Continental margins mark the Earth’s principal locus of past and present lithospheric deformation, as represented by both their structural architecture and the preserved basin infill. Most importantly, the preserved basin stratigraphy is a proxy for the history of vertical motion of the lithosphere surface and thus represents a “tape recording” of the deformation of the lithosphere but of “variable fidelity.” Consequently, to understand the nature and origin of continental margins requires an understanding of the physical processes associated with lithosphere-scale deformation. Limits to understanding deformational phenomena and margin evolution have arisen in two ways. First, since the mid-1960s, when the plate tectonic paradigm was advanced, we have witnessed an enormous growth in our descriptive knowledge of the many expressions of deformational processes, but our understanding of the responsible mechanisms has not grown apace. In many instances we have no theory at all, or an incomplete physical theory to account for observed phenomena. The converse situation has also arisen—considerable advances have been made in describing deformational processes in rock physics laboratories and through computer simulations of prescribed model structures, but the specific predictions they make about the Earth are currently unsupported, in general, by the observational data base. In both cases major paradoxes have arisen—the Earth appears to behave in ways we believe to be theoretically impossible.
The “Rupturing Continental Lithosphere (RCL)” scientific community has accepted and strongly endorsed, through a series of workshops and Theoretical Institutes, the MARGINS paradigm of concentrating resources in selected critical or focus regions in order to define a process-oriented approach to study active extensional systems. Although much progress has been made over the last few years in understanding the kinematics of continental lithosphere deformation, it is agreed by the community that the mechanics by which the continental lithosphere deforms is not well understood, nor is the manner in which strain is partitioned and magma distributed. Five overarching research themes have been recognized:
• How does the strength of the lithosphere evolve during rupturing?
• How is strain partitioned during lithospheric rupturing?
• What is the role of magmatism (and volatiles) during extension and in the transition to sea-floor spreading and what is the relationship between magma petrogenesis and the deformation magnitude and history?
• What is the stratigraphic response to lithospheric rupturing?
• How are fluid fluxes modified or controlled by lithospheric rupturing?
The respective focus sites are the Gulf of California/Salton Trough, a region of orogenic crustal rifting, and the complementary central and northern Red Sea/Gulf of Suez, a rift system characterized by cratonic continental crustal extension. Defining criteria for selecting the focus sites require that:
• active continental rifting culminates laterally in seafloor spreading,
• conjugate margins are identifiable,
• pre-rift sediments and or basement and syn-rift sediments and associated fault geometries can be adequately imaged and sampled,
• the crustal structure of the entire rift system and its transition to seafloor spreading can be imaged at kilometer scales,
• highly thinned continental, transitional and young oceanic crust is accessible to sampling,
• the plate-tectonic kinematic framework is/can be well-resolved.
Allied field studies for “Rupturing Continental Lithosphere” will concentrate on outcrops that can be used to map and/or define the deformation, metamorphism and magmatism of the lower continental crust and upper mantle exhumed within rift settings. Geological, geophysical and petrological characterization of the focus sites will be accomplished by a series of transects, one group focussing on the early stages of continental rifting and the other group concentrating on the transition from late-stage continental extension to the earliest stages of oceanic crust generation and seafloor spreading.
The “Rupturing Continental Lithosphere” initiative will proceed by focused investigations combining seismic reflection and refraction imaging across the entire zone of extension and initial seafloor generation—the geophysical characterization phase. This regional phase will comprise active and passive seismic experiments to image the deformed upper and lower crust and underlying mantle structure of the rift and adjacent regions. Subsequent phases will consist of detailed geological field studies of the deformed zone, primarily to map the surface expression of deformation, to map the distribution, style and timing of deformation, and to ascertain the changing role of lowangle and high-angle fault systems. Additional studies will consist of regional and detailed geophysical, geological, geochemical, petrological and geodetic studies. Gravity and geoid data, heat flow measurements and GPS-determined deformation rate estimates will provide crucial constraints on sub-surface crustal and mantle flow and plastic deformation. ODP and IODP drilling in both shallow and deep water environments will allow the bounding surfaces of stratigraphic packages to be sampled in order to define unconformity and correlative conformity age, sedimentary facies, and depositional environment (including information on paleo-water depth), and bounding fault gouge composition and geotechnical information. Drilling will also allow direct sampling of rift-related volcanics and basement composition to be determined. Laboratory studies will be crucial in providing calibrations and a framework for the rheological properties (i.e., constitutive laws) and scaling relationships of deforming crustal and mantle rocks as functions of composition, temperature, and strain rate, in addition to controls on rock strength and permissible ranges in stress difference as a function of depth. These various disparate data sets will be integrated through numerical simulations of the four-dimensional style, distribution, and depth partitioning of extension within continental lithosphere to determine the spatial and temporal rheology of the lithosphere, why rifts form where they do, and the forces required to sever continental lithosphere.
The RCL initiative is a decadal-scale program, initiated by geophysical characterizations of the focus sites using regional geological and geophysical field experiments, and converted to a research program through MARGINS workshops, AGU and GSA special sessions, town meetings and theoretical institutes. Implementation of the RCL Science Plan, via the NSF MARGINS program, primarily facilitates competitive proposal-driven research in those focus areas considered by the community to have the best promise of solving the first-order problems related to the extensional deformation of the lithosphere. A website listing of on-going programs and successful proposals will form a basis for other synergistic projects in the RCL focus areas. The MARGINS website will also serve as data custodian for project data well before publication, in addition to providing a bibliography of work completed by US and foreign collaborators. Results will be communicated through international and national meetings and workshops.
