ridge:Depositionalhistory,Geo Mar.
Lett., 9, 11-18, 1989.
Huerfano, V., Crustal structureand stress regime nearPuertoRico, M.S. thesis,Univ. of
PuertoRico, Mayagtiez, 1995.
Jany,I., A. Mauffret, P. Bouysse,A. Mascle,B.
Mercier de L6pinay, V. Renard, and J. F.
Stephan,Relev6bathym6trique
Seabeam et
tectoniqueen d6crochement
au sud des Iles
Vierges [Nord-Est Caraibes], C R. Acacl.
Sci., Ser. II, 304, 527-532, 1987.
Ladd,J. W., and J. S. Watkins,Active margin
structureswithin the north slope of the
MuertosTrough,Geol. My'nbouw,57, 225260, 1978.
Ladd, J. W., J. L. Worzel, and J. S. Watkins,
Multifold seismic reflection records from
the northern Venezuela Basin and the north
bean underthrustingof the Greater Antilles
along the Muertos Trough, Trans. Caribb.
Geol. Conf, 7, 235-242, 1976.
Mauffret, A., and I. Jany, Collision et tectoniqued'expulsionle long de la frontiere
Nord-Caraibe, Oceanol. Acta, 10, 97-116,
1990.
McCann,W. R., On the earthquakehazardsof
Puerto Rico and the Virgin Islands, Bull.
Seismol.Soc. Am., 75, 251-262, 1985.
McCann, W. R., and W. D. Pennington,Seismicity, largeearthquakes,
andthe marginof
the Caribbean plate, in, The Geology of
North America, vol. H, The Caribbean Re-
gion, edited by G. Dengo and J. Case, pp.
291-306, Geol. Soc. Am., Boulder, Colo.,
1990.
McCann, W. R., and L. R. Sykes,Subductionof
aseismic ridges beneath the Caribbean
plate: Implicationsfor the tectonics and
seismicpotentialof the northeastern
Caribbean, d. Geophys.Res., 89, 4493-4519,
slopeof MuertosTrench,in, MauriceEwing
Ser., Island Arcs, Deep-Sea Trenchesand
1984.
Back-Arc Basins,vol. 1, editedby M. Talwani and W. C. Pitman,pp. 41-56, AGU, Meltzer, A. S., Fault structureand earthquake
Washington,D.C. 1977.
potential,LajasValley, SW PuertoRico, U.
Larue, D. K., and H. F. Ryan, ExtensionaltecS. Geol. Surv. Rep., 16 pp., 1997.
tonismin the Mona Passage,Puerto Rico Mercado, A., and W. McCann, Numerical
simulation of the 1918 Puerto Rico tsunami,
and Hispaniola:a preliminarystudy:Trans.
Caribb.Geol. Conf, 12, 301-313, 1990.
J. Nat. Hazards, 18, 57-76, 1998.
Larue, D. K., and H. F. Ryan, Seismicreflec- Molnar, P., and L. R. Sykes,Tectonicsof the
tion profiles of the Puerto Rico Trench;
Caribbean and Middle America regions
shorteningbetween the North American
from focal mechanisms and seismicity,
Geol. Soc. Am. Bull., 80, 1639-1684, 1969.
and Caribbeanplates, in, Tectonicsand
geochemistry
of the northeasternCaribbean, Moussa,M. T., G. Seiglie, A. A. Meyerhoff,
editedby E.G. Lidiak and D. K. Larue,
and I. Taner, The Quebradillas Limestone
Spec. Pap. Geol. Soc. Am., 322, 193-210,
[Miocene-Pliocene],northern Puerto Rico,
1998.
and tectonics of the northeastern
Larue, D. K., J. Joyce, and H. F. Ryan,
Neotectonicsof the Puerto Rico Trench;
extensional
tectonism
and forearc
subsi-
dence,Trans. Caribb. Geol. Conf, 12, 231247, 1990.
Leroy, S., and A. Maulfret, Intraplate deformationin the Caribbeanregion,d. Geodyn.,
21,113-122, 1996.
Lichten,S. M., Estimation
andfilteringfor high
precision GPS applications, Manuscr.
Geod., 15, 159-176, 1990.
Malfait, B., and M. B. Dinkelman, CircumCaribbeantectonicandigneousactivityand
the evolutionof the Caribbeanplate,Geol.
Soc.Am. Bull., 83, 251-272, 1972.
Mann, P., and K. Burke, Neotectonics of the
Caribbean, Rev. Geophys.,22, 309-362,
1984.
Mann, P., C. Schubert, and K. Burke, Review
of Caribbeanneotectonics,
in TheGeology
of North America, vol. H, The Caribbean
Region,editedby G. Dengo and J. E. Case,
pp. 307-308, Geol. Soc. Am., Boulder,
Colo., 1990.
Mann, P., F. Taylor, L. Edwards, and T. Ku,
Caribbean
margin,Geol. Soc. Am. Bull., 99, 427-439,
1987.
Murphy,A. J., andW. R. McCann, Preliminary
results from a new seismic network
in the
northeasternCaribbean, Seismol.Soc. Am.
Bull., 69, 1497-1513, 1979.
Muszala, S. P., N. Grindlay, and R. T. Bird,
Three-dimensional
Euler deconvolution
and
tectonic interpretationof marine magnetic
anomaly data in the Puerto Rico trench, d.
Geophys.Res., 104, 29,175-29,188, 1999.
Nemec, M. C., A two-phasemodel for the tectonic evolution of the Caribbean, Trans.
/
1037
zone, northeasternHispaniola,J. Geophys.
