Marine Geophysical Researches 22: 17–45, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

17

Prolonged Magmatic and Tectonic Development of the Caribbean Igneous Province Revealed by a Diving Submersible Survey Alain Mauffret1,∗ , Sylvie Leroy, Jean-Marie Vila2 , Erwan Hallot3 , Bernard Mercier de L´epinay4 & Robert A. Duncan5 1 D´ epartement

de G´eotectonique, CNRS ESA 7072, Universit´e P. et M. Curie, Paris, France; 2 Lab. de G´eologie S´edimentaire et Pal´eontologie, Universit´e Paul Sabatier, Toulouse, France; 3 G´eosciences Rennes, Universit´e de Rennes, France; UMR-GeoAzur, Valbonne, France; College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, USA; ∗ Corresponding author (Tel: +33-144-275176; Fax: +33-144-275085; E-mail: [email protected]) Received 1 October 1999, accepted 27 November 2000

Key words: Large Igneous Province, Carillean Plate, diving

Abstract Using the diving submersible survey NAUTICA we investigated the central part of the Caribbean large igneous province (CLIP) to observe and sample internal portions of this proposed oceanic plateau. Most of the samples are gabbroic and doleritic rocks; basalts are scarce. Radiometric dating by 40 Ar/39 Ar incremental heating experiments indicate that the intrusive rocks are Campanian in age (81–75 Ma). In some places these intrusive rocks underlie older Santonian (85–83 Ma) extrusive basaltic rocks, suggesting that the Campanian rocks represent a sill injection and an underplating episode. Results of the diving program supplemented by information from ODP and DSDP drilling sites document a 20 m.y. period (94–75 Ma) of igneous activity in the submerged portion of the Caribbean large igneous province (CLIP). In the northern part of the Beata Ridge late Campanian and/or post Campanian uplift is documented by prominent Maastrichtian (71–65 Ma) erosion and the establishment of a Paleocene-middle Eocene (65–49 Ma) carbonate platform. During and after the uplift an extensional period is indicated by seismic images and the subsidence (3 km depth) of the carbonate platform. Paleocene ages (55–56 Ma) determined on some volcanic samples are attributed to localised decompression mantle melting that accompanied the extension. We document a prolonged period of magmatic and tectonic events that do not fit with the current models of shortlived plateau formation during mantle plume initiation but shares many similarities with the constructional histories of other oceanic large igneous provinces.

Introduction The Caribbean Plate has been inserted between the larger North and South American Plates and currently has an eastward motion relative to these plates (Figure 1; Pindell and Barrett, 1990). North-south compression is also indicated by the deformed zones located in the northern and southern boundaries of the Caribbean Plate. The results of DSDP and ODP drilling (Edgar et al., 1973; Sigurdsson et al., 1997b) demonstrated that much of the central Caribbean region bas an igneous basement that formed during a short period in Late Cretaceous time (late Turonian-

Campanian, 90–81 Ma). The Caribbean Plate is currently interpreted to have formed as an oceanic large igneous province (LIP) covering a 600,000 km2 area formed in the Pacific realm (Farallon Plate), possibly above the Galapagos hotspot (Duncan and Hargraves, 1984; Pindell and Barrett, 1990). Recent geophysical studies (Leroy, 1995; Mauffret and Leroy, 1997) indicate that the Caribbean plate has a composite crustal thickness that varies from 20 km to 5 km. Moreover, several pieces of the primitive Caribbean LIP that have been accreted to the adjacent continents (Figure 1), should be included in the surface extent of

18

Figure 1. The Caribbean large igneous province is formed by the volcanic basement beneath the Caribbean Sea and several fragments accreted to the adjacent lands. On the Pacific side The Santa Nicoya complex and the Komatiites of Gorgona have been dated by the 40 Ar/39 Ar method 88–90 Ma (early Coniacian-late Turonian) and 86–88 Ma (Coniacian) respectively (Sinton et al., 1998). A younger age (Canipanian) have been obtained with the same tecbnique for the Serrania de Baudo (74–78 Ma; Kerr et al., 1997). North of the Caribbean Sea the Dumisseau formation (La Selle massif, Hispaniola) has been dated by 40 Ar/39 Ar method 88–92 Ma (early Coniacian-late Turonian; Sinton et al., 1998) and south of Cribbean Sea the Curaçao lavas are 78–88 Ma (Campanian-Coniacian) old. Within the Caribbean Sea the basalt collected at ODP Site 1001 is dated by 40 Ar/39 Ar method 81 Ma (Campanian); the dolerites sampled at DSDP Site 146 are 90–92 Ma (Turonian) and at DSDP Site 150–94 Ma (late Cenomanian) (Sinton et al., 1998). On Beata Ridge 90 Ma is the late Turonian age determined by biostratigraphy of the sediments resting upon the basalts drilled at DSDP Site 153 (Edgar et al., 1973). 75 Ma (late Canipanian) age is based on the 40 Ar/39 Ar method applied to the samples of the NAUTICA diving survey although younger ages (55 Ma) were also obtained. A 2000-km diameter circle has been drawn to compare with other volcanic provinces. Note that this circle is tangent to the komatiites of Gorgona and picrites of Nicoya Peninsula (Alvarado et al., 1997), Curaçao and Duarte (Dupuis et al., 1997) but does not include all the Caribbean LIP.

the LIP (Kerr et al., 1997). For a comparison with other similar igneous province (White and McKenzie, 1989; White and McKenzie, 1995) we draw a 2000-km diameter circle centered on the Colombia basin (Figure 1). This circle is not enough wide to include all the Cretaceous magmatic rocks but it shows that some parts of the magmatic plateau are missing

particularly in the southeast quarter. Therefore, the surface and the volume of the Caribbean LIP cannot be accurately estimated. Radiometric dating by the 40 Ar/39 Ar incremental heating method shows that the igneous complexes of Gorgona, Nicoya, Dumisseau (La Selle Massif in Hispaniola), Curaçao and Colombia (Serrania de Baudo and western Cordillera) ac-

19 creted terranes are approximately coeval (92–74 Ma, Turonian-Campanian, Figure 1) with the submerged, intact portion of the Caribbean LIP. In the Caribbean Basin, DSDP and ODP holes sampled only the top of the Cretaceous igneous basement and basaltic rocks were recovered in most of the sites except in the Venezuela basin where dolerites sills and massive flows were sampled at DSDP Sites 146 and 150. Geophysical data suggest (Leroy et al., 1996; Mauffret and Leroy, 1997) that the extrusive basaltic top layer is thin and maybe absent in tectonic windows. The 20-km thick Beata Ridge (Ewing et al., 1960; Edgar et al., 1971; Case et al., 1990), in the central Caribbean Sea (Figure 2), could be one of these tectonic window and could be the footwall of a major normal fault that exposes deep levels of the plateau. In 1996 we investigated structure and composition of igneous rocks and sediments that crop out along the western, faulted-bounded escarpment of the Beata Ridge using the NAUTILE submersible on the R/V NADIR.

Bathymetry, geological sampling, tectonic and geophysical settings

300 km wide from the western Haiti sub basin to the eastern Venezuela basin or Puerto Rico and Dominican sub basins. The W–E profile shape of the Beata Ridge is strongly asymmetrical with relatively steep scarps bounding the Haiti sub basin, whereas the eastern slopes are more gentle and interrupted by several hills and N–S trending ridges. We have named some of these secondary topographic features from north to south: Diorys Hill rising to 2.7 km water depth; Tairona Ridge, rising to 1 km below sea level; and DSDP 151 Ridge, rising to 2 km water depth both at its northern and southern parts (Figures 2 and 3). Although topographic profiles of some of these secondary features (Diorys and Tairona) are consistent with volcanic constructions, the present survey shows them to be of tectonic origin. Preparation for the NAUTICA diving survey on the Beata Ridge (dives NB04 to NB14) was based on a seabeam survey (Seacarib 1) carried out in 1985 on R.V. Jean Charcot (Figure 3) and published recently (Mauffret and Leroy, 1999). A short seabeam survey, performed during a transit of the R.V. Jean Charcot from Panama to Fort de France, made possible the two southernmost dives (NB15 and NB16). Geological sampling

Bathymetry The Venezuela basin and the Puerto Rico sub basin are limited by the 4 km bathymetric contour, but depths greater than 5 km are reached in the Muertos trough on the north and the Venezuela basin on the south margin of the plateau (Figure 2). The floor of the Puerto Rico sub basin is punctuated by several conical seamounts. One of these features, Kathy’s seamount, was surveyed during the R/V M. Ewing 9501 cruise (Diebold et al., 1999) and the hydrosweep swath acquired during that cruise was used to navigate our dive NB01 (Figure 2). The Muertos trough was surveyed during the Seacarib seabeam cruise (Figure 3; Mercier de Lépinay et al., 1988) and this map was used to locate dives NB02 and NB03. The trough extends westward up to the northern part of the Beata Ridge that we named the North Beata Hill (where NB02 was located, Figure 2). Extending southward from the southernmost tip of Hispaniola, and rising from depth of 4 km, the Beata Ridge is a major structure whose summits, far from the south coast of Hispaniola, rise to between 1 and 2 km water depth (Figure 3). The ridge is about 450 km long, trending in a NNE–SSW direction from Hispaniola to the Colombia basin. It is up to about

The Venezuela basin, the Dominican and Puerto Rico sub basins are floored by doleritic sills that intrude late Turonian-Santonian sediments drilled at DSDP Sites 146 and 150 (Donnelly et al., 1973) (Figure 2). These dolerites have been subsequently dated by 40 Ar/39 Ar method at 90–94 Ma (Sinton et al., 1998). The Pleistocene to late Oligocene sedimentary sequence at Site 151 located at the crest of the ridge, consists of foraminifera-nannofossil ooze and chalk that unconformably overlies a early Eocene Paleocene carbonate pelagic sequence with ash layers (Edgar and Saunders, 1973; Holcombe et al., 1990). This sequence rests upon a siliceous breccia that is interpreted as a hard ground at an unconformity. Beneath the hard ground a Santonian sequence consists of carbonaceous day (black shales), foraminiferal sandstone and volcanic ash. The underlying basalt is vesicular but apparently not pillowed. DSDP Site 153 is located in the Aruba Gap at the southern boundary of the Beata Ridge (Figure 2). The upper sequence consists of Pleistocene to middle Oligocene day rich foraminifera-nannofossil chalk. Early Eocene silicified limestone and cherts correlated with a Caribbean-wide seismic reflector known

20

Figure 2. Location map of the NAUTICA diving survey.

as horizon A00 . Coring below horizon A00 recovered siliceous limestone and marlstone and minor volcanic ash layers. At the Cretaceous-Tertiary boundary a siliceous breccia resembles the hard ground of Site 151. A hiatus between the middle Maastrichtian pelagic foraminiferal himestone and the middle Paleocene pelagic nannoplankton marl corresponds to this hard ground. The lowermost sediments from the hole are late Turonian-Coniacian limestone, interlayered with carbonaceous day and ash layers. The fine grained texture of the uppermost basalt suggests that it represents a quenched flow top. At DSDP Site 31 Leg 4, located on the eastern flank of the Beata Ridge (Figure 2), (Bader, Gerard

et al., 1970) the uppermost strata are PleistocenePliocene and contain corroded foraminefera deposited beneath the carbonate compensation depth (CCD). Below this unit lies a Miocene chalk which shows no sign of dissolution. In contrast the basal Miocene-upper Oligocene indurated chalk contains poorly preserved corroded planktonic foraminifera. Coring along the western flank of Beata Ridge recovered middle Eocene neritic chalk and early Oligocene to early Miocene and Pleistocene deep water carbonates oozes (Fox et al., 1970; Fox and Heezen, 1975). Dredge hauls from the same escarpment recovered basalts and dolerites at depth ranging from 4 to 2.3 km. A K-Ar age of 64 Ma (Maas-

21

Figure 3. Bathymetric map of the central part of the Caribbean Sea from (Leroy, 1995). All the bathymetric data were compiled but the most accurate data were acquired during the Seacarib seabeam surveys. These surveys, utilized during the diving cruise, are shaded.

trichtian) was determined from feldspars separated from the least weathered basalt (Fox et al., 1970; Fox and Heezen, 1975). A re-study of the thin sections (T.W. Donnelly, written communication) indicated that several samples, in addition to the basalts and dolerites previously identified, are fine grained gabbros and coarse cumulate gabbros. Tectonic seffing The Venezuela basin and the northern sub basins are bounded to the east by the Aves Ridge, interpreted to be a remnant island arc (Fox and Heezen, 1975; Holcombe et al., 1990). The southern part of the Venezuela basin is underlain by an older (JurassicEarly Cretaceous?) oceanic crust (Diebold et al., 1999). The Muertos trough marks the northward active convergence of the Caribbean crust underneath Hispaniola and Puerto Rico (Figures 1 and 2).

