Shear-wave splitting beneath the southern Western Canada Sedimentary Basin: A snapshot of the interaction betweencratons and terrane?
Yu Jeffrey Gu1, Kenny Kocon1, Ahmet Okeler1, William Menke2
- Department of Physics, University of Alberta, Edmonton, AB, T6G2G7, Canada.
- Lamont-Doherty Earth Observatory, Columbia University, 61 Route 9W, Palisades, NY10964, USA.
Abstract
Western Canadian Sedimentary Basin(WCSB) marks the transition from the old North American continental lithosphere (east) to young accreted ‘terranes’ (west). Geologic, seismic and magnetic data in this region have suggested various crustal domains and strong conductive anomalies and seismic velocity gradients in the mantle. Extensive deformation is further evidenced by seismic anisotropy in the southern WCSB based onteleseismic earthquakedata recorded by the Canadian Rockies and Alberta Network (CRANE). Our shear-wave splitting measurements show 1-1.5 sec splitting times and a fast propagation direction along the northeast-southwest orientation near the Rocky Mountain foothills, approximately parallel to the absolute plate motion of the North American continent. This single-layer anisotropic pattern suggests the alignment of the crystalographic axis of olivine due to shear deformation at the base of the lithosphere. On the other hand, the spatial distribution of the SKS orientations is significantly more complex east of the Rocky Mountain. Several stations display two-layer anisotropic variations and hightthe contrasting mantle structures and histories between the Rockies and its adjacent domains. Disrupted mantle flow near the edge of the migrating continental root east of the provincemay be largely responsible for the complex shear-wave splitting fast directions in this region.
1. Introduction
The Western Canada Sedimentary Basin (for brevity, WCSB) is a relatively thin, northeastward-trending wedge of supracrustal rocks tapering on, or juxtaposed with, Precambrian crystalline rocks (Bally et al., 1966; Price, 1981; Baumont, 1981; Hoffman, 1988; Mossop, G.D. and Shetsen, 1994). This elongated geological structurebegan its formation during thetectonic development of western Laurentia, and continued its evolutionthrough recent interactions between the North American craton and Cordilleran orogen (Hoffman, 1989; Ross et al., 1991, 2002; Clowes et al., 2002). Today, this diverse geological framework consists of Archaencraton(s), Proterozoicorogens and associated accretionary margins (also known as ‘terranes’) (e.g., Price 1981; Hoffman, 1998; Ross et al., 1991, 2002).
An astute example of geological diversity is the region surrounding the Alberta Basin, located in thesouthern half of WCSB. Beneath the sedimentary cover arejuxtaposed tectonic domains (Figure 2) (Hoffman, 1988; Ross et al., 1991; Clowes et al., 2002; Shragge et al., 2002) that likely have undergone substantial thermal and tectonic overprinting (Ross and Eaton, 2002; Aubachet al. 2004; Mahan and Williams, 2005; Beaumountet al., 2010). This basin is bounded in the east by the Trans-Hudson Orogenic Belt, a controversial geological structure that extends well into eastern Canada(Hoffman, 1988; Banks et al., 1998; Zelt and Ellis, 1999). Directly west of this basin is the northern Rockies, a section of the western Cordillera presumably originated from the Laramideorogeny (Livaccariet al., 1981; Bird, 1998; Maxson and Tickoff, 1996; Cook et al., 2002; English and Johnson, 2010; Liu et al., 2010). Strong seismic velocity gradients (van der Lee and Frederiksen, 2005; Nettles and Dziewonski, 2008; Mercier et al., 2010) and anisotropy (e.g., Shragge et al. 2002; Marone and Romanowicz, 2007; Courtier et al., 2010; Yuan and Romanowicz, 2010) have been proposed for the mantle beneath this region, which accentuates the sharp transition from thestable continental mantle east of the Alberta Basin to the accreted terranes west of it.