Research, 100, 6265-6280, 1995.
Sillard, P., Z. Altamimi, and C. Boucher, The
ITRF-96
realization
and its associated ve-
locity field, Geophys.Res. Lett., 25, 32223226, 1998.
Speed, R. C., and D. K. Larue, Extension and
transtensionin the plate boundary zone of
the northeasternCaribbean,Geophys.Res.
Lett., 18, 573-576, 1991.
Stein, S., and R. G. Gordon, Statisticaltests of
additionalplate boundariesfrom plate motion, Earth Planet. Sct. Lett., 69, 401-412,
1984.
Stein, S., C. DeMets, R. G. Gordon, J. Brodholt,
D. Argus, J. F. Engeln, P. Lundgren, C.
Stein, D. S. Weins, and D. F. Woods, A test
of alternativeCaribbeanplate relative motion models, J. Geophys.Res., 93, 30413050, 1988.
Sykes,L. R., W. R. McCann, and A. L. Kafka,
Motion of Caribbeanplate duringlast 7 million years and implicationsfor earlier Cenozoic movements, J. Geophys.Res., 87,
10,656-10,676, 1982.
Van Fossen,M. C., J. T. Channel, and J. H.
Schellekens,Paleomagneticevidence for
Tertiary anticlockwiserotation in southwest
Puerto Rico, Geophys.Res. Lett., 16, 819822, 1989.
van Gestel, J., P. Mann, J. Dolan, and N.
Grindlay,Structureandtectonicsof the upper Cenozoic Puerto Rico-Virgin Islands
carbonate platform as determined from
seismicreflectionstudies,J. Geophys.Res.,
103, 30,505-30,530, 1998.
Ward, S. N., Pacific-North America plate motions: new resultsfrom very long baseline
interferometry, J. Geophys. Res., 95,
21965-21981, 1990.
Zumberge, J. F., M. Heftin, D. Jefferson, M.
Watkins, and F. Webb, Precisepoint positioningfor efficient and robust analysisof
GPS data from large networks,J. Geophys.
Res., 102, 5005-5017, 1997.
E. Calais, Geosciences Azur, Centre Na-
tional de RecherchesScientifique,06560 Val-
bonne, France.
Caribb. Geol. Conf, 9, 23-34, 1980.
Officer, C. B., Jr., R. S. Edwards,J. I. Ewing,
C. DeMets, Department of Geology and
and H. R. Johnson,Geophysicalinvestigations in the eastern Caribbean: Venezuelan
Geophysics,
Universityof Wisconsin,Madison,
Basin, Antilles Island Arc, and Puerto Rico
trench, Geol. Soc. Am. Bull., 68, 359-378,
1957.
WI 53706.
T. H. Dixon, Rosenstiel School of Marine
Pacheco,J. F., and L. R. Sykes, Seismicmoand AtmosphericSciences,University of Mimentcatalogof large shallowearthquakes,
1900 to 1989, Bull. Seismol. Soc. Am., 82,
1306-1349, 1992.
ami, Miami, FL 33149.
P. E. Jansma,Departmentof Geosciences,
Prentice,C., P. Mann, F. W. Taylor, G. Burr,
and S. Valastro, Paleoseismicityof the University of Arkansas, Fayetteville, AR
North
American
plate
boundary 72701.
[SeptentrionalFault], Dominican Republic,
P. Mann, Institutefor Geophysics,
UniverGeology,21, 49-52, 1993.
Reid, H., and S. Taber, The Virgin Islands sity of Texasat Austin,Austin,TX 78759-8500.
Mann, P., C. S. Prentice, G. Burr, L. R. Pena,
earthquakesof 1867-1868, Bull. Seismol.
andF. W. Taylor, Tectonic geomorphology
Soc. Am., 10, 9-30, 1920.
G. S. Mattioli andA. Lopez,Department of
and paleoseismicityof the Septentrional
Reid, J., P. Plumley, and J. Schellekens,Pa- Geology, Universityof PuertoRico, P.O. Box
fault system,DominicanRepublic,in Active
leomagnetic evidence for late Miocene
strike-slipand collisional tectonicsof the
9017, Mayagtiez, PR 00681-9017.
counterclockwise
rotation of the North
northern Caribbean Plate boundary zone,
Coast carbonate sequence, Puerto Rico, (p_jansma•comp.uark.edu;
editedby J. Dolan, andP. Mann, Spec. Pap.
Geophys.Res.Lett., 18, 565-568, 1991.
Geol. Soc. Am., 326, 63-124, 1998.
glen•geology.uprm.edu)
Rosencrantz, E., and P. Mann, SeaMARC I I
Mao, A., C. G. A. Harrison, and T. H. Dixon,
mappingof transformfaults in the Cayman
Noise in GPS coordinate time series, d.
Trough,CaribbeanSea, Geology,19, 690Geophys.Res., 104, 2797-2816, 1999.
693, 1991.
Masson, D. G., and K. M. Scanlon, The
(Received October 18, 1999;
Russo,R. M., and A. Villasefior, The 1946 Hisneotectonicsetting of Puerto Rico, Geol.
revised
May 8, 2000;
Soc.Am. Bull., 103, 144-154, 1991.
paniolaearthquakeand the tectonicsof the
Matthews, J., and J. Holcombe, PossibleCaribNorth America-Caribbean plate boundary acceptedJune8, 2000.)
Actively evolvingmicroplateformationby
obliquecollisionandsidewaysmotionalong
strike-slipfaults: An example from the
northeastern
Caribbeanplate margin, Tectonophysics,
246, 1-69, 1995.