The fluctuation of the CCD observed at DSDP Site 31 suggested that the Beata Ridge subsided to near the CCD by the Oligocene, was then uplifted 1 km in the Miocene and has subsided since the Miocene to its present depth (Benson et al., 1970). On the eastern flank of the Beata Ridge multichannel reflection seismic profiles show several structures that are interpreted as reverse and strike-slip faults (Mauffret et al., 1994; Leroy, 1995; Leroy and Mauffret, 1996; Mauffret and Leroy, 1999). These structures suggest a Miocene to Present day reactivation of the Beata Ridge related to plate collision occurring in Hispaniola (Mercier de Lépinay et al., 1988; Mann et al., 1991). In this hypothesis the western escarpment of Beata Ridge could be a recent major strike-slip fault. In contrast, Driscoll and Diebold (1999) propose that this fault is extensional and formed as the Beata Ridge rose by isostatic unloading during or after the forma-

22 tion of the Cretaceous volcanic plateau.The 5-km thick crust of the Haiti sub basin may have been thinned during this extensional event. Alternatively, it is a relict of the old and thin oceanic crust similar to that of the Southeastern Venezuela basin. Geophysical setting Seismic refraction data show that the crust of Beata Ridge is 20 km thick (Ewing et al., 1960; Edgar et al., 1971; Case et al., 1990). Crustal thickness decreases to the east, in the Dominican sub basin, to 10–15 km and less than 5 km in the Venezuela basin. Such variations are interpreted as resulting from the thickening of a normal oceanic crust during the emplacement of the late Cretaceous plume-related magmatism. A compilation of the refraction data (Leroy, 1995; Mauffret and Leroy, 1997) suggests that the thickening of the Beata Ridge crust has been accomplished mainly by injection at deep levels whereas the uppermost level (horizon B00 ), that consists of basaltic flows sampled at the ocean drilling sites, is thin. Seismic reflection profiles show in the Venezuela basin a smooth reflector (horizon B00 ) that is correlated with the top of the Cretaceous igneous event sampled by drilling (Edgar and Saunders, 1973). This reflector cannot be identified in the rough topography of the crest of the Beata Ridge. The outcrop of gabbros and dolerites on the western flank of the ridge, demonstrated by the dredging (Fox et al., 1970; Fox and Heezen, 1975) and our NAUTICA diving survey, confirms that the basaltic upper level is thin or/and missing.

NAUTICA diving survey Location and objectives 16 dives were performed during our survey with the principal objective to explore the Beata Ridge (Figures 2 and 3). The ridge is bounded by steep scarps (more than 20◦ ) and probably exhibits some of the best submarine outcrops of the Caribbean LIP. Moreover, in the area of the Beata Ridge, complete multibeam bathymetric surveys were the basis for the dive tracks. Seven dives (from north to south: NB05, NB06, NB07, NB08, NB12, NB09 and NB10, Figure 4) explored the northwestern steep scarp that bounds the Beata Ridge. The other dives on the Beata Ridge are located on the eastern and southern parts of the ridge (Figure 3). Dives NB02 and NB03 investigated the deformed accretionary prism in the inner side of the Muertos

trough, and the most northeastern slope of the Beata Ridge in the outer side of the Muertos trough, respectively. Dives NB04 and NB11 were located on Diorys Hill and Tairona Ridge. Southwards, four dives investigated the northern (NB13 and NB14) and southern (NB15 and NB16) parts of the large north-south trending DSDP 151 Ridge that forms the backbone of the southern Beata Ridge. The easternmost dive (NB01, Figure 2) investigated the northern flank of one of the biggest seamounts, called the Kathy’s seamount, that pierces the sediments of the Puerto Rico sub basin (Matthews and Holcombe, 1976). The Nautica diving survey provided about 200 samples both of igneous and sedimentary rocks and recorded tectonic and petrographic observations on 77 hours of video films. Compositions and petrographic descriptions of the igneous rocks are reported separately by Révillon et al. (2000). Radiometric ages from igneous rocks were determined using the 40 Ar/39 Ar method, (Révillon et al., 2000) and are reported Table 1. Biostratigraphic ages were derived from fossil assemblages in sedimentary rocks.

Results The western Beata escarpment (NB05 to NB1O and NB12) The main results of the dives on the western Beata escarpments are presented in Figure 5. To the north the western Beata escarpment merges with the western Haitian slope (Figure 3). Topography is poorly known in the area of dive NB05. In the area of dives NB06 and NB07, the seabeam map shows a steep slope between 4.3 and 3 km water depth (Figure 4A). South to dive NB07, a deep terrace with a smooth slope ranging from 4 to 3.5 km water depth divides the escarpment. A lower step bounds the Haitian sub basin between 4.3 and 3 km water depth. The dip of the slope ranges from 16◦ (NB07) where the rocks are poorly exposed to 24◦ (NB10) where the sedimentary cover is almost absent. NBOS (Figure 5B) This dive began at mid-slope at 3.2 km water depth where a 78.7±0.5 Ma (Campanian) dolerite outcrops (sample NB05-01). Paleocene to Eocene nenitic limestones, containing echinoderm, Dasycladaceae and melobesia algae fragments, overlie the igneous rocks. Above the limestones, early Miocene fine grained carbonate and carbonate turbidite indicate deeper water depth. In the upper

23 Table 1. 40 Ar/39 Ar ages of Beata Ridge (R´evillon et al., 2000) and comparison with ODP, DSDP results and adjacent areas. On Beata Ridge the magmatic intrusions are Campanian (81 Ma–75 Ma) whereas the first sediments resting upon the basalts drilled at DSDP Sites 153 and 151 are late Turonian-Coniacian (90–86 Ma) and Santonian (86–83 Ma), respectively. The significance of the young ages (Paleocene) on the Beata Ridge is discussed in the text. 40 Ar/39 Ar ages of the DSDP Sites 146 and 150 and of the ODP 1001 are from Sinton et al. (1998, 2000). Volcanic ash beds and high percentage of plagioclase in the DSDP and ODP sections (Donnelly et al., 1973; Sigurdsson et al., 1997a) suggest a volcanie activity around the Caribbean Sea from the early Paleocene (65 Ma) to the middle Eocene (45 Ma). However, the occurrence in an island arc context of the 45 Ma old volcano-clastic turbidites in ODP Site 998 (Sigurdsson et al., 1997a) is discussed in the text. In the Presqu’île du Sud of Haiti (Hispaniola) the Caribbean magmatic crust outcrops in La Hotte and La Selle massifs (Maurasse et al., 1979; Calmus, 1983; Bien-Aime Momplaisir, l986). A 105 Ma (Albian) K-Ar age were obtained in La Selle massif (Bien-Aime Momplaisir, 1986). Dumisseau formation, in La Selle massif, was dated 89 Ma (early Coniacian) and 92 Ma (Turonian) by 40 Ar/39 Ar method (Sinion et al., 1998). The Bath massif in Jamaica is a fragment of the Caribbean volcanic province (Lewis and Draper, 1990). From the Maastrichtian (72 Ma) to the early Eocene (49 Ma) calc-alkaline and alkaline rocks are well represented in the adjacent areas of Beata Ridge. Sietas Cabezas formation in Dominican Republic (Hispaniola) are dated (Sinton et al., 1998) 68 Ma (Maastrichtian) and 60 Ma (Paleocene). However, these young ages of a formation that resembles the magmatic Caribbean province maybe caused by a reheating by an adjacent pluton. The K-Ar ages for the Cayman Ridge are from (Perfit and Heezen, 1978). In Jamaica the Above Rocks granodiorite have yielded ages of 60–63 Ma (early Paleocene; Ahmad et al., 1985). The granodiorite drilled on the Pedro bank is 53 Ma (late Paleocene-early Eocene; Holcombe et al., 1990). Alkaline rocks probably related to an extensional event (Calmus, 1983) in the Presqu’île du Sud of Haiti (Hispaniola) are dated by K-Ar method 61 Ma (early Paleocene) and 50 Ma (late Eocene).

24

Figure 4. A. Bathymetric map of the central part of the Caribbean Sea from (Leroy, 1995). B. Enlarged portion of the bathymetric map where the dives are located. Note the dive NB05 on the Haitian (south of Hispaniola) slope where no seabeam data was available. C. Seabeam map of the western escarpment that bounds the Beata Ridge and the Haiti sub basin. The position of the dives are indicated as well as those of the dredges and cores performed by (Fox et al., 1970; Fox and Heezen, 1975). Note also the Tairona Ridge where is located the dive NB11. The position of the southwestern part of the seismic line Ewing 1323 (Figure 10) is indicated.

part of the dive, outcrops of Paleocene to middle Eocene (Lutetian) reef enclosing echinoderms, lamellibranchs, algaes and bryozoairs were observed. In between 2.8 and 2.4 km depth, a chalk containing early Miocene (Aquitanian) nannofossils was recovered below an outcrop of late Eocene platform carbonate. Thus, along dive NB05, a sedimentary sequence composed of shallow water, Paleocene to Eocene, platform carbonates underlying deeper Miocene carbonates appears repeated at least three times.

NB06 (Figure 5C) Between 4.2 km and 3.6 km below sea level, dive NB06 surveyed an avalanche path along a canyon axis and most of the samples are sedimentary with doleritic and gabbroic pebbles. A small outcrop of Paleocene polygenic microconglomerate was sampled. It consists of remains of echinoderms, ostrea, bryozoair, melobesia algae and benthic foraminifera, indicating a neritic environment. From 3.6 km to 3.2 km below sea level, the scarp is formed by doleritic rocks. Late Miocene pelagic muds were recovered in the upper part of the slope.

25

Figure 5. Dives NB05 to NB10 on the north western scarp that bounds the Beata Ridge and Haiti. A. Log of the dives. B to H. Dive tracks. No vertical exaggeration. The symbols of the outcrops are dipping by convention and the true dips have not been measured. The position of the dives are indicated in the inset I and Figure 4.

NB07 (Figure 5D) In the lower part of dive NB07 the slope was smooth (16◦) and covered by recent mud. At 3.6 km below sea level an outcrop of typical pillowed basalt (NB07-3) was recovered. The vesicular basalt is dated from separated plagioclase at 76.9±2.1 Ma (Campanian). A Tertiary age is indicated by the biostratigraphic data of a carbonate rock interlayered between the pillowed basalts. Miocene (early and late Miocene, Langhian) chalk and carbonate from moderate to deep water environment were recovered near the top of the scarp. NB08 (Figure 5E) In the lower slope, ranging from 3.8 to 3.4 km water depth, breccias and conglomerates were observed. Early Miocene and Burdigalian marls were recovered. The conglomerates are Late Cretaceous to Paleocene with a neritic facies (rudists, melobesia algae, lamellibranch). Late Cretaceous silty marls were also recovered. A doleritic pebble was

sampled (NB08-8) from a breccia and produced age of 55.3±1.2 Ma (late Paleocene). Gabbros and dolerites were sampled between 3 km and 2.8 km water depth. The NB08-14 bis sample is a doleritic clast belonging to a breccia that is likely tectonic. This dolerite is 81.1±0.5 Ma (Campanian). Late Miocene calcareous marl was recovered at the top of the dive track. NB12 (Figure 5F) The objective of the NB12 dive was to observe the complete escarpment from 3.6 km to 2 km. Calcareous man and conglomerates with volcanic pebbles were recovered between 3 km and 2.6 km water depth. These samples are Maastrichtian to Paleocene in age with a neritic facies similar to the samples of the same age recovered in the previous dives. Gabbros and dolerites were also recovered. Aquitanian and Burdigalian (lower Miocene) carbonates were sampled at the top of the escarpment.