2. The MARGINS Philosophy
The MARGINS research program can be encapsulated by the following mission statement: “To understand the complex interplay of processes that govern continental margin evolution globally." The program was initiated by the scientific community and the National Science Foundation and has been designed to elevate our present largely descriptive and qualitative knowledge of continental margins to a level where theory, modeling and simulation, together with field observation and experiment, can yield a clearer and quantitative understanding of the processes that control margin genesis and evolution.
Although continental margins have been traditionally assigned to three distinct tectonic settings, i.e., convergent, divergent and translational, the approach used by the MARGINS program recognizes that a range of fundamental physical and chemical processes that form and deform the surface of the Earth operate at all margins. Tectonic setting may govern the specific expression of a particular process that may vary in different environments. However, a relatively small number of processes, i.e., lithospheric deformation, magmatism, other mass/energy fluxes, sedimentation, and fluid flow, are fundamental to the evolution of margins. Study of these basic processes, wherever they are best expressed, provides a more logical line of inquiry for understanding the complex nature of continental margins. This process-oriented approach to understanding the entire system of margin evolution requires broadly based interdisciplinary studies and a new class of major experiments. The Rupturing Continental Lithosphere MARGINS science plan, developed from a series of well attended workshops over the past decade and finishing with the Puerto Vallarta and Sharm el-Sheikh workshops, advocates concentrating on several study areas (Focus Sites) targeted for intensive, multidisciplinary programs of research in which interaction between field experimentalists, numerical modelers and laboratory analysts would occur. Geological, geophysical and petrological characterization of the focus sites should be accomplished by a series of transects, one group focussing on the early stages of continental rifting and another concentrating on the transition from late-stage continental extension to the earliest stages of oceanic crust generation and seafloor spreading.
MARGINS plans to foster the involvement of a broad cross-section of research in focused, multidisciplinary experiments, to achieve the objectives that could not be accomplished otherwise.
3. What are the main processes
to be studied in the “Rupturing Continental Lithosphere” initiative and how will they be studied?
Although much progress has been made over the last few years in understanding the kinematics of continental lithosphere deformation, it is agreed that the physics by which the continental lithosphere deforms is not well understood, nor is the manner in which strain is partitioned (either spatially or temporally) or the timing, composition, spatial distribution and melting depth of rift-related magmas. These processes control the fundamental architecture of margins and hence the location and magnitude of resources and geologic hazards. They are best studied by a suite of nested and multidisciplinary investigations at various space and time scales utilizing field studies, laboratory experiments and numerical simulations.
Five overarching themes, also part of the original MARGINS Draft Science Plan and re-endorsed by the RCL community, comprise:
• How does the strength of the lithosphere evolve during rupturing?
• How is strain partitioned during lithospheric rupturing?
• What is the role of magmatism (and volatiles) during extension and in the transition to sea-floor spreading and what is the relationship between magma petrogenesis and the deformation magnitude and history?
• What is the stratigraphic response to lithospheric rupturing?
• How are fluid fluxes modified or controlled by lithospheric rupturing?
A particularly powerful way to address and solve these problems is to focus a comprehensive investigation on faulting, strain partitioning and magma emplacement at sites of active continental rifting where there is a lateral transition to initial seafloor spreading that will provide a spatial proxy for temporal variability. Structural targets within extensional systems include: 1) Determine the local and regional states of stress, the distribution and rate of strain, the pressures and temperatures, and the physical and chemical properties of rocks and fluids associated with a well-imaged and seismically active low-angle normal detachments (the extreme case of the weak fault paradox). Measurements of these in situ parameters made by drilling, instrumenting and long-term monitoring will be used to determine how such faults move at resolved shear stresses far smaller than those expected based on laboratory observations and Coulomb rheologies. 2) Determine the spatial and temporal distribution of strain by (i) mapping the geometry and offset of faults, (ii) inverting and modeling the stratigraphic and structural record to resolve the history of strain variation and its control on topography/erosion/deposition, (iii) using seismic, gravity/geoid and geothermal methods to obtain an integrated sum of the deformation and a measure of the ductile thinning of the lower crust, and (iv) evaluating the heterogeneity of the continental lithosphere prior to rifting. 3) Determine the pattern of mantle flow, the extent of melt generation, and the style of melt migration and emplacement during continental rifting and the early stages of seafloor spreading by imaging with seismic and electromagnetic methods an active rift-spreading transition, by measuring the heat flow distribution, and by analyzing the chemistry and petrology of magmas emplaced in these regions.
4. What do we need to know about “Rupturing Continental Lithosphere”?
In more detail, the fundamental questions that the MARGINS “Rupturing Continental Lithosphere” program is attempting to answer include:
• What are the driving forces of rift initiation and continuation?
How do these forces evolve during rifting? What are the positive and negative feedbacks during rifting that cause some rifts to succeed and others to fail? What controls the locus and conditions of initial rifting (intra-continental versus intraorogen/ intra-arc)?
• How do rifts behave as thermo-mechanical systems?
What mechanisms allow continental lithosphere to extend and rupture (e.g., what are the specific rheological flow laws and what is the role of low-angle normal faults)? What are the interactions between upper mantle thermo-mechanical processes and rifting of the continental lithosphere? What is the scale of deformation of the lower crust? How is heat transferred into and within the lithosphere during rifting? How is extensional strain partitioned, both in depth and in map view (including strain localization, the possible role of low-angle normal faults, the role of strike-slip faults and vertical-axis rotation in oblique extension)? What controls the amount, location and episodicity of strain and magmatism?
