Untitled Document

One of the fundamental questions concerning lithospheric behavior in zones of active tectonics is how relative motion between blocks or plates is accommodated. Is motion taken up along a few major faults or narrow belts of deformation separating discrete crustal blocks or is strain broadly distributed throughout the deforming region as similar displacements occur on many closely spaced faults of comparable size? The two end-members have different implications for seismic hazard. Larger earthquakes within narrow zones are likely to occur in the former, whereas smaller events over a broad region are more probable in the latter. Which end-member more closely characterizes Puerto Rico and the northern Virgin Islands has long been controversial, i.e. do Puerto Rico and the northern Virgin Islands define a discrete, rigid block within the North American-Caribbean plate boundary within which little motion occurs or do several faults capable of accommodating significant displacement cross the region? Recent results from GPS geodesy provide upper bounds on the potential intrablock displacement.

Puerto Rico and the northern Virgin Islands define the eastern terminus of the Greater Antilles, which extend eastward from offshore eastern Central America to the Lesser Antilles volcanic arc and mark the boundary between the Caribbean and North American plates. Tectonic models for the northern Caribbean (e.g., Byrne et al., 1985; Mann et al., 1995) propose active microplates within the boundary zone on the basis of geologic and earthquake evidence. Seismicity along the east-west trending boundary between the North American and Caribbean plates is consistent with evolution of the boundary zone from a relatively simple set of transform faults in the west to a more complex deformation zone approximately 250 km wide in the east (Figure 1). Motion along the predominantly east-west striking major structures of the northern Caribbean is primarily left-lateral. In the west, the Swan and Oriente transform faults define the EW-trending Cayman trough and bound the short (~100 km), NS-trending Mid-Cayman spreading center. In Hispaniola, Puerto Rico and the Virgin Islands, the Puerto Rico trench and the Muertos trough define the northern and southern limits of the plate boundary zone, respectively. Three microplates lie within this diffuse boundary zone (Figure 1): 1) the Gonave in the west (Mann et al., 1995); 2) the Hispaniola in the center (Byrne et al., 1985); and 3) the Puerto Rico-northern Virgin Islands (PRVI) in the east (Masson and Scanlon, 1991). Such a microplate model assumes that nearly all of the deformation associated with North America-Caribbean motion is concentrated along the faults that bound the three rigid blocks: the Oriente, Septentrional, Enriquillo-Plantain Garden, and Anegada faults, the Muertos trough and North Hispaniola deformed belt, and the Mona rift faults northwest of Puerto Rico (Figure 1).

Hundreds of earthquakes per year occur within and around Puerto Rico and the Virgin Islands (Figure 2). The majority of events are small and located offshore. Several large events have occurred during historic time, including the 1916, 1918, and 1943 Mona Passage earthquakes (Ms=7.2, 7.3, and 7.5 respectively), the 1867 Anegada earthquake (Ms=7.3), the 1787 Puerto Rico trench earthquake (M=7.5?) and the 1670 San German earthquake (M=6.5?) (Pacheco and Sykes, 1992). The concentration of both current seismicity and historic events offshore is consistent with a Virgin Islands block (PRVI) in the northeastern corner of the Caribbean (Byrne et al., 1985; Masson and Scanlon, 1991).