26 NB09 (Figure 5G) A vesicular basalt was recovered at the bottom of dive NB09. Then, ranging between 3.6 and 3.2 km water depth, dolerites, gabbros and altered basalts were recovered. A gabbro (NB09-7) gave a 74.8±1.3 Ma age (Campanian). From 3 km to 2.5 km, Paleocene reef with coral, Melobesia algae, lamellibranchs and mollusks was recovered. One sample consists of calcareous breccia with gabbroic clasts. A calcareous man yields a Late Eocene age. NB10 (Figure 5H) The slope investigated by the NB10 dive is the steepest part of the Beata escarpment (24◦ ). The lower part of the scarp consists of outcrop of gabbros and dolerites. A dolerite (NB10-9) produced an age of 79.9±0.8 Ma (Campanian). Late Paleocene-early Eocene micritic carbonates showing a neritic facies were recovered in the upper part of the dive. Oligocene-early Miocene pelagic limestone were also sampled. The uppermost sample is an Eocene carbonate mantled by a polymetallic crust as thick as 15 cm. Kathy’s seamount (NB01, Figure 6B, location in Figure 2) This seamount, in the Puerto Rico sub basin, nanges from 5.1 to 4.3 km depth. The upper part of the slope has a smooth topography and is mantled by soft sediment. In the lower part, pillowed basalts were observed. These basalts are highly weathered and mantled by a thick polymetallic crust. No samples were suitable for radiometric dating. North Beata Hill (NB02, Figure 6C) This structure is the northernmost feature of the Beata Ridge. It is located on the outer side of the Muertos trough. At the base of the hill, 4.2 km below sea level, a breccia with calcitic cement contains rounded and highly vesicular basaltic clasts. The scarp is formed by Serravallian (middle Miocene), late Miocene and late Pliocene marls. Muertos prism (NB03, Figure 6D) This dive was performed on the seaward-most anticline of the Muertos accretionary prism. The apparent disorder of the marls sampled (Pliocene; late Miocene; early Miocene) and repetitions of series (late Miocene) suggest active reverse faulting. We failed to detect hydrothermal activity although the frequent occurrence of shells of Solemya sp. is probably related to this activity. A basaltic pebble was sampled in the lower slope of the anticline. This structure is separated from

the main slope of the Muertos accretionary prism by a syncline. This syncline prevents any transport from the continent. Therefore, the basaltic pebble has been deposited on the floor of the Muertos trough then recently uplifted by the formation of the anticline. Diorys Hill (NB04, Figure 6E) At water depths between 4 and 2.7 km, Diorys Hill forms a small north-south trending ridge standing on the lower part of the eastern Hispaniola slope. The profile of the hill is asymmetric with a moderate western slope and a steeper eastern scarp where dive NB04 was performed. Down the scarp, gabbros in slope breccia were sampled. Plagioclases separated from fine-grained gabbro NB04-02 produced an age of 77.9±0.2 Ma (Campanian). Then, from about 3.5 km water depth, diorites and quartz-diorites, overlain by a succession of gabbros and dolerites, were recovered. Very often doleritic outcrops appear pillowed by weathering and are covered by a polymetallic crust. Near the top of the scarp, a very fine-grained dolerite, possibly represents the part of a massive, nonvesicular and well-crystallized basaltic flow. The crest of the hill is covered by an Aquitanian (early Miocene) calcareous marl. Tairona Ridge (NB11, Figure 6F) This feature trends north-south (Figure 4), on the eastern margin of Beata Ridge at depths ranging from 2 to 1 km. This structure is the shallowest feature investigated during our survey and stopped at 1.5 km water depth. The shape is asymmetrical with a smooth western slope and a steep eastern slope that were investigated by dive NB11. Dolerites and gabbros outcrop in the lower part of the slope; none produced a reliable cristallization age. A late Santonian siliceous marly limestone is interlayered in the doleritic rocks. Burdigalian (lower Miocene) carbonate was recovered below dolerites and pillow basalts. DSDP 151 Ridge (NB13 to NB16, Figure 7) This 130 km long ridge, that trends north-south, is surrounded by the 3.5 km bathymetric contour. Two crests rise to 2 km below sea level. These summits are located in the northern part and southern part of the structure, respectively. The northern tip of DSDP 151 ridge was surveyed by the Seacarib seabeam cruise (Figure 7C). DSDP Site 151 is located on the northwestern part of the ridge that is oriented north-south. North of DSDP Site 151 this ridge changes direction to a NE–SW orientation. It is evident that this orientation

27

Figure 6. Dives NB01, NB02, NH03, NH04 and NB11 These dives are located in the eastern part of the Beata Ridge and east of this structure (NB01 and NB03). A. Log of the dives and DSDP Sites.. B. B. to E. Dive tracks. No vertical exaggeration. The symbols of the outcrops are dipping by convention and the true dips have not been measured. The position of the dives are indicated in the inset G except for NB01 on the Kathy’s seamount. This feature is located in the Puerto Rico sub basin (see Figure 2). The location of NB11 is also indicated in Figure 4.

28

Figure 7. Dives NB13, 14, 15 and 16, ODP and DSDP Sites A. Bathymetry of the central part of the Caribbean Sea (Leroy, 1995). B. Bathymetry of the DSDP Ridge. Enlarged portion of the bathymetric map where the dives NB13, NB14, NB15 and NB16 are located. The short seabeam survey performed in the southern part of the ridge during a transit of the R/V Jean Charcot is not shown but was utilized during the dives NB15 and NB16. C. Seabeam map of the northern part of the DSDP 151 ridge. Note the change of trend from north-south for the ridge to northeast-southwest to the escarpment that bounds the ridge and the adjacent basin. Observe the steep scarp investigated by the dive NB14. The position of the dives NB13 and NB14 are indicated as well as the basalt dredged by (Fox et al., 1970; Fox and Heezen, 1975). The position of the seismic profile Seacarib 169, shown in Figure 9, is indicated. D. Log of the dives, DSDP and ODP Sites. E. to G. Dives NB13 to NB16 on the DSDP 151 ridge. No vertical exaggeration. The symbols of the outcrops are dipping by convention and the truc dips have not been measured. The position of the dives are indicated in the insets C and D. The first part of the dive NB13 in the smooth and muddy slope is not represented.

29 has the same trend as the western escarpment of Beata Ridge (Figure 3). NB13 (Figure 7E) This dive investigated the western slope of the ridge. Dolerite from the base of the scarp (NB13-03) was dated at 78.8±0.5 Ma (Campanian). The marly matrix of a conglomerate is dated late Maastrichtian by nannofossil determination and two other pelagic marly limestone are late Cretaceous in age. The sedimentary rocks are overlain by gabbros and dolerites. One dolerite (NB13-09) yielded an age of 75.4±1.0 Ma (Campanian) while a gabbro (NB13-10) is 77.3±0.4 Ma (Campanian). NB14 (Figure 7F) This dive investigated the western slope of the DSDP 151 Ridge. This escarpment is the steepest (40◦ ) of the NAUTICA diving survey and vertical normal faults were observed. The scarp consists of fine grained dolerites, dolerites and gabbros. A diorite enclave was found in a dolerite. A gabbro (NB14-05) is 56.2±0.9 Ma (late Paleocene). NB15 (Figure 7G) NB15 and NB16 dives investigated the eastern and western slope of the southernmost crest of the Beata Ridge, respectively. The top of the southernmost hill is 1.8 km below sea level. The eastern slope is smooth (11◦) and is mostly covered by mud. A lower step was missed and an upper step, near the top, was sampled. Dolerites and gabbros were recovered near the crest of the DSDP 151 Ridge between 2.15 and 2.08 km water depth. NB16 (Figure 7H) The western slope, investigated during the NB16 dive, of the southern part of the DSDP 151 Ridge is steeper (15◦) than the eastern slope. Therefore the outcrops ranging between 3.1 km and 2 km water depth are of a good quality. Fine grained dolerite, dolerites and fine grained doleritic breccia were recovered at the base of the slope. No reliable crystallization ages were produced for samples from this dive. Late Campanian-early Maastrichtian soft marl overlies the igneous rocks. A red chert was sampled in the middle of the slope, embedded in a Campanian soft man. The two Mesozoic samples have a pelagic facies. Nannofossil determinations yield early Burdigalian and Aquitanian-early Burdigalian ages (early Miocene) for a calcareous marl and chalky man, respectively. The sampling of dolerites and gabbros at the top (NB15) and at the base (NB16)

Figure 8. Synthetic log of the NAUTICA dives. 77.9±3.2 Ma and 55–56 Ma represent non weighted means of the ages obtained by 40 Ar/39 Ar method (Table 1). Two clusters of ages have been obtained: one late Campanian (81–75 Ma, non weighted mean 77.9 Ma) the other Paleocene (55–56 Ma, non weighted with only two values). Scarce vesicular and pillowed basalts have been recovered at the top and base of the western Beata scarp. The other magmatic samples are dolerites and gabbros except few diorites. Maastrichtian conglomerates suggest a prominent erosion and an uplift. This uplift is confirmed by a carbonate platform and reefs Paleocene-early Eocene (65–49 Ma) and maybe middle Eocene. This shallow water carbonate platform subsided and is presently 2.5 km below sea level. The Miocene, Pliocene and Pleistocene marls are pelagic.

suggests that the southern DSDP 151 Ridge is 1.2 km high intrusive massif.

Interpretation and discussion We present a synthetic log (Figure 8) of the NAUTICA dives mainly derived from the results of the survey of the western escarpment of the Beata Ridge but

30 including the observations made in other areas. The radiometric ages (Révillon et al., 2000) are presented Table 1. Igneous rocks Igneous rocks On the western escarpment of Beata Ridge an igneous section is found below 2.8 km water depth. However, in other parts of the ridge igneous rocks were recovered or observed up to 1.8 km below sea level (e.g., Tairona Ridge, NB11). The main surprise of the diving survey on the Beata Ridge is the paucity of volcanic relative to intrusive rocks. Volcanics represented by pillowed and vesicular basalts were sampled on only three dives (NB07, NB09 and NB11). They are located at the base of the escarpment on the western part of the ridge (dives NB07 and NB09) or, eastwards, at the crest of the ridge (Tairona Ridge, NB11). Such different positions for lavas flows at the crest or base of the Beata Ridge were also observed at the DSDP Sites 151 or 153, respectively. Thus, we conclude that such a feature is a general characteristic of the Beata Ridge (Figure 8). Most of the igneous rocks sampled or recovered during the diving survey are intrusive (about 90 samples from dives NB04 to NB16). They are represented by a group of microgranular, sometimes cryptogranular, to granular gabbro-dioritic rocks (gram size up to about 5 mm), ranging from dolerites and gabbros to more rare diorites and quartz-diorites. Textures vary from intergranular to ophitic, sometimes over relatively short distances on some dives, and indicate that the magmas probably crystallized under variable cooling rates such as is expected in dike or sill-like intrusions. We interpret the finer grained dolerites as representing magmas almost chilled under high cooling rates on the margins of dikes or sills and the coarser grained gabbros as representing magmas crystallizing relatively slowly in the core of the intrusions. We conclude that the environment was mainly hypovolcanic but we cannot exclude the possibility that some of the finer grained dolerites of this group could be massive basaltic flows rather than chilled margins of sills (e.g. summit of dive NB04). None of our gabbroic sample shows clear cumulative features. T. Donnelly (person. comm.), from examination of the thin sections of the dredged rocks obtained by Fox et al. (1970), suggests that cumulate coarse gabbros also outcrop on the ridge. Nevertheless, the relative high quartz-content of some diorites, may indicate that the group forms a magmatic suite which underwent substantial differentiation, pos-

