Collaborative Research:
Crust-Mantle Interactions During Continental Growth
and High-Pressure Rock Exhumation at an Oblique Arc-Continent Collision Zone: The SE Caribbean margin
Introduction
The interiors of the continents are thought to be constructed of islands arcs, small continent fragments, oceanic plateaus, marginal basins, and associated metamorphosed sedimentary rocks (e.g., Karlstrom and Bowring, 1988; Hoffman, 1989). Theories for the post-Archean formation of the continents (e.g., Nelson, 1991; Durrheim and Mooney, 1994), and explanations for the bulk chemistry of the continental crust (e.g., Rudnick and Fountain, 1995; Rudnick, 1995) rely on collision and accretion of island arcs to continental margins as a primary element in continental development. Whether the total continental mass is still growing, or is instead remaining constant while continental materials are continuously recycled by subduction, accretion, and magmatism is still unknown (e.g., McCulloch and Bennett, 1994; Taylor and McLennan, 1985; Albarede, 1998). What is well known is that the bulk chemistry of the continental crust is intermediate in composition (Taylor and White, 1965a, b; Taylor and McLennan, 1985; Rudnick and Fountain, 1995; Christensen and Mooney, 1995) and any Post-Archean growth or large scale continuous recycling mechanisms must produce an intermediate bulk composition crust. Recent work in several island arc terranes including the Aleutians and the Izu-Bonin arc has shown that their composition is mafic (Holbrook et al., 1999; Suyehiro et al., 1996). This observation makes it difficult to reconcile the concept that intermediate-composition crust has formed simply by progressive welding of Aleutian-type island arc terranes to a central continental core. The questions arise: 1) How can arc accretion to continents produce intermediate composition crust? and 2) Are the accretion processes we observe today the same as those that have operated throughout history to create the continents (Lyell, 1830; Taylor, 1967; McCulloch and Bennett, 1994)? This proposal addresses both of these fundamental questions on the origin of the continental crust.
Island arc-continent collision is currently occurring in Taiwan (Kao et al., 1998), northern Australia (Genrich et al., 1996), the Aleutians-Kamchatka Peninsula (Geist et al., 1994), and along the South American-Caribbean plate boundary (Avé Lallemant, 1997). The Aegean (Reilinger et al., 1997) is in a pre/early collision stage. All of the Indonesian and related arc systems are viewed as in an early pre-collision stage of accretion to southern Asia analogous to Proterozoic growth of the southwestern U.S. (Karlstrom, personal communication).
None of these are simple collisions. The first is a complex interaction between the Eurasian plate and the Philippine plate, and occurs only over a short distance with little study area. The second is in the early stages of development, with the continental Australian plate still subducting northward beneath the Java trench and the Timor arc, and hence does not lend itself to understanding the true stages of accretion. The third is also an immature collision of extremely limited geographic extent, with the mafic Aleutian islands colliding with Eurasia at a high angle. The latter two are pre-collisional.
In contrast, the South Caribbean Plate Boundary Zone where the Leeward Antilles Arc is colliding with South America (Figs. 1-3) covers a very large area (~1,600 km east-west; ~106km2), and is an eastwardly-diachronous (time transgressive) collision occurring over the last 50 million years, that allows us to study the processes and geometries of arc-continent collision over a range of stages of development. Specifically, this example allows us to analyze the common phenomena in plate tectonics of arc-polarity reversal due to trench choking by the continental plate: Along the eastern end of the plate boundary near Trinidad, choking and flipping has not yet occurred, whereas further west it was achieved over 20 million years ago. In the central segment, east of Margarita Island, the process is active and ongoing today. This setting allows us to see what can happen within the accretion zone after accretion has been completed, which is as important to the understanding of continental growth as the accretion itself. The Leeward Antilles arc appears to be felsic to intermediate, built on oceanic plateau crust. Lastly, one of the products of accretion in this setting is the still poorly understood process of exhumation of high-pressure/low-temperature (HP/LT) metamorphic rocks. This arc-continent collision setting allows us to trace the exhumation process for 50 million years and determine if large lateral displacements provide an important path in the exhumation process. Such attributes make the southern Caribbean boundary ideal for a modern study of arc accretion, arc polarity reversal, and eclogite and blueschist facies metamorphism tectonics.
