Pergamon

INTRAPLATE

0264-3707(95)00037-2

DEFORMATION

J. Geodymmics Vol. 21, No. 1, pp. 113-122, 1996 Elsevier Science Ltd. Printed in Great Britain 0264-3707/96 $15.OO+O.Otl

IN THE CARIBBEAN REGION

S. LEROY and A. MAUFFRET Department de Gtotectonique, Case 129, Univ. P. et M. Curie, CNRS-URA 1759, 4 Place Jussieu, 75252, Paris Cedex 05, France (Received 11 September 1995; accepted 25 September 1995)

Abstract-The Caribbean plate is commonly described as a single feature surrounded by large plates. A new seismic survey indicates that intraplate defomation occurs in the area of Beata ridge, a structure that lies between the Colombian and Venezuelan basins. The Pecos fault zone, at the southern tip of the Beata ridge, has been an active transpressive feature since the early Miocene. The deformation increases towards the North and cannot be related to the subduction of the Caribbean oceanic plateau beneath South America. A NE-SW to E-W stress is inferred from the shape of the structure and determination of focal mechanisms. We conclude that relative displacement exists between Colombian and Venezuelan microplates and we examine the implication on the plate motion of the Caribbean region.

INTRODUCTION

The main component of the Caribbean plate is a large volcanic plateau probably formed in the Pacific Ocean. This plate moves towards the east relative to the North America (NOAM) and South America (SOAM) plates (Pindell and Barrett, 1990). Consequently the Caribbean plate is limited (Fig. 1) to the north by a left-lateral strike-slip fault zone (North America-Caribbean plate boundary) and to the south by a complex set of right-lateral faults. In addition, the NOAM and SOAM plates are slowly converging with a present pole situated near the Mid-Atlantic ridge (Pindell and Barrett, 1990; Muller and Smith, 1993; Miiller et al., in press). Deformed sedimentary belts (Ladd and Watkins, 1978; Ladd and Watkins, 1980; Ladd et al., 1990) north of South America and south of Puerto-Rico result from this compression which increases towards the west. The 4000 to 5000 m deep Colombian and Venezuelan basins are separated by the 2000 m deep Beata ridge. Many speculative models have been proposed for the formation of this ridge: normal faulting (Fox et al., 1970; Fox and Heezen, 1975; Holcombe et al., 1990); a buoyant thick oceanic plateau which resists subduction (Burke et al., 1978) or reverse faulting related to a transpressive motion (Vitali, 1985; Mauffret et al., 1994). New multichannel seismic profiles, acquired during the CASIS cruise (April 1992) in the Aruba Gap area (Figs 1 and 2) demonstrate strong transpressive tectonic features along this ridge. 113

Fig. 1. Plate tectonic framework of the Caribbean regron. The pole for NOAM-Venezuelan microplate and for NOAM-Colombian microplate are from Heubeck and Mann (1991). The pole of NOAM-SOAM is from Mtiller and Smith (1993). We suppose that Gonave. Hispaniola, Peurto-Rico and the Venzuelan basin can be considered as a single microplate moving eastwards relative to the NOAM plate at a velocity of 2 cm/a. The velocity rate of the Colombian basin could be 2.8 cm/a. The differential motion is absorbed in collision in southern Hispaniola and transpression in Beata ridge area.

ARUBA

TECTONIC

SETTING

The Aruba Gap is located between the Beata ridge and the South American deformed belt and forms a sill between the Colombian and Venezuelan basins. The Beata ridge is a 20 km thick Cretaceous volcanic plateau in its northern part but it thins to 10 km in the Aruba Gap area (Edgar et al., 1971). The western Venezuelan basin has about the same crustal thickness (10 km; Ladd and Watkins, 1980) whereas the crust of the Colombian basin is probably thinner. The acoustic basement of this basin is deeper than that of the Venezuelan basin. The acoustic basement of the eastern Colombian basin and northwestern Venezuelan basin, is relatively smooth and we can identify the typical Caribbean B” reflector; therefore, rough true oceanic crust is localized in the western part of the Colombian basin (Bowland and Rosencrantz, 1988) and in the southeastern part of the Venezuelan basin (Diebold et al., 1981). DSDP Site 153 (Edgar and Saunders, 1973) allows calibration of the seismic profiles (Hoipkins, 1973) and identification of a prominent middle Miocene horizon, an early Miocene marker at the top of a disturbed layer, the middle Eocene A” reflector, and the top of the Cretaceous volcanic plateau (B” reflector, Fig. 3). The interval velocity between the B” marker and the prominent sub-B” reflector is 4.5-5 km/s (Stoffa et al., 1981; Fig. 4). The acoustic basement of the Colombian and Venezuelan basins dips to the south towards the South American accretionary prism, which has been active since the early to middle Miocene (Biju-Duval et al., 1982).

