Shear-wave Splitting in the Lower Crust Beneath the Archean Crust of Southwest Greenland

William P. Clement, Ramon Carbonell, and Scott B. Smithson
Department of Geology and Geophysics
University of Wyoming
Laramie, WY USA 82071


Slant stacks of seismic data from rifted ancient Archean crust along the southwest coast of Greenland indicate that the lower crust is strongly seismically anisotropic. The slant stack data show that a lower crustal, wide-angle S-wave reflection has a different intercept time and ray parameter on the radial and transverse components from a receiver gather recorded on PASSCAL REFTEK three-component seismometers. From the S-wave analysis, the continental crust is clearly seismically anisotropic above a high velocity wedge in the lower crust. Possibly, magmatic underplating during Late Cretaceous Labrador Sea rifting heated the pre-existing lower crust promoting plastic flow and enabling alignment of anisotropic minerals to produce the seismic anisotropy.


A deployment of a marine airgun source and land based receivers along the southwest coast of Greenland provided high-resolution, wide-angle seismic reflection data that show shear wave seismic anisotropy beneath the attenuated continental crust. The seismic survey, conducted by the University of Wyoming in 1989, consisted of a marine airgun array recorded by PASSCAL REFTEK 2-Hz three-component seismometers (Figure 1). The research vessel cruised by the land-based stations firing its airgun at 100 to 120 m intervals to record continuous common-receiver gathers out to offsets of 150 to 250 km to generate an unusual data set. Several shear phases were identifiable on the receiver gathers. In particular, a strong lower crustal shear phase was evident on both horizontal components recorded at Kangak station on the coast near Godthaab. This lower crustal shear phase is evident on other sections along the coast, but is not as prominent. The data presented in this paper were recorded at the Kangak fjord as the marine array was towed parallel to the southwest coast of Greenland. Preliminary analysis of the other receiver gathers confirms the conclusions based on the Kangak data. The strong shear energy in this densely sampled data set enables a detailed study of the shear-wave response of the crust beneath southwest Greenland.

Geology and Geophysics

The Archean block of southwest Greenland is a rifted part of the Nain province that also crops out in Labrador, Canada (Hoffman, 1989). The Archean rocks in southwest Greenland consist of gneissic, granitic, and supracrustal complexes with ages ranging between 3800 to 2750 Ma (Bridgewater et al., 1976; Brown et al., 1981; McGregor, 1973, 1979; McGregor et al., 1986). Highly deformed, thin mylonitic zones (sutures) occur at the boundaries of these complexes, suggesting that four Archean subterranes assembled at about 2750 to 2650 Ma (Nutman et al., 1989). After this period, the Nain province remained tectonically stable until the opening of the Labrador Sea starting in the Late Cretaceous.

Studies of oceanic magnetic anomalies and the delineation of fracture zones have determined the tectonic evolution of the Labrador basin (Srivastava, 1983; Srivastava and Tapscott, 1986), starting with rifting in the southern Labrador Sea about 90-85 Ma. Rifting progressed northward into the northern Labrador Sea and ceased around 36 Ma in the Davis Strait and Baffin Bay area (Srivastava, 1983; Srivastava and Tapscott, 1986). No evidence exists for volcanism associated with the Labrador Sea rifting event along the southwest Greenland coast; however, in the Davis Strait and Baffin Bay region, basaltic volcanics erupted between 61 to 58 Ma (Soper et al., 1982). Hyndman (1973, 1975) suggests that, after the initial formation of oceanic crust, an active hot spot formed around 60 Ma under the Davis Strait. The hot spot caused the extensive volcanism and created the overthickened oceanic crust beneath the Davis Strait (Keen and Barrett, 1972; Hyndman, 1975). White and McKenzie (1989) argue that the extensive volcanism and crustal thickening resulted from the northward progressing rift that intersected an asthenospheric plume. In their model, decompression melting of the mantle plume as the overlying lithosphere stretched and thinned caused the extensive volcanism.

Hinz and others (1979) conducted an extensive geophysical survey across the Labrador Sea from Labrador to southwest Greenland. Based on combined analysis of seismic, gravity, and magnetic data, they observed a steeply westward dipping basement with a thin sediment veneer along the southwest Greenland coast. The sediment cover thickens to the north into the Davis Strait. Gohl and Smithson (1993) analyzed the vertical component of the wide-angle seismic reflection data discussed in this paper and found a five to eight km thick high-velocity wedge at the base of the continental crust along the southwest Greenland coast. The high-velocity wedge thins inland toward the Archean continental crust. Gohl and Smithson (1993) postulate that the wedge was emplaced during the Labrador Sea rifting event and possibly represents underplated igneous rocks. The crust/mantle boundary along the coast deepens to the north. Detailed gravity modeling by Speece (1992) supports a steep, eastward dipping transitional oceanic-continental crust/mantle boundary and a northward crustal thickening along the continental margin.


