Invited review articleThe Cascadia Paradox: Mantle flow and slab fragmentation in the Cascadia subduction system

- Discussion of present-day mantle flow in the Cascadia subduction system.
- I review observations of seismic anisotropy and other constraints on flow.
- Observations in different parts of the system yield conflicting interpretations.
- Evidence for fragmentation of the slab in the upper mantle is mixed.
- I discuss possible avenues for resolving the Cascadia Paradox.

The pattern of mantle flow in subduction systems and the processes that control the mantle flow field represent fundamental but still poorly understood aspects of subduction dynamics. The Cascadia subduction zone is a compelling system in which to understand the controls on mantle flow, especially given the dense geophysical observations recently provided by EarthScope, GeoPRISMS, the Cascadia Initiative, and related efforts. Observations of seismic anisotropy, which provide relatively direct constraints on mantle flow, are particularly intriguing in Cascadia because they seem to yield contradictory views of the mantle flow field in different parts of the system. Specifically, observations of seismic anisotropy on the overriding plate, notably in the central portion of the backarc, apparently require a significant component of three-dimensional, toroidal flow around the slab edge. In contrast, new observations from offshore stations are compellingly explained with a simple two-dimensional entrained flow model. Recent evidence from seismic tomography for the likely fragmentation of the Cascadia slab at depth provides a further puzzle: how does a fragmented slab provide a driving force for either two-dimensional entrained flow or three-dimensional toroidal flow due to slab rollback? Resolution of this apparent paradox will require new imaging strategies as well as the integration of constraints from seismology, geodynamics, and geochemistry.