The highest levels of onshore seismicity are in the Lajas Valley (Asencio, 1980), an EW-trending feature in southwestern Puerto Rico, which continues offshore to the west and passes south of the southern termination of the Mona Canyon, where offshore faults are mapped (Figure 3). Quaternary offsets along the southern edge of the Lajas Valley were interpreted from seismic reflection profiles (Meltzer et al., 1995; Meltzer, 1997) and active faulting was inferred from the “basin-and-range” style topography of the region (Joyce et al., 1987). Displacements along the east-west striking faults are inferred to be normal with components of strike-slip (Meltzer, 1997; Almy et al., 2000; Prentice et al., 2000). In addition, the island of Puerto Rico is traversed by two northwest-southeast striking fault zones: 1) the Great Northern Puerto Rico fault zone (GNPRFZ) and 2) the Great Southern Puerto Rico fault zone (GSPRFZ) (Figure 3). The fault zones were active during the Eocene and record predominantly thrust and left-lateral displacement in response to amalgamation of discrete island arc terranes at the leading edge of the Caribbean plate into PRVI (Glover and Mattson, 1960; Glover, 1971; Erickson et al., 1990; 1991). Both the GNPRFZ and the southern end of the GSPRFZ are covered by little deformed Neogene strata. The two fault zones, however, represent large areas of weakness within PRVI along which active displacements may be localized. Indeed, the southern end of the GSPRFZ immediately offshore may cut and disturb Recent shelf sediments (McCann, 1985; Joyce, personal communication). The projection of the northern end of the GSPRFZ, which continues offshore into Mona Canyon, is sub-parallel to faults of similar orientation (NW/SE), which are seismically active (McCann, 1985; Joyce et al., 1987). In addition, an EW-striking splay of the GSPRFZ, the Cerro Goden fault, cuts across to the west coast of Puerto Rico about 10 kilometers north of the city of Mayagüez (Figure 3). Whether Quaternary motion occurred along the Cerro-Goden fault is unknown, although the offshore projection of the fault merges with other mapped structures that presumably are Quaternary in age. Lao-Davila et al. (2000) infer Recent displacement with components of normal motion and left-lateral strike-slip on the basis of offset stream drainages and terraces. Some workers have identified the surficial expressions of the Cordillera and Joyuda faults, which they argue may correspond to a WNW/ESE trend across southwestern Puerto Rico that is defined by a series of epicenters of small earthquakes that were recorded by the Puerto Rico Seismic Network in 1995 (Moya, personal communication). Geological estimates for displacements along the fault are unconstrained.

GPS measurements were first collected in the northeastern Caribbean in 1986 at six locations (Dixon et al., 1991), which were re-occupied subsequently as part of CANAPE (CAribbean-North American Plate Experiment) in 1994. The original sites were: 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). The network was densified during CANAPE and each subsequent year (for details, see Dixon et al., 1998; Jansma et al., 2000; Calais et al., 2002; Mann et al., 2002). Since 1994, measurements have been made on subsets of the entire network each year. A permanent IGS station was established in St. Croix in 1995 (CRO1) and a vector tie to the original 1986 site, STCX, was established (Dixon et al., 1998), which extended the time series by nearly a decade and therefore improved the CRO1 velocity estimate. 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-Mayagüez), ZSUA (San Juan), MONA (Mona island), DSCH (Desecheo island), ADJU (Adjuntas), ARC2 (Arecibo), CCM5 (Ponce), FAJA (Fajardo), LAJ1, LAJ2, and LAJ3 (Lajas Valley), SALN (Salinas), VIEQ (Vieques), and ANEG (Anegada, British Virgin Islands) and continuous sites GEOL in Mayagüez, FAJA in Fajardo, UPRR in Rio Piedras, and UPRH in Humacao operated by the Department of Geosciences, University of Arkansas, and PUR3 in Aguadilla maintained by the U.S. Coast Guard.

Results from GPS geodesy confirm the presence of an independently translating PRVI whose motion is 2.6±2.0 mm/yr toward N82.5°W±34° (95%) with respect to the Caribbean (Figure 4). Geodetic data are consistent with east-west extension of several mm/yr from eastern Hispaniola to the eastern Virgin Islands. Extension increases westward with the most, 5±3 mm/yr, accommodated in the Mona rift, confirming earlier GPS geodetic results. East-west extension of 3±2 mm/yr per year also is observed across the island of Puerto Rico, consistent with composite focal mechanisms and regional epicentral distributions. Although the loci of extension are not known, similarity of GPS-derived velocities among sites in eastern Puerto Rico suggest the active structures lie west of the San Juan metropolitan area. Re-activation of the Great Northern and Southern Puerto Rico fault zones as oblique normal faults with right-lateral slip is a possibility. East-west extension of 2±1 mm/yr also must exist between eastern Puerto Rico and Virgin Gorda, which likely is attached to the Caribbean plate. These extensional belts allow eastward transfer of slip between North America and the Caribbean from the southern part of the plate boundary zone in the west to the northern segment in the east. Motions along or across any of the individual subaerial structures of Puerto Rico are ≤ 2 mm/yr. The Lajas Valley in the southwest, where microseismicity is greatest, is the locus of highest permissible on-land deformation. Northwest-southeast to east-west extension of 2±1 mm/yr is also observed across the Anegada Passage.