Faulting and episodes ofmagmatism, accretion and subduction (e.g., Hoffman, 1988; Ross et al., 2000) can inflictpermanent deformation at both crustal and mantle depths beneath the southern WCSB. In particular, olivine’s crystollographic fast axis is known to preferentially align with the direction of maximum shear or least compression (e.g., Vinnik ??; Anderson, 1989; Silver, 1993). The strength of radial and/or azimuthal anisotropyduo to lattice preferred orientation (LPO)is therefore a direct reflection of ‘order’ in mantle rocks (Anderson, 1989; Silver, 1993; Long and Silver, 2009). In southern WCSB, strong azimuthal anisotropy has been inferred from 1+ sec SKS wave splitting times from temporary arrays (Shragge et al., 2002; Courtier et al., 2010), suggesting extensive deformation in the region. However, the orientations of fast SKS directionsare questionable, as significant complexities have been previously documented from the regional permanent stations (Kendall, ??; Currie and Hyndman, 2006). Undesirable resolution caused by the restrictive linear or semi-linear array geometries, which limited the effectiveness of existing shear-wave splitting measurements (e.g., Shragge et al., 2002; Courtier et al., 2010) and anisotropic tomography (Marone and Romanowicz, 2007; Nettles and Dziewonski, 2008; Yuan and Romanowicz, 2010), remains the greatest challenge in determining the seismic anisotropy beneath the southern WCSB.
Since early 2006, the regional seismic data coverage in the southern WCSB improved significantly from the establishment of the Canadian Rockies and Alberta Network (nicknamed CRANE), the first semi-uniform broadband seismic array in Alberta and parts of Saskatchewan, Canada (see Figure 1). Continuous seismic signals from this array enabled a detailed examination of both regional seismicity and crust/mantle structures. This study uses SKS splitting measurements to constrain the azimuthal anisotropy beneath CRANE and nearby permanent (EDM, WALA) broadband seismic stations (see Figure 1). With vastly improved data coverage and measurement accuracies (based on four different techniques, see Section 2), we provide an updated model of past and ongoing mantle processes near the western boundaryof the North American craton.
2. Data and Method
The study analyzes earthquake records from eleven CRANE stations and two permanent stations monitored by the Canadian National Seismic Network (CNSN). These thirteen stations aresemi-uniformly distributed in central and southern Alberta withan average spacing of ~150 km between adjacent ones, (see Figure 1). Most of the stations operated continuously for 2+ years, accumulating enough large earthquakes for our examination of SKS splitting (Figure 3), a phenomenon often interpreted as the consequence of anisotropy similar to optical birefringence of minerals under polarized light (Bowman and Ando, 1987; Silver and Chan, 1991; Menke and Levin, 2003). The resulting splitting parameters, which consist of delay time (between the fast and slow shear waves) and fast polarization azimuth, are sensitive functions of the strength and direction of receiver-side anisotropy (see Long and Silver, 2009).
We restrict our data set to Mw>6.5 earthquakes with source-receiver distances between 85 and 115 arc deg. The resulting three-component (East-west, North-south, Vertical) time series are band-pass filtered with corner frequencies of 1secand 15 sec, and then subjected to a signal-noise ratio (SNR) test; records with SNR<?? are automatically rejected (or Visual inspection?? KENNY?). The average number of earthquakes that survive the above selection process is ~12 per station (Table 1; Figure 4a). We retain HYLO,the most recently installed station, due to its unique position within the array and the robustness of the single SKS waveform. The overall distribution of the source-receiver pairs is highly non-uniform, displaying a strong northwest–southeast orientation (Figure 4b).
In this analysis each measurement is made primarily based on the cross-convolution method for multiple earthquakes (Menke and Levin, 2003). We convolve the observed radial and tangential component seismograms with the impulse responses predicted by an isotropic background model, and then introduce anisotropic perturbations to one or multiple layer(s) to minimize the misfit between the observed and predicted cross-convolution functions(see Menke and Levin, 2003). This multi-layer approach considers all earthquakes arriving at a single station, which is more flexible than methods based on a single anisotropic layer assumption (for instance, the rotation correction (Bowman and Ando, 1987), minimum energy (Silver and Chan, 1991) or eigenvalue (Silver and Chan, 1991) methods (see Long and Silver, 2009 for a review)). For each station we distinguish one- or two-layer anisotropy based on the average cross-correlation coefficient between observed and predicted convolution functions for all earthquakes.