sibly explaining the formation of cumulates at greater depth than the emplacement level of the dikes and/on sills (Révillon et al., 2000). Radiometric dating on 10 dolerites and gabbros of this group yielded ages ranging from 81.1 to 74.8 Ma (Campanian) and 55–56 Ma (late Paleocene). A basalt (NB07) is dated 76.9 Ma (Campanian). Sedimentary rocks Sedimentary intercalations In dive NB11 (Tairona Ridge, Figure 6F), a late Santonian siliceous marly limestone is interlayered in the doleritic rocks. On the eastern flank of the DSDP 151 Ridge (NB13, Figure 7E), the marly matrix of a conglomerate yields a late Maastrichtian age. Two samples of nannoplankton marly limestones yield a Late Cretaceous age. These pelagic sediments are interlayered in Campanian dolerites (78.8 Ma beneath and 75.4 Ma above). The micro-paleontological determinations and the absolute dating are concordant to attribute a Late Cretaceous age to the igneous rocks and the sedimentary intercalation sampled during the NB13 dive. Nevertheless, the late Maastrichtian age of the matrix of the conglomerate is too young relative to the radiometric dating and the late Maastrichtian conglomerate may cover Campanian dolerites. Maastrichtian marly carbonate beneath dolerite and Maastrichtian silt with volcanic material were sampled between gabbroic and doleritic rocks (NB12, Figure 5F) but we do not know if these Maastrichtian samples cover the igneous series or is intercalated in it. The first hypothesis is more compatible with the results of the others dives. Conglomerate and breccias A sequence of detrital rocks containing igneous debris and ranging from silt to conglomerates and breccias with pebbles up to few tens of cm often outcrop at the base of the western Beata slope. This late Maastrichtian to Paleocene-Eocene unit may also contain shallow water fossils (oyster, algae, rudist) suggesting that the Beata Ridge was emerging or at shallow water depth and underwent a vigorous erosion at that time (e.g., NB08, Figure 5E). The fact that very often the gabbros and dolerites of the intrusive unit are pillowed by spheroidal weathering is consistent with the hypothesis of emergence of the ridge. Indeed such weathering structures that we observe now covered by a polymetallic crust are not likely to have formed in deep marine environments.

31 Polymetallic crust The Paleocene carbonate rocks sampled during the NAUTICA survey are covered by a thick metallic crust (up to 15 cm on NB1O samples, Figure 5H). The basaltic pillows observed on the flanks of Kathy’s seamount are also covered by a thick crust and are deeply weathered. Such weathering and thick crust indicate a long time exposure to seawater. In contrast the dolerites and gabbros display a very thin (few millimeters) polymetallic crust and these observations suggest that the sea water weathering is a recent event for these rocks. Consequently the dolerites and gabbros that outcrop along the lower slope of Beata Ridge have been probably exposed by a recent faulting. Sedimentary cover In the southernmost dive (NB16, Figure 7H) we recovered Campanian-Maastrichtian nannofossil marls that directly overlie fine grained dolerites at the base of the scarp. Campanian pelagic marls with chert were also recovered in the same dive. In the other dives the Maastrichtian rocks contain volcanic pebbles, conglomerate and breccias with remains of shallow water carbonates (oyster, algae, rudist). Paleocene to middle Eocene and late Eocene (in NB05) reefs are now as deep as 2.5 km. The Oligocene limestone to Pleistocene mud are pelagic and deposited in deep water although some bathyal facies cannot be excluded. Several repetitions of series can be observed (NB05, NB09, NB12) which implies a tectonic disturbance subsequent to the Maastrichtian-Paleocene sedimentation. Paleo-environment The Cretaceous sediments, except the Maastrichtian conglomerates and breccias, are pelagic. The Maastrichtian-Paleocene sediments reveal a contrasting facies between the northern and southern part of the Beata Ridge. In the northern part, the presence of reefs, shallow facies fossils and evidence of prominent erosion indicate that this area was shallow and maybe locally emergent; in the southern part, the CampanianMaastrichtian sediments are pelagic and there is no evidence of shallow water although the present depth of the DSDP 151 ridge (2–3 km) is moderate. However, Maastrichtian conglomerates (NB13) and hard grounds discovered at DSDP Site 151 and 153 suggest erosion, slumping and non deposition on the top and the flanks of the ridge. Erosion is unlikely because we have no evidence of shallow environment necessary to permit an erosion process. The northern part

of the Beata Ridge was probably still shallow during the middle Eocene (NB05 and middle Eocene neritic chalk recovered by coring (Fox et al., 1970; Fox and Heezen, 1975). Deep water facies prevailed since the early Oligocene (Fox et al., 1970; Fox and Heezen, 1975). The Miocene, the Pliocene and the Pleistocene marls recovered by the NAUTICA dives show always a deep sea facies. Comparison of dives with DSDP and ODP drilling Submersible diving and deep ocean drilling have provided two complementary methods of sampling. Dive sampling along low angle, normal faults scarps has produced a composite section much greater than the drilled section (Figures 6A and 7D; Table 1). In the case of the central Caribbean the nature and the age of the igneous basement sampled by the two methods are surprisingly different and paradoxically the diving survey seems to have investigated deeper levels of the volcanic plateau than at previous sites. A disadvantage of the dive sampling, however, is that the relationship between the different samples along a dive track is not always as clear as in the drilled column, and the samples are more weathered. Radiometric dating yields two age groups for the samples recovered during the NAUTICA diving survey (Révillon et al., 2000). The younger ages will be discussed later. The older (from 81.1 Ma to 74.8 Ma, Campanian) is clearly related to the construction of the thickest crust in the Caribbean LIP. The basalts sampled during DSDP Leg 15 (Edgar and Saunders, 1973) at the base of Beata Ridge (DSDP Site 153, Figure 6A) are late TuronianConiacian and at the crest (DSDP Site 151, Figure 7D) Santonian, respectively. However, these ages were determined from micro-paleontological data of the lowermost sediments that rest upon the volcanic basement and consequently these basalts can be slightly older. However, the basalts recovered along the Hess escarpment (DSDP 152 and ODP 1001) are Campanian (81 Ma; Figure 7D; Sinton et al., 2000) and the dolerites sampled in the Venezuela basin are 94 Ma (Cenomanian, DSDP Site 150, Figure 6A) and 92– 90 Ma (Turonian, DSDP 146, Figure 6A), respectively (Sinton et al., 1998). In DSDP Site 146 the upper dolerite sill bas chilled against a section of Coniacian (89–85 Ma) marble which was recrystallized by the beat of the intrusion (Donnelly et al., 1973). In all DSDP sites basaltic ash beds were identified in Santonian (85–83 Ma) limestones (Donnelly et al., 1973). We conclude that the formation of the Caribbean LIP

32

Figure 9. A. The seismic profile Seacarib 169 illustrates the DSDP 151 ridge. Observe the steep slope where is located the dive NB14. Normal faults have been observed during the diving survey. Note the Santonian (83–86 Ma) age of the first sediment resting on the basalt drilled at the DSDP Site 151 whereas the dolerites and gabbros sampled by the NB13 dive yield Campanian age (79–75 Ma). A Campanian underplating is suggested. The young age (56 Ma, late Paleocene) of a dolerite recovered during the diving NB14 is attributed to some localized decompression mantle melting that accompanied the extension. The detailed logs of NB13 and 14 can be seen Figure 7. The location of the profile is indicated in Figure 7C and this figure (B). B. Seabeam map of the northern part of the DSDP 151 nidge. See also Figure 7C. This map shows the location of the seismic profile shown in A and the DSDP Site 151.

33 was not short and catastrophic (Sinton et al., 1998) but occurred during a period of about 20 Ma (Cenomanian to Campanian, 94–75 Ma). The previous interpretation proposed two region-wide igneous events, one 92–88 Ma (Turonian-Coniacian) major event and the younger 76 Ma (Campanian) minor event (Sinton et al., 1998). The Campanian magmatism on the Caribbean plateau is recorded also in Curaçao (Beets et al., 1984; Kerr et al., 1995; Sinton et al., 1998) Haiti (Maurasse et al., 1979; Bien-Aime Momplaisir, 1986) and in the Serrania de Baudo on the Pacific coast of Colombia (Kerr et al., 1997). Therefore, the Campanian activity was not a small magmatic event restricted to the Hess escarpment but it is well represented across the Caribbean LIP although its extent is not as large as the Turonian-Coniacian event (Figure 1). In Haiti an earlier magmatic activity could be as old as 105 Ma (Albian, K-Ar age; (Bien-Aime Momplaisir, 1986). Rare basalts were recovered near the base of the escarpment (NB07; NB09) or at the top of the ridge (Tairona Ridge, NB11, Figure 6F). Elsewhere we found exclusively gabbros and dolerites. The seismic profile displayed in Figure 9A illustrates the relationships between the results of the dives NB13 and NB14 and those of the DSDP Site 151. The nanno-calcareous marl sampled during the dive NB13 is Late Cretaceous and more precisely late Maastrichtian for the matrix of one conglomerate whereas a hard ground separates the Paleocene nanno-chalk from the Santonian foraminiferal sandstone at the crest of the ridge (DSDP Site 151). Basaltic rocks were sampled by dredging (Fox et al., 1970; Fox and Heezen, 1975) near the location of dive NB14 (Figure 9B) although the location and the depth of a dredged sample cannot be determined precisely. Dive NB14 investigated the slope from 3.15 km to 2.45 km below sea level where we observed and sampled only doleritic and gabbroic rocks. DSDP Sit e 151 reached basalt at 2.407 km below sea level. Therefore the basaltic layer on the top of the DSDP 151 ridge is only 50 meters thick and confirms the hypothesis deduced from the geophysical studies (Mauffret and Leroy, 1997) that this extrusive layer is very thin. However, the DSDP Site 151 is projected from the southern top of the ridge (Figure 9) and the situation can be different in the dive NB14 area and the site. Late Maastrichtian conglomerates deposited in deep sea marls and the hard ground at the crest of the ridge suggest an erosional period. However, this erosion did not reach the basaltic layer that is covered

by 12 m of Santonian sediments at the top of the ridge. Therefore, the erosion cannot be responsible of the thinning of the basaltic layer. The three age determinations obtained for gabbroic and doleritic rocks recovered during the NB13 dive are nearly concordant (78.8 to 75.4 Ma, Campanian) and are younger than the basalt sampled by at DSDP Site 151. The age of the overlying sediments (Santonian) provides a minimum age for the basalt. We conclude that at this place the intrusive rocks are younger than the extrusive rocks and have contributed to the formation and uplift of DSDP 151 Ridge. A gabbro from the western scarp of the DSDP 151 Ridge (NB14) yields an age of 56.2 Ma (Paleocene). We noted already that the scarp investigated by the NB14 dive is very steep. The seismic profile (Figure 9A) illustrates a prominent scarp and the visual observations from the submersible show that vertical faults form the scarp. This NE–SW escarpment is obviously a fault that limits northward the extent of DSDP 151 Ridge that trends north-south. The young age (56.2 Ma, Figure 9) determined for a gabbro sampled during the dive NB14 could be related to the fault activity. The Santonian sediments (NB11, Figure 6F) interlayered between the igneous layers and the lowermost Campanian sediments (NB16, Figure 7H) resting upon the top of the plateau have been deposited in deep water. The basalts of the Hess escarpment show large vesicles and could have been erupted in shallow water depth (1 to 2 km). The moderate depth of the Hess escarpment during the Campanian is confirmed by the limestones above the basalts that contain a Campanian foraminifera assemblage characteristic of a neritic or bathyal environment (Sigurdsson et al., 1997a). However, the basalts recovered during the NAUTICA diving survey are always vesicular including those of the Kathy’s seamount. It is evident that the Venezuela basin and the Puerto Rico sub basin, where Kathy’s seamount is located, were deep during the plateau formation (Diebold et al., 1999) and the presence of vesicles indicates a magma rich in volatiles, rather than a shallow level eruption. We conclude that, during the Campanian formation of the plateau, the Hess escarpment area was shallow and Beata Ridge deep. Nevertheless, we observe an inversion of the tectonic situation during the Maastrichtian between the Beata ridge and the Hess escarpment. In the first area, Maastrichtian to middle Eocene shallow water carbonate reefs have been sampled during the NAUTICA diving survey, whereas in the second area, Maastrichtian limestones and Paleocene chalks