The experimental design permitted the acquisition of densely spaced, spatially unaliased wide-angle seismic reflection data. For the entire seismic experiment, we deployed fifteen three-component PASSCAL REFTEK 72A digital seismic recording instruments at 35 locations along the coast of southwest Greenland and inland along the Ameralik and Godthaab fjords (Figure 1). The PASSCAL REFTEK instruments recorded over 10,000 marine air-gun array shots ranging in offset from 2.5 to 400 km. The source array consisted of five air-guns (19.7 l each) fired at 15 m depth with a spacing of about 100 m between shots. The large number of shots allowed for reliable correlation of the arrivals, especially for phases with low signal-to-noise ratios. The shear waves are probably generated at the water-basement interface by conversion from P-waves; S-wave energy is strong. The experimental design provides an excellent data set to study crustal seismic anisotropy.


Processing of the seismic data consisted of several steps to enhance the coherency of the arrivals. The horizontal components of the receiver gathers were rotated to obtain the radial and transverse component sections (Figure 2). At offsets greater than about 40 km, the energy recorded on the radial and transverse components arrives from azimuths approximately parallel to the coast. To enhance phase correlation, static corrections were made to these data to remove undulations in the arrivals caused by sea-floor topography. Seismic deconvolution was applied to decrease source-generated noise such as water reverberations and air-bubble noise to temporally sharpen the phases. Band-pass frequency filtering between 4 and 13 Hz removed low frequency water wave energy and high frequency ambient noise. These processing steps removed unwanted energy from the data and better aligned the arrivals, thus making phase correlation easier and more reliable.

After filtering, the data were then transformed to the tau-p domain via the Radon transform for further analysis (Figure 3). Tau is the zero-offset intercept time of the seismic phase and p, the ray parameter, is the reciprocal of the phase's apparent velocity. The slant stacks show strong P arrivals with ray parameters ranging from 0.165 s/km in the upper crust to 0.090 s/km in the lower most crust. The most prominent P phases are the upper crustal arrival (Pg), a wide-angle reflection from the lower crust (PlP), and the wide-angle Moho event (PmP). The small ray parameter values are from wide-angle reflections for deeper events and show an apparent velocity higher than the true lower crustal velocity. No arrivals from the upper mantle appear in the slant stacks. The S-wave arrivals on the slant stacks are not as strong or as prominent as the P arrivals. However, the upper crustal shear arrival (Sg) is evident, as well as a strong arrival from the lower crust (SlS). Analysis of the seismic phases in the tau-p domain increases the signal-to-noise ratio of the seismic phases by suppressing random noise (Russell and Hampson, 1990). Thus, phases that are difficult to recognize such as shear phases are better identified in tau-p sections.


Data from offsets between 17 to 80 km recorded a large-amplitude, lower crustal shear event. We interpret this phase as a wide-angle reflection from the top of a lowermost crustal, high-velocity layer (SlS) of Gohl and Smithson (1993). A shear reflection from the Moho (SmS) is not recorded in the receiver gather probably because of inadequate offset and recording time and the Moho may not be a strong shear-wave reflector. Mid-crustal shear events are difficult to see in the receiver gather because of their weaker amplitudes and coda from the earlier arriving phases.

The SlS phase has a different intercept time and slowness on the radial component than on the transverse component. Figure 3 shows the SlS phase as a smeared point on the slant stacks of the radial and transverse components. Wide-angle reflections would usually be ellipses on these plots. However, figure 3 shows only the strongest amplitude events to emphasize the coherent arrivals so the weaker elliptical branch of the SlS phase is not shown. The radial component records a p value of 0.2025 +/-0.003 s/km and a tau of 10.1+/-0.1 s. The transverse component has a p value of 0.2200+/-0.003 s/km and a tau of 9.55+/-0.1. These p values correspond to apparent velocities of 4.9+/-0.1 km/s for the radial component and 4.5+/-0.1 km/s for the transverse component. The error bounds on the p values are determined from the scatter about the center of the arrival from the slant stacks. The measured ray parameter values represent the apparent velocities of the arrivals, higher than the true velocities, because we are interpreting wide-angle reflections and because of structure on the reflecting interface. The significant difference in intercept time and apparent velocity between the two horizontal components indicates that the crust beneath the coast of southwest Greenland is seismically anisotropic.

