R. N. Hey

Tectonic evolution of the Cocos-Nazca spreading center

Magnetic and bathymetric data from the eastern Pacific have been analyzed and a model for the evolution of the Galapagos region developed. The Farallon plate appears to have broken apart along a pre-existing Pacific-Farallon fracture zone, possibly the Marquesas fracture zone, at about 25 m.y. B.P. to form the Cocos and Nazca plates. This break is marked on the Nazca plate topographically by the Grijalva scarp and magnetically by a rough-smooth boundary coincident with the scarp. The oldest Cocos-Nazca magnetic anomalies parallel this boundary, implying that the early Cocos-Nazca spreading center trended east-northeast. This system soon reorganized into an approximately east-west rise–north-south transform configuration, which has persisted until the present, and the Pacific-Cocos-Nazca triple junction has since migrated north from its original location near lat 5°S. If correct, the combination of these simple geometric constraints produced the “enigmatic” east-trending anomalies south of the Carnegie Ridge. The axes of the Cocos-Nazca spreading center and the Carnegie Ridge are essentially parallel; this can lead to paradoxical conclusions about interpretation of the Cocos and Carnegie Ridges as hotspot tracks. Hey and others (1977) have shown that recent accretion on the Cocos-Nazca spreading center has been asymmetric, resulting at least in part from small discrete jumps of the rise axis. I show here that the geometric objections to both the “hot-spot” and “ancestral-ridge” hypotheses on the origin of the Cocos and Carnegie Ridges can be resolved with an asymmetric-accretion model. However, all forms of the ancestral-ridge hypothesis encounter more severe geometric difficulties, and these results support the hotspot hypothesis. After further elaboration of the hotspot hypothesis by Johnson and Lowrie (1972) and Hey and others (1973), Sclater and Klitgord (1973) examined both the hotspot and ancestral-ridge hypotheses and decided that both should be rejected, concluding that the Cocos and Carnegie Ridges “are not tectonically related” (p. 6973). The difficulty in the Galapagos area arises primarily because the older magnetic anomalies have proven extremely difficult to correlate, surprisingly so considering the high data density and the ease with which the very young anomalies are correlated. An important problem is the reason for this difficulty in correlating older anomalies — either clearly recognizable anomalies were never formed here, or some mechanism has acted in this area to destroy them after they were formed. On the basis of all available evidence, including new data presented here, I conclude that a model based on the hotspot hypothesis, with the modification of asymmetric accretion resulting at least in part from discrete jumps of the rise axis as discussed by Hey and others (1973) and demonstrated by Hey and others (1977), successfully meets the objection of Sclater and Klitgord (1973) and allows us to outline the history of the area from the break-up of the Farallon plate and birth of the Cocos-Nazca spreading center to the present. The “instantaneous” (a term Hey and others, 1977, have examined) configuration of plate boundaries and motions (Fig. 1) has generated a wedge of crust spread from the Cocos-Nazca spreading center, which is characterized by a slow spreading rate, rough topography, and strong magnetic anomalies. This wedge, termed the Galapagos gore by Holden and Dietz (1972) and discussed in detail by Hey and others (1977), is surrounded by crust spread from the Pacific-Cocos and Pacific-Nazca spreading centers which has the smooth morphology common to fast-spreading rises and low-amplitude magnetic anomalies, as both these segments of the East Pacific Rise are oriented nearly parallel to the Earths magnetic field vector. My model explains the location and orientation of the magnetic and bathymetric rough-smooth boundaries that thus bound the gore and implies that there are two genetically different magnetic and bathymetric boundaries in the area.

A new class of “pseudofaults” and their bearing on plate tectonics: A propagating rift model

The pattern of magnetic anomaly offsets striking obliquely to the Blanco fracture zone near the Juan de Fuca spreading center appears to be incompatible with the rigid-plate hypothesis. Previous workers have thus called upon complex, or anomalous, mechanisms to explain the tectonic evolution of this area. According to the “propagating rift” model developed here, the basic observations that previous hypotheses have attempted to explain, i.e., the oblique trends of the fracture zones, are in fact misinterpretations. The previously proposed faults are instead “pseudofaults” consisting of en echelon sets of fracture zones frozen into progressively younger crust; these en echelon fracture zones have resulted from sequences of spreading center jumps propagating down the Juan de Fuca spreading center. Although the overall trends of the en echelon fracture zones are oblique to the Blanco transform fault, the strike of each individual fracture zone is quite different, and is compatible with transform motion between the rigid Pacific and Juan de Fuca plates.