34 at ODP Site 1001 have been deposited in a deep sea environment (Sigurdsson et al., 1997a). DSDP Site 151 results identified a prominent hiatus between Paleocene and Santonian time and a hard ground marks the base of the Cenozoic (Figure 7D). The same hard ground is observed at DSDP Site 153 (Figure 6A) but the hiatus is restricted to late Maastrichtian to early Paleocene time. In contrast, no hiatus is recorded in the Venezuela basin (DSDP 146, Figure 6A) or on the Hess escarpment (DSDP 152 and ODP 1001, Figure 7D). The DSDP and ODP results show that Maastrichtian, Paleocene and Eocene sediments are pelagic and have been deposited in deep water. The Late Cretaceous sediments of dive NB13 have been also deposited in deep water. The presence of hard ground overlain by pelagic Paleocene chalk and the basin facies of the Santonian sandstone and carbonaceous day that underlie this hard ground at the top of the southern DSDP 151 ridge suggest an uplift and an inversion of basin between the Santonian and the Paleocene (Edgar et al., 1973). This uplift occurred between late Maastrichtian and early Paleocene time as suggested by the hiatus and the hard ground of DSDP Site 153 and the Maastrichtian conglomerate of dive NB13. Because we do not have any evidence of uplift of the DSDP 151 ridge above sea level, we suggest that the hard grounds and conglomerates are the result of gravity slides along steep scarps and a period of non deposition on topographic highs rather than erosion near sea level. The 56.2 Ma gabbro of NB14-05 (Figure 7F) suggests that the steep scarp that bounds the DSDP 151 Ridge (Figure 9B) was active during the late Paleocene. The uplift of the DSDP 151 Ridge probably began during the late Campanian with the intrusion of gabbroic and doleritic rocks. In contrast to the southern part of Beata Ridge, Maastrichtian, Paleocene and early Eocene reefal limestones reveal a shallow water environment for the northern part of Beata. The Paleocene reefs of the Beata escarpment are now deeper than the pelagic Paleocene chalk of DSDP Site 151 and a post-Paleocene differential movement between the two areas is evident. The late Santonian siliceous limestones of the crest of the northern Tairona Ridge (NB11, Figure 6F) suggest a Maastrichtian or younger uplift of the northern part of the Beata Ridge, while the present depth of the reefs indicates a large subsidence subsequent to the uplift and the coeval formation of the carbonate platform. We conclude that the Beata Ridge underwent a Maastrichtian uplift much more pronounced in the northern part than in the southern part and conversely

post early Eocene subsidence was moderate in the south and larger in the north. The tectonic boundary between the two regions is in the approximate position of the NE–SW fault where is located NB14, and the adjacent basin that bounds to the north DSDP 151 ridge (Figure 9B). Tectonic observations and young ages A tectonic analysis of the Beata Ridge bas been presented elsewhere (Leroy, 1995; Mauffret and Leroy, 1999). The results of the NAUTICA diving survey give new constraints on the tectonic history. Figure 10A shows the dive tracks NB10 and NB11 projected on the seismic profile Ewing 1323 (95-01 R/V M. Ewing cruise; Diebold et al., 1999). The western slope is divided into two parts by a deep terrace. This feature disappears to the north where the slope is more continuous (Figures 4 and 10D). The lower part of the northern slope that is equivalent to the deep terrace was investigated by the dive NB07 (Figures 4 and 5D). A 76.9 Ma (Campanian) basalt was recovered (NB07-03) and this basalt may be representative of the uplifted basement of the Haiti sub basin but this hypothesis is unlikely. This basin has a very thin crust (Leroy, 1995; Mauffret and Leroy, 1997) and a Moho reflection is evident in Figure 10A. The Haiti sub basin may be a relict of the original oceanic crust formed before the construction of the Caribbean volcanic plateau and moderately thickened and injected by igneous material during the formation of the adjacent volcanic plateau (Mauffret and Leroy, 1997). Alternatively it is a new basin formed by extension between the Nicaragua Rise and the Beata Ridge during or after the volcanic event (Driscoll and Diebold, 1999). The seismic profile illustrated Figure 10A shows a dipping reflector that could be an extensional detachment fault that cuts the crust up to 9 seconds TWTT. NB10 was projected on a migrated part of the seismic line (Figure 10B). The contact between the gabbros and the sedimentary cover is shown by the seismic profile (Figure 10C) although the dive is projected from 12 km to the north-east (Figure 4 and 10D). Two wedges look like gravity slides although the upper wedge can be interpreted also as a pro gradational carbonate platform constructed during the Paleocene-Eocene interval. Dive NB10 was located on the steepest part of the Beata escarpment and normal faults were identified by visual observations. Dive NB11 (Figure 10A) investigated the eastern slope of

35

Figure 10. A. Seismic profile Ewing 1323 kindly provided by J. Diebold (R/V M. Ewing 95-01 cruise). In the Haiti sub basin a Moho reflection at 10 seconds two way travel time (TWTT) and the basement at 7.5 sec. TWTT indicates that the crust is only 6 km thick. A detacbment fault can be seen up to 9 sec. TWTT. The logs of the dives NB10 and NB11 have been reported at the same scale. The detailed logs of NB10 and NB11 can be seen Figures 5 and 6. Extension is documented by the steep scarps that bounds the Taino Ridge, the Beata ridge and the Haiti sub basin; the thinness of the crust of this basin; the blocks that seem tilted and the detachment fault. B. The seismic profile and the dive track NB10 are at the same scale and without vertical exaggeration. The lower part of the slope is formed by magmatic rocks (dolerites and gabbros) the upper part is a Paleocene-early Eocene (Ypresian) carbonate platform (65–49 Ma). C. Zoom of the contact between the gabbro and the sedimentary cover. An almost horizontal reflector corresponds to the summit of the gabbros. D. Seabeam map of the western escarpment that bounds the Beata Ridge and the Haiti sub basin. The position of the dives are indicated as well as those of the dredges and cores performed by (Fox et al., 1970; Fox and Heezen, 1975). Note also the Tairona Ridge where is located the dive NB11. The position of the southwestern part of the seismic line Ewing 1323 (Figure 10A) is indicated. See also Figure 4 for a larger scale.

the Tairona ridge. This feature is asymmetrical with a smooth western slope and a steep eastern one. Such a structure is typical of tilted block. DSDP 151 Ridge is also highly asymmetrical with a steeply dipping western scarp and a more gently dipping eastern slope. The samples recovered on the western scarp yield young ages: 56.2 Ma (Paleocene) for NB14 (Figure 7F). The NB08 dive track (Figure 5E) is not particularly steep and yielded igneous rocks of two ages: a 81.1 Ma (Campanian) old dolerite at the top, characteristic of the volcanic plateau and a 55.3 Ma (Paleocene) age for a dolerite sampled at the base of the slope.

The young ages (Paleocene) may correspond to a localized decompression mantle melting that occurred during a late extensional event that reheated the volcanic plateau. Early Paleocene (61 Ma) to early Eocene (50 Ma) alkaline volcanic rocks have been reported from the Presqu’île du Sud (Haiti, Hispaniola; Table 1; Calmus, 1983). The young igneous rocks and the old ones look the same and are adjacent (NB13 and NB14; Figure 7E and F) or even in the same dive (NB08, Figure 5E). In summary, there are indications of several episodes of extension in the Haiti sub basin and along the Beata western scarp following the initial plateau-forming magmatic

36 event (Driscoll and Diebold, 1999). The first extension that occurred on north-south trending faults is related to deep injection of sills before a thickening by underplating. The subsequent second extension that occurred on northeast-southwest trending faults is documented by young volcanism, the coeval uplift of the northern Beata Ridge, then the collapse of the carbonate platform. The thickened underplated crust has been subsequently thinned and the original crust may have been exhumed in the deep part of the Haiti sub basin. Plate tectonic framework Plate reconstructions (Pindell and Barrett, 1990) include a Maastrichtian (71 Ma) to middle Eocene (49 Ma) magmatic arc that extended from the upper Nicaragua Rise to Aves Ridge (Table 1 and Figure 11). Pre-middle Eocene andesites and granodiorites have been sampled by industrial wells on the upper Nicaragua rise (Arden, 1975; Holcombe et al., 1990). Granodiorites from the Pedro bank are 53 Ma (early Eocene; K-Ar method; Holcombe et al., 1990). In Jamaica, calc-alkaline rocks (granodiorite of Above Rocks, Table 1) of the are crop out and have produced (Ahmad et al., 1985) radiometric ages of 60–63 Ma (early Paleocene). Late Paleocene-early Eocene (59– 49 Ma) alkaline and calc-alkaline volcanic rocks are interlayered in the sedimentary fill of the Wagwater trough (Lewis and Draper, 1990). Metamorphic rocks dredged along the Cayman Ridge yielded an age of 69 Ma (Maastrichtian; K-Ar method) and the ages of granodiorites, sampled on the same ridge, range from 64 Ma to 59 Ma (early Paleocene; K-Ar method; Perfit and Heezen, 1978). The triangular shape of the Yucatan back-arc basin suggests that the Cretaceous Cuban island are underwent a clockwise rotation relative to the Cayman Ridge along transform margin bordering the Yucatan margin (Pindell and Barrett, 1990; Rosencrantz, 1990). Cuba collided with the Bahamas platform from west to east and was progressively accreted to the North America plate while the motion of the Caribbean plate changed from northeast to east (Malfait and Dinkelman, 1972; Mann et al., 1995). The final accretion of Cuba occurred (Lewis and Draper, 1990) at 46 Ma (late Eocene) whereas the onset of seafloor spreading in the Cayman trough is 49 Ma (early Eocene; Leroy, 1995; Leroy et al., 2000). Ypresian (early Eocene; 53–49 Ma) to Lutetian (middle Eocene, 49–41 Ma) redeposited volcanic ash and

volcano-clastic turbidites drilled at ODP Site 998 (Table 1) suggest that the Cayman Ridge was an island are at this time with a subduction to the north (Sigurdsson et al., 1997a). However, the Cayman trough was already opened during the Lutetian (middle Eocene, 49–41 Ma; Leroy, 1995; Leroy et al., 2000) and we suggest that the Nicaragua-JamaicaCayman-Sierra Maestra are was definitively extinct at this time. The Eocene volcano-clastic turbidites of the ODP Site 998 may correspond to the erosion of the 69–59 Ma (Maastrichtian-early Paleocene) old Cayman are. In Hispaniola, the activity of the calcalkaline are persisted during the Paleocene-Eocene (Lebron and Perfit, 1994). The Siete Cabezas formation is a basaltic formation that compositionally resembles the Caribbean volcanic plateau. Radiolarians from intercalated sediments yielded Cenomanian to Santonian age (98–83 Ma; Mercier de Lépinay, 1987). However, 40 Ar/39 Ar dating indicates a 68– 69 Ma age (Maastrichtian; Sinton et al., 1998) and a reheating by the intrusion of granodiorites maybe inferred. The position of the subduction zone relative to the Aves arc and its polarity are known, but the subduction related to Jamaica-upper Nicaragua Rise is not yet identified. The Pedro escarpment is a possible location for this subduction zone having a northwest vergence (Pindell and Barrett, 1990). However, if the volcanic plateau of the lower Nicaragua Rise is thick it should have resisted subduction especially just after its formation when it was still hot and buoyant. Nevertheless, the moderately thick (10–15 km) Caribbean crust subducts presently beneath the Muertos trough and the south Caribbean deformed belt and the crust must be particularly thick (15–20 km, Beata Ridge) to resist subduction. Extension in the Haiti sub basin was coeval with the formation of the Yucatan and Grenada basins (Paleocene-early Eocene) but its position relative to the subduction zones (Figure 1) does not favor a backarc origin. However we do not know the position of the Presqu’île du Sud d’Haiti and Beata Ridge relative to Hispaniola north block. In our reconstruction (Figure 11) we assume a large amount of subduction in the Muertos trough, but if the Beata Ridge- Presqu’île du Sud block was closer to the Hispaniola island arc, the Haiti sub basin may have had a back are position. Extension in the Haiti sub basin is probably related to the uplift of the Beata edge of the volcanic plateau and its subsequent subsidence. The Haiti sub basin is divided into two parts (Figure 11). The northern part is narrow, bounded by steep faults, and underlain by a