The seismically anisotropic zone is in the lower crust. A one-dimensional velocity-depth model for the receiver gather was obtained from extremal inversion of the slant stacks (Bessonova, 1974; Hawman and Phinney, 1992). Figure 4 shows the results of the inversion and compares the shear wave velocity-depth functions to the two-dimensional velocity-depth function from the P-wave interpretation of Gohl and Smithson (1993).

Even though the crust beneath southwest Greenland is not one-dimensional in structure, we can compare the results from the inversion of the radial and transverse components. Along the coast, the crust thickens to the northwest from 33 km beneath Grad receiver (GR) to about 40 km north of Uman station (UM), a distance of 250 km. The P-wave velocities in the lower crust remain the same along the entire profile. The seismically anisotropic layer in shear velocity has a P-wave velocity of 6.9 km/s and increases in thickness from about 5 km to about 13 km thickness to the north (Gohl, 1991, Gohl et al., 1991; Gohl and Smithson, 1993). We can not determine if the P velocities are anisotropic from the data. Azimuthal differences in Vp may be due to structural variations in the crust and not to seismic anisotropy of the crust. However, because our shear-wave velocity results are obtained from a single receiver gather, any travel time distortions due to the two- or three-dimensionality of the crust are common to each component. Thus, the comparison of tau-p data from the radial and transverse components from the Kangak station is still valid and the conclusions are consistent between the horizontal components.

On the time-offset plots, the shear-wave splitting is difficult to observe (Figure 2). Figures 2c and 2d show the receiver gathers at a smaller scale. On these expanded-scale plots, a 0.2 second delay of the lower crustal phase between the radial and transverse components is more apparent. The slant stacks of the horizontal components (Figure 3) more clearly show the difference in this phase between the radial and transverse components compared to the time-offset plots (Figure 2). A particle motion diagram (Figure 5) showing the normalized amplitudes of the radial and transverse components clearly demonstrates that the SlS phase changes with time from arriving predominantly on the radial component (up-down) to arriving predominantly on the transverse component (left-right). The shear-wave anisotropic nature of the crust beneath the coast of southwest Greenland is interpreted on the basis of the shear wave splitting of the SlS phase as evidenced by the tau-p transformed Kangak record sections and the particle motion diagram.

From extremal inversion, shear-wave velocities are determined to about 30 km depth. The one-dimensional velocity-depth function for the radial and transverse components overlap to about 22 km depth. Between 22 to 24 km, the radial component has a higher Vs at a given depth than the transverse component. The velocity discrepancy is largest between 23 to 29 km depth, the zone of greatest seismic anisotropy. There are no arrivals in the slant stacks to determine the velocity below about 30 km. The zone of seismic anisotropy is above the 7.4 km/s Vp layer from Gohl and Smithson (1993). From the P-wave interpretation, the Moho in this area is about 37 km deep (Gohl and Smithson, 1993). The difference in shear-wave velocities indicates that the lower crust beneath southwest Greenland is seismically anisotropic.

The percent anisotropy in shear velocity calculated for the anisotropic zone ranges from about 5 to 15 percent (percent anisotropy =(Vfast - Vslow)/ Vmean) (Figure 6). The range of percent anisotropy is bounded from the extremal inversion; the actual percent of seismic anisotropy is likely about 6 to 8 percent based on the 0.2 second time difference of the SlS phase between the radial and transverse components. Figure 6 shows seismic anisotropy of a few percent to depths of about 20 km. This small percent of anisotropy is probably not significant. The bounds on the velocity-depth functions from the radial and transverse component have about the same velocities at similar depths and show a similar trend of gradually increasing velocity with depth. In the lower crust, the velocity-depth functions for the different components have very different velocities at the same depth. Between 22 to 24 km depth, the velocity-depth functions do not overlap. Below 24 km, the resolution of the transverse component is much less than the radial component because fewer data points are used in the inversion, but the transverse component generally shows a slower velocity for this depth than the radial component. Until the velocity-depth functions for the horizontal components differ greatly below 22 km, the seismic anisotropy represents limitations of the extremal inversion and not seismic anisotropy due to velocity differences in the crust.