115

Fig. 2. Structural map of Aruba Gap. The boundary between the Colombian basin and Beata ridge is outlined by the Pecos fault zone. This structure is low towards the south and is buried beneath the South American accretionary prism, then the Pecos fault zone rises towards the north and a second deformation zone appears along the southern flank of the structure. The V reflector shows the characteristics of volcanic sills and is different from the smooth reflector B”. A focal mechanism solution indicates a northeast-southwest stress (Kafka and Weidner, 1981). The location of the DSDP Site 153 and the seismic profile shown in Fig. 3 arc indicated.

PECOS

FAULT

ZONE

The Pecos fault zone (Fig. 2) was first described by Hopkins (1973) but its orientation was misinterpreted. The northwest-southeast trend was observed in the seismic data (Stoffa et al., 1981) but because these early data were not migrated, the reverse faults were not identified. These faults, with a throw of 100-200 m, are now clearly displayed on the Casis seismic lines (Figs 3 and 4). The reverse faults are facing towards the northeast and southwest (Figs 2, 3 and 4) and the Pecos fault zone is clearly a transpressive feature (positive flower structure; Harding, 1974). The calibration of the main reflectors by the DSDP Site 153 (Hopkins, 1973) and the identification of the early Miocene reflector in the eastern Colombian basin at the base of a well layered seismic unit (Fig. 3) by seismic correlation with the western Colombian basin, indicates that the main tectonic event occurred during the early Miocene (23 Ma). The post early Miocene layers of the Colombian basin onlap the flank of the Pecos fault zone (Fig. 3) but some recent reactivations are evident and the seafloor is deformed

S. Leroy and A. Mauffret

Intraplate deformation in the Caribbean region

117

NE

t4

Fig. 4. Structural cross-sections of the Pecos fault zone. The structure rises towards the north and the seafloor is deformed in the northernmost seismic profile (Cas AO6). Observe the difference in thickness from north to south of the post-A” layer. This southwards thickening is caused by the general dip towards the South American deformed belt.

(Fig. 4; profile Cas A06). In the southern part of the study area, near and below the accretionary prism, the deformation is weak and the Pecos structure is low (less than 0.4 km of throw relative to the Colombian basin; Cas AlO, Figs 4 and 5). Towards the north, the deformation increases (Cas A07, Fig. 3; Cas A08, Fig. 4) and a new area of deformation appears at the base of the Pecos fault zone (Figs 3 and 4). The top of the Pecos fault zone is 1 km higher than the floor of the Colombian basin (5.4 km and 6.4 km deep respectively; Fig. 5). The isobath of basement map indicates clearly, a deepening of the Pecos Fault zone towards the south (6.4 km maximum on the south and 4.6 km on the north). Nevertheless, we do not know the initial topography in this area. The A”-B” interval has about the same thickness elsewhere but this layer is pelagic (Edgar and Saunders, 1973) and may drape a pre-existent structure. The early Miocene-A” interval (Fig. 3) is much thicker (800 m) in the Colombian basin than in the Venezuelan basin (300 m) and this observation indicates that some topography (500 m?) existed prior to the Miocene deformation. In addition, the map of post-B” sediments thickness indicates a greater thickness into the eastern Colombian basin (1200 to 3000 m; Fig. 6) than into the western Venezuelan basin (800 to 1000 m; Fig. 6). The B”-subB” interval and the crust are thinner in the Colombian basin than in

S. Leroy and A. Mauffret

118

72”(

72”Jp’W

"W

/

/ /

(

h g

14”00’N

13”WN

‘Pecos

faultzonr

Fig. 5. Basement isobath map. Depths to basement, m m, arc calculated from multichannel seismic profiles (contour interval 100 m) with the velocities obtained by the refraction study (Edgar ef al.. 1971) and by the seismic processing (Dix interval velocities).

the Venezuelan basin and this is again an argument to suppose a different initial level between the two basins. In conclusion, the 1 km step between the Pecos fault zone and the Colombian basin results from an initial topography (500 m?) and a Miocene reactivation (500 m?). We made our estimation where the Beata ridge is moderately uplifted, but northwards, the ridge was uplifted probably about 1000 m (Benson et al.. 1970). Taking into account the 500 m initial topography, we estimate the shortening of Cas A08 and Cas A06 (Fig. 4), 13.5 km and 20 km respectively. The distance between these two lines being 10 minutes, the shortening is about 40 km per degree. From 13”40N, where we note the first evidence of compression, to 18”N (south coast of Hispaniola) the shortening can be roughly estimated to 170 km.