Although arc-continent collision and arc accretion are universally accepted processes in the South Caribbean Plate Boundary Zone, a variety of plate tectonic models for the evolution of this zone have been proposed which imply a broad range of possibilities for (1) the number, age, geometry and polarity of subducted slabs, and (2) the chemical nature, origin, and degree of arc maturity of accreted fragments, along northern South America (Malfait and Dinkelman, 1972; Maresch, 1974; Ladd, 1976; Burke et al., 1978; Dickinson and Coney, 1980; Pindell and Dewey, 1982; Beets et al., 1984; Burke, 1988 ; Snoke et al., 1990; Pindell and Barrett, 1990). Since the development of a quantitative regional plate-kinematic framework (Pindell et al., 1988), the strong points of each of these models have been increasingly incorporated into a fairly standard and kinematically robust history of development recently summarized by Pindell et al. (1998), who refined it further using the regional history of sedimentation. In addition, this kinematic model has now been used to preliminarily test specific processes such as displacement partitioning, and associated arc-parallel extension and uplift of HP/LT rock assemblages (Avé Lallemant and Guth, 1990; Avé Lallemant, 1997). We now have an understanding of the paleogeographic history and associated processes involved in this complex zone of arc-continent accretion.
However, this model implies certain crust- lithosphere inter-relationships within the South Caribbean Plate Boundary Zone. We are still far from having a clear picture of the geometry of the various slabs and deep crustal/upper mantle elements which have played primary roles in this 4-dimensional history of arc-continent accretion and HP/LT metamorphic rock exhumation, especially in central and eastern Venezuela and Trinidad where accretion and reversal are ongoing. With carefully designed experiments to define these geometries within the plate boundary zone, we can learn the manner in which lithospheric slabs and crustal fragments interact during the plate boundary evolution, and therefore gain a dynamic, time-transgressive understanding of the detailed processes of ongoing arc-continent accretion, and attendant HP/LT rock exhumation.
The Caribbean-South America plate boundary has been accreting the Antilles Arc (the Venezuelan-Dutch archipelago) to the South American craton along an ever-lengthening plate boundary consisting of a trench at either end of a right-lateral transpressional fault system (Figs 1-7; Table 1). The right-lateral compressional plate boundary zone connects the NW-dipping Lesser Antilles subduction zone where the American plate is subducting beneath the Caribbean plate in the east, to the SE-dipping subduction zone off western Venezuela and Colombia in which the Caribbean plate subducts beneath South America. Teleseismic tomography and seismicity patterns show the Caribbean plate overriding the Atlantic seafloor southwest of Trinidad beneath the Gulf of Paria (Figs 8-9), whereas to the west the Caribbean plate is seen subducting beneath continental South America (Figs 9-10). Thus, subduction polarity reverses across the length of the right-lateral transpressional zone. The Aves Ridge/Lesser Antilles Arc, built on plateau-like, 12km thick Caribbean oceanic crust, is accreting to South America along this boundary, with a 200-500 km-wide fold and thrust belt/foreland basin system developing between the two subduction zones. The core of the folded belt lies within the transpressional boundary, exposing HP/LT accretionary metamorphic rocks in a series of boundary parallel belts. Cessation of arc magmatism, exhumation of the metamorphic belts, and development of the fold and thrust belt/foreland basin system all appear to have developed diachronously from west to east. In a larger context, it has been suggested that the Caribbean Plate and the Scotia Plate formed and migrated relatively eastward from deep mantle corner flow around the Nazca Plate as a result of Nazca-South American convergence (Russo and Silver, 1996). Plate tectonic/terrane reconstructions of the Caribbean-South American plate boundary qualitatively support this hypothesis, showing lithospheric fragments being swept around the northern margin of South America and accreting to the continental landmass (Norton, Exxon Production Research).
As we describe in subsequent sections, aspects of this plate boundary make it an excellent study area for the processes associated with island-arc-continent accretion, subduction polarity reversal, and HP/LT rock exhumation.
In the west, the plateau-like Caribbean plate is subducting beneath the felsic-intermediate Leeward Antilles arc and the metamorphic belts of the Caribbean Mountain system, both of which were formerly part of the Caribbean plate and are now accreting to South America. Passive and active seismology will image the subduction polarity reversal, and zones of crustal strain associated with arc accretion.
The Antilles arc is being dismembered laterally by arc-parallel extension along the margin and its fragments are being accreted to the continent. This extension is the result of displacement partitioning within the curved arc (see Fig. 11) and displacement along en echelon strike-slip faults. Available geochronology suggests that arc magmatism ceased from west to east as subduction polarity reversed and the arc began accreting to the continent. U-Pb and Ar-Ar dating will provide precise timing of magmatic history and cessation along the length of the Leeward Antilles arc, either confirming or refuting the diachronous cessation of magmatism as a function of arc-continent interaction age.
