Anisotropy Variations in the Alaska Subduction Zone Based on Shear‐Wave Splitting From Intraslab Earthquakes

Abstract

The Cook Inlet region of south-central Alaska is located toward the eastern end of the 4,000 km-long Aleutian-Alaska subduction zone, where the Pacific plate subducts under North America. Plutonic rocks and basin-deposited strata indicate that subduction has persisted in this region for more than 200 Ma (Fisher & Magoon, 1978;Jarrard, 1986). The Cook Inlet segment of the subduction zone exhibits abundant seismicity down to approximately 200 km depth, as well as active volcanoes (Figure 1). Here, we use recordings from these intraslab earthquakes to characterize the anisotropic structure of the subduction zone.Plate motion has often been linked to shearing and flow in the upper mantle (Long & Wirth, 2013), including viscous coupling between the downgoing slab and the overlying mantle (van Keken, 2003). Shearwave splitting has been utilized to study anisotropy in both the crust and upper mantle in various regions around the globe (Savage, 1999;Silver & Chan, 1991). In subduction zones, it is predicted that multiple types and layers of olivine fabrics and other sources of anisotropy may be present (Jung & Karato, 2001;Karato et al., 2008;Silver & Savage, 1994). Changes in olivine fabric have been evoked to explain sharp transitions of shear-wave splitting fast directions from arc-parallel in the arc and forearc to arc-perpendicular in the backarc without requiring a change in mantle flow direction (Nakajima & Hasegawa, 2004;Kneller et al., 2005). However, even a small presence of other mantle minerals, such as antigorite, can have a strong influence on the observed splitting pattern (Horn et al., 2020). Furthermore, organized melt channels can be

Mantle Wedge Seismic Anisotropy and Shear Wave Splitting: Effects of Oblique Subduction

“…In Sumatra, the margin-parallel fast direction is attributed to a strike-slip fault system (Collings et al, 2013). In both subduction zones, the delay times that are associated with margin-normal fast directions increase with the thickness of the mantle wedge, and these observations have been interpreted to indicate a 2-D mantle wedge flow (Collings et al, 2013;Richards et al, 2021). However, the subduction direction in these two regions is oblique to the margin, and the margin-normal fast directions and increasing delay times with the mantle wedge thickness are consistent with the SWS results for 3-D mantle wedge flow patterns due to oblique subduction.…”

“…Our modeling results show that 3‐D mantle wedge flow due to oblique subduction can also results in margin‐normal fast directions that cannot be distinguished from those that result from 2‐D flow patterns (i.e., margin‐normal flow). An SWS study that uses local S waves for the Alaska subduction zone, for example, indicates margin‐parallel fast direction above the cold mantle wedge nose and predominantly margin‐normal fast directions with some variability in the arc and backarc regions (Richards et al., 2021). Another local SWS study on the Sumatra subduction zone indicates margin‐parallel fast directions in the arc region and the margin‐normal fast directions in the backarc (Collings et al., 2013).…”

“…Other possible mechanisms that may cause margin‐parallel fast directions include the shape‐preferred orientation of melt lenses (e.g., Holtzman et al., 2003) and anisotropy in the overriding crust (e.g., Uchida et al., 2020). The margin‐normal fast directions in the backarcs of Alaska and Sumatra have been interpreted to indicate 2‐D mantle wedge corner flow based on the assumption that the mantle flow direction is parallel to the fast direction (Collings et al., 2013; Richards et al., 2021). Other studies invoke either more complex flow patterns or other sources of anisotropy for the interpretation of the observed SWS parameters (e.g., Long & Wirth, 2013).…”

“…The fast directions that are measured in subduction zones are spatially variable, both within and among subduction zones (Figure 1). However, regardless of subduction obliquity, the fast directions in the forearc and arc regions are more commonly subparallel to the margin than those normal to the margin, and many subduction zones, including oblique subduction zones, exhibit margin‐normal fast directions in the backarc (e.g., Abt et al., 2010; Collings et al., 2013; Greve & Savage, 2009; Long & Silver, 2008; Long & Wirth, 2013; Nakajima & Hasegawa, 2004; Richards et al., 2021). Observations of margin‐parallel fast directions in the forearc and the arc have been interpreted to indicate margin‐parallel mantle flow by assuming A‐type olivine CPO (e.g., Hoernle et al., 2008; Long & Silver, 2008).…”

Shear‐Wave Splitting in the Mantle Wedge: Role of Elastic Tensor Symmetry of Olivine Aggregates

“…Previous local SWS studies indicate that the fast direction in the Alaska and Sumatra subduction zones is margin-normal in the arc and back-arc and margin-parallel in the forearc and that the delay times range from ∼0.05 to 0.8 s with a weak positive correlation with the travel distance through the mantle wedge (Collings et al, 2013;Richards et al, 2021). Our study also indicates an overall increase in delay times with the depth to the slab surface regardless of the variation in the fast direction with depth as the delay time accumulates along the ray path unless the fast axis of a given layer coincides with the slow axis of another (i.e., the fast axis orientation is offset by ∼90°), canceling out each other's anisotropy.…”