Petrologic consequences of rift propagation on oceanic spreading ridges

Abstract The production of anomalously differentiated lava compositions at several mid-ocean spreading centers can be attributed to magmatic processes associated with propagating rifts. The degree of differentiation attained by magmas beneath oceanic spreading ridges depends mainly on the balance between cooling rate and the supply rate of new magma to shallow chambers. Low supply rates and moderate cooling rates allow advanced degrees of closed-system fractionation to occur. High supply rates result in open systems in which magma compositions are buffered by frequent replenishment with new hot magma. Propagating rift tips are a special class of ridge-transform intersection in which the balance between cooling and supply rates is conducive to the development of advanced degrees of differentiation over an expanded length of ridge. This balance is affected by the spreading rate, the propagation rate of the rift, the length of the bounding transform and proximity to hotspots. Maximum compositional variability and maximum degree of differentiation occur within 50 km of propagating rift tips and subsequently diminish with increasing distance. Rifts that propagate through plates in directions approximating their absolute motion relative to the lower mantle are characterized by the presence of anomalously differentiated lavas over longer ridge segments than are rifts that propagate against their absolute motion. Geochemical anomalies may persist, though changing in degree and extent, for several million years on ridge segments that stop propagating. The concept of “magnetic telechemistry” is generally supported by our study, but in the vicinity of hotspots, magnetic anomaly amplitude may be controlled more by bathymetric and/or thickened magnetic layer effects than by geochemistry.

History of rift propagation and magnetization intensity for the Cocos‐Nazca sspreading Center

Analysis of magnetic anomaly profiles collected nearly parallel to tectonic flow lines allows detailed interpretation of the complicated tectonic history of the Cocos-Nazca spreading center. Forward models of the magnetic anomalies accounting for spreading rate variations, ridge jumps, asymmetric spreading, magnetization intensity variations, and bathymetry show excellent agreement with observed anomalies. Spreading rates can be constrained to a common finite rotation history through anomaly 4A with three changes in rates. Rate changes at about 1.5 Ma and 4.1 Ma correspond to changes in rate gradients and occur during the well-calibrated part of the reversal timescale, so they can unquestionably be identified as true changes in plate motion. A ∼15% rate decrease at about 5.2 Ma could be interpreted either as a change in plate motion or as an artifact of poor calibration of the older part of the timescale. The change at 4.1 Ma is especially important because many timescales are based on the assumption of constant spreading rate for this plate pair for 0–6 Ma. Rift propagation has played a dominant role in the continuous reorganization of the geometry of the ridge axis. Propagation has been predominantly away from the hotspot, with jumps predominantly south-ward. Propagation rates have ranged from 30 to 120 mm/yr, commonly near 70 mm/yr. Origin of most propagation sequences is difficult to interpret, but many appear to involve discrete southward ridge jumps forming a new segment near the hotspot. Magnetic anomaly amplitude appears to be a reliable tracer of Fe content of lavas. Several generalizations can be drawn about along-axis variations in magnetization intensities since 8 Ma: high magnetizations are only observed at the far ends (relative to the Galapagos hotspot) of segments at least 150 km long; offset at the end of a high-magnetization segment is at least 15 km; and there are no offsets larger than 30–45 km between high-magnetization segments and the reconstructed position of the hotspot. We interpret these patterns to indicate that fractionated lavas erupt where gradients in magma supply cause along-axis flow of evolved magma. The gradients in supply result from subaxial flow of hotspot-derived asthenosphere in a narrow conduit. This flow is only partly obstructed by an offset of 20–30 km but entirely blocked by an offset of 50 km.

Recent plate motions in the Galapagos area

Recent accretion on the Cocos-Nazca spreading center has been asymmetric, with more material (along most of the rise) added to the Cocos than to the Nazca plate. There is evidence in the magnetic record that some of this asymmetric accretion has resulted from small discrete jumps of the rise axis to the south, forming and destroying transform faults. Relative and absolute models of instantaneous plate motions derived here provide an accurate representation of recent motions in the east Pacific. The existence of a self-consistent model that fits all the relative-motion data provides strong support for the hypothesis that plates behave rigidly. In addition, the exceptional agreement between the relative-motion and absolute-motion models provides strong support that the Wilson-Morgan hotspot hypothesis holds for the recent past. In particular, the agreement of the predicted instantaneous azimuths of the Cocos and Carnegie Ridges with their observed azimuths strongly suggests that at least the young parts of both aseismic ridges were formed by the motion of the Cocos and Nazca plates over a Galapagos hotspot.