Intraplate deformation in the Caribbean region 73’1

W -

119 72”1 ‘W

72”3O’W

I

\

\

VP

23” 3O’N

South America De Fig. 6. Total sediments isopach map. Sediment thickness, in m, is calculated as the basement isobath. Contour interval is 100 m.

DISCUSSION

This short description of the Pecos fault zone indicates that the deformation increases towards the north and consequently is not directly related to the subduction of the volcanic plateau (Burke et al., 1978). Moreover, the trend of the Pecos fault zone is almost perpendicular to the strike of the South America accretionary prism. A first-motion solution for an earthquake in the southern Beata ridge by Molnar and Sykes (1969) was reinterpreted by Kafka and Weidner (1981). Strike-slip and reverse motions were determined respectively. The main stress is oriented northeast-southwest in the two interpretations (Fig. 2). However, the focal solution of this earthquake is poor and a better determination in the Venezuelan basin indicates an east-west trending compressive stress field (Kafka and Weidner, 1979; Kafka and Weidner, 1981). These seismological studies indicate that the intraplate deformation of the Caribbean

120

S. Leroy and A. Mauffret

plate is not directly related to the Muertos and South American compressional zones and that north-south to northeast-southwest compressional features may exist within the Caribbean plate. The northeast-southwest orientation of the Pecos fault zone is compatible with a left-lateral strike-slip motion with a component of compression. We estimate the shortening of the Pecos structure and we extrapolate this estimation to about 170 km, south of Hispaniola (Haiti), where the Caribbean plateau is presently colliding with the northern part of Hispaniola (Mercier de Lepinay et al., 1988; Heubeck and Mann, 1991). We failed to identify any subduction zone and the compression is mainly localized along the western edge of the Beata ridge whereas the eastern edge is bounded by a steep northeast-southwest scarp that is probably a right lateral strike-slip fault. The Cayman strike-slip system is divided into two branches (Fig. 1): a northern branch from the Cayman spreading centre to the Puerto-Rico trench and a southern branch from Central America to Haiti. The western part of the southern branch (Plantain Garden-Enriquillo fault; Mann and Burke, 1984) is clearly active (Sykes et al., 1982; Rosencrantz and Mann, 1991). A seismic slip rate of 0.8 cm/a has been estimated in Haiti by Mocquet and Aggarwal (1983). This rate corresponding to 160 km of motion since the early Miocene, is compatible with the 180 km of shortening estimated above if the compressional component is oblique to the main motion. The Beata ridge is an oceanic plateau whose edges have been reactivated by differential motion between a Colombian and a Venezuelan microplate. In this interpretation, the Gonave (Rosencrantz and Mann, 1991), northern Hispaniola and Puerto-Rico microplates are attached to the Venezuelan microplate, but this first approximation is not perfect because Hispaniola and Puerto-Rico are separated from the Venezuelan basin by the Muertos trough (Ladd and Watkins, 1978; Bryne et al., 1985; Ladd et al., 1990). Nevertheless, compression and rotation are observed at the boundaries of the Hispaniola and Puerto-Rico microplates (Jany et al., 1990; Ladd et al., 1990; Mauffret and Jany, 1990; Masson and Scanlon, 1991) and these perturbations do not affect the eastwards motion of the Caribbean plate relative to NOAM. In the same way, the Colombian microplate may be divided into several blocks by the Pedro and Hess escarpments (Fig. I). However, the movement of the Plantain Garden fault in Jamaica is estimated to 0.4 cm/a (Burke et al., 1980) whereas the velocity along the Enriquillo fault in Haiti could be as high as 0.8 cm/a (Mocquet and Aggarwal, 1983). The difference from 0.8 cm/a to 0.4 cm/a may reflect a strike-slip motion along the Pedro and Hess escarpments. We agree with Heubeck and Mann (1991) who placed a pole for NOAM-Colombian microplate in the south (Stein et al., 1988) and a pole for NOAM-Venezuelan microplate in the north (Sykes et al.. 1982; Fig. I). The differential motion between the two plates may result from the greater convergence of NOAM-SOAM in the western part of the Caribbean zone than in the eastern part (Heubeck and Mann, 1991; Miiller et al., in press) or from the influence of the large convergent motion