• How does the rift architecture evolve? What are the processes that control the locus of initial rifting? How do fluids (magma, volatiles) affect the lithosphere during rifting and in the transition to sea-floor spreading? What controls basin architecture/geometry, including segmentation, asymmetry, and its evolution during rifting? How do erosion and sedimentation affect tectonism and vice versa? Can correlating “strain markers” be unequivocally identified on either side of the rift system, thereby providing an absolute constraint on the amount, transport direction, and location of total strain across the rift system?
• What processes are important in the transition from rifting to initial seafloor spreading? How are these processes reflected in the structures/ geology of the continent-ocean transition? What controls the width of rifting and its ultimate focussing to seafloor spreading? What controls the locus of the continent-ocean boundary, the character and origin of transitional lithosphere (crust and/or mantle), and the manner in which extension is transferred from continental lithosphere to the mid-ocean ridge?
5. Scientific strategies required in studying “Rupturing Continental Lithosphere”
5.1. Mapping the rheological zonation of the lithosphere, the role of décollements (detachments) and accommodation zones, and the mechanical behavior of deforming crust and mantle
The scientific objectives outlined above build on a large body of continuing research that has been conducted with core funding from NSF and other agencies. For example, geophysical data from many conjugate margins document the existence of large regional subsidence with only minor accompanying brittle deformation and erosional truncation. To explain the amplitude of the regional subsidence with little or no attendant brittle deformation requires significant lower crustal and mantle extension across these margins and may possibly involve mantle dynamic effects. A suitable model requires that a diffuse zone separates the brittle and ductile deformation in the crust (i.e., an intracrustal décollement), which shoals in the region of maximum heat input. Therefore, depending on the location of athenospheric upwelling (e.g., the future ocean/continent boundary), the detachment will dip towards both margins, thereby solving a long-lived problem of passive margins, dubbed the “upper plate paradox,” in which geological and geophysical studies suggested that rift geometries and structure were inherently asymmetric but whose subsidence patterns both tended to be symmetric and identified them as upper plate margins. The balancing brittle deformation is focused in a narrow region adjacent to the continent/ocean boundary and soles into the detachment. The deformed continental crust in this region is highly intruded and overprinted by volcanism associated with rift-induced decompression melting. The depth of the detachment migrates throughout the history of the rifting in response to the input of heat. The lower crustal extension appears to be most dominant during the late stages of the rifting phase just prior to continental breakup because the upwelling of asthenospheric heat causes the lower crust to deform plastically. This is itself paradoxical as modeling studies using yield-stress envelopes suggest that the late stages of rifting should be characterized by brittle deformation irrespective of the amount of advected heat.
Requiring a major change in the behavior of extending continental crust as breakup in approached is a tantalizing hypothesis. It is nevertheless unclear if this is a common feature of rift systems that have experienced large degrees of extension. For example, while this same non-brittle, supposedly plastic deformation-induced subsidence is a feature of the western Woodlark basin, the measured present-day heat flow across the extending region completely fails to support the idea of enhanced lower crustal temperatures required to modify the mechanical properties of the extending crust. The late stages of Woodlark basin extension is indeed characterized by brittle deformation, as evidenced by the late installation of the Moresby low-angle fault system.
Addressing this rheology problem requires complete three-dimensional seismic, tomographic and geologic mapping across, along and through the zone of actively extending continental lithosphere. Achieving MARGINS research objectives such as the example above will require new experimental approaches that include developing multidisciplinary case studies, focusing on active systems, studying whole systems, establishing scaling relations and developing comparative global studies. The system volume may be on the order of 100’s by 100’s km (width and length) by 100 km (depth). Within this volume, sedimentary basins will have a characteristic rift structure and architecture coupled with a stratigraphic record of the deformation history of the continental lithosphere. Normal and low-angle faults help to define the upper crustal, brittle deformation associated with the collapse of the hanging wall blocks and ultimately the thinning of the continental crust. The footwalls of these faults, where exhumed, will provide dating, geothermometry, geobarometry, structural fabric, and hydrologic histories, all of which bear upon the nature of the deformation mechanisms at depth as modified by the exhumation process. The region would include along-strike variations in the magnitude of fault slip, as well as non-slipped regions, so that stress and strength parameters can be studied as a function of fault development and total slip.
Low-angle normal faults continue to be a source of controversy in that a number of studies of large fault structures of extensional systems indicate fault movement at resolved shear stresses smaller than those expected to cause failure. Further, it is not clear at what time during the rifting process low-angle faults play a crucial role in the crustal thinning process. Observations thus far are conflicting. Whereas low-angle faulting occurs in the late stages of extension in the western Woodlark basin, deformation in the Gulf of California/Salton Trough region has low-angle faults active both early and late in the history. Although in some locations, low-angle faults are thought to be active now, in other places the extension appears to have initiated along low-angle faults that have since been abandoned and dissected by high-angle normal faults. Although fault zone lithologic composition may play a pivotal role in facilitating slip on low-angle fault systems, the time-temperature-pressure relationships that are conducive to low-angle fault movement need to be incorporated into a viable theory to account for this mode of failure. Help with this apparent paradox will require studies of major active faults that characterize their in-situ properties and determine their evolution in space and time, together with studies of the exhumed roots of major fault systems, experimental studies of rock mechanical properties, plus modeling and theoretical studies of deforming systems. Normal slip on low-angle faults represents the extreme case of the low stress paradox and would therefore be the focus of an ideal case study. Such a study would first obtain a 3-D image of the active low-angle fault and the encompassing geological system. This would include:
• geologic and topographic/bathymetric mapping, as well as visual/radar/sonar imaging of the surface,
• heat flow, gravity and magnetic measurements,
• geodetic measurements of strain, including tilt,
• geophysical images (seismic reflection-refraction; earthquake seismicity, tomography, and shear-wave splitting; electrical resistivity) of the subsurface.
Following the initial characterization of the study volume, drilling to and through the active fault zone will be required to make in-situ measurements to determine the state of stress, fluid pressure and composition, permeability, deformation and P-T history. These measurements are required both within and away from the fault zone to understand the factors that control fault zone localization and the development of low stress, and to understand how the active fault zone compares with exhumed sections (both locally and globally). The origin, type and flux of fluids are critical parameters to be determined. Direct sampling of the fault zone at depth will provide materials for lab-based studies of physical properties and failure modes. Boreholes will also allow long term monitoring of fluid geochemistry, seismicity and strain.
In addition to this field-based case study, laboratory studies would focus on the details of the behavior of fault zone rock types at confining pressures, strain rates, temperatures and fluid pressures that are appropriate for both shallow and deep segments of major active faults.
Understanding the kinetics of mineral precipitation out of the fluid phase may be crucial in evaluating the role of rapid void production in maintaining high permeability near active faults. An important component of the laboratory studies would be determination of the physical processes responsible for the observed laboratory behavior to aid in assessing whether the processes operate on natural faults. Detailed three-dimensional mechanical models would be constructed of any areas chosen for intensive field studies. These should include constitutive relations based on laboratory data and the distribution of rock types, and physical conditions inferred from field studies. These models will be useful in guiding experimental and field work as well as in interpreting and integrating results from these other approaches. In addition, modeling studies will play an important role in planning experiments that can best differentiate between competing hypotheses about the processes active in fault zones.
5.2. Determining how strain is
partitioned as a function of space
(across and through the lithosphere)
and with time
Understanding the partitioning of strain in the lithosphere requires multidisciplinary experiments with the objective of establishing the deformational history of a margin over the full volume of the deforming structures. Such experiments must bring to bear a diversity of geological and geophysical measurements in a single region, at scales appropriate to the active processes and at denser spatial sampling than currently available. The experiments must extend across-strike from undeformed crust inboard of the margin to uninvolved oceanic crust outboard, along-strike over distances sufficient to characterize more than one segment of the margin, and vertically to the base of the lithosphere.
The overall goal of these experiments will be to determine the spatial and temporal distribution of strain in an active rift and young passive margin with the ultimate goal to identify the parameters that control strain distribution. This goal implies specific experimental targets at all levels within the lithosphere requiring:
• the mapping of definitive deformation structures in three dimensions;
• locating boundaries between regions of contrasting rheological and structural behavior, such as décollements and transfer zones;
• measuring the stress field that drives deformation;
• defining the mass balance of the deforming system, including magmatism, delamination, sedimentation, and erosion;
• determining how strain is partitioned by deformation mechanism as well as in space; and
• quantifying the roles of the parameters that control deformational processes, including thermal state, strain rate, state of stress, lithology, and lithospheric structure.
Clearly it will be difficult to directly map and observe the mode, amplitude and depth-partitioning of extension in actively deforming extensional systems, a task necessarily relegated to allied field studies. However, a variety of experimental and field approaches will be necessary to overcome these observational difficulties in order to achieve the above goals, and will include:
• Onshore geologic studies, including surface mapping, petrological, geochronological, and paleo-elevation studies, to provide relative and absolute dating, structural and uplift/subsidence history, Holocene fault motions, and the fundamental lithological framework.
• Acquiring high-resolution DEMs via SPOT (or equivalent) images and transformation software for structural mapping; draping of DEMs with LANDSAT, ASTER and other remote sensing bands to measure absolute dips, plunges, etc.
• Active-source seismic methods (multichannel and wide-angle) to define basic lithospheric geometries and Moho depth, constrain magmatic additions to the system, and image faults, offset geological markers, and (perhaps) strain fabrics within plastically deforming layers.
• Passive seismic arrays to provide insight into local seismicity, the ambient stress field and rheology, as well as lithospheric thickness, mantle velocity structure, and anisotropy.
• Surface geodetic studies, where possible, to provide boundary conditions on deformation.
• Downhole stress measurements to provide information on the stress field.
• Potential field data, including gravity, magnetic and electrical measurements, to constrain lithospheric structure and lithological interpretations.
• Heat flow and magnetotelluric surveys,
• Onshore and offshore drilling may be required, to provide detailed sampling to define sediment and basement type, volcanic rock composition, the age of the surfaces bounding stratigaphic packages, paleo-waterdepth measurements, sediment and basement porosity and permeability, and stress measurements.
Stratigraphic proxies provide first-order information on the amplitude, timing and depth-partitioning of extension, but say nothing about mechanism. For example, extensional strain partitioning at mid-crustal levels compared to depth dependent extension leads to completely different syn-rift stratigraphic geometries (Figure 3); regional sag basin development characterized by a lack of brittle deformation in contrast to the more usual configuration of a basin-bounding normal fault, hangingwall collapse and rollover, and footwall uplift, respectively. Recognizing depth-dependent extension or a strain partitioning component can be complicated during poly-phase extension, when earlier brittle deformation of the crust is overprinted by a later “plastic” phase of rifting. Most importantly, lower crust and mantle extensional thinning and depth-dependent lithospheric extension are both equivalent representations of the rifting process except that one involves plastic deformation of the lower crust and the other brittle deformation of the upper crust/middle crust (and possibly plastic deformation of the lower crust). The standard definition of rifting, that is, the active period of normal faulting, needs to be broadened to include possible plastic deformation of the lower crust. Rifting is thus more appropriately viewed as the process during which the crust is being thinned, albeit by upper (brittle) and/or lower crustal (plastic) thinning. Why brittle deformation of the crust should be abandoned and replaced by plastic deformation during the rifting process is not clear, and understanding and predicting this mechanical behavior and its rheological implications will be a major goal of the RCL initiative.
In addition to the various field approaches outlined above, laboratory and theoretical work will be crucial to designing the field experiments and understanding the data collected in them. Laboratory studies of the mechanical properties of rocks under a wide range of conditions must be undertaken to constrain such poorly understood variables as the effects of fluids, diagenesis, strain history and partial melt. Theoretical analysis will include numerical simulation of geodynamic processes and three-dimensional palinspastic reconstructions of margin geometries. Ideally, these will be jointly applied to a given extensional system. Recent advances in computational capabilities should allow numerical simulations at a small enough scale to attempt to reproduce the detailed geological evolution of extensional systems, and to thus provide constraints on the range of dynamic parameters and rheological behavior that are consistent with the observed geological history.
6. Where will we study “Rupturing Continental Lithosphere”?
The research objectives outlined earlier can best be addressed at margins where the following criteria can be met:
• Sites of active continental rifting that culminate laterally in seafloor spreading.
• Identifiable conjugate margin segments.
• Syn-rift stratal and associated fault geometry can be imaged, and the sediments sampled.
• Pre-rift surface/strata can be imaged at 100-m scales and sampled.
• Entire crustal structure can be imaged at kilometer scales.
• Pre-rift continental basement is accessible to sampling.
• Transitional crust is accessible to sampling.
• Oceanic basement is accessible to sampling.
• Plate-tectonic kinematic framework can be well-resolved.
• Access to geological and geophysical data (reflection and refraction seismics, potential field, drilling and logging data, and field observations).
• Accessibility (logistically, politically and culturally).
Using these criteria, the RCL community selected the Gulf of California and the central/northern Red Sea region as focus sites for detailed and integrated research on rifting processes. Some discussion should be made concerning why the Afar and Gulf of Aden are not part of the RCL focus sites. During the Snowbird meeting (January, 2000) in which the community deliberated on the RCL focus sites, the Red Sea and the Afar/Gulf of Aden were kept separate because of the nature of the science considered important in these two areas. The MARGINS program seeks to concentrate resources on several study areas targeted for intensive, multidisciplinary programs of research, implying that the physical size of the focus site needs to be amenable to research efforts within a reasonable timeframe. With the unanimous choice of orogenic rifting in the Gulf of California, the cratonic rifting candidates become the central/northern Red Sea and the Afar/Gulf of Aden.
Workshop participants were asked to further vote between the two cratonic focus sites again using the following criteria: Which sites will produce science that will have the greatest impact? Which sites are ready now to start work (as opposed to first enhancing an aging or poorly developed geological and/or geophysical data base and framework)? Which sites are logistically viable (i.e. if the best science sites are located in war ravaged regions, is there any point in investing MARGINS funds in such regions)? The Afar/Gulf of Aden candidate was rejected based primarily on perceived logistical problems of working onshore in Somalia and Aden, the political instability of Aden, and the then-war between Eritrea and Ethiopia.
The MARGINS community is very aware of the political and potentially negative implications of a large US program suddenly appearing in a foreign country in order to conduct so-called collaborative research. For this, our foreign colleagues must be involved at all levels of the science planning and implementation stages of the MARGINS work. Workshops in Mexico and Egypt have already informed and involved colleagues from these countries concerning the overall goals of the MARGINS program, the types of problems to be addressed by the Rupturing of Continental Lithosphere initiative, and our need for meaningful collaborations and interactions with Mexican, Egyptian, and Saudi universities and institutes. “Meaningful collaborations” imply the building on foundations already established, especially by foreign researchers, and engaging interested foreign collaborators in the formulation of both proposals and research objectives. It is imperative therefore that proposed work in Mexico, Egypt, and Saudi Arabia should involve co-principal investigators from these countries. Although a MARGINS post-doctoral fellowship scheme has been implemented, funds should also be included within proposals for the engagement of foreign senior and post-doctoral researchers in addition to students and the visit of foreign principal investigators.
6.1 Focus Site 1: Gulf of California/Salton Trough
The Gulf of California/Salton Trough focus site encompasses the zone of the Pacific-North America plate boundary, in southern California and western Mexico, where the boundary changes along strike from oceanic to continental in nature. In the southernmost sections, the continental lithosphere has been fully ruptured, leading to the creation of new seafloor, whereas in the more northern sections new seafloor is not yet forming and the plate displacement is occurring by continental extensional deformation.
The study area encompasses the modern, active plate boundary and the surrounding region where rift-related processes have been operating during the time of opening of the Gulf (since about 12 Ma). The region of continental extension surrounding the Gulf has been labeled the “Gulf Extensional Province." However, it merges with the somewhat older Basin and Range province of southern California and western Mexico at its northern edge (in California, Arizona, and Sonora) and along a part of its southeastern margin (in Nayarit; Henry and Aranda-Gómez, 2000).
