Figure 2.    Probability for damaging ground motion for San Juan and Puerto Rico in comparison to other high seismic risk areas in the U.S. such as Seattle.  (Source: A. Frankel, USGS)

Eighty-eight tsunamis have been reported for the entire Caribbean Region in the past 500 years, including 14 tsunamis reported from Puerto Rico and the Virgin Islands (Lander et al., submitted). 30 tsunamis caused significant damage including reports of as many as 9600 fatalities, which can be attributed to tsunamigenic earthquakes and tsunamis combined. 1922 deaths are confirmed as being specifically related to tsunamis during the last 150 years. A 1918 M 7.5 earthquake resulted in a tsunami that killed at least 116 people in northwestern Puerto Rico. A runup of about 6 m has been documented by mapping, and sedimentary evidence for at least two earlier tsunamis in the area has been cited. Eyewitness reports of the St. Croix tsunami give a maximum wave height of 7.6 m in Frederiksted. Immediately after the 1946 Hispaniola earthquake a tsunami struck northeastern Hispaniola, followed by a series of smaller tsunamis, probably caused by continuing landsliding. The initial tsunami was as much as 4-5 m high and propagated inland for several kilometers, drowning, according to some reports, nearly 1,800 people.

A seismic network was established in Puerto Rico in 1979 and was operated by the Lamont-Doherty Earth Observatory of Columbia University for the first 5 years. The network has since been operated and by the University of Puerto Rico, Mayaguez. A sparse strong motion instrument network has operated in Puerto Rico upgraded by the University of Puerto Rico, Mayaguez since the 1970s. Starting in 1995, the network has been modernized and densified to include many more sites, both rural and within large buildings in urban areas.

Figure 3.    Numerical simulation of the 1918 Puerto Rico Tsunami showing calculated inundation of the Agmadilla coast, NW Puerto Rico due to an earthquake-generated Tsunami from Mona Passage to the Northwest. heavy black line = coast line thin lines = topographic contours (Source: E. Mercado, U. Puerto Rico)

The only published earthquake hazard map for Puerto Rico is by W. McCann (1994). This map considers different source areas, such as the Puerto Rico trench, the Mona Passage, the Muertos Trough, and the Anegada Passage. It does not calculate the risk of rupture of specific faults, because of lack of knowledge. Preliminary hazard maps were presented during the workshop by the USGS, Dames and Moore Consultants, and Risk Management Solutions, Inc. These maps follow the assumptions stated in McCann (1994).

Global Positioning System (GPS) geodesy is an essential technique to constrain complex strain within the Caribbean-North American plate boundary zone. GPS studies in the eastern Greater Antilles and Virgin Islands have been ongoing since 1986 and results obtained in 1994 (CANAPE project) yielded velocities for six sites separated by ~800 km in east-west extent within the boundary. Dixon et al. (1998) used these preliminary GPS-derived velocities in concert with geological interpretation and simple elastic half-space models to conclude the following:

1. The relative motion of the Caribbean plate with respect to the North American plate is essentially east at ~2 cm/yr, a rate twice as high as predicted by the NUVEL-1A global plate motion model (DeMets et al., 1990).
2. The majority of motion east of Cabo Rojo, located in the southernmost Dominican Republic, must be accommodated to the north of Puerto Rico, since the measured velocity at Isabela in Puerto Rico was similar to that at Cabo Rojo at the 95% confidence level.
3. Relative motion between North America and the Caribbean plate is localized along one offshore fault north of the Dominican Republic and at least two discrete onshore faults in the Dominican Republic.

Since 1994, the regional network has been steadily improved with the addition of several continuous GPS sites, CR01 in St. Croix (10/95), GEOL in Mayaguez, Puerto Rico (9/96), PUR3 in Aguadilla, Puerto Rico (6/97), and over 30 campaign sites installed in the Dominican Republic in 1994 and 1998 and 12 campaign sites installed in Puerto Rico and adjacent islands. In addition, reoccupation in 1998 of San Andres Island and Aves Island in the western and eastern Caribbean plate interior, respectively, permitted development of a well constrained Caribbean plate velocity (Jansma et al., 1999). The current density of sites within the boundary zone should allow us to address: 1) the existence of a single, rigid block consisting of Puerto Rico and the Virgin Islands; 2) the relative motion between the eastern Dominican Republic and western Puerto Rico and concomitant opening rate of the Mona Canyon; 3) possible elastic strain accumulation along mapped structures in the Dominican Republic and identification of other currently unrecognized zones, where displacement is being accommodated; and 4) offshore displacement to the north and south of western Puerto Rico.

Results to date have produced a well constrained regional velocity fields referenced to either stable North America or the Caribbean plates. Additional sites and longer time series on existing sites, however, are required to constrain velocities along potentially seismogenic structures both offshore and onshore. Unpublished GPS data and analysis do allow certain broad conclusions to be made without the benefit of the recently densified Dominican Republic and Puerto Rico networks. Major conclusions directly relevant to seismic hazard assessment are:

1. The majority of the slip between North America/Caribbean plate is taken up offshore of northern Puerto Rico, but some also is accommodated to the south, probably along the Muertos Trough.
2. Relative motion of Puerto Rico/Virgin Islands relative to North America is northeast (more northerly than that between North America /Caribbean plate), yielding a component of convergence across the Puerto Rico Trench (PRT), i.e. PRT is transpressional.
3. Tectonic setting of the Anegada Passage/Whiting Basin/Virgin Islands Basin is left-lateral transtension.
4. Mona Canyon is characterized by left-oblique opening with east-west extension.
5. GPS velocities from campaign sites in western and eastern Puerto Rico agree within 1 sigma (~1 mm/yr), implying that Puerto Rico is rigid; data from the CGPS site at GEOL show a small residual in the longitudinal velocity component at the 95% confidence level, the origin of which is yet to be determined.

NASA and the Department of Defense have recently funded laser altimetry and IFSAR surveys by the University of Puerto Rico to generate improved digital elevation models for the eastern Greater Antilles (Dominican Republic, Puerto Rico, Virgin Islands) for a variety of applications, including morphotectonic analysis and fault identification. Laser altimetry data from an airborne Lidar instrument were collected in February 1997.

Several marine geophysical surveys have been conducted around Puerto Rico and the Virgin Islands during the last 30 years. A Gloria side-scan survey with coincident single-channel seismic recording mapped giant submarine slides north of Puerto Rico. An extensive multibeam and single-channel seismic survey by the R/V Ewing in 1996 (van Gestel et al., 1998) mapped the deep water north of Puerto Rico. A striking finding from this survey was a continuous fault at the base of the north slope of the Puerto Rico margin.

Figure 4.    Seismic reflection profile showing block faulting near St Croix, USVI, collected by the USGS during GLORIA sidescan survey of the U.S. Exclusive Economic Zone.  (Source: K. Scanlon, USGS)

Deep seismic data in the northeastern Caribbean are sparse. A few short refraction lines using old recording and analysis technology were collected in the Puerto Rico Trench in 1959 (Bunce and Fahlquist, 1962). Multichannel seismic reflection data were collected both by academia and by industry in the 1970s. Deep reflections were not observed on these data because of the short receiver array (2400 m) and the small volume of the airguns (<2000 cu. in.) used at the time.

References: