The Caribbean lithospheric plate as a whole is now
bounded east and west by subduction zones, the west-dipping Lesser Antilles and
east-dipping Middle America subduction zones, respectively (Figs 1, 8, 10), and
to the north and south by transcontractional (transpressional), mostly
strike-slip fault zones. Most of
the boundary (Figs 1-3) is reasonably well defined by seismicity (e.g., van der
Hilst, 1990; Pennington et al., 1990; Deng and Sykes, 1995). However, in several places the boundary
is poorly defined, because the relative velocity between the Caribbean and the
North and South American continents is quite small [about 2.5 cm/year (Dixon
and DeMets, 1997; Dixon and Mao, 1997)], and the boundary is a wide zone of
distributed shear.
The east-west-trending Caribbean-South American
Plate Boundary Zone in northern and off-shore Venezuela (Figs. 2-3) is a shear
zone between 300 and 500 km in width, consisting of four EW-trending
lithotectonic
belts: (1) The South Caribbean Deformed Belt in the north; (2) The Leeward
Antilles volcanic arc; (3) The Caribbean Mountains system in northern
Venezuela; and (4) The Serran’a del Interior foreland fold and thrust
belt in the south (Figs 2-3).
Northern South America comprises an autochthonous, Cretaceous-Early Tertiary passive continental margin, which has been overthrust by an accreted belt of allochthonous rocks formerly belonging to the Cretaceous-Early Tertiary offshore Leeward Antilles arc complex (Figs 2, 4; Table 1). This arc complex formed on an oceanic plateau in the Pacific, possibly as an intra-oceanic arc stretching from Mexico to Ecuador. After the Cretaceous the arc converged upon northern South America. Collision and accretion have progressed from the west in the Paleocene to the east in modern times forming a right lateral oblique collision zone resulting from Caribbean and South American relative motion. Thus we can study this zone of arc-continent accretion at a variety of stages by looking at different meridians through the Plate Boundary Zone. Figure 4 shows a series of plate tectonic reconstructions outlining this history, a summary of which is given in Table 1. Figs 12 & 13 show the active source corridors that represent different stages of the collision-accretion history.
Rifting and Drifting phases. The autochthonous continental margin of Venezuela formed during the Middle and Late Jurassic (e.g., red beds/basalts of 162 Ma; Feo-Codecido et al., 1984) as a result of the break-up of Pangea (Table 1; Fig. 4). Presumably, the continental crust was initially stretched and thinned (Lugo and Mann, 1997; di Croce and Bally, 1999). The paleo-coastline is shown in Fig. 5. During the drift phase (Late Jurassic to end Cretaceous), sea floor was created by NW-SE divergence between North America and South America (Proto-Caribbean lithosphere). A fairly thick sequence of sediments was deposited on the passive Venezuelan margin (e.g., Villamil and Pindell, 1998; Fig. 4).
NE migration of the Caribbean. In
the west before mid-Cretaceous time, the "Great Arc of the Caribbean"
(Burke, 1988) consisting at least of the Greater Antilles, the Aves Ridge, the
basement of the Leeward Antilles, and parts of western Colombia and Ecuador,
was part of the west-facing, western North and South American arc system (Fig.
4). Between 120 and 110 Ma, the polarity of this arc reversed and the arc began
to migrate northeastward, overriding and consuming Proto-Caribbean lithosphere
and eventually colliding, along its southern part, with western Colombia in
Late Cretaceous (Fig. 4). The HP/LT metamorphic rocks of northern Venezuela,
described in a later section, are thought to have formed in the Cretaceous
subduction zone after the polarity reversal. By the late Cretaceous, a new arc
had formed in the Pacific (Costa Rica-Panama arc) far to the west of the first
arc, which first gave the Caribbean Plate its own identity by separating it
from its parental Farallon plate (e.g., Pindell et al., 1988). In the
Eocene the
Great Caribbean Arc collided with the Bahamas Bank, coincident with a change in
South American-Caribbean relative motion.