6.1.2 Location and nature
of the modern plate boundary
The modern Pacific-North America plate boundary in the Gulf of California extends from the Pacific-North America-Rivera triple junction (near the Tamayo transform fault) north-northwestward along the length of the Gulf. It is a highly oblique boundary, comprising short spreading centers or pull-aparts separated by long transform faults. For a very complete summary of this region, see Lonsdale (1991). The segment of the Pacific-North America plate boundary north of the Tamayo transform fault, the Alarcon basin, has a typical mid-ocean ridge crest (the Alarcon Rise) and symmetric marine magnetic anomalies back to chron 2a (3.4 Ma) (corresponding to a total width of 180 km of oceanic crust formed in this segment). Basement depths in this basin are also consistent with it being normal oceanic crust.
Basins in the central Gulf (e.g., Farallon and Guaymas) have narrow axial magnetic anomalies, inferred to be produced by young, mainly intrusive mafic igneous rocks, such as those drilled on DSDP leg 64. Nevertheless it has proved difficult to recognize symmetric magnetic anomalies within these basins. Basins in the northern Gulf (Wagner, Upper Delfín, Lower Delfín) have very shallow bathymetry, with water depths generally less than 1000 m. These basins are sediment-filled sags with distributed normal faults controlling them, and only rare exposures of volcanic rocks at the seafloor. They lack the symmetric pattern of magnetic anomalies that would be typical of oceanic crust (Figure 5). The Salton Trough region of the plate boundary, which is the northern limit of the area discussed here, contains basins that are so filled with sediment that they are now subaerial. Here, the process of new crustal formation consists of mafic dikes intruding and metamorphosing sediment, and normal oceanic crust is absent.
The nature of the transform boundaries also changes character from south to north. The transform faults of the southern Gulf appear to be “oceanic” in character (in the sense that most of the plate boundary slip is localized along them) from the Tamayo transform fault northwards to the Ballenas transform system. However, northward of the Ballenas transform system, the various extensional basins are not separated by discrete transform structures. Rather, the basins appear to partly overlap, with complicated fault zones separating them, often with an orientation oblique to the direction of plate motion.
6.1.3 Locations and styles
of lithospheric rupture
The architecture of the rift is probably best known (at the present time) for the southernmost and northernmost segments. For the southernmost segment (the Alarcon basin) oceanic crust has been forming since 3.4 Ma. However since there is a maximum of 180 km of oceanic crust in this segment, there must be additional motion that was taken up as continental extension during the rifting process, and the details of this strain partitioning remain a first-order objective of the RCL initiative.
In the Salton Trough, geophysical and geological observations (both at the surface and at depth) are consistent with new crust having been formed here, but with a thickness exceeding 20 km and consisting of mafic magmas intruding a very thick sedimentary succession. Further, the amount and distribution of new crust produced is not well constrained (Lewis et al., 2001). Thus, the partitioning of the plate motions in space and time between the formation of new crust and the extension of the continental lithosphere adjacent to this region is still not well constrained (Oskin et al., 2001).
In the northern Gulf basins (Wagner and Delfín), the nature and composition of the crust and lithosphere are also not well known. Limited information on Moho depths suggests thicknesses of 15 km or more. The material forming the crust of the basins may be similar to the new crust in the Salton Trough, or it may be extended continental crust, or some combination of the two. However, it appears that at least across the upper Delfín basin there cannot be much area of submerged continental upper crust because of the constraints provided from geological re-constructions of the Miocene rocks that correlate on either side of the gulf (Figure 5).
The segmentation of the present plate boundary system in some cases may be related to the structural segmentation that developed during the earliest stages of rifting in the region. Axen (1995) proposed a model for structural segmentation of the western rift margin. The connection between the structural segmentation of the rift margin and the subsequent evolution of the fracture zones has been discussed by Stock (2000) for the northern Gulf basins but remains to be addressed for much of the length of the Gulf.
6.1.4 Amount of slip in the gulf
It is generally thought that the amount of opening in the Gulf of California is about 300 km. This is based on several lines of reasoning:
• The offset along the San Andreas fault system in central California, where it is best determined, is 315±10 km. This fault system is kinematically connected to the northern Gulf; although there are various faults, zones of vertical-axis rotations, etc. that may cause the net amount of displacement to differ between here and the northern Gulf (see discussion by Dickinson, 1996).
• A distinctive Tertiary conglomerate deposit is correlated from the Santa Rosa Basin (in NE Baja California) to the coast in Sonora near Isla Tiburón (Gastil et al., 1973). The locations of these outcrops suggest about 290 km of offset across the Upper Delfin basin. This is consistent with more recent studies of Miocene ignimbrites that correlate from near these two locations and suggest 255±10 km of shoreline separation (Oskin et al., 2000).
• The Alarcon basin has opened by seafloor spreading at an average rate of 45 mm/a since 3.58 Ma (DeMets, 1995). If it is assumed that this rate can be extrapolated back to 6 Ma, then the amount of total opening would be 270 km.
• There is a greater width of seafloor present south of the Tamayo Transform fault, in the mouth of the Gulf (e.g., Lonsdale, 1989). However since much of this seafloor was formed by Pacific-Rivera motion and not Pacific-North America motion, it is not necessarily straightforward to relate the amount of spreading in the mouth of the Gulf of California to the kinematics farther north in the gulf.