37

Figure 11. Early Eocene (53 Ma) reconstruction based on (Pindell and Barrett, 1990) modified. Calc-alkaline and alkaline volcanic rocks, related to an island arc and back-arc extension respectively, are widespread during the Maastrichtian (71 Ma) to the early Eocene (49 Ma) on Nicaragua upper rise, Jamaica, Cayman Ridge, Sierra Madre in Cuba, Hispaniola and Aves Ridge. Granodionites from the Pedro bank are 53 Ma old (early Eocene; K-Ar method; Holcombe et al., 1990). In Jamaica the Above Rocks granodiorite have yielded ages of 60–63 Ma (early Paleocene; Ahmad et al., 1985). Metamorphic rocks dredged along the Cayman Ridge yielded an age of 69 Ma (Maastrichtian; K-Ar method) and the ages of granodiorites range from 64 Ma to 59 Ma (early Paleocene; K-Ar method; (Perfit and Heezen, 1978). Early Paleocene-early Eocene (65–49 Ma) back arc extension is observed in the Yucatan and Grenada basins (Pindell and Barrett, 1990; Rosencrantz, 1990). The onset of oceanic crust in the Cayman trough is 49 Ma old (early Eocene; Leroy, 1995; Leroy et al., 2000) and the Nicaragua-Jamaica-Cayman-Sierra Maestra are was definitively extinct at this epoch. We assumed that the rifting in the Haiti sub basin is coeval to the opening of the Yucatan and Grenada basin (early Paleocene- early Eocene (65–49 Ma) but a back arc extension is unlikely. Extension of the Haiti sub basin and Beata Ridge is probably related to a magmatic event that postdated the formation of the volcanic plateau. Extent of thick plateau, thin magmatic crust and oceanic crust are from (Leroy, 1995; Mauffret and Leroy, 1997; Diebold et al., 1999). The transition from magmatic crust to oceanic crust is outlined by southwards dipping reflectors (Mauffret and Leroy, 1997; Diebold et al., 1997). Observe the probable subduction of a large surface of magmatic crust west of the Western Cordillera of Colombia that corresponds to an obducted portion of the Caribbean volcanic province. Position of the wells on the upper Nicaragua rise are from (Arden, 1975) and (Holcombe et al., 1990). DSDP Sites (Edgar et al., 1973) as well as ODP Sites 998 and 1001 (Sigurdsson et al., 1997a) are replaced.

thin crust; the southern part is wide and its crust is relatively thick. The margin of the southern Beata Ridge is formed by several blocks and the Haiti sub basin is separated from the Colombia basin by the Warao Rise that connects the Nicaragua Rise and the southern Beata Ridge (Leroy, 1995; Mauffret and Leroy, 1997). Southwards dipping reflectors mark the transition between the thick crust of the Warao Rise and thin crust of the eastern Colombia basin (Figure 16; Mauffret and Leroy, 1997). A rough acoustic base-

ment suggests that an old oceanic crust may exist in the deepest part of the Colombia basin (Bowland and Rosencrantz, 1988) but this zone is narrow and the thin crust subducts and disappears beneath the South America accretionary prism (Figure 1). A 5-km thick and 10-km long region of southwards dipping reflectors overlies the oceanic crust of the south Venezuela basin (Diebold et al., 1997). This oceanic crust is probably a relict of the Farallon plate formed at the Farallon-Pacific spreading center during the Jurassic

38

Figure 12. Evolution of Beata Ridge and adjacent areas. From late Cenomanian (94 Ma) to Santonian (83 Ma) and particularly during a short period between 90 Ma (late Turonian) and 88 Ma (early Coniacian) a widespread volcanic event covered the Farallon oceanic crust around the Galapagos plume. Dipping reflectors have been evidenced in the Venezuela basin (Diebold et al., 1997, 1999; Driscoll and Diebold, 1999) and the smooth top of sheet flow basalts have been drilled by the DSDP Sites Leg 15 (Edgar et al., 1973). A moderate underplating thickened the crust of the Puerto Rico sub basin that formed the hangingwall north of the Venezuela basin. During the Campanian (75 Ma) the lower Nicaragua rise was shallow (ODP Site 1001; Sigurdsson et al., 1997a) and the Beata Ridge deep but an underplating thickens the crust beneath this feature (this study). From the Maastrichtian to Paleocene (72–53 Ma) the lower Nicaragua rise was deep whereas the northern part of the Beata Ridge was shallow with a prominent erosion, edification of a carbonate platform and reefs. The Caribbean crust is thinned by extension and the former Farallon oceanic crust may have been exhumed in the Haiti sub basin. A part of the uplift of Beata Ridge maybe generated by the isostatic rebound of the Beata Ridge footwall (Driscoll and Diebold, 1999) but we attribute the most part of this uplift to the Campanian underplating. Extension persisted during the early Eocene (49 Ma) and the carbonate platform subsided and is presently 2.5 km deep.

39 or the Early Cretaceous. The thin crust in the Colombia and Venezuela basins are now separated by the Beata Ridge but they are probably two small remnants of a large area of Farallon oceanic crust that has subducted beneath South America (Figure 11). However, if the Cretaceous plume head was centered in the western part of the Colombia basin (Figure 1) the Farallon oceanic crust should have been covered by the Cretaceous volcanic flows. Moreover, the obducted volcanic plateaus of the western Cordillera of Colombia (Figures 1 and 11) indicate that the thick igneous crust extended well to the south. We conclude that the original oceanic crust is preserved in a narrow window (Figure 11) and that the volcanic event did not have a radially-symmetric distribution but was dispersed as several volcanic plateaus separated by unmodified deep basins. Evolution of the eastern Caribbean Sea and Beata Ridge During the late Cenomanian-Santonian (94 Ma– 83 Ma) the Caribbean plate was affected by a widespread volcanic event (Sinton et al., 1998). Based on current age determinations, the peak of activity occurred at 88–92 Ma and a 2500 km diameter plume head was probably centered over the present location of the Galapagos hotspot (Duncan and Hargraves, 1984; Mauffret and Leroy, 1997). Beneath the Caribbean Sea, the top of the volcanic section is represented by basaltic flow overlying, in the Venezuela basin, a 5-km thick body of dipping reflectors whose age is unknown but is contemporaneous with or older than the Cenomanian-Santonian igneous rocks and ash beds drilled at DSDP 146 and 150 (94–83 Ma; Diebold et al., 1997b). This sequence rests upon a thickened lower crust probably formed by underplating, e.g., Puerto Rico sub basin (Figure 12, upper right corner). However, the igneous crust is relatively thin in this part of the Caribbean Sea (10 km thick) and the uplift relative to the old crust that lies in the southern part of the Venezuela basin, is moderate (2 km). During the formation of the volcanic province the original oceanic crust was extended (Diebold et al., 1997) and sandwiched between the upper volcanic layer and the underplated lower crust (Leroy, 1995; Mauffret and Leroy, 1997). The dipping reflectors that show many similarities with a syn-rift wedge have been formed by rotation of the Puerto Rico basin hanging wall (Figure 12v; Diebold et al., 1999; Driscoll and Diebold, 1999).

During the Campanian (83–71 Ma), the lower Nicaragua Rise was shallow whereas the Beata Ridge was deep. The crust of the upper Nicaragua Rise and Beata Ridge are 15 and 20 km thick, respectively (Case et al., 1990). Basalts drilled at ODP site 1001 and the ages of the intrusive rocks from western Beata Ridge escarpment indicate volcanic activity and amount of a large underplating in the Campanian. Moreover, a conglomerate, indicating an uplift and a correlative erosion is dated Campanian-Maastrichtian by micro-paleontological studies. Therefore, it appears certain that the rise of Beata Ridge occurred during the Campanian. The Campanian dolerites and gabbros sampled along the scarp of DSDP 151 Ridge and the Santonian basalt at the top suggest a Campanian underplating but in this southern part of the Beata Ridge the uplift bas been moderate. The DSDP 151, Diorys and Tairona Ridges are oriented northsouth whereas the western Beata escarpment shows a northeast-southwest trend. The features of this orientation seems to cut the north-south structures (Figure 9) that are probably older. We showed that the northeastsouthwest faults are probably related to the Paleocene rifting of the Haiti sub basin. Consequently we propose that the north-south orientation is Campanian (83–71 Ma) in age and represents the trend of the deep faults and feeder dikes. In the other hand the basement highs that are located north of the dipping reflectors in the Venezuela basin show a northeast-southwest direction (Diebold et al., 1999) although some north-south structures have been also described in this area (Mauffret and Leroy, 1997). The DSDP Site 150 is located near the top of one of these NE-SW basement highs and we conclude that the emplacement of these highs is contemporaneous or older than early Cenomanian (94 Ma; Sinton et al., 1998). The uplift of the northern part of the Beata Ridge must have been rapid and was coeval with a prominent phase of erosion from the late Campanian to Maastrichtian. This erosion is evidenced by a wedge at the base of the western Beata scarp formed by conglomerate and breccias. A carbonate platform in a neritic environment was established during the MaastrichtianPaleocene-early Eocene. During this period the northern Beata Ridge was higher than the lower Nicaragua Rise where the sediments are pelagic. This situation suggests a rifting where the Beata Ridge was the footwall and the Nicaragua Rise the hanging wall (Driscoll and Diebold, 1999). The thin crust of the Haiti sub basin may have resulted from this extensional process. The present situation of the deep reefs and carbon-

Figure 13. The volcanic basement beneath the Caribbean Sea is reported on a Cretaceous reconstruetion in the hot spot reference frame (Duncan and Hargraves, 1984; Engebretson et al., 1985; Pindell and Barrett, 1990). Nicoya complex and Gorgona Island are on a 2500-km diameter circle and are close to be obducted. This Turonian (90 Ma) circle is supposed to be larger than the Campanian (80 Ma, 2000 km) circle. Note that the Galapagos hot spot is off axis and not placed on the Pacific-Farallon spreading center.

40

41 ate platform indicates that the thinning of the western Beata margin persisted into the middle Eocene. The late Eocene to early Miocene was a period of quiescence when the Beata Ridge subsided. A reactivation with compression and transpression may have occured from early Miocene to Present (Leroy and Mauffret, 1996; Mauffret and Leroy, 1999). Visual observations made during the submersible survey suggest that reverse faults disturb and repeat the igneous and sedimentary series. Cause of the uplift of Beata Ridge and relation with the Galapagos plume A surfacing plume may generate three kinds of uplift (White and McKenzie, 1989, 1995; Nadin et al., 1997): a dynamic uplift, an uplift due to heating and an uplift due to magmatic underplating. Dynamic uplift must be excluded for the Campanian formation of Beata Ridge because it should have produced a regional effect which is not seen for the original oceanic crust of the Venezuela basin (Mauffret and Leroy, 1997; Liebold et al., 1997). Moreover, this dynamic uplift and that due to heating should have begun during Turonian (88–92 Ma) associated with the main magmatic event (Sinton et al., 1998). Uplift due to magmatic underplating is documented by geophysical and diving data and we favor this hypothesis although a local thermal effect cannot be excluded. From the Turonian to the Campanian the Caribbean plate moved towards the east with the Farallon plate (Figure 13; 5 cm/yr; Engebretson et al., 1985; Pindell and Barrett, 1990). We reconstructed the Caribbean LIP in the hotspot reference frame (Duncan and Hargraves, 1984) above the Galapagos plume during the Campanian (80 Ma). Nicoya, Gorgona, Curaçao and Duarte complexes are within the Turonian (90 Ma) 2500-km large diameter circle and are close to be obducted. We observe that during the Campanian the Caribbean LIP was still whithin the influence of the Galapagos plume (2000-km large diameter circle) and not separated from the Farallon plate by the Central Americas subduction zone as suggested by some studies (Pindell and Barrett, 1990). Comparison with other oceanic LIPs A common characteristic of the volcanic provinces in rifted magin settings is the presence of thick sequences of seaward dipping reflectors. However, on the volcanic margins these reflectors develop from a subaerial spreading center and precede the formation of the deeper normal oceanic crust. In the Caribbean

