Untitled Document

Along obliquely convergent plate boundaries, the relative plate motion vector frequently is partitioned into arc-normal and arc-parallel components. Complete separation of the normal and parallel components leads to arc-parallel thrusts and folds in the forearc and arc-parallel strike-slip faults in the volcanic arc. The forearc thus translates relative to the interior of the overriding plate. Partial separation of the two components, however, is more common, particularly when coupling occurs across the plate interface. In addition, changes in obliquity of convergence along the plate boundary can create arc-parallel gradients in shear stress to produce internal deformation of the forearc. Either arc-parallel extension or compression may occur.

The northeastern Caribbean corner affords an excellent opportunity to use Global Positioning System (GPS) geodesy to examine strain partitioning, coupling along the plate interface, and arc-parallel extension in a slow subduction zone. Caribbean-North American relative plate motion is ~20 mm/yr. The trend of the plate boundary changes from NNW in the Lesser Antilles to EW along the eastern Greater Antilles (Figure 1). The ENE azimuth of the convergence vector predicts near normal subduction at Martinique and near strike-slip motion at Puerto Rico. Horizontal slip directions derived from shallow-thrust earthquakes east of the Lesser Antilles lie between the predicted convergence direction and the normal to the subduction trace, suggesting that some partitioning occurs.

We focus on the northern Lesser Antilles for a number of reasons. First, we have been collecting data in the region since 1994 and have a reasonably dense mixed-mode network (Figure 2). Second, GPS-derived velocities for the southern Lesser Antilles agree within error with those predicted for Caribbean motion relative to North America at those locations, implying little or no deformation of the overriding plate. In contrast, a simple elastic half-space model shows that preliminary GPS geodetic results from the northern Lesser Antilles are consistent with coupling along the plate interface (Figure 3). Third, the northern Lesser Antilles arc consists of two concentric island chains that provide subaerial exposure closer to the trench and over a greater arc transverse distance than do the islands in the south. Fourth, NE-trending structures transverse to the arc in the northern Lesser Antilles accommodate a component of arc-parallel extension. In addition, volcanic centers on individual islands have migrated parallel to the arc at rates equivalent to those of the arc-parallel component of convergence. Finally, EW-oriented extension along NS-trending structures has been measured by GPS geodesy for the EW-trending eastern Greater Antilles.

Following the formulation of Savage (1983), a simple elastic half-space model was developed to examine coupling along the plate interface in the northern Lesser Antilles along a cross-section oriented N45E from the trench through Aves Island in the Caribbean interior at a distance of 475 km from the trench (Figure 4). Shallow slab dip of 30° and down-dip length of the locking zone of 150 km were assumed from seismicity (Speed et al., 1984). The expected vertical displacements were calculated for various degrees of coupling (25%, 50%, and 100%). The GPS-derived vertical displacement rate for the CGPS station in Antigua (BGGY) is shown for comparison (Figure 4 and inset). BGGY was selected as it has the longest time series and thus provides the current best estimate of the vertical signal. The vertical displacement rate at BGGY is consistent within error with some coupling along the plate interface. Longer time series at BGGY are required in addition to more stations to constrain coupling further. The model predicts the maximum vertical signal where the new CGPS site for Barbuda (BUDA) is scheduled for installation.

The questions we address are:
Is oblique slip partitioned into trench parallel and normal components in slow subduction zones, i.e. the northern Lesser Antilles?
If partitioning occurs, does it depend upon obliquity of convergence?
What is the degree of coupling along the plate interface in the northern Lesser Antilles?
Is arc-parallel extension of sufficient magnitude for detection by GPS geodesy? If so, does it increase in magnitude as obliquity of subduction increases?

Determining whether the leading edge of the Caribbean plate corner in the eastern Greater Antilles and the Lesser Antilles is composed of distinct block(s) that move relative to the Caribbean interior also will constrain seismic hazard in the northeastern Caribbean, an area that is little studied and particularly at risk for seismogenic tsunamis.

Current funding provided by NSF-EAR, Tectonics. Previous funding from NASA-URC.

