Pulling the Rug out From Under California: Seismic Images of the Mendocino Triple Junction Region

The active tectonic regime of northwestern California changes abruptly from transform motion to subduction at the Mendocino Triple Junction. Northward migration of the triple junction has been a major factor in the tectonic history of the continental margin of California since the Oligocene and continues at present. Understanding the effects of triple junction migration on the structure of the crust and upper mantle in this region is therefore necessary for reconstructing the geologic evolution of the continental margin of California and accurately assessing seismic hazards associated with the San Andreas fault system and the Cascadia subduction zone.

In 1993 and 1994 a network of large-aperture seismic profiles was collected to image the crustal and upper-mantle structure beneath northern California and the adjacent continental margin. The data include approximately 650 km of onshore seismic refraction/reflection data, 2000 km of offshore multichannel seismic (MCS) reflection data, and simultaneous onshore and offshore recording of the MCS airgun source to yield large-aperture data. Scientists from more than 12 institutions were involved in data acquisition.

Mendocino Triple Junction

Fig. 1. a) Locations of the 1993 and 1994 seismic experiments overlain on a simplified geological map of northern California. BoS, Base of Slope; BSF, Barlett Springs Fault; CLV, Clear Lake Volcanics; CSZ, deformation front of the Cascadia Subduction Zone; MF, Maacama Fault; MTF, Mendocino Transform Fault; PA, Point Arena; SAF, San Andreas Fault. Orange arrows show the portion of lines 1, 6, and 9 shown schematically in Figure 2
b) Schematic three-dimensional diagram of plate interactions in the region of the Mendocino triple junction showing development of a slab window as the Gorda plate migrates north (not to scale). GR, Gorda Ridge; LP, Lassen Peak; MS, Mount Shasta; other abbreviations are the same as in Figure 1a. Pale shades represent crust; darker shades represent lithospheric mantle. Pale yellow "Pacific plate" indicates crustal material that has been transferred from the North American to the Pacific plate; gradual shading from yellow to green indicates crust within the San Andreas fault zone that has a motion intermediate between that of the North American and Pacific plates. Arrows represent directions of asthenospheric flow.

Since Atwater's [1970] pioneering work on the influence of plate tectonics on the geological history of western North America, consideration of the time-transgressive dynamic effects of triple-junction migration has led to an increased understanding of the geological evolution of western North America. Dickinson and Snyder [1979] introduced the concept of a slabless window resulting from northward migration of the subducted Gorda/Juan de Fuca plate relative to North America and correlated major tectonic events with the movement of the slab-free zone. Furlong [1984] developed this concept into a thermomechanical model of the upper plate in the wake of the migrating subduction zone. The predictive capability of more recent detailed modeling studies, which incorporate thermal effects of the variation in age of the subducted plate during the history of the margin or calculations of the amount of melt generated by asthenospheric upwelling in the slabless window, has been limited by uncertainties about the structure of the crust overlying the slabless window.

The implications of the slabless window scenario are illustrated in Figure 1b. As the Gorda plate moves to the northeast relative to North America, it leaves in its wake a wedge of North American lithosphere that thins to the west and is underlain by asthenospheric upper mantle. This scenario leads to a number of predictions about spatial variation. For example, arc magmatism should end at the triple junction. Elevation of the coast ranges should increase rapidly south of the MTJ as hot, asthenospheric mantle flows into the slab-free zone; elevation should decrease further south as this material cools. The asthenospheric upwelling should also result in high heat flow and crustal and subcrustal magmatism, and the maximum in heat flow and the surficial expression of volcanism should lag the highest topography. Velocity anomalies in the crust and upper mantle are associated with heating, melting, and magmatic underplating. These predicted patterns have been confirmed by modeling of heat flow data, seismic P wave delays, and surface wave dispersion. The slabless window model also explains observed patterns of recent volcanism and uplift and deformation of the Coast Ranges north of San Francisco. A list of references to these studies can be found in the Mendocino World-Wide Web homepage (http://quakes.oce.orst.edu/mendocino).