While we have unequivocally established that individual site velocities derived from the preliminary GPS dataset have measurable and statistically significant residuals relative to the Caribbean plate, the mechanism responsible has not yet been established. The current surface deformation field may be treated as a manifestation of 1) a kinematically distinct microplate between the North American and Caribbean plates; 2) a reflection of elastic strain accumulation along major offshore faults or the subduction interface; or 3) a viscoelastic response to large historical earthquakes. Determining which of these possibilities accounts for the residual velocities is an essential constraint to seismic hazard and risk analysis for the region. For example, in the case of elastic strain accumulation along the plate interface, the potential of large, devastating earthquakes, for which Puerto Rico is not adequately prepared, is significant. The probability of frequent, large events is less likely for a freely translating microplate. Quantitative modeling of the results, therefore, is a key objective of future research.

The questions we address are:
• What are the permissible displacement rates along the mapped faults given the GPS geodetic results?
• Do these rates agree with those derived from the geologic data?
• Are these rates consistent with significant and, therefore, possibly destructive earthquakes occurring along subaerial faults within PRVI?

This is just one small piece of the northern Caribbean. Lots of people have worked with us on this and other GPS projects in the Caribbean/North America plate boundary zone over the years as part of CANAPE, including Eric Calais, Chuck DeMets, Tim Dixon and Paul Mann. Check the following links for more information on GPS in the northern Caribbean in general and on GPS for specific regions of the northern Caribbean:

Eric Calais, Purdue University (Hispaniola)

Chuck DeMets, University of Wisconsin (Jamaica)

Tim Dixon, RSMAS (original CANAPE work)

Paul Mann, University of Texas (Caribbean geology)

Funding provided by USGS-NEHRP; Puerto Rico Sea Grant; NSF-EAR, Geophysics; NSF-CREST; and the College of Arts and Sciences, University of Puerto Rico, Mayagüez.