Geodynamics of the Lesser Antilles arc

Along obliquely convergent plate boundaries, the relative plate motion vector frequently is partitioned into arc-normal and arc-parallel components. Complete separation of the normal and parallel components leads to arc-parallel thrusts and folds in the forearc and arc-parallel strike-slip faults in the volcanic arc. The forearc thus translates relative to the interior of the overriding plate. Partial separation of the two components, however, is more common, particularly when coupling occurs across the plate interface. In addition, changes in obliquity of convergence along the plate boundary can create arc-parallel gradients in shear stress to produce internal deformation of the forearc. Either arc-parallel extension or compression may occur. The northeastern Caribbean corner affords an excellent opportunity to use Global Positioning System (GPS) geodesy to examine strain partitioning, coupling along the plate interface, and arc-parallel extension in a slow subduction zone. Caribbean-North American relative plate motion is ~20 mm/yr. The trend of the plate boundary changes from NNW in the Lesser Antilles to EW along the eastern Greater Antilles (Figure 1). The ENE azimuth of the convergence vector predicts near normal subduction at Martinique and near strike-slip motion at Puerto Rico. Horizontal slip directions derived from shallow-thrust earthquakes east of the Lesser Antilles lie between the predicted convergence direction and the normal to the subduction trace, suggesting that some partitioning occurs. We focus on the northern Lesser Antilles for a number of reasons. First, we have been collecting data in the region since 1994 and have a reasonably dense mixed-mode network (Figure 2). Second, GPS-derived velocities for the southern Lesser Antilles agree within error with those predicted for Caribbean motion relative to North America at those locations, implying little or no deformation of the overriding plate. In contrast, a simple elastic half-space model shows that preliminary GPS geodetic results from the northern Lesser Antilles are consistent with coupling along the plate interface (Figure 3). Third, the northern Lesser Antilles arc consists of two concentric island chains that provide subaerial exposure closer to the trench and over a greater arc transverse distance than do the islands in the south. Fourth, NE-trending structures transverse to the arc in the northern Lesser Antilles accommodate a component of arc-parallel extension. In addition, volcanic centers on individual islands have migrated parallel to the arc at rates equivalent to those of the arc-parallel component of convergence. Finally, EW-oriented extension along NS-trending structures has been measured by GPS geodesy for the EW-trending eastern Greater Antilles. Following the formulation of Savage (1983), a simple elastic half-space model was developed to examine coupling along the plate interface in the northern Lesser Antilles along a cross-section oriented N45E from the trench through Aves Island in the Caribbean interior at a distance of 475 km from the trench (Figure 4). Shallow slab dip of 30° and down-dip length of the locking zone of 150 km were assumed from seismicity (Speed et al., 1984). The expected vertical displacements were calculated for various degrees of coupling (25%, 50%, and 100%). The GPS-derived vertical displacement rate for the CGPS station in Antigua (BGGY) is shown for comparison (Figure 4 and inset). BGGY was selected as it has the longest time series and thus provides the current best estimate of the vertical signal. The vertical displacement rate at BGGY is consistent within error with some coupling along the plate interface. Longer time series at BGGY are required in addition to more stations to constrain coupling further. The model predicts the maximum vertical signal where the new CGPS site for Barbuda (BUDA) is scheduled for installation. The questions we address are: Is oblique slip partitioned into trench parallel and normal components in slow subduction zones, i.e. the northern Lesser Antilles? If partitioning occurs, does it depend upon obliquity of convergence? What is the degree of coupling along the plate interface in the northern Lesser Antilles? Is arc-parallel extension of sufficient magnitude for detection by GPS geodesy? If so, does it increase in magnitude as obliquity of subduction increases? Determining whether the leading edge of the Caribbean plate corner in the eastern Greater Antilles and the Lesser Antilles is composed of distinct block(s) that move relative to the Caribbean interior also will constrain seismic hazard in the northeastern Caribbean, an area that is little studied and particularly at risk for seismogenic tsunamis. Current funding provided by NSF-EAR, Tectonics. Previous funding from NASA-URC.