This large amount of seismic anisotropy can constrain the composition and structure of deep layers, especially in conjunction with the P-wave velocity information. From the slant stack data and the extremal inversion, the seismic anisotropy occurs in the lower crust just above the high-velocity, lowermost crustal layer. Mineral alignment (Christensen, 1965; Babuska, 1981; Christensen and Szymanski, 1988) and fluid-filled microcracks (Nur and Simmons, 1969; Crampin, 1984) are the most likely causes for seismic anisotropy in the crust. Fluids in the lower crust can maintain open microcracks by reducing the effective confining pressure at those depths (Walder and Nur, 1984). However, fluids may react rapidly with surrounding rock and be consumned. Also, fluid-free microcracks close at pressures greater than 4 kbar (13 km) (Christensen, 1965) and a source of fluid to keep the microcracks open is unlikely at these depths and metamorphic conditions. Nur and Walder (1990) determined that fluid filled cracks can remain open in the deep crust for no more than about 107 years without fluid replenishment. Also, fluid-filled microcracks in the lower crust would reduce the P-wave velocity (Spencer and Nur, 1976). The 7.4 km/s Vp is higher than usually found in the lower crust (Holbrook et al., 1992), so the lower crust probably is free of open microcracks. Fluid-filled microcracks probably are not the cause of seismic anisotropy in the lower crust.

The more likely cause for seismic anisotropy in the lower crust is alignment of anisotropic minerals. Common anisotropic minerals for lower crustal rocks are plagioclase, biotite, amphiboles, and pyroxenes (Babuska, 1981; Ji and Mainprice, 1988; Mainprice and Nicolas, 1989; Siegesmund and Kruhl, 1991; Siegesmund and Vollbrecht, 1991). From laboratory studies on amphibolites from the Ivrea zone, plagioclase shows weak mineral alignment (Siegesmund and Kruhl, 1991; Siegesmund and Vollbrecht, 1991) and is likely to contribute less to the seismic anisotropy than amphiboles or biotite (Siegesmund and Kruhl, 1991). Biotite is highly anisotropic (Babuska, 1981), but is a small constituent in amphibolites and most granulites (Christensen, 1965; Manghnani et al, 1974). Velocity measurements on a granulite containing ten percent biotite had a shear-wave anisotropy of only three percent (Manghnani et al, 1974). Hornblende has the highest shear-wave anisotropy in the foliation plane with the maximum shear velocity parallel to the developed lineation (Siegesmund and Vollbrecht, 1991; Ji and Salisbury, 1993). Measurements on pyroxenes show that shear-wave anisotropy is greatest oblique to the lineation and the amount of anisotropy is about half that of hornblende in rocks with the same fabric (Siegesmund and Vollbrecht, 1991). The presence of seismic anisotropy narrows the possible composition to those rocks that are strongly anisotropic and suggests a foliated and possibly layered metamorphic rock assemblage.

The lower crust along the southwest coast of Greenland has a Vp of 6.9 km/s (Gohl and Smithson, 1993), which provides further constraints on the composition. Laboratory velocity measurements indicate that this P velocity corresponds to a range of rocks such as hornblende-pyroxene gneiss (Fountain and Christensen, 1989), stronalite (Burke and Fountain, 1991), mafic granulite (Holbrook et al., 1992) or anorthosite (Birch, 1960; Christensen, 1989; Holbrook et al., 1992). Anorthosites, common in exposed rocks in southwest Greenland, have P- and S-wave seismic anisotropies of around 4 percent (Table 1) (Manghnani et al., 1974; Christensen, 1989b). Many granulite facies rocks have a Vp of about 6.9 km/s, but have shear-wave and P-wave seismic anisotropies of around 2 to 5 percent (Manghnani et al., 1974; Christensen, 1989b) because granulite facies rocks are characterized by minerals with low anisotropy. The high shear-wave seismic anisotropy indicates that amphibolites are the most likely rock type that can cause the anisotropy, yet their Vp is too high, about 7.2 km/s (Christensen, 1965; Christensen, 1966). Other rocks, however, are often interlayered with amphibolites. Drill cores from the Inner Piedmont of South Carolina contain amphibolite with thin layers of gneiss that have a Vp of about 6.8 km/s (Christensen, 1989a). A wide range of rocks could be interlayered with amphibolite to produce a 6.9 km/s velocity. In southwest Greenland, tonalitic gneiss is common in surface outcrops. Tonalitic gneiss has a Vp of about 6.5 km/s (Fountain et al., 1990) and contains significant amounts of plagioclase. Amphibolite with fine layers of tonalitic gneiss could have the appropriate P-wave velocity of about 6.9 km/s and the alignment of hornblende and plagioclase would create a highly anisotropic rock. The high shear seismic anisotropy, combined with the 6.9 km/s Vp for the seismically anisotropic layer, strongly favors an interlayered amphibolite/gneiss composition at the seismically anisotropic depth (Birch, 1960; Babuska, 1981; Fountain and Christensen, 1989; Fountain et al., 1990; Siegesmund and Vollbrecht, 1991).