Microplate tectonics in the northeastern Caribbean

One of the fundamental questions concerning lithospheric behavior in zones of active tectonics is how relative motion between blocks or plates is accommodated. Is motion taken up along a few major faults or narrow belts of deformation separating discrete crustal blocks or is strain broadly distributed throughout the deforming region as similar displacements occur on many closely spaced faults of comparable size? The two end-members have different implications for seismic hazard. Larger earthquakes within narrow zones are likely to occur in the former, whereas smaller events over a broad region are more probable in the latter. Which end-member more closely characterizes Puerto Rico and the northern Virgin Islands has long been controversial, i.e. do Puerto Rico and the northern Virgin Islands define a discrete, rigid block within the North American-Caribbean plate boundary within which little motion occurs or do several faults capable of accommodating significant displacement cross the region? Recent results from GPS geodesy provide upper bounds on the potential intrablock displacement. Puerto Rico and the northern Virgin Islands define the eastern terminus of the Greater Antilles, which extend eastward from offshore eastern Central America to the Lesser Antilles volcanic arc and mark the boundary between the Caribbean and North American plates. Tectonic models for the northern Caribbean (e.g., Byrne et al., 1985; Mann et al., 1995) propose active microplates within the boundary zone on the basis of geologic and earthquake evidence. Seismicity along the east-west trending boundary between the North American and Caribbean plates is consistent with evolution of the boundary zone from a relatively simple set of transform faults in the west to a more complex deformation zone approximately 250 km wide in the east (Figure 1). Motion along the predominantly east-west striking major structures of the northern Caribbean is primarily left-lateral. In the west, the Swan and Oriente transform faults define the EW-trending Cayman trough and bound the short (~100 km), NS-trending Mid-Cayman spreading center. In Hispaniola, Puerto Rico and the Virgin Islands, the Puerto Rico trench and the Muertos trough define the northern and southern limits of the plate boundary zone, respectively. Three microplates lie within this diffuse boundary zone (Figure 1): 1) the Gonave in the west (Mann et al., 1995); 2) the Hispaniola in the center (Byrne et al., 1985); and 3) the Puerto Rico-northern Virgin Islands (PRVI) in the east (Masson and Scanlon, 1991). Such a microplate model assumes that nearly all of the deformation associated with North America-Caribbean motion is concentrated along the faults that bound the three rigid blocks: the Oriente, Septentrional, Enriquillo-Plantain Garden, and Anegada faults, the Muertos trough and North Hispaniola deformed belt, and the Mona rift faults northwest of Puerto Rico (Figure 1). Hundreds of earthquakes per year occur within and around Puerto Rico and the Virgin Islands (Figure 2). The majority of events are small and located offshore. Several large events have occurred during historic time, including the 1916, 1918, and 1943 Mona Passage earthquakes (Ms=7.2, 7.3, and 7.5 respectively), the 1867 Anegada earthquake (Ms=7.3), the 1787 Puerto Rico trench earthquake (M=7.5?) and the 1670 San German earthquake (M=6.5?) (Pacheco and Sykes, 1992). The concentration of both current seismicity and historic events offshore is consistent with a Virgin Islands block (PRVI) in the northeastern corner of the Caribbean (Byrne et al., 1985; Masson and Scanlon, 1991). The highest levels of onshore seismicity are in the Lajas Valley (Asencio, 1980), an EW-trending feature in southwestern Puerto Rico, which continues offshore to the west and passes south of the southern termination of the Mona Canyon, where offshore faults are mapped (Figure 3). Quaternary offsets along the southern edge of the Lajas Valley were interpreted from seismic reflection profiles (Meltzer et al., 1995; Meltzer, 1997) and active faulting was inferred from the “basin-and-range” style topography of the region (Joyce et al., 1987). Displacements along the east-west striking faults are inferred to be normal with components of strike-slip (Meltzer, 1997; Almy et al., 2000; Prentice et al., 2000). In addition, the island of Puerto Rico is traversed by two northwest-southeast striking fault zones: 1) the Great Northern Puerto Rico fault zone (GNPRFZ) and 2) the Great Southern Puerto Rico fault zone (GSPRFZ) (Figure 3). The fault zones were active during the Eocene and record predominantly thrust and left-lateral displacement in response to amalgamation of discrete island arc terranes at the leading edge of the Caribbean plate into PRVI (Glover and Mattson, 1960; Glover, 1971; Erickson et al., 1990; 1991). Both the GNPRFZ and the southern end of the GSPRFZ are covered by little deformed Neogene strata. The two fault zones, however, represent large areas of weakness within PRVI along which active displacements may be localized. Indeed, the southern end of the GSPRFZ immediately offshore may cut and disturb Recent shelf sediments (McCann, 1985; Joyce, personal communication). The projection of the northern end of the GSPRFZ, which continues offshore into Mona Canyon, is sub-parallel to faults of similar orientation (NW/SE), which are seismically active (McCann, 1985; Joyce et al., 1987). In addition, an EW-striking splay of the GSPRFZ, the Cerro Goden fault, cuts across to the west coast of Puerto Rico about 10 kilometers north of the city of Mayagüez (Figure 3). Whether Quaternary motion occurred along the Cerro-Goden fault is unknown, although the offshore projection of the fault