3. Shear-Wave Splitting Measurements
3.1 Data Fit and Uncertainty
Of the 15 stationsanalyzed in this study, clear evidence of shear wave splitting is observed under 14 stations. With the exception of WALA, where reasonable splitting parameters could not be determined, most of the stations exhibit significant anisotropy that requires minimization of the differences between the two ‘corrected’ horizontal components (see Menke and Levin, 2003). The inverted splitting parameters improve the overall correlation and the linearity of particle motions for SKS phases after correcting for anisotropy (see Figures 5a and 5b). The best results are obtained for stations CLA, HON, JOF, LYA and PER (see Figures 5a and 5b), while highly linear particle motions prior to the inversions (e.g., CZA, EDM, NOR, REC) are generally retained by the cross-convolution analysis. StationsFMC and DOR remain problematic: in particular, the visibly nonlinear particle motions and disagreements among four different shear-wave splitting methods (see latter part of this section) at DOR indicate complexities in the anisotropic mantle. Furthermore, similar data fits are obtained for EDM and JOF for 1- vs. 2-layer anisotropic models.
The delay times and azimuth uncertainties are determined by a bootstrap re-sampling algorithm (Efron and Tibshirani, 1991). At each station, we randomly select the same number of earthquakes from the list of observations and follow the same procedure detailed in Section 2 to obtain a bootstrapped shear-wave splitting measurement. This sampling process is repeated for 300 times and the standard deviations of distribution of the bootstrapped measurements can be used as effective indicators of the timing and split angle uncertainties at each station (see Table 1). Most of the stations exhibit Gaussian distributions centered near the meanvalues (Figure 6). The splitting times range from 1.1 to1.9 sec, with uncertainties of 0.06 - 0.50 sec (see Table 1). The distributions of the fast splitting directions are generally Gaussian, which suggests self-consistent measurements from individual earthquakes. Notable exceptions are DOR and FMC, each containing two distinct angles, and JOF that exhibits greater variations (hence uncertainty) of fast azimuth than other stations. Among these three stations, the large number of earthquakes (14) recorded by DOR argues against thepossibility of increased error due to data shortage. Uncertainties associated with splitting times (see Figure 6)closely track those of fast orientations: for instance, significant timing uncertainties exist beneath the same three stations under which substantial angular variations are identified.
To verify the stability of the measurements, the splitting parameters are independently determined using the rotation correction (Bowman and Ando, 1987), minimum energy (Silver and Chan, 1991) and eigenvalue (Silver and Chan, 1991) methods (see Long and Silver, 2009 for a review) for the majority of the stations in the array. For all three approaches, the median of the splitting parameters from individual events is used to account for multiple earthquakes recorded by a given station.
3.2Splitting Parameters and Lateral Variation
The splittimes and orientations vary systematically across the CRANE array (Figure 7). The majority of the measurements along the Rocky Mountain foothills (e.g., CLA, LYA, NOR, BRU) show consistently large (~1.5-1.9 sec) split times(between fast and slow SKS arrivals) withfast directions along a northwest-southeast orientation. These measurements correlate strongly with those of Courtier et al. (2010) a few degrees northwest of this region along the Cascadia Deformation Front (for short, CDF; Courtier et al., 2010; Figure 8). The amount of azimuthal anisotropy decreases from the northern part of the array (1.5+ sec) to ~1.1 sec US-Canada border, where the orientati with stations WALA that However, relatively small splitting time delays are observed beneath eastern-central Alberta where the fast axes. While .
The orientations are, to first order, consistent with the direction of the absolute plate motion (~deg southwest) of the North American continent. These value
Similar directions have been previously reported by Shragge et al. (2002) using a linear, approximately north-south trending temporary array in this region (see Figure 1).
Bottom Left: Alternatively, the complex SKS splitting directions (and
times) may also reflect more localized mantle upwelling. The heat flow
map shows enhanced activities near HLO [Blackwell & Richards, 2004]
and the stations around this geographical location appear to track the
geometry of the regional hotspot.
Local heat-flow anomalies (circled region) are supported
by the reduced values of Bouguer gravity. It is unclear whether the SKS
splitting observations reflect a larger-scale tectonic movement/history
(see Top Left) and/or more localized thermal variations (bottom figures).