Eastward migration of the Caribbean. After
the Eocene collision with the Bahamas, the Caribbean plate migrated easterly
relative to North America but southeasterly relative to South America in
response to convergence between the Americas that continues today (Pindell et
al., 1988; Figs 4-5), making the Caribbean-South American plate boundary zone a
long-lived zone of right lateral convergence (Figs. 5 and 6). In the early part
of this oblique migration, back-arc spreading within the Great Arc of the
Caribbean produced the Grenada Basin between the Lesser Antilles arc and the
Aves Ridge (Bird et al., 1993), which coincided with the arc wrapping around
the continental corner of Colombia (Pindell, 1993). The southern part of the
arc, southeast of the Grenada Basin (Fig 4), collided with South America first
in the west during the Paleocene to Eocene (Pindell et al., 1988; Audemard,
1991) and then progressively to the east, like a wave breaking obliquely along
a beach (Fig. 4), leaving behind arc fragments and accreted metamorphic
terranes along the length of the margin. This wave of arc collision with South
America resulted in a succession of sequenced tectonic events along the margin
and in northern South America, including cessation of arc magmatism, formation
of strike-slip structures and associated basins, arc fragmentation, metamorphic
terrane accretion and exhumation, and folded belt/foreland basin
formation.
Initially, the Caribbean plate (with arc and HP/LT metamorphic belts) was obducted with from west to east onto South America giving rise to the Caribbean Mountain system and the Serran’a del Interior foreland fold and thrust belt. The obduction and consequent loading of the South American crust resulted in the formation of the eastward-younging (Pindell et al., 1988; 1998) foreland basins. After a time by unclear processes subduction polarity reversed:, The Caribbean Plate now subducts beneath the Leeward Antilles arc and South America starting in the west and extending now possibly as far east as 62o, as indicated by tomographic images (van der Hilst ,1990; van der Hilst and Mann, 1994; Figs 7, 8; also see Ysaccis, 1998).
When subduction polarity reverses, it signifies
the completion of arc accretion at that point along the margin: The point of
initial collision, obduction of the Caribbean terranes onto South America, and
the completion of arc accretion then migrates eastward with the relative
eastward motion of the Caribbean plate. After the polarity reversal, the
entire fold and thrust belt (Caribbean Mountain system and Serran’a del
Interior) became a zone of orogenic float (Oldow et al., 1990; Fig. 14 &
15), that is, the Caribbean Mountains and the Serran’a del Interior are
detached, with detachment faults either extending under the entire plate
boundary from the Leeward Antilles arc to the front of the Serran’a, or
extending from the metamorphic belts along the coast to the front of the folded
belt. Within this zone of float, right-lateral transforms with pull-apart
basins and restraining bends have formed (Fig. 2). These faults are the primary
locus of the right lateral strike-slip component of the ongoing relative plate
motion.
Russo and Silver (1996) have hypothesized that deep mantle resistance to trenchward motion during South America-Nazca plate convergence has resulted in the formation of the Andean Cordillera, its modern geometry, and Cenozoic eastward relative motion of the Caribbean and Scotia Sea plates. This concept predicts a slab-parallel flow field in the upper mantle under the Nazca plate. Corner flow around the southern and northern ends of the subducting Nazca plate results in the easterly migration of the Scotia Sea and Caribbean Plates relative to South America during the Oligocene-Miocene. Plate tectonic/terrane reconstructions of the structural elements of the Caribbean-South American plate boundary (e.g., Fig. 4) qualitatively support this hypothesis as they show lithospheric fragments such as the Maracaibo block and the Leeward Antilles arc being swept around the northern edge of South America along the Bocon— Fault and into the Mor—n and El Pilar and related faults during the Cenozoic (Norton, Exxon Production Research presentation at Rice workshop, October 1997). If this hypothesis is true, a broad (200-500km) zone beneath the plate boundary will show S-wave anisotropy oriented in the flow direction. Available S-wave splits in northeastern South America support this model (Silver, 1996; Fig 16). However, neither the mantle tomography models (e.g. Grand 1994, van der Hilst, 1990, White et al., 1998), nor the S-wave splitting data are extensive enough to clearly define the northernmost edge of the South American craton and its tectospheric root, or details in the plate boundary zone.