Layering in the seismically anisotropic zone is difficult to determine from wide-angle reflection data (Braile and Chiang, 1986); however, in our data, there is some indication that the seismically anisotropic zone may be layered. The waveform of the SlS phase is longer in duration than the waveform of the PlP phase. Shear-wave splitting may cause this longer duration waveform by energy leaking between components, but thin layering may account for some of the waveforms duration. Shear waves can resolve finer scale features than P-waves if the frequency content of the arrivals is the same. In the Greenland data, the shear-wave and P-wave arrivals from the lower crust have a peak frequency of around 8 Hz. The observed shear-waves have shorter wavelengths because of their slower velocity and similar frequency content, and thus can discern finer layering in the lower crust than the P-waves. The waveform of the shear phase may indicate that there is layering on a scale too fine for the P-wave data to resolve.

The alignment of minerals causing the shear wave seismic anisotropy probably is the result of lattice- and shape-preferred orientation of the anisotropic minerals (mainprice and Nicolas, 1989). During the opening of the Labrador Sea, magmatic underplating of the lower crust supplied heat to the overlying rocks (White and McKenzie, 1989; Gohl, 1991). The increased heat flow could have increased the plasticity of the pre-existing rocks, making them deform plastically. From borehole breakout measurements in the North Atlantic, the stress field is oriented such that the maximum horizontal compressive stress is approximately perpendicular to the southwest coast of Greenland (Zoback, 1992). Plastic deformation of the lower crust under the current stress field could develop lattice- and shape-preferred orientations within the lower crust (Tullis and Yund, 1987; Mainprice and Nicolas, 1989). Feldspar, deformed under compressive stress, developed a strong preferred orientation of elongate grains perpendicular to the applied stress (Tullis and Yund, 1987). To attain the high seismic anisotropy, the anisotropic minerals must have aligned with the maximum anisotropy parallel to the coast. The seismic anisotropy increases between 22 and 26 km depth, indicating that the minerals are more strongly aligned closer to the Moho. The 7.4 km/s lowermost crustal wedge is probably seismically anisotropic, although no shear wave energy was recorded from this layer. This high-velocity wedge is attributed to mafic magmatic underplating during Late Cretaceous rifting (Gohl and Smithson, 1993). Viscous flow in the magma or solid state flow probably aligned the minerals to develop shape-preferred orientation in this wedge (Mainprice and Nicolas, 1989). Alternatively, the seismic anisotropy may have developed during the formation of crust in the Archean; however, the Archean block was formed from the juxtaposition of several subterranes (Nutman et al., 1989). These terranes formed separately from each other and the rocks in the subterranes have different strain histories (Nutman et al, 1989). An early, east-west oriented folding phase is observed in the Faeringehaven and Tre Brodre terranes, but not in the Tasiusarsuaq and Akia terranes. Later, northeast-southwest oriented folding took place. These later folds are most pronounced near the boundary between the other terranes and the Akia terrane and the intensity of the folding lessens away from this boundary (Nutman et al., 1989). The crust formed by the amalgamation of these blocks would not necessarily have a common mineral alignment. Without the plastic deformation caused by the increase in temperature from the magmatic underplating, mineral alignment would not be readily reset and seismic anisotropy related to rifting strain would not develop in the lower crust.


Slant stacking and extremal inversion of the Kangak fjord receiver gather indicate that the lower crust beneath the southwest coast of Greenland is seismically anisotropic in shear velocity by 6 to 8 percent. The presence of shear-wave splitting in the lower crust constrains the crustal composition to interlayered amphibolite/gneiss. The high shear wave seismic anisotropy and the 6.9 km/s P-wave velocity of the anisotropic zone indicates an interlayered amphibolite/tonalitic gneiss composition for the lower crust.

The presence of seismic anisotropy in the lower crust also constrains the interpretation of the tectonic history of the southwest coast of Greenland. The development of the seismic anisotropy probably occurred during the rifting of the Labrador sea. Possibly, magmatic underplating during the late Cretaceous sufficiently heated the pre-existing lower crust to allow mineral re-alignment during continental rifting. Since the anisotropic zone is believed to have developed in the lower crust beneath southwest Greenland in the Late Cretaceous and Cenozoic, the lower crust, at least here, has evolved through geologic history and bears a rifting imprint.

The recognition of seismic anisotropy in the crust beneath Greenland can lessen the amount of seismic anisotropy attributed to the upper mantle. Savage and others (1990) found that the upper mantle beneath the Basin and Range province has a seismic anisotropy of about four percent for vertical shear wave propagation. They based their interpretation on a 0.9 second time difference between the horizontal components. From our analysis, showing a 0.2 second time difference between the horizontal components for energy propagating through 6 km of the lower crust, seismic anisotropy in the crust could account for part of this time difference, thus reducing the amount of seismic anisotropy attributable to the upper mantle.