3.2 Regional variations in fast splitting directions
4. Interpretation and Discussion
The splitting parameters in the vicinity of the Canadian Rockies suggest strong ‘order’ in mantle mineralogy (Long and Silver, 2009). The northeast-southwest trending fast direction is consistent with previously reported values utilizing Lithoprobe data (Shraggeet al., 2002) and the direction of maximum horizontal stress – a proxy for the ‘fossil’ strain field within the lithosphere in response to the past episodes of northwest-southeast plate convergence and subduction of Farallon and Kula plates (e.g., Helmstaedt and Schulze, 1989; Ross et al. 2000). The combination of pre-existing fabric within the mantle lithosphere and the present-day absolute plate-motion that is, coincidentally, northeast-southwest, offers an attractive explanation for the large SKS delays and 3-6% azimuthal anisotropy in this region. Without further data constraint it is difficult to resolve the full history of mantle deformation or the number of anisotropic layers (e.g., Shraggeet al. 2002; Marone and Romanowicz, 2007), however.
The origin of the complex shear-wave splitting pattern beneath eastern-central Alberta remains unclear. It could potentially be linked to the adjacent Buffalo Head Terrane, a region that has attracted national attention in recent years due to the discovery of precious minerals. The vicinity of the anisotropic anomaly exhibits enhanced heat-flow (Blackwell and Richards, 2004) and below-average seismic velocity (van der Lee and Frederiksen, 2005) and bouguer gravity values. The presence of a divot (Fouchet al., 2000) or an abandoned plume conduit (Bank et al., 1998) on the continental root offers a viable explanation. On a local scale, geometrical imperfection associated with past plate interactions could trap hot asthenospheric material and disrupt the mantle flow around it. Within a larger tectonic framework, the anomalous shear-wave splitting observations in eastern-central Alberta could signal a hidden tectonic boundary between stable continents (east/northeast) and accreted terranes (west). For instance, streamlined mantle flow around the edges of moving continental `keels' (e.g., Gaherty and Jordan, 1995; Ben Ismail and Mainprice, 1998; Bokelmann and Silver, 2002) can induce strong north-south oriented horizontal strain. In other words, shear deformation base of the lithosphere (~200 km) and disrupted flow at shallower depths could both be present, hence producing complex, multi-layered anisotropy in this region. Furthermore, due to the substantial topographical relief on the base of the lithosphere (Hyndman et al., 2005), both radial and azimuthal anisotropy would be expected in this transitional region.
5. Conclusions
Our SKS splitting analysis providesfirst-order evidence for strong mantle anisotropy beneath the southern WCSB. Large delay times and coherent SKS fast directions are observed along the Rocky Mountain foothills, which are roughly consistent with the direction of present-day plate motion but contrast with the highly variable splitting directions east of the northern Rockies. Our SKS measurements do not correlate strongly with local Bouguer gravity or heat-flow data, but the general pattern may be explained by recent models of shear velocity variation and suggest relatively sharp change of mantle deformation mechanism from simple, plate-motion induced shear in the southwestern WCSB to complex, disrupted asthenospheric flow east of it. Based on this interpretation, the spatial distribution of SKS splitting variations across the stations in the southern WCSB could mark the transition betweenold, seismically fast cratons east of the WCSBandthe significantly younger accreted terranes in western Canada. Overall, the broadband data from CRANE has undoubtedly offered a window of opportunity to probethe history, present state and geometry of the continental lithosphere. The continued data acquisition and development of this arrayin the near future may, ultimately, contribute to the discussion of the lithosphere evolution from thin convergent margins to thick, depleted continental roots.
3.3 Measurement uncertainties
The t
To validate the measurements further we compare
Above: Tabulated splitting measurements with the available data constraints.
5. Conclusions
Key Observations:
The SKS wave splitting pattern in the transition region between cratons (east) and the
accretedterranes (west) are more complex than previously documented.
The direction of fast-splitting axis coincide with present-day plate motion near the Can-
adian Rockies, but is nearly horizontal or northwest-southeast trending east of the Rockies.
Enhanced heat flow is observed at ~55N, which may be partially responsible for complex-
ities in SKS spitting measurements away from the Rockies.
Acknowledgement:
We thank our collaborator, Bill Menke, for technical help and suggestions (see reference below). The work presented
is supported by the Canadian Foundation for Innovation (CFI), Alberta Ingenuity and the University of Alberta. Some
of the data used in this analysis was distributed by IRIS and the Canadian National Seismographic Network (CNSN).
We are also grateful to the hosts of our seismic instruments in the province of Alberta.
Figure Caption:
Top Right: Event distribution for the SKS splitting analysis. All earthquakes have magnitudes (MB, MS, MW)