merges with other mapped structures that presumably are Quaternary in age. Lao-Davila et al. (2000) infer Recent displacement with components of normal motion and left-lateral strike-slip on the basis of offset stream drainages and terraces. Some workers have identified the surficial expressions of the Cordillera and Joyuda faults, which they argue may correspond to a WNW/ESE trend across southwestern Puerto Rico that is defined by a series of epicenters of small earthquakes that were recorded by the Puerto Rico Seismic Network in 1995 (Moya, personal communication). Geological estimates for displacements along the fault are unconstrained. GPS measurements were first collected in the northeastern Caribbean in 1986 at six locations (Dixon et al., 1991), which were re-occupied subsequently as part of CANAPE (CAribbean-North American Plate Experiment) in 1994. The original sites were: 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). The network was densified during CANAPE and each subsequent year (for details, see Dixon et al., 1998; Jansma et al., 2000; Calais et al., 2002; Mann et al., 2002). Since 1994, measurements have been made on subsets of the entire network each year. A permanent IGS station was established in St. Croix in 1995 (CRO1) and a vector tie to the original 1986 site, STCX, was established (Dixon et al., 1998), which extended the time series by nearly a decade and therefore improved the CRO1 velocity estimate. 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-Mayagüez), ZSUA (San Juan), MONA (Mona island), DSCH (Desecheo island), ADJU (Adjuntas), ARC2 (Arecibo), CCM5 (Ponce), FAJA (Fajardo), LAJ1, LAJ2, and LAJ3 (Lajas Valley), SALN (Salinas), VIEQ (Vieques), and ANEG (Anegada, British Virgin Islands) and continuous sites GEOL in Mayagüez, FAJA in Fajardo, UPRR in Rio Piedras, and UPRH in Humacao operated by the Department of Geosciences, University of Arkansas, and PUR3 in Aguadilla maintained by the U.S. Coast Guard. Results from GPS geodesy confirm the presence of an independently translating PRVI whose motion is 2.6±2.0 mm/yr toward N82.5°W±34° (95%) with respect to the Caribbean (Figure 4). Geodetic data are consistent with east-west extension of several mm/yr from eastern Hispaniola to the eastern Virgin Islands. Extension increases westward with the most, 5±3 mm/yr, accommodated in the Mona rift, confirming earlier GPS geodetic results. East-west extension of 3±2 mm/yr per year also is observed across the island of Puerto Rico, consistent with composite focal mechanisms and regional epicentral distributions. Although the loci of extension are not known, similarity of GPS-derived velocities among sites in eastern Puerto Rico suggest the active structures lie west of the San Juan metropolitan area. Re-activation of the Great Northern and Southern Puerto Rico fault zones as oblique normal faults with right-lateral slip is a possibility. East-west extension of 2±1 mm/yr also must exist between eastern Puerto Rico and Virgin Gorda, which likely is attached to the Caribbean plate. These extensional belts allow eastward transfer of slip between North America and the Caribbean from the southern part of the plate boundary zone in the west to the northern segment in the east. Motions along or across any of the individual subaerial structures of Puerto Rico are ≤ 2 mm/yr. The Lajas Valley in the southwest, where microseismicity is greatest, is the locus of highest permissible on-land deformation. Northwest-southeast to east-west extension of 2±1 mm/yr is also observed across the Anegada Passage. While we have unequivocally established that individual site velocities derived from the preliminary GPS dataset have measurable and statistically significant residuals relative to the Caribbean plate, the mechanism responsible has not yet been established. The current surface deformation field may be treated as a manifestation of 1) a kinematically distinct microplate between the North American and Caribbean plates; 2) a reflection of elastic strain accumulation along major offshore faults or the subduction interface; or 3) a viscoelastic response to large historical earthquakes. Determining which of these possibilities accounts for the residual velocities is an essential constraint to seismic hazard and risk analysis for the region. For example, in the case of elastic strain accumulation along the plate interface, the potential of large, devastating earthquakes, for which Puerto Rico is not adequately prepared, is significant. The probability of frequent, large events is less likely for a freely translating microplate. Quantitative modeling of the results, therefore, is a key objective of future research. The questions we address are: • What are the permissible displacement rates along the mapped faults given the GPS geodetic results? • Do these rates agree with those derived from the geologic data? • Are these rates consistent with significant and, therefore, possibly destructive earthquakes occurring along subaerial faults within PRVI? This is just one small piece of the northern Caribbean. Lots of people have worked with us on this and other GPS projects in the Caribbean/North America plate boundary zone over the years as part of CANAPE, including Eric Calais, Chuck DeMets, Tim Dixon and Paul Mann. Check the following links for more information on GPS in the northern Caribbean in general and on GPS for specific regions of the northern Caribbean: Eric Calais, Purdue University (Hispaniola) Chuck DeMets, University of Wisconsin (Jamaica) Tim Dixon, RSMAS (original CANAPE work) Paul Mann, University of Texas (Caribbean geology) Funding provided by USGS-NEHRP; Puerto Rico Sea Grant; NSF-EAR, Geophysics; NSF-CREST; and the College of Arts and Sciences, University of Puerto Rico, Mayagüez.