High-pressure metamorphic assemblages are being exposed in the metamorphic belts forming the transpressional zone in the core of the Caribbean Mountain system. These rocks appear to show a west-to-east decrease in metamorphic ages, we conjecture that exhumation is proceeding west-to-east with accretion of the island arc. Field studies and metamorphic thermobarometry will confirm this pattern of exhumation as a function of arc location.
The available images of mantle structure and seismicity patterns suggest that the lithospheric plate geometry along this boundary controls development of surface structures, although available tomographic images are of too low resolution to provide details. The passive source seismology will provide considerably higher resolution tomographic images of the lithosphere, as well as better constraint on seismicity patterns.
The corner flow hypothesis predicts that seismic anisotropy patterns will follow the trend of the plate boundary in a broad region up to 400km wide. The passive source seismology will map mantle flow as an anisotropy signature across the length of the plate boundary.
Portions of slabs may have broken off during choking of the original Benioff zone by South America. We will image these as well as the edge of the South American craton against which they are subducting.
Plate geometries beneath the strike-slip boundary and particularly near the zones of transition from subduction to strike-slip are poorly known (Figs 9-10), and thus the details of polarity reversal and its consequences for crustal evolution are unknown. Imaging this region will assist with interpretations of regional seismicity. In the eastern zone we will be able to understand the origin of the subduction related Matur'n Basin-Gulf of Paria gravity low, the largest on-land negative Bouguer gravity anomaly in the world. This low appears over the region where a tear between nonsubducting South America and the subducting Atlantic plate has been hypothesized.
The zone of crustal-scale orogenic float between the two plates is currently cut by E-W transcurrent faults that participate in the strain partitioning. In addition to establishing the strain partitioning, these faults and their relations with flanking sedimentary basins can be assessed to understand large-scale isostatic rebound resulting from plate truncation and overthrusting, and creating regional unconformities.
Logistically, this boundary also makes an excellent study area. Marine seismic acquisition is possible over most of the Antilles arc and into the center of the metamorphic belts comprising the core of the fold and thrust belt. The islands and peninsulas enable field study of the nature and chemistry of basement. Onshore-offshore seismic profiling, seismic velocity determination, and correlation with surface geology are also facilitated by the islands in the arc, and by islands and peninsulas of exhumed high-pressure rocks forming the interior of the folded belt system. The coastal geology of much of the plate boundary is well exposed and well studied in reconnaissance. (Venezuela and the archipelago, despite the tropical latitude, is an arid, semi-desert region with excellent exposures for geologic studies and sampling). Reconnaissance geophysics in the form of seismicity and regional tomography studies have been completed, providing considerable guidance in locating of active source seismic transects and passive seismic arrays. Petroleum exploration provides a wealth of sedimentary and upper crustal information on the foreland fold and thrust belt and the offshore regions, further allowing us to pick our seismic transects carefully. We have thousands of kilometers of shallow petroleum exploration data to examine to aid in choosing the exact locations of our transects, and for interpretation of shallow structure. We have developed important working relationships with PDVSA-INTEVEP, FUNVISIS, and a number of Venezuelan universities. The former is a Venezuelan oil company conglomerate. FUNVISIS is the Venezuelan seismological institute. PDVSA has agreed to undertake and pay for permitting the experiment with the Venezuelan government (A number of Rice, UH, and UT graduates work for PDVSA). FUNVISIS is proposing a number of seismic investigations to complement our continental margin passive and active source seismic investigations, described below and in an appendix.