LIP the dipping reflectors have been formed in deep water and overlie older oceanic crust (Diebold et al., 1997). The dipping reflectors of the Caribbean LIP resemble those of the Kerguelen plateau (Rotstein et al., 1990; Schlich et al., 1993) although the latter was partly emergent during the volcanic construction. On the Ontong Java plateau, the basalts have 40 Ar/39 M ages of about 90 Ma and 122 Ma (Mahoney et al., 1993). The younger basalt age has been interpreted as a second volcanic event caused by separation of the original plume head from its trailing conduit (Bercovici and Mahoney, 1994; Gladczenko et al., 1997). Two discrete ages (110 Ma and 85 Ma) have also been determined for the construction of the Kerguelen plateau (Coffin and Gahagan, 1995). We note that, because the Kerguelen and Ontong Java plateaus are much larger than the Caribbean volcanic province, the sampling has been scarce and intermediate ages between the two identified events cannot be excluded. Moreover the deep portions of these plateaus have never been investigated and the previous results of drilling on the Beata Ridge, dated the top of the basaltic flows that are older than the intrusive rocks forming the core of the ridge. The results of the NAUTICA diving survey clearly show that the igneous event was continuous from the early Campanian (81.1 Ma) to the mid-Campanian (74.8 Ma). The new radiometric dates from DSDP Site 150 in the Venezuela basin yield a late Cenomanian-early Turonian (94.3±2.8 Ma Ma) age that is the oldest age of the Caribbean LIP determined by 40 Ar/39 Ar technique. The Caribbean igneous event thus lasted at about 20 Ma and we do not take into account the paleontological and K-Ar dates for a possible Albian volcanic episode (105 Ma; Bien-Aime Momplaisir, 1986). On Beata Ridge rifting was coeval with magmatism while uplift and subsidence post-date the igneous activity. The young (Paleocene, 55–56 Ma) intrusive rocks are probably related to a localized decompression mantle melting related to a late extensional event and this situation mimics those of East Greenland. An East Greenland plume that apparent lack of a broad initial uplift during break-up of the North Atlantic and coeval magmatism (58–57 Ma) and the late regional margin uplift all seems at odds with several current plume models (Larsen and Marcusen, 1992). 40 Ar/39 Ar data for young (32 Ma) picritic dolerites indicating that these are rather more than later than the main volcanic episodes invite the speculation that some localized decompression mantle melting accom-

42 panied the late Eocene-early Oligocene uplift and rejuvenescent activity only occurring along the faulted coastal zone (Upion et al., 1995). A sequence of magmatism followed by synchronous crustal extension and uplift for Yemen does not fit with the traditional categories of active (uplift-magmatism-rifting) and passive (rifting-uplift-magmatism) rifting. Extension and uplift (1 km) post date magmatism related to the 30 Ma Afar plume by some 10–15 m.y. (Menzies et al., 1992). The volcanism in West Greenland started at an early stage (Paleocene, 60–61 Ma) approximately 5 m.y. before the start of sea-floor spreading in the northern North Atlantic. The presence at the periphery of a plume (West Greenland, Hebrides and northwest of the British Isles) of picrites with eruptives temperatures of as much as 1500 ◦ C does not indicate proximity of a narrow plume head (Larsen et al., 1992; Chalmers et al., 1995). The same observation can be made for the Caribbean plume. The picrites of Curaçao and the komatiites of Gorgona (Kerr et al., 1995); the picrites discovered near St. Helena, in Costa Rica (Alvarado et al., 1997); and in the Duarte complex of Hispaniola (Dupuis et al., 1997) are located over a broad region with a diameter of about 2000 km (Figure 1). Although all these terranes are tectonized and displaced it is unlikely that they have been formed in a single narrow hot head of the plume. It implies a high temperature (1500 ◦ C) over a large surface or several plumes (Chalmers et al., 1995). A very large hot plume head is a more acceptable explanation than several plunes producing the picritic melts. It is implied in several studies that plumes carinot break the lithosphere and utilized the ruptures, thin spots and spreading centers, to emerge. Icelandic and Afar hot spots are good examples of this relationship between lithospheric structures and plumes and many Pacific oceanic plateaus have been supposed to be formed on spreading centers although volcanic flows are contemporaneous on very large surfaces and may rest upon older original oceanic crust. The plate reconstructions of the eastern Mesozoic Pacific Ocean are quite uncertain because the Mesozoic parts of the Farallon, Kula and Phoenix Plates have been subducted and the magnetic anomalies of the Caribbean Sea have a structural origin (Leroy, 1995). In Figure 13, based on (Duncan and Hargraves, 1984) and (Engebretson et al., 1985) reconstructions, the Galapagos Cretaceous plume is formed off axis within the Farallon plate, with no obvious thin spot. However, if the Gala-

pagos plume is older than 90 Ma, as suggested by Albian (112–98 Ma) fossils and K-Ar dates in Haifi (Bien-Aime Momplaisir, 1986) and >100 Ma volcanic rocks of Colombia (Kerr et al., 1997) the primitive hot spot maybe, indeed, located on a spreading center or a triple-junction.

Conclusions and summary The Caribbean LIP is composed of volcanic basement underlying the central Caribbean Sea and several fragments accreted to the adjacent continental masses. The main volcanic event was apparently relatively brief (88–92 Ma). At the beginning of the magmatism, from evidences in the Puerto Rico sub basin and the Venezuela basin, the thickening of the crust is moderate (about 5 km) and the uplift was 1–2 km. This early magmatism is characterized by sequences of southward dipping reflectors, probably formed in deep waters, and piercement of volcanoes. Basement highs that trends northeast-southwest, are probably underplated by sills and feeder dikes. The dipping reflectors and basaltic lavas flow towards the southeast (south Venezuela basin) over the original Farallon oceanic crust. The crust beneath the Beata Ridge and the lower Nicaragua Rise is thicker (10 to 15 km) than in the east and the associated uplift have reached 5 km. These basement highs are oriented north-south. On the Beata Ridge, the paucity of basalts relative to the dolerites and gabbros is the main surprise of the NAUTICA diving survey. Radiometric dating shows that this volcanic event occurred in the Campanian (6 m.y. activity) and the first sediments that rest upon the volcanic surface are Campanian-Maastrichtian. Given the previous basement ages (Late TuronianSantonian; 89–83 Ma) of the DSDP Sites on the Beata Ridge these results are surprising and document the Campanian underplating that caused a 5-km uplift. The DSDP and ODP results and the new radiometric and biostratigraphic ages obtained from our samples indicate that the plateau forming volcanic events were not short and catastrophic but occurred over a protracted period (94–75 Ma). During the MaastrichtianPaleocene, the northern crest of Beata Ridge was close to sea level, as shown by a prominent erosion surface and the establishment of a carbonate platform and reefs. Seismic evidence, the present 2.5-km depth of the Paleocene reef, and young ages (55– 56 Ma) suggest rifting continuing after the formation of the plateau. Young ages are explained by localized

43 decompression mantle melting related to a late extensional event that allowed magmas to rise from the base of thinning lithosphere. The basaltic rocks may have been erupted during this extension but doleritic and gabbroic rocks with young ages have the same composition as the Campanian intrusive rocks. The crust of the northern Haiti sub basin is very thin and similar in structure to typical oceanic crust. The proto-Caribbean crust, modified by the subsequent volcanic event, may have been exhumed by the late Mesozoic-Early Tertiary extension of the Nicaragua-Beata volcanic plateau. As is true for several others oceanic plateaus, the development of the Caribbean LIP does not fit with current active (uplift-magmatism-rifting) and passive (rifting-upliftmagmatism) plume models because uplift and extension post-date initial magmatism. The present distribution of the fragments of the Caribbean LIP suggests a 2000–2500-km diameter melting region for a Cretaceous plume head centered on the Galapagos hotspot. Large parts of the volcanic province have been obducted and subducted. We observe that the crust must be very thick (15–20 km) in places to resist subduction. The Caribbean LIP is composite and includes thickened crust like Beata Ridge and thin, largely unmodified, oceanic crust presently located in the southwestern and southeastern parts of the Caribbean Sea. This crust is probably a relict of the Farallon plate.

Acknowledgements This work has been supported by the ATP Géosciences Marine CNRS-INSU. We thank officers and crew of the R/V NADIR. We are grateful to the technical staff and the pilots of the NAUTILE for their efficient work. J. Diebold is thanked for providing a seismic line of the R/V M. Ewing cruise and for useful discussions on the formation of the plateau. H. Feinberg determined biostratigraphic ages of the nannofossil assemblages. Two anonymous reviewers and J. C Sibuet made helpful suggestions.dd Contribution of the CNRS-ESA 7072.

References Ahmad, R., Lal, N., and Sharma, P., 1985, A fission-track age for the Above Rocks granodiorite, Jamaica, Workshop on recent advances in Jamaican geology, Geol. Soc. Jamaica 3.

Alvarado, G. E, Denyer, P. and Sinton, C. W., 1997, The 89 Ma Tortugal komatiitic suite, Costa Rica: Implications for a common geological origin of the caribbean and Eastern pacific region from mantle plume, Geology 25: 439–442. Arden, D. D, 1975, Geology of Jamaica and the Nicaragua rise in A. E. M. Nairn and T. G. Stehli (eds.), The ocean basins and margins, V 3, The Gulf of Mexico and the Caribbean, New York, 617–661. Bader, R. G., Gerard, R. D. et al., 1970, Initial Reports of the Deep Sea Drilling Project, Leg 4, U.S. Government Printing Office, Washington D.C. Beets, D. J., Maresch, W. V., Klaver, G. T., Mottana, A., Boccio, R., Beunk, F. F. and Monen, H. P., 1984, Magmatic rocks series and high pressure metamorphism as constraints on the tectonic history of the southern Caribbean, in W. D. Bonini (ed.), The Caribbean-South American Plate Boundary and Regional Tectonics, Geological Society of America, pp. 95–130. Benson, W. E., Gerard, R. D. and Hay, W.W., 1970, Summary and conclusions, in R. G. Bader, R. D. Gerard (eds.) Initial Reports of the Deep Sea Drilling Project, U.S. Government Printing Office, Washington D.C. Bercovici, D. and Mahoney, J., 1994, Double flood basalts and plume head separation at 660-kilometer discontinuity, Science 266: 1367–1369. Bien-Aime Momplaisir, R., 1986, Contribution à l’étude géologique de la partie orientale du massif de la Hotte (presqu’île du sud d’Haïti). Syntèse structurale des marges de la presqu’île à partir de données sismiques. Thèse de l’Université de Paris, 223 p. Bowland, C. L. and Rosencrantz, E., 1988, Upper crustal structure of the western Colombian Basin, Geol. Soc. Am. Bull. 100: 534– 546. Calmus, T., 1983, Contribution à l’étude géologique du massif de Macaya (Sud-Ouest d’Haïti, Grandes Antilles), sa place dans l’évolution de l’orogène Nord-Caraïbe, Doctorat 3ème cycle thesis, Univ. Pierre et Marie Curie, 162 p. Case, J. E., MacDonald, W. D. and Fox, P. J., 1990, Caribbean crustal provinces; seismic and gravity evidence, in G. Dengo and J. E. Case (ed.), The Geology of North America, Vol H, The Caribbean Region. (A decade of North American Geology), Geological Society of America, Boulder, Colorado, 15–36. Chalmers, J. A., Larsen, L. M. and Pedersen, A. K., 1995, Widespread Paleocene volcanism around the northern North Atlantic and Labrador Sea: evidence for a large, hot, early plume head, J. Soc. Geol. London 152, 965–969. Coffin, M. F. and Gahagan, L. M., 1995, Ontong Java and Kerguelen Plateaux, Cretaceous Icelands? J. Geol. Soc. London 152: 1047– 1052. Diebold, J., Driscoll, N., Abrams, L., Buhl, P., Donnelly, T., Laine, E., Leroy, S. and Toy, A., 1999, New insights on the formation of the Caribbean basalt province revealed by multichannel seismic images of volcanic structures in the Venezuelan Basin, in P. Mann (ed.), Caribean Sedimentary Basins, 561–589. Diebold, J., Mauffret, A. and Leroy, S., 1997, Seismic images of the Venezuela volcanic plateau (Caribbean), Eur. Un. Geosci. Stasbourg, 9: 525. Donelly, T. W., Beets, D., Carr, M. J., Jackson, T., Klaver, G., Lewis, J., Maury, R., Schellenkens, H., Smith, A. L., Wadge, G. and Westercamp, D., 1990, History and tectonic setting of Caribean magmatism, in G. Dengo and J. E. Case (Ed.) The Geology of North America, Vol H, The Caribbean Region. (A Decade of North American Geology), Geological Society of America, Boulder, Colorado, 339–374. Donelly, T. W., Melson, W., Kay, R. and Rogers, J. J. W., 1973, Basalts and dolerites of Late Cretaceous age from the Central

