Seismic Constraints on Slab Interaction With the Transition Zone

Over the past decade, seismic tomography has revealed that subducting lithospheric slabs interact with the transition zone in a variety of ways, directly penetrating into the lower mantle in some locations, while stagnating in others. Here, we present preliminary results of attempts to characterize and quantify the stagnation of slab material in the transition zone initiated at the 2006 Cooperative Institute for Deep Earth Research (CIDER) workshop. Providing seismic constraints on slab interaction with the transition zone is essential for verifying dynamic calculations that examine to what degree slabs are hindered from penetrating through the 660 km seismic discontinuity. First we compute the tomographic signature of an end-member mantle model in which 100 km thick slabs descend from the upper to lower mantle without deformation / stagnation in the transition zone. We then compare the amplitude of the predicted shear velocity anomaly with that observed in the most recent Scripps, Berkeley, Harvard, Caltech, and UT Austin global tomographic models. We find that in the western Pacific slab material is accumulating within the transition zone, while under South America, the slabs appear to enter the lower mantle unhindered. This accumulation of slab material in the transition zone indicates that some mechanism is temporarily delaying it from passing into the lower mantle. This finding is consistent with comparisons of power spectra of the observed models in and below the transition zone, which indicate that the pattern of seismic heterogeneity changes drastically across the 660 km discontinuity. Furthermore, the focal mechanisms of deep (>400 km) earthquakes from the Harvard Centroid Moment Tensor project provide a wealth of information on slab deformation within the transition zone. We have systematically compared the orientations of earthquake compressional axes to the slab orientations (as defined by the Wadati-Benioff zone) for all regions of deep seismicity. The results support the findings of tomography that slabs are held up in the western Pacific, but are being pulled into the lower mantle in South America. Dynamic models of slabs should strive to match the seismic observations that slabs exhibit highly variable behavior as they progress from the upper to lower mantle.