We are proposing detailed land and marine passive seismic imaging from the Guyana shield to the central Caribbean to investigate upper mantle structure in northern South America and the Leeward Antilles. The imaging will be based on travel time and waveform tomography, receiver function analysis and shear-wave splitting. Detailed seismic reflection and onshore-offshore profiling along the length of the boundary and along four boundary normal corridors representing the plate boundary at different times of development will provide high-resolution, velocity-reflectivity images of the crust and upper mantle. Field geologic studies, U-Pb and Ar-Ar age dating, geobarometry and geothermometry will date and characterize the timing and cessation of arc volcanism and the exhumation histories of the metamorphic belts in the study corridors. The formation and history of the E-W through-going strike-slip fault systems and their relation to the collisional orogen will be assessed to understand the role of strain partitioning and isostatic response resulting in new basins and uncomformities. Geodynamic modeling of coupled continental tectonics and mantle dynamics in the arc-continent collision setting will constrain the physical conditions that can cause subduction polarity reversal (e.g., the buoyancy and volume of arc crust required, the influence of rheologic factors such as a weak lower crust). Modeling will also be used to see how structural elements in the mantle lithosphere, e.g., cratonic roots, can effect patterns of crustal deformation in an arc-continent setting. This aspect of the modeling will be able to map the conditions under which mantle structures are related directly to deformation patterns in the upper crust and the conditions under which some form of decoupling can occur. By comparing model predictions to the proposed geologic and seismic datasets, physical factors that are not accessible by direct observations alone can be constrained; specifically, the nature of deformation within the lower crust and, by association, its rheology.
These geodynamic, geophysical, and geologic studies will concentrate along 5 of key transects (Figs. 12-13) which are characteristic for 0 Ma (Present) to 50 Ma (Eocene) accretion and collision. To spatially sample the region well, we have added additional MCS seismic lines between the main transects, as suggested by previous reviewers of this proposal. Geologic studies will of course cover the entire study region. Geodynamic modeling will be done for 2-D and 3-D structures. Broad spatial sampling is necessary for understanding the time-transgressive nature of arc accretion, HP/LT rock exhumation, and crust-mantle interplay.
1) Trinidad/0Ma (Present): A NW-SE trending transect through Trinidad will establish the pre- and syn-collisional crust and upper mantle geometry (Atlantic lithosphere dips to the west and northwest) and will ascertain the cause of the negative Bouguer gravity anomaly over the Matur'n Basin/Gulf of Paria (Fig. 3). This is the initial condition of the collision/accretion process.
2) 65W/15Ma (Miocene): This NS trending line through Eastern Venezuela will examine the Miocene-Pliocene collisional crust-upper mantle geometry. West- to northwest-dipping Atlantic crust is hypothesized to be cut by incipient south-dipping Caribbean lithosphere subduction.
3) 67.5W/30Ma (Oligocene): The Central Venezuela line will establish Oligocene collisional crust-upper mantle geometry: How long and how far the South American crust has been subducted northward and the Caribbean southward. The two are recognized by the expected opposite vergence of structures in the upper crust.
4) 70W/50Ma (Eocene): The Western Venezuela line will show the Eocene collisional slab geometry similar to the 30 Ma line but influenced by the northward extrusion of the Maracaibo block. This and the 30Ma line are the final condition.
5) Arc/12N/0-50 Ma: This EW trending line along the Dutch and Venezuelan Leeward Antilles islands examines possible crustal thickness changes resulting from heterogeneous arc-parallel stretching (Fig. 11). Seismic velocities and fieldwork will constrain the proportions of mafic-versus-sialic rocks; deformation structures will help elucidate the processes responsible for the exhumation of high-pressure metamorphic rocks.
In addition, there are a number of collaborating projects by Venezuelan scientists: FUNVISIS is proposing three seismic investigations to complement the studies in this proposal. Some funding is already in hand, other funds are being requested from the Venezuelan government and oil industry. 1) FUNVISIS is now upgrading the permanent Venezuelan seismic network with 20 BB seismographs (Guralp CMG-40T sensors) and 30 short-period stations (1 Hz, vertical component S13s) installed in the seismogenic zones in northern Venezuela. Seven additional portable BB seismographs will be used to augment the passive experiments proposed here. The U.S. effort is designed to make use of the FUNVISIS permanent and temporary stations. 2) FUNVISIS is requesting funds from the Venezuelan National Council on Scientific and Technological Investigations (CONICIT) that will add three land shotpoints along each of four onshore-offshore transects we are proposing to record them as land refraction profiles, thus extending their landward the coverage. 3) FUNVISIS is requesting funds from PDVSA-INTEVEP to conduct a seismic refraction/reflection profile across the Serran'a del Interior fold and thrust belt, tying one of our onshore-offshore transects from the metamorphic hinterland near the coast to the foreland basin on the edge of the craton in the continental interior. FUNVISIS will also undertake detailed gravity measurements along all seismic profiles. 4) PDVSA-INTEVEP has agreed to undertake and pay for all permitting for the experiment. Additionally, they will provide financial support for two Venezuelan students to pursue Ph.D. studies at the U.S. institutions. (See attached letter).