The seismic anisotropy extends from about 23 to 29 km depth. The depth range is limited and, importantly, is constrained to the lower crust. This limited depth range of the seismic anisotropy demonstrates the high resolution of crustal velocity structure that is possible from densely sampled seismic surveys and opens new vistas for lithospheric imaging. Additionally, the combined interpretation of P- and S-wave data provide stronger constraints on crustal composition and structure than is possible from either data set alone.

Seismic anisotropy in the crust strengthens interpretations about the nature of the crust. The direction of the fast and slow seismic velocities can determine the fabric orientations of the rocks in the crust. From this information, we can infer the stress orientation at the time of the development of the seismic anisotropy. The amount of anisotropy also limits the rock types present in the anisotropic zone. A high amount of seismic anisotropy constrains composition to rocks with highly anisotropic minerals such as amphibole and biotite. Seismic anisotropy, along with seismic velocity, can indicate the presence of fluids in the crust. Slow seismic velocities and seismic anisotropy may indicate that similarly oriented, fluid-filled cracks are present in the crust. The finding of seismic anisotropy in the crust reduces the range of possible interpretations for the geologic character of the crust.


This work was supported by NSF grants DPP 8821974 and DPP 9023847. We also thank R. Clowes, R.P. Meyers, N.I. Christensen, and an anonymous reviewer for their helpful comments.


Babuska, V., 1981. Anisotropy of Vp and Vs in rock-forming minerals. J. Geophys., 50: 1-6.

Bessonova, E.N., Fishman, V.M., Ryaboyi, V.Z. and Sitnikova, G.A., 1974. The tau method for inversion of travel times-I. Deep seismic sounding data. Geophys. J. Int., 36: 377-398.

Birch, F., 1960. The velocity of compressional waves in rocks to 10 kbar, part 1. J. Geophys. Res., 65: 1083-1102.

Braile, L.W. and Chiang, C.S., 1986. The continental Mohorovicic Discontinuity: Results from near-vertical and wide-angle seismic reflection studies. In: M. Barazangi, L. Brown (Editors), Reflection Seismology: A Global Perspective. Geodyn. Ser., Am. Geophys. Union, 13: 257-272.

Bridgewater, D., Keto, L., McGregor, V.R. and Myers, J.S., 1976. Archean gneiss complex of Greenland. In: A. Escher and W.S. Watt (Editors), Geology of Greenland, Groen. Geol. Unders., 18-74.

Brown, M., Friend, C.R.L., McGregor, V.R. and Perkins, W.T., 1981. The late Archean Qorqut granite complex of southern West Greenland. J. Geophys. Res., 86: 10617-10632.

Burke, M.M. and Fountain, D.M., 1990. Seismic properties of rocks from an exposure of extended continental crust - new laboratory measurements from the Ivrea zone. Tectonophysics, 182: 119-146.

Christensen, N.I., 1966. Shear wave velocities in metamorphic rocks at pressures to 10 kilobars. J. Geophys. Res., 71: 3549-3556.

Christensen, N.I., 1989a. Reflectivity and seismic properties of the deep continental crust. J. Geophys. Res., 94: 17793-17804.

Christensen, N.I., 1989b. Seismic velocities. In: R.S. Carmichael (Editor) Physical Properties of Rocks and Minerals, 431-546.

Christensen, N.I. and Szymanski, D.L., 1988. Origin of reflections from the Brevard fault zone. J. Geophys. Res., 93, 1087-1102, 1988.

Crampin, S., 1984. Effective elastic constants for wave propagation through cracked solids. Geophys. J. Roy. Astron. Soc., 76: 135-145.

Fountain, D.M. and Christensen, N.I., 1989. Composition of the continental crust and upper mantle; a review. In: L.C. Pakiser and W.D. Mooney (Editors), Geophysical Framework of the Continental United States, Geol. Soc. Am. Mem. 172: 711-742.

Fountain, D.M., Salisbury, M.H. and Percival, J., 1990. Seismic structure of the continental crust based on rock velocity measurements from the Kapuskasing Uplift. J. Geophys. Res., 95: 1167-1186.

Gohl, K., 1991. Seismic Wide-Angle Studies of Early Archean and Proterozoic Crust in Greenland, Minnesota, and Wyoming. Ph.D. thesis, University of Wyoming, Laramie, WY, 189 pp.