While the slab-free window model is consistent with the large-scale thermomechanical behavior of the lithosphere and asthenosphere, a number of problems arise when it is used to explain the observed crustal architecture. While the model can explain why the Pacific/North American plate boundary has migrated east with time, resulting in the transfer of the leading edge of the North American plate to the Pacific plate, it does not explain why seafloor magnetic anomalies offshore southern and central California indicate that the Pacific/Farallon spreading center was deformed and slowly extinguished before it reached the subduction zone between the Farallon and North American plates, welding fragments of the Farallon plate to the Pacific plate and influencing the evolution of the Pacific/North American plate boundary [ Nicholson et al. , 1994]. Nor does it explain why oceanic crust and/or sediments of the Pacific and/or Farallon plates appear to have been tectonically underplated beneath rocks of North American affinity that are now found on the Pacific plate. These observations have led Bohannon and Parsons [1995] to challenge the slab-free window model and suggest that the entire Farallon plate is welded to the Pacific plate, and it remains in the lower lithosphere and asthenosphere beneath western North America, controlling Basin and Range extension. Their model, however, provides no explanation for the topographic, tomographic, and volcanic observations.

The primary objective of the Mendocino triple-junction seismic experiment was to image the lower crust and upper mantle immediately before and after triple-junction passage to help resolve arguments for and against the slab-free window model and understand the processes shaping western North America.

Regional Seismicity

Fig. 2. a) Cross section north of the MTJ showing seismicity of the Cascadia subduction zone along a transect approximately coincident with line 6 (adapted from Smith et al. [1993]). The top and bottom of the crust of the subducted Gorda plate and the boundary between the Franciscan and Klamath terranes, as determined from our data, are also shown. Solid lines show interfaces constrained by seismic observations; dashed lines show the inferred extensions of these interfaces.
b) Cross section south of the MTJ along the western half of onshore line 1, showing a simplified line drawing of the anomalously high-amplitude lower crustal reflectivity. Earthquakes (1984–1994) occurring within 7.5 km of the profile are also shown. Hypocentral locations for earthquakes with magnitude > 1.5 are from the NCSN catalog available from the University of California, Berkeley; symbol size increases as a function of magnitude. This pattern of seismicity, which shows steeply dipping faults in the upper 15 km of the crust, is typical of the San Andreas fault system.
c) Cross section of seismicity (1984–1994) within 7.5 km of line 9 from the NCSN catalog, showing the transition from a strike-slip to a convergent seismotectonic regime. The top and bottom of the Gorda plate and the position of anomalously strong lower crustal reflectivity south of the triple junction are also shown. Crust labeled as "transitional" is in the process of being transferred from the Pacific to North American plates; because offsets in lower crustal reflectivity underlie faults of the San Andreas fault zone, we infer that the underlying mantle also is transitional.

North of the triple junction, a Benioff zone is defined, which suggests that the Gorda slab is at a depth of about 15 km near Cape Mendocino and dips about 11° inland (Figure 2a). In this zone, seismicity apparently doubles; three explanations have been proposed. Smith et al. [1993] suggested that the subduction zone jumped east. In this view, the double seismic zone represents reactivation of the old plate boundary and, at the same time, motion along the present plate boundary caused by north-south compression of the Gorda plate. Wang and Rogers [1994] suggested that the double seismic zone represents rheological layering: a layer of ductile oceanic lower crust is sandwiched between brittle upper oceanic crust and upper mantle. Alternatively, the double seismic zone simply may be an artifact (D. Oppenheimer, personal communication, 1995).

South of the MTJ (Figure 2b), the pattern of seismicity changes abruptly. Here, most earthquakes are less than 15 km deep and occur along steeply dipping faults parallel to, but east of, the San Andreas Fault. This change is compatible with geodetic observations and geodynamic models that suggest the main locus of North American/Pacific motion is presently east of the San Andreas fault. An important unresolved question is whether these faults are coupled through amid crustal detachment zone, as has been suggested for analogous faults in the San Francisco Bay area [Brocher et al. , 1994]. Such coupling has implications for models of stress accumulation and release.