Hans Avé Lallemant met with representatives of FUNVISIS, PDVSA-INTEVEP and Simón Bolívar University on May 4, 2000, in Caracas, and worked out a preliminary protocol for cooperation with the U.S. investigators and the Venezuelan government, universities, and oil industry. Alan Levander, Colin Zelt, and Terry Wallace will visit Venezuela in the fall of 2000 to plan the experiments, and develop student-faculty exchanges with our Venezuelan colleagues. We anticipate that a number of Venezuelan students will enroll in the Universities of Texas, Arizona, and Rice University as part of this collaboration.
This proposed project and the cooperating investigations study island arc-continental accretion, the mechanics and P-T-d-t paths of burial and exhumation of HP/LT rocks, and the development of folded belts. To do this we are proposing extensive geological investigations in the arc and metamorphic belts to provide dates on volcanic activity and uplift histories along the length of the margin, as well as the sequencing of sedimentary basin development. We will image the crust and mantle structure beneath the Caribbean-South American continent, the transpressional regime, and the Caribbean Sea. The deep seismic reflection/refraction investigations and the mantle tomography studies will jointly image the crust-mantle interactions which have produced the surface geology. The 1000km by 500km area for the active source studies encompasses a large part of the SE Caribbean plate margin. The land and marine teleseismic imaging we are proposing encompasses an even greater area than the active source investigations, and will image the lithospheric and sub-lithosphereic mantle beneath the southern Caribbean plate and as far south as the Guyana shield of the South American craton.
This project is an outgrowth of a two-day workshop on the Caribbean-South American plate boundary organized by Professor H.G. Avé Lallemant, and held at Rice University in October 1997, which was attended by approximately 40 academic scientists (including most of the U.S. PIs) and about 60 oil industry representatives, including employees of PDVSA-INTEVEP. This proposal involves 10 PIs at six U.S. universities, and one Caribbean expert who will be paid through Rice (James Pindell). Additionally, this project is coordinated with investigators at three Venezuelan institutions: The Universidad Central de Venezuela, Simón Bolívar University, and FUNVISIS. The geologic investigations are designed to begin in the first year and continue through the third year. Active source seismic experiments are planned for the second year (winter 2002), with data analysis in the third year. Passive source seismic experiments are planned for the second year and first half of the third year, with data analysis extending through the third year.
We note that criticisms of last year's submission of this project included inadequate PI support, and inadequate attention to program coordination and management. We have addressed these criticisms: PI support has been increased for the project managers. Project coordination will be enhanced by annual workshops, following the highly successful example set by the CDROM project. All PIs have budgeted for annual workshops. Management will be enhanced by additional secretarial support, and addition of a post-doc for coordinating seismic studies, and helping organize workshops. We have also included a passive OBS experiment in this request that was lacking from last year's submission.
| Direct participants in this proposal are: |
||
|---|---|---|
| Rice University | Alan Levander | PI: Project Coordinator/ MCS profiles |
| Hans Avé Lallemant | PI: Geology Coordinator/Venezuela field studies | |
| Colin Zelt | PI: Active Source Seismology Coordinator/ Onshore-Off-shore profiling, wide-angle data analysis |
|
| Adrian Lenardic | PI: Geodynamic modeling | |
| James Pindell | Research scientist: Cenozoic basins and kinematic studies | |
| Inci Ertan | Post-doc: Thermobarometry | |
| University of Houston | Peter Copeland | PI: Antilles field studies Ar-Ar thermochronology on Antilles and Venezuela |
| University of Georgia | James E. Wright | PI: Antilles field studies, U-Pb thermochronology |
| University of Arizona | Terry Wallace | PI: Passive Seismology Coordinator Land passive seismology |
| UC, San Diego | Frank Vernon | PI: Marine passive seismology |
| University of Texas | Paul Mann | PI: Seismic interpretation and tectonics |
| Gail Christeson | PI: OBS seismology and crustal structure | |
Participating Scientists (see accompanying letters in Supplementary Information Section)
- Universidad Central de Venezuela: Professor Marino Ostos - Stratigraphy in Serran'a
- Universidad Simón Bolívar: Professor J. Rigueiro - Onshore seismology
- Professor Professor J. Castillo - Onshore Seismology
- Dr. Jesus Castill - Gravity
- FUNVISIS - Dr. Michael Schmitz, Dr. Gustavo Malave - Land-active Seismology
- Dr. Herbert Rendon - Land-passive seismology
- Dr. Frank Audemard - Onshore Geology
Facebook