44 Caribbean, in N. T. Edgar, J. B. Saunders et al. Initial Report of the Deep Sea Drilling Project, U.S. Government Printing Office, Washington, D.C. Driscoll, N. W., and Diebold, J. D., 1999, Tectonic and stratigraphic development of the Caribbean: new constraints on the CaribbeanSouth American plate, in P. Mann, Caribbean Sedimentary Basins, 591–626. Duncan, R. A. and Hargraves, R. B., Plate tectonic evolution of the Caribbean region in the mantle reference frame, in W. D. Bonini, R. B. Hargraves and R. Shagam (eds.), The Caribbean-South American Plate Boundary and Regional Tectonics, Geological Society of America, Memoir 162, 81–94. Dupuis, V., Lapierre, H., Mercier de Lepinay, B., Maury, R. C., Tardy, M. and Hernandez, J., 1997, Duarte complex (Hispaniola) revisited: evidence for three structural units in a remmant of the Caribbean oceanic plateau, Europ. Un. Geosc. 9: 525. Edgar, T. N., Ewing, J. and Hennion, J., 1971, Seismic refraction and reflection in the Caribbean Sea, Am. Assoc. Petrol. Geol. Bull. 55: 833–870. Edgar, T. N. and Saunders, J. B., 1973, Initial Report of the Deep Sea Drilling Project, D.S.P. 15, 1137 p. Engebretson, D. C., Cox, A. and Gordon, R. G., 1985, Relative motions between oceanic plates in the Pacific Ocean, Geol. Soc. Am. Spec. Pap. 206: 1–206. Ewing, J., Antoine, J. and Ewing, M. 1960, Geophysical Measurements in the Western Caribbean Sea and in the Gulf of Mexico, J. Geophys. Res. 65: 4087–4125. Fox, P. J. and Heezen, B. C., 1975, Geology of the Caribbean crust, in A. E. M. Nairn and T. G. Stehli (eds.) The ocean basins and margins, V 3, The Gulf of Mexico and the Caribbean, Plenum Press, New York, Vol. 3, 421–426. Fox, P. J., Ruddiman, W. F., Ryan, W. B. F. and Heezen, B. C., 1970, The geology of the Caribbean crust, 1: Beata Ridge, Tectonophys. 10: 495–513. Gladczenko, T., Coffin, M. F. and Eldhom, O., 1997, Crustal structure of the Ontong Java plateau: modeling of new gravity and existing seismic data, J. Geophys. Res. 102: 22 711–22 729. Holcombe, T. L., Ladd, J. W., Westbrook, G., Edgar, T. and Bowland, C. L., 1990, Caribbean marine geology and basins of the plate interior, in G. Dengo and J. E. Case (eds.), The Geology of North America, Vol. H, The Caribbean Region. (A Decade of North American Geology), Geological Society of America, Boulder, Colorado, 231–260. Kerr, A. C., Marriner, G. F., Tarney, J., Nivia, A., Saunders, A. D., Thirlwall, M. F. and Sinton, C.W., 1997, Cretaceous basaltic terranes in Western Colombia: elemental, chronological and Sr-Nd isotopic contraints on petrogenesis, J. Petrology 38: 677–702. Kerr, A. C., Tarney, J., Marriner, G. F., Nivia, A, Klaver, G. T. and Saunders, A. D., 1995, The geochemistry and tectonic setting of late Cretaceous Caribbean and Colombian volcanism, J. South Amer. Earth. Sci. 9: 11–120. Larsen, L. M., Pedersen, A. K., Pedersen, G. K. and Piasecki, S., 1992, Timing and duration of early Tertiary volcanism in the North Atlantic: new evidence from West Greenland, in B. C. Storey, T. Alabaster and R. J. Pankhurst (eds), Magmatism and the causes of continental break-up, Geol. Soc. Special Pub. 68: 321–333. Lebron, M. C. and Perfit, M. R., 1994, Petrochemistry and tectonic significance of Cretaceous island-arc rocks, Cordillera Oriental, Dominican Republic. Tectonophys. 229: 69–100. Leroy, S., 1995, Structure et Origine de la plaque Caraïbe, Implications Géodynamiques, Thèse de l’Université thesis, P & M Curie, Paris, 240 p.

Leroy, S., Bitri, A. and Mauffret, A. 1996, Migration velocity analysis based on common shot depth migration applied to the seismic data of the Caribbean oceanic plateau, Geophys. J. Int. 125: 199–213. Leroy, S. and Mauffret, A., 1996, Intraplate deformation in the Caribbean region, J. Geodynamics 21: 113–122. Leroy, S., Mauffret, A., Patriat and Mercier de Lépinay, B., 2000, An alternative interpretation of the Cayman trough evolution from a reidentification of magnetic anomalies Geophys. J. Int. 141: 539–557. Lewis, J. F. and Draper, G., 1990, Geology and tectonic evolution of the northern Caribbean margin, in G. Dengo and J. E. Case (eds.), The Geology of North America, Vol. H, The Caribbean Region. (A Decade of North American Geology.), Geological Society of America Boulder, Colorado, 77–140. Mahoney, J. J., Storey, M., Duncan, R. A., Spencer, K. J. and Pringle, M., 1993, Geochemistry and geochronology of Leg 130 basement lavas: nature and origin of the Ontong Java Plateau, in W. H. Berger and L. H. Kroenke (eds.), Proc. of the ODP Sci. Res., 3–22. Malfait, B. T. and Dinkelman, M. G. 1972, Circum-Caribbean tectonic and igneous activity and the evolution of the Caribbean plate, Geol. Soc. Amer. Bull 83: 251–272. Mann, P., Draper, G. and Lewis, J. F. 1991, An overview of the geologic and tectonic development of Hispanolia, in P. Mann, G. Draper and J. F. Lewis (eds.), Geologic and Tectonic Development of The North America-Caribbean Plate Boundary in Hispaniola, Geol. Soc. Am., Spe. Pap. 162, 1–28. Mann, P., Taylor, F. W., Edwards, R. Lawrence and Teh-Lung KU, 1995, Actively evolving microplate formation by oblique collision and sideways motion along strike-slip faults: an example from the northeastern Caribbean plate margin, Tectonophysics 246: 1–69. Matthews, J. and Holcombe, T. 1976, Possible Caribbean underthrusting of the Greater Antilles along the Muertos Trough, 7th Caribbean Geol. Conf. Trans. Guadeloupe 235–242. Mauffret, A. and Leroy, S., 1997, Seismic Stratigraphy and structure of the Caribbean Sea, Teconophysics 283: 61–104. Mauffret, A. and Leroy, S., 1999, Neogene intraplate deformation of the Caribbean plate of the Beata Ridge, in P. Mann (ed.), Caribbean sedimentary basins, 627–669. Mauffret, A., Mercier de Lepinay, B., Leroy, S., Vila, J. M., Campan, A., Green, C., Gorini, C., Marton, G. and Reyes, J. R., 1994, Premier résultats de la campagne Casis dans le bassin Caraïbe, C.R. Acad. Sci. Paris 318: 1379–1386. Maurasse, F., Husler, G., Georges, G., Schmitt, R. and Damond, P., 1979, Upraised Caribbean Sea Floor below acoustic reflector B00 and the Southern Peninsula of Haïti, Geo. Minj. 8: 71–83. Menzies, M. A., Baker, J., Bosence, D., Dart, C., Davison, I., Hurford, a., Al’Kadast, M., McClay, K., Nochols, G., Al’Subbary, A. and Yelland, A., 1992, The timing of magmatism, uplift and crustal extension: preliminary observations from Yemen, in B. C. Storey, T. Alabaster and R. J. Pankhurst (eds.), Magmatism and the causes of continental break-up, Geol. Soc. Special Pub., 68: 293–304. Mercier de Lépinay, B., 1987, Evolution de la bordure nord-caraïbe: l’exemple de la transversale d’Hispaniola de Doctorat d’Etat thesis, Univ. P. et M. Curie. Mercier de Lépinay, B., Mauffret, A., Jany, I, Bouysse, P., Mascle, A., Renard, V., Stephan, J.F. and Hernandez, E. 1988, Une collision oblique sur la bordure nord-caraïbe à la jonction entre la ride de Beata et la fosse de Muertos, C.R. Acad. Sci. Paris 307: 1289–1296.

45 Nadin, P. A., Kusznir, N. J. and Cheadle, M. J., 1997, Early Tertiary plume uplift of the North Sea and Faeroe Shetland Basins, Earth Planet. Sci. Lett. 148: 109–127. Perfit, M. R. and Heezen, B. C., 1978, The geology and evolution of the Cayman Trench, Geol. Soc. Am. Bull. 89: 1155–1174. Pindell, J. L. and Barrett, S. F., 1990, Geological evolution of the Caribbean Region, in G. Dengo and J. E. Case (eds.), The Geology of North America, Vol. H, The Caribbean Region. (A Decade of North American Geology), Geological Society of America, Boulder, Colorado, 405–432. Révillon, S., Hallot, E., Ardnt, N. T., Chauvel, C. and Duncan, R. A. 2000, A complex history for the Caribbean plateau/petrology, geochemistry and geochronology of the Beata Ridge, southern Hispaniola, Geological Magazine, in Press. Rosencrantz, E., Structure and tectonics of the Yucatan basin, Caribbean sea, as determined from seismic reflection studies, Tectonics 9: 1037–1059. Rostein, Y., Schaming, M., Schlich, R. and Colwell, J. B., 1990, Basin evolution in oceanic volcanic plateau: seismic evidence from the Kerguelen Plateau, South Indian Ocean, Mar. Petrol. Geol. 7: 1–12. Schlich, R., Rostein, Y. and Schaming, M., 1993, Dipping basement reflectors along volcanic passive margins, new insight using data from the Kerguelen Plateau, Terra Nova 5: 157–163.

Sigurdsson, H, Leckie, R. M., Acton, G. D. et al., 1997a, In: Proc. ODEP, Init. Repts., V. 165. College Station, TX (Ocean Drilling Program). Sigurdsson, H., Leckie, R. M., Acton, G. D. et al. 1997b, Caribbean volcanism, cretaceous/tertiary impact and ocean-climate history: synthesis of leg 165, in O.D. Program Initial Report, Proc. ODP, College Station (TX), 377–400. Sinton, C. W., Duncan, R. A., Storey, M., Lewis, M. and Estrada, J. J., 1998, An oceanic flood basalt province at the core of the Caribbean plate, Earth Planet. Sci. Lett 115: 221–235. Sinton, C. W., Sigurdsson, H. and Duncan, R. A., 2000, Radiometric date of ODP Site 1001. In: Proc. ODP, Sci. Results, V. 165, College Station, TX (Ocean Drilling Program), in press. Upton, B. G. J., Emeleus, C. H., Rex, D. C., Thirlwall, M. F., 1995, Early Tertiary magmatism in NE Greenland., J. Geol. Soc. London 152: 959–964. White, R. S. and McKenzie, D., 1995, Mantle plumes and flood basalts, J. Geophys. Res. 100: 17543–17585. White, R. S. and McKenzie, D. P., 1989, Magmatism at rift zones: the generation of volcanic continental margins and flood basalts, J. Geophys. Res. 94: 7685–7729.