The boundary between the North American and Caribbean plates is characterized primarily by left-lateral motion along predominantly east-west striking faults. Seismicity and marine geophysical survey data suggest that at least two, and possibly three, microplates occupy the diffuse boundary zone in the northeastern Caribbean (Figure 1): 1) the Gonave (Rosencrantz and Mann, 1991); 2) the Hispaniola (Byrne et al., 1985; Mann et al., 1995); and 3) the Puerto Rico-Virgin Islands block (PRVI) (Byrne et al., 1985). We have acquired GPS geodetic measurements since 1994 at several sites (Figure 2) to test the microplate hypothesis, define PRVI translation and rotation within the boundary zone, and constrain PRVI neotectonics. GPS-derived velocities are analyzed with respect to both North American and Caribbean plate reference frames. The motions of PRVI sites relative to North America and the Caribbean are 16.9±1.1 mm/yr toward N68°E±3° (1sigma) and 2.4±1.4 mm/yr toward S79°W±26° (1sigma), respectively (Jansma et al., 2000). All sites in PRVI have velocities that are equivalent at the 95% confidence limit (Figure 2). The discrepancy in the velocities between Puerto Rico sites and one in Virgin Gorda at the eastern extreme of PRVI is close to 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 velocities on PRVI permits several mm/yr of slip, the absence of any obvious geographic pattern in the velocities argues against substantial organized deformation within PRVI and suggests that any deformation is slower than the several mm/yr resolution of the present GPS velocities. We solved for a stronger upper bound on intra-PRVI deformation by examining the evolution of baseline length between the two continuous sites in western Puerto Rico, GEOL and PUR3. Since continuous measurements began at both stations in June 1997, the baseline length has remained constant within error at -0.5±0.3 mm/yr (1sigma). Intrablock deformation, therefore, is limited and deformation is accommodated along bounding structures of the microplate.

a) Process ISAB-PARG baseline in point position mode with GIPSY. This baseline extends north-south across western Puerto Rico and traverses all the onland structures that are proposed to be active. Students at the undergraduate level will run UNIX scripts that effectively treat GIPSY processing as a “black box.” If the course is modified for the graduate level, students will make a substantial investment in learning the mathematics and motivation behind the software.
b) Determine the change in velocity of each site by the addition of data from successive campaigns. Students will compare the calculated velocity with the “law of diminishing returns” for geodetic measurements and the impact of time since first observation. For all PRVI sites, we have a minimum of 3 epochs of data. For some, we have as many as 6 epochs. Concepts of precision and accuracy are illustrated. Error analysis is critical to the interpretation. As each set of campaign data is added, the amount of internal deformation possible in PRVI decreases.
c) Compare the CGPS time series for PUR3, GEOL, and CRO1 with those from sites in the dry, western US. The noise in the data from the three sites in the humid tropics is greater than that for sites in arid regions. Seasonal variations also are more significant in the tropics. Students will learn about different types of noise and that running campaigns at different times of the year will potentially yield inaccurate velocities and increase errors.