• There are no major Miocene or younger trans-peninsular faults on the Baja California peninsula south of the Agua Blanca fault. Thus the Baja California peninsula is inferred to have behaved as a rigid block (south of the Agua Blanca fault) during the time of opening of the Gulf of California (e.g., Umhoefer and Dorsey, 1997). Any differential extension along strike in the Gulf and its Extensional Province would thus be due to geometric effects (e.g., distance from the pole of relative plate motion).
If it is assumed that all of these observations are correct, then the amount of opening of the Gulf would be approximately the same along strike and the variations in structural style of the rift system and the nature of the continent-ocean transition are necessarily independent of spreading. In this case, exactly what is controlling the structural variation along strike remains to be identified.
6.1.5 Timing of opening of the Gulf of California/Salton Trough rift
The timing of opening of the rift is constrained from various observations: timing of extensional faulting around the margins; magmatic history; and history of marine sedimentation. There has long been a notion of a “Proto-gulf” (Karig and Jensky, 1972), an early marine incursion in middle Miocene time, followed by the “modern Gulf phase” starting at ca. 6 Ma. For the peninsular side of the Gulf, Helenes and Carreño (1999) recently summarized the Neogene sedimentary history. Studies presently underway suggest that most of the well-constrained dates for initiation of marine sedimentation in the Gulf region are 8 Ma and younger. One site that had long been considered good evidence for the proto-gulf, with marine strata thought to be ~12 Ma, is Isla Tiburón (Gastil et al., 1999) but a reinterpretation of geological relationships there suggests that the marine rocks more likely are latest Miocene in age (Oskin and Stock, 2003). Basin histories have been reconstructed from Pliocene sedimentary rocks in the Salton Trough (Winker and Kidwell, 1996), the Puertecitos area (Martin-Barajas et al., 1997) and in the Loreto region (Dorsey and Umhoefer, 2000).
Both the rift-related volcanism and the extensional faulting around the Gulf predate the majority of the known marine sedimentary rocks. The history of volcanism and extension has been recently summarized by Martín-Barajas (2000). Extensional faulting within the Gulf Extensional Province has been active in various localities prior to 6 Ma and as far back as pre-11 Ma in Baja California (Lee et al., 1996). Much of this was high-angle faulting although some low-angle normal faults have been identified in northern Baja California and the Salton Trough (Bryant, 1986; Axen and Fletcher, 1998; Axen et al., 2000). Early Miocene detachment faulting has been documented in the Basin and Range province of Sonora (Nourse et al., 1994, Gans, 1997); in the Mexican Basin and Range province, there is also middle-late Miocene normal faulting that is “proto-Gulf” in age (Henry and Aranda-Gomez, 2000).
The magmatic history of the Gulf region has been the subject of several extensive summaries (Martín-Barajas, 2000; Sawlan, 1991). Prior to the opening of the Gulf, there was a subduction-related andesitic magmatic arc along the eastern side of the peninsula, from perhaps 22 to 16 Ma in the northern part of the peninsula and from about 22 to 12 Ma in the southern part of the peninsula. The earlier cessation of this arc volcanism in the north is attributed to the earlier cessation of the subduction of Farallon plate fragments there. The andesitic arc volcanism was followed by bimodal volcanism (basalt-rhyolite, including ignimbrite eruptions) as well as by locally diverse volcanic fields containing both alkalic and tholeiitic compositions. Some of these have continued to be active into Pliocene and even Quaternary time. In addition, locally, calc-alkaline volcanism has continued to the present day (e.g., Puertecitos Volcanic Province, Tres Virgenes volcano).
6.1.6 Rift evolution in the context
of plate boundary slip
Onland geological observations suggest that the direction of opening of the Gulf of California rift has changed through time, with the early rift phase (late Miocene) being more orthogonal and the later rift phase being more oblique (e.g., Angelier et al., 1981; Umhoefer et al., 1994). It has been proposed that the Gulf has a two-stage tectonic history (e.g, Stock and Hodges, 1989). During the first phase, Baja California acted as a microplate caught between North America and the Pacific plate, so that the opening of the Gulf represented part of the total Pacific-North America plate motion, with the remainder of the motion being strike-slip in nature and accommodated elsewhere. During the second phase, Baja California would have been essentially attached to the Pacific plate, so that the later opening in the Gulf represents Pacific-North America plate motion.
One constraint on the overall kinematics of the region is the motion between the Pacific and North America plates, determined from global plate circuits (e.g., Atwater and Stock, 1998). The plate circuit calculations show a total of 600 km of relative displacement of the Pacific plate relative to North America since 12 Ma (for a point now near the Salton trough). If the Gulf only contains about 300 km of opening during that interval, the remainder of the plate motion must have been accommodated somewhere else. A common assumption is that several hundred km of this missing motion occurred by margin-parallel strike-slip displacement in the California Borderland west of the Peninsula, on structures of the Tosco-Abreojos-San Benito fault system (Spencer and Normark, 1979). This is inferred from the displacement of the Magdalena fan from its probable source area (Yeats and Haq, 1981). However, the exact amount of displacement of the Magdalena fan, as well as the timing of this displacement, has come under scrutiny (Fletcher et al., 2000). Even if the total offset of several hundred km is correct, there is still some additional extension, younger than 12 Ma, that is needed between Pacific and North America at the latitude of the Gulf of California. This motion may be represented by extension within the Mexican Basin and Range province, as suggested by Henry and Aranda- Gomez (2000).