Propagating rift explanation for the tectonic evolution of the northeast Pacific—the pseudomovie

Abstract We have developed a new computer graphic technique to use in reconstructing plate tectonic evolution and to make movies illustrating this evolution. Using this technique, we have deciphered the evolutionary history of the Juan de Fuca area in the northeast Pacific, and show that a combination of seafloor spreading and rift propagation can explain this seemingly complex area as a simple consequence of the interaction of two rigid plates.

Paleoceanography of the Tropical Eastern Pacific Ocean

The East Pacific Barrier (EPB) is the most effective marine barrier to dispersal of tropical shallow-water fauna in the world today. The fossil record of corals in the eastern Pacific suggests this has been true throughout the Cenozoic. In the Cretaceous, the EPB was apparently less effective in limiting dispersal. Equatorial circulation in the Pacific then appears to have been primarily east to west and the existence of oceanic atolls (now drowned guyots) in the eastern Pacific probably aided dispersal. Similarly, in the middle and early Mesozoic and late Paleozoic, terranes in the central tropical Pacific likely served as stepping stones to dispersal of tropical shelf faunas, reducing the isolating effect of an otherwise wider Pacific Ocean (Panthalassa).

Spreading center jumps and sub-axial asthenosphere flow near the Galapagos hotspot☆

Abstract We have identified a sequence of small rise-axis jumps on the Cocos—Nazca spreading center between 93° and 95.5°W. The locus of jumps has migrated 150 km west along the rise axis, away from the Galapagos Islands, during the last three million years, at an average rate of 50 mm/year. The linear increase in jump distance during this sequence of jumps has resulted in a change in regional azimuth of the rise axis from about 085° to 095°. We visualize this sequence of jumps as a new rift propagating through the Cocos plate, forming a new Cocos—Nazca spreading center. The region affected by these rise jumps appears to correlate with an area of exceptionally high-amplitude magnetic anomalies. The high-amplitude region seems to result from Fe-Ti-rich (FeTi) basalts of high remanent magnetization. We speculate that the development of the new accretion axis and concomitant rise jumps are related to the flow of FeTi basalt-producing asthenosphere away from the Galapagos hotspot. The snout of anomalous asthenosphere has remained nearly stationary, with respect to the Galapagos hotspot, during the last 3 m.y. A northwestward component of flow, reflecting the southward position of the plume center with respect to the spreading axis, might explain why the new spreading center is developing along a more northwesterly azimuth. The rise jumps have resulted in the sort of pattern of asymmetric accretion which is required to substantiate the hotspot hypothesis for the origin of the Cocos and Carnegie ridges. Several other puzzling platetectonic phenomena may be explained by the propagating rift model developed here.

Speed limit for oceanic transform faults

Oceanic transform faults with slip rates greater than ∼145 km/m.y. do not currently exist along the East Pacific Rise where sea-floor spreading rates range from 145 to 160 km/m.y. Instead, offsets of the the very fast spreading East Pacific Rise are accommodated by microplates, propogating rifts, or overlapping spreading centers. This suggests that there might be a speed limit above which transform faults do not exist. A physical reason for a speed limit is not known, but it might be related to unstable stress fields near the rifts tips, causing them to episodically propagate and prevent a transform fault from being formed. The spreading rates quoted are from our new (0-0.73 Ma) relative-motion model for the Pacific and Nazca plates.

Synchronous reorientation of the Woodlark Basin spreading center

Abstract A sidescan and multibeam bathymetry survey of the Woodlark Basin reveals that its 500 km long spreading center reoriented synchronously, without propagation, about 80 ka. There is no evidence of the V-shaped pseudofault geometry typical of spreading center propagation, nor of the progressive fanning of seafloor fabric characteristic of spreading center rotation. The reorientation is recognized by a sharp contact between two seafloor fabric trends, and ruptured off-axis lithosphere formed up to 0.7 Myr previously. The length of the reoriented spreading segments and the tendency to fault pre-reorientation seafloor fabric are controlled by the strength of the lithosphere, the angle of the reorientation, and the length of pre-existing spreading and transform segments. We document the process of synchronous reorientation in the Woodlark Basin and propose that it may occur in other ocean basins.