The 1993 and 1994 Seismic Experiments

In June 1994, ~2000 km of multichannel seismic (MCS) profiles were collected on the Northern California continental margin. The MCS data are complemented by large-aperture data recorded at 34 seafloor sites and 477 onshore sites (Figure 1a). In addition, a shot point from 1993 (near the intersection of lines 1 and 9) was reoccupied. Data reports documenting these experiments are being prepared.

Each of these field programs represented a formidable logistical challenge. The 1993 experiment required a considerable investment in preexperiment surveying to find nearly straight paths through the rugged, heavily forested northern California coast ranges. The 1994 field season required coordination among two different land parties and two research ships with different, and sometimes conflicting, logistical constraints and problems.

Preliminary Results

North of the triple junction (Figure 2a), the new data indicate that the crust of the Gorda plate lies above both layers of seismicity: it is about 10 km beneath the deformation front and dips about 7° to the east beneath the continental margin. This indicates that the earthquakes in a three-dimensional model must be relocated before inferences can be made about the thermomechanical behavior of the lithosphere. An earthquake offshore Cape Mendocino was recorded on several instruments during the experiment and will be useful for relocating the earthquakes. Onshore, dip increases to about 11° and most of the seismicity appears to be in the subducted oceanic crust and uppermost mantle of the Gorda plate; shallow crustal seismicity is concentrated in the Franciscan terrane.

South of the triple junction (Figure 2b), the lower crust is anomalously reflective in the region where the slabless window is predicted by other investigative methods. A single-fold seismic reflection section constructed from the large-aperture data shows a discontinuous layer of lower-crustal reflectivity that deepens to the east and extends from the coast to the western edge of the Great Valley. The high-reflectivity layer appears to be offset beneath the Maacama (MF) and Bartlett Springs (BSF) fault zones, which suggests that these faults extend through the entire crust and into the upper mantle. This result contrasts with recent work in the San Francisco Bay region that indicated a midcrustal detachment zone linking the Hayward and San Andreas faults in the San Francisco Bay region [ Brocher et al. , 1994]. The high reflectivity beneath the Coast Ranges may result from the presence of either magmatic or aqueous created by thermal effects of the slab window. The crust thickens as the triple junction is approached from both north and south (Figure 2c) [ Verdonck and Zandt , 1994]. The new data suggest that the thickening may represent a tectonic effect resulting from deformation of Franciscan rocks overlying the Gorda plate as the triple junction migrates north. South of the triple junction, Franciscan rocks may have been thinned through assimilation into slab window melts and are underlain by the material that generates the high-amplitude reflections.

The 1994 data also provide spectacular new images of the offshore San Andreas and Mendocino transform faults, the Cascadia subduction zone "megathrust" decollement, and the crust of The Gorda and Pacific plates. A very strong lateral change in velocity and reflectivity is associated with the San Andreas fault, which has been seismically quiet in this region since the 1906 earthquake. Considerable along-strike variation is observed in the reflectivity of the decollement and in the faulting pattern within the accretionary complex, which may be associated with backstop structure and with the type and thickness of sediments on the subducting Gorda plate. Large basement and seafloor offsets are observed on the Gorda plate, which will be useful for constraining the history of intraplate deformation.

In conclusion, although a detailed reconstruction of crustal and upper-mantle evolution in response to triple-junction migration awaits an integrated analysis of the complete 1993 and 1994 data sets, data from the Mendocino Triple Junction seismic experiment are leading to a more accurate three-dimensional model of the region's lithospheric structure and the related geodynamic processes at work. The scientific community is encouraged to contribute to the data analysis through participation in a special poster session at the 1995 AGU Fall Meeting and a workshop in Eureka, Calif., that is being planned for spring 1996. Updates about this workshop and the status of publications documenting the new data can be obtained from the Mendocino World-Wide Web homepage .

Mendocino Working Group

Acknowledgments

References

Wang, K., and G. C. Rogers, An explanation for the double seismic layers north of the Mendocino triple junction, Geophys. Res. Lett., 21 , 121, 1994.

Return to Science and Society
Return to Starting Point