Gohl, K. and Smithson, S.B., 1993. Structure of the Archean crust and passive margin of southwest Greenland from seismic wide-angle data. J. Geophys. Res., 98: 6623-6638.

Gohl, K., Hawman, R.B., Smithson, S.B. and Kristoffersen, Y., 1991. The structure of the Archean crust is sw Greenland from seismic wide-angle data: a preliminary analysis. In: R. Meissner, L. Brown, H.-J. Durbaum, W. Franke, K. Fuchs and F. Seifert (Editors), Continental Lithosphere: Deep Crustal Reflections. Geodyn. ser. Am. Geophys. Union, 22: 53-57.

Hawman, R.B. and Phinney, R.A., 1992. Structure of the crust and upper mantle beneath the Great Valley and Allegheny Plateau of eastern Pennsylvania 1. Comparison of linear inversion methods for sparse wide-angle reflection data. J. Geophys. Res., 97: 371-391.

Hinz, K., Schlueter, H.-U., Grant, A.C., Srivastava, P.S., Umpleby, D. and Woodside, J. 1979. Geophysical transects of the Labrador Sea: Labrador to southwest Greenland. Tectonophysics, 59: 151-183.

Hoffman, P., 1989. Precambrian geology and tectonic history of North America, In: A.W. Bally and A.R. Palmer (Editors), The Geology of North America-An overview. The Geology of North America, Geol. Soc. Am. A: 447-551.

Holbrook, W.S., Mooney, W.D., and Cristensen, N.I., 1992. The seismic velocity structure of the deep continnental crust, In: D.M. Fountain, R.J. Arculus and R.M. Kay (Editors). The Lower Continental Crust, Elsevier, 1-43.

Hyndman, R.D., 1973. Evolution of the Labrador Sea, Can. J. Earth Science, 10: 637-664.

Hyndman, R.D., 1975. Marginal basins of the Labrador Sea and the Davis Strait hot spot. Can. J. Earth Science, 12: 1041-1045.

Ji, S. and Mainprice, D., 1988. Natural deformation fabrics of plagioclase: Implication for slip systems and seismic anisotropy. Tectonophysics, 147: 145-163.

Ji, S. and Salisbury, M.S., 1993. Shear-wave velocities, anisotropy and splitting in high-grade mylonites. Tectonophysics, 221: 453-473.

Keen, C.E. and Barrett, D.L., 1972. Seismic refraction studies in Baffin Bay: an example of a developing ocean basin. Geophys. J. Roy. Astron. Soc., 30: 253-271.

Kern, K. and Wenk, H.-R., 1990. Fabric-related velocity anisotropy and shear wave splitting in rocks from the Santa Rosa mylonite zone, California. J. Geophys. Res, 95: 11213-11223.

Mainprice, D. and Nicolas, A., 1989. Development of shape and lattice preferred orientations: application to the seismic anisotropy of the lower crust. J. Struct. Geol., 11: 175-189.

Manghnani, M.H., Ramananantoandro, R. and Clark, Jr., S.P., 1974. Compressional and shear wave velocities in granulite facies rocks and eclogites to 10 kbar. J. Geophys. Res., 79: 5427-5446.

McGregor, V.R., 1973. The early Precambrian gneisses of the Godthaab district, West Greenland. Phil. Trans. Roy. Soc. London, Ser. A, 273: 343-358.

McGregor, V.R., 1979. Archean gray gneisses and the origin of the continental crust: Evidence from the Godthaab region, West Greenland. In: F. Barker (Editor), Trondhjemites, Dacites and Related Rocks, 169-205,.

McGregor, V.R., Nutman, A.P. and Friend, C.R.L., 1986. The Archean geology of the Godthaabsfjord region, southern West Greenland. Lun. Planet. Inst. Tech. Rep. 86-04, 113-169.

Nur, A. and Simmons, G., 1969. Stress-induced velocity anisotropy in rock: an experimental study. J. Geophys. Res. 74: 6667-6674.

Nur, A.M. and Walder, J., 1990. Time-dependent hydraulics of the Earths crust. In: Bredehoeft, J.D. and Norton, D.L. (Editors) The Role of Fluid in Crustal Processes, 113-127.

Nutman, A.P., Friend, C.R.L., Baadsgaard, H. and McGregor, V.R., 1989. Evolution and assembly of Archean gneiss terranes in the Godthaabfjord region, southern west Greenland: structural, metamorphic, and isotopic evidence. Tectonics, 8: 573-589.

Russell, B. and Hampson, D., 1990. Noise elimination and the Radon transform, The Leading Edge, 9: 18-23.