Another overall kinematic constraint is that geological observations on the conjugate margins of the Upper Delfín basin require most of the displacement of the two sides to be post-6.1 Ma (Oskin et al., 2000). Thus it appears that Baja California was moving at nearly the Pacific rate relative to North America through most of Pliocene time, but that between 6 and 12 Ma the average rate of opening of the Gulf must have been very slow.
6.1.7 Geodynamic setting of the Gulf of California at the start of rifting
Prior to the development of the Gulf of California, there was a long history of eastward subduction at these latitudes, with the oceanic Farallon plate converging with western North America. The Pacific-Farallon ridge approached North America during Tertiary time, until it reached the trench and the Pacific and North America plates came into direct contact at about 28 Ma. At this point the subduction zone contained two independent plates (Juan de Fuca and Nazca) separated by the Pacific-North America boundary. This boundary zone lengthened with time because its northern end (the Mendocino triple junction) migrated northwards with the Pacific plate, and its southern end (the Rivera triple junction) jumped discontinuously southward due to progressive extinction of spreading between the Pacific plate and the microplates. As the Rivera triple junction migrated southward, subduction west of Baja California would have ceased, and the subduction-related volcanic arc in Mexico was progressively extinguished from north to south.
Plate reconstructions show that by about 16 Ma the Rivera triple junction was adjacent to Central Baja California (Stock and Lee, 1994). At this time, the Baja California continental borderland north of the Vizcaino peninsula was taking up much of the deformation of the Pacific-North America boundary. Arc volcanism in northern Baja was waning, whereas subduction related volcanism was still continuing in the region of southern Baja California. The Magdalena and Guadalupe plates, which were being subducted adjacent to southern Baja California, had broken from the Cocos plate, which itself had broken from the northern part of the Nazca plate (Lonsdale, 1991).
The Magdalena and Guadalupe microplates stopped spreading with respect to the Pacific, and became attached to the Pacific plate, at about 12 Ma. Thus, the Rivera Triple Junction jumped down to approximately its present location, and the entire length of boundary west of southern Baja California experienced a major tectonic change, from a microplate-North America subduction zone to a transtensional zone of motion between the Pacific and North America plates. Nearly all of the extension in the Gulf extensional province in Baja California happened after this tectonic change. However, some extension east of the Gulf occurred prior to this time and hence would have been taking place in a back-arc setting.
After this major tectonic change, the Pacific-North America plate boundary motion at the latitude of the Gulf of California must have been accommodated largely outside the Gulf region (either west of the Baja California Peninsula, or east of the Gulf in mainland Mexico, or both). There was a delay of roughly 6 m.y. from when subduction stopped adjacent to the Baja California peninsula until the Gulf began to move at nearly the full Pacific-North America plate rate. The kinematics and dynamics of the plate boundary history during this interval (12-6 Ma) are debated but are certainly important to a full understanding of how the Gulf of California rift developed.
6.1.8 Outstanding problems of the Gulf of California/Salton Trough region
Numerous aspects of the Gulf of California/Salton Trough region lend it to investigation within the Rupturing of the Continental Lithosphere Initiative. This brief discussion cannot hope to touch on everything but rather highlights a few of the important issues. First, the Gulf of California contains a transition along strike from what is clearly oceanic spreading, in the south, to active continental extension that is clearly not oceanic spreading, in the north. This transition does not appear to be related to the timing, the net amount, or the rate, of the extension, as these parameters do not vary much along the length of the rift (insofar as we know at the present time). This transition needs to be better characterized and the factors controlling it need to be identified and understood.
The crust in the “transitional” region has only been studied in the Salton Trough area. Elsewhere it may be similar in nature to that found in the Salton Trough, but this is not known. One of the outstanding questions for this rift system concerns the degree to which lower continental crustal material may have flowed out from under the marginal areas into the region of the rift, as has been documented farther north in the Basin and Range province.
The lithospheric architecture of the margins of the rift is incompletely known, but identifiable conjugate margin segments can be studied to determine the complete architecture of given segments of the rift and their evolution through time. These conjugate margins are accessible on land on both sides. The distribution of water and land offers much potential for well-designed studies to constrain the Moho depth, the seismic velocity structure, and other parameters that will be necessary for modeling the rift evolution and for understanding its present geometry and structure.
Syn-rift stratal and associated fault geometries can be imaged in the Gulf of California/Salton Trough region, and the sediments are within reach to be sampled by the drill in order to constrain the rift evolution, both in surface exposures and in submarine settings. The syn-rift sedimentation rate generally decreases from north to south due to the high sediment input of the Colorado River at the northern end of the system, and the character of the sediments changes accordingly, providing, at the north end of the Gulf, a very high fidelity record of the faulting and sedimentation history, and farther south in the Gulf, thinner sequences that can provide information farther back in time. The region has not been significantly overprinted by any post-rifting tectonics or buried by excessive volcanism. The accessibility of the onland geology means that the evolution of rift segmentation with time, and the changes in strain partitioning with time during rifting, can be addressed.
Finally, there is great potential for addressing the thermomechanical history of the rift and its present thermal state. Although the rift is not yet blanketed by volcanics, there are pre- and syn-rift volcanic rocks in various areas in and around the rift. These have the potential to provide information on the nature of mantle source regions beneath the rift, the nature of the mantle lithosphere (if xenoliths are present) as well as constraints on the thermomechanical evolution of the rift. Very few heat flow measurements are available for the rift but the correct conditions certainly exist for such measurements to be made. In addition, there are definite (slow) upper mantle shear velocity anomalies beneath the Gulf of California and Basin and Range Range that may reflect the present-day thermal structure of the lithosphere (Figure 6).