Intraplate

deformation

in the Caribbean

region

121

of the Cocos plate. In the eastern Colombian basin, intraplate deformation has been observed and the results of the DSDP Site 154 (Edgar and Saunders, 1973) indicate that the uplift of the high where the DSDP site 154 was drilled, is caused by an early Pliocene reverse fault. However, the convergent motion of the Cocos plate is mainly absorbed by the Middle America trench and we favour the first hypothesis. CONCLUSIONS

The southern tip of the Beata ridge has been uplifted relative to the Colombian basin since the early Miocene, and in this area the acoustic basement may be representative of the basement of the Colombian basin that is deeply buried beneath recent sediments. The deformation results from an east-west stress between two microplates. However, the Beata ridge is an oceanic plateau and the thick crust was formed during the Cretaceous volcanic event and has not been thickened during the compressional deformation. The initital topography was totally disturbed during the early Miocene by the formation of the South American deformed belt and the uplift of the edges of Beata ridge. Acknowledgemenrs--Supported by grants INSU ATP 733 and 780. We thank the officers and crew of the R/V Nadir for assistance in this project. We are especially indebted to the technical crew of GENAVIR who greatly helped in the acquisition of multichannel data. These data were processeed at Institut de Physique du Globe de Strasbourg and we thank R. Schlich and M. Schaming who have facilitated access to the processing centre and helped us to use the Geovector software. The authors are grateful to Paul Mann (Univ. of Texas Institute for Geophysics), and to John Diebold (Lament Doherty Earth Observatory of Columbia University) for helpful comments and improvements to the text. Contribution of URA 718 and URA 1759.

REFERENCES Benson W. E., Gerard R. D. and Hay W. W. (1970) Summary and conclusions. in initial Reports ofthe Deep Sea Drilling Project (Bader R. G. and Gerard R. D.. (eds). Washington D.C., U.S. Government Printing Office. pp. 659-673. Biju-Duval B., Mascle A., Rosales H. and Young G. (1982) Episutural Oligo-Miocene basins along the north Venezuela margin, In Margin Association of Petroleum Geoligists. Memoir (Watkins J. S. and Drake C. L., eds.), Am. Assoc. Petrol. Geol. pp. 347-358. Bowland C. L. and Rosencrdntz E. (1988) Upper crustal structure of the western Colombian Basin. Geol. Sot. Am. Bull. 100, 534-546. Bryne D., Saurez G. and McCann W. R. (1985) Muertos Trough subduction; Microplate tectonics in the northern Caribbean. Nature 317. 420-421. Burke K.. Fox P. and Sengor A. M. C. (1978) Buoyant ocean floor and the evolution of the Caribbean. Journul ofGeophysical Research 83, 3949-3954. Burke K., Grippi J. and Sengor A. M. C. (1980) Neogene structure in Jamaica and the tectonic style of the northern Caribbean plate houndary zone. J. Geol. 88, 375-386. Dichold J. B., Stoffa P., Buhl P. and Truchan M. (19X1) Venezuela Basin crustal structure. Journal oj Geophysicul Reseurch 86, 7901-7923. Edgar T. N., Ewing J. and Hennion J. (1971) Seismic refraction and reflection in the Caribbean Sea. Am. Assoc. Petrol. Geol. Bull. 5.5, 833-870. Edgar T. N. and Saunders J. 9. (1973) Initial Report of the DSDP. Leg 15. D.S.D.P., 15. Fox P. J. and Heezen 9. C. (1975) Geology of the Caribbean crust. In The ocean basins and murgins. V-3, The Gulf of Mexico and the Caribbean (Nairn A. E. M. and Stehli T. G., eds.), New York. pp. 421-466. Fox 1’. J.. Ruddiman W. F.. Ryan W. 9. F. and Hcezcn 9. C. (lY70) The geology of the Carihbcan crust. I: Beata ridge. Tectorrophys. 10. 3YS-513. llarding T. P (1074) Petroleum traps associated with wrench faults. Am. Assoc. Petrol. Geol. Bull. 5X. 12YO1304.

I Ic,ubcch ( .III~ M:I~II I’. (I99I) (ic.oIt):~c ~\~ILI,III~III 01 plalc hlncmatlc nroclcl\ t