Savage, M.K., Silver, P.G. and Meyers, R.P., 1990. Observations of teleseismic shear-wave splitting in the Basin and Range from portable and permanent stations. Geophys. Res. Letts., 17: 21-24.

Siegesmund, S. and Kruhl, J.H., 1991. The effect of plagioclase textures on velocity anisotropy and shear wave splitting at deeper crustal levels. Tectonophysics, 191: 147-154.

Siegesmund, S. and Vollbrecht, A., 1991. Complete seismic properties obtained from microcrack fabrics and textures in an amphibolite from the Ivrea zone, Western Alps, Italy. Tectonophysics, 199: 13-24.

Soper, N.M, Dawes, P.R. and Higgins, A.K., 1982. Cretaceous-Tertiary magnetic and tectonic events in North Greenland and the history of adjacent ocean basins, Nares Strait and the drift of Greenland: A conflict in plate tectonics. Geoscience, 8: 205-220.

Speece, M.A., 1992. Geophysical Studies of Precambrian Regions: The Laramie Mountains, Wyoming and the Godthaabsfjord Area, South West Greenland. Ph.D. thesis, University of Wyoming, Laramie, WY, 110 pp.

Spencer, J. W. Jr. and Nur, A. M., 1976. The effects of pressure, temperature and pore water on velocities in Westerly granite. J. Geophys. Res. 81: 899-904.

Srivastava, P.S., 1983. Davis Strait: Structures, origin and evolution, In M.H.P. Bott, S. Saxov, M. Talwani and J. Thiede, (Editors), Structure and Development of the Greenland-Scotland Ridge, 159-189.

Srivastava, P.S. and Tapscott, C.R., 1986. Plate kinematics of the North Atlantic. In: P.R. Vogt and B.E. Tucholke (Editors), The Western North Atlantic Region, The Geology of North America, M, Geol. Soc. Am. M: 379-404.

Tullis, J. and Yund, R.A., 1987. Transition from cataclastic flow to dislocation creep of feldspar: Mechanisms and microstructures, Geology, 15: 606-609.

Walder, J. and Nur, A., 1984. Porosity reduction and crustal pore pressure development, J. Geophys. Res., 89: 11539-11548.

White, R. and McKenzie, D., 1989. Magmatism at rift zones: The generation of volcanic continental margins and flood basalts. J. Geophys. Res., 94: 7685-7729.

Zoback, M.L., 1992. First- and second-order patterns of stress in the lithosphere: The world stress map project. J. Geophys. Res., 97: 11,703-11,728.

Figure Captions

Figure 1) Map of southwestern Greenland showing the geometry of the wide-angle seismic experiment. Stations occupied during the wide-angle experiment are indicated by the dots in the figure. The ship's path is indicated by the solid line. All data shown in this paper are from the station labeled KA. The major subterranes are labeled in the figure legend.

Figure 2) Time-offset record sections from the Kangak receiver gather. Figure 2a shows the radial component and figure 2b displays the transverse component. The prominent phases are labeled in each figure. Figures 2c and 2d are enlarged sections to show the 0.2 sec travel time delay of the SlS phases between the radial and transverse components. The arrows in the data indicate the SlS phase. The arrow along the top points to the trace used in the particle motion diagram of figure 5.

Figure 3) Semblance filtered tau-p sections highlighting the shear phases. Numbers along the horizontal axis are the ray parameter (p) in 10-3 s/km. Tau (vertical axis) is the zero-offset intercept time in seconds. The weaker amplitude events on the section are removed to highlight the major crustal phases. Figure 3a is from the radial component and figure 3b is the transverse component. The heavy solid line in figure 3b below and to the left of the SlS phase is the SlS phase from the radial component.

Figure 4) The velocity-depth bounds for the crust a long the southwest coast of Greenland. The velocities from the radial (solid lines) and the transverse (dashed lines) are calculated from the one-dimensional extremal inversion of the tau-p transformed Kangak receiver gather. The bounds indicate the range of the velocity-depth function resolvable from the tau-p inversion. The P-wave (dotted line) velocity-depth function is from the two-dimensional inversion of Gohl and Smithson (1993).

Figure 5) A particle motion diagram of the radial and transverse component data from a trace about 70 km from the source. The diagram starts at 24.2 seconds and ends at 24.6 seconds, encompassing the SlS phase in figures 2c and 2d. The axes are the normalized amplitudes of the radial and transverse components. The seismic trace used to determine the particle motion was processed the same as the data in figure 2.

Figure 6) A plot of the percent anisotropy with depth from the extremal inversion of figure 4. The squares and circles are the percent anisotropy from the minimum and maximum velocities. The largest seismic anisotropy is between 23 to 29 km depth.