Roughness of the mantle transition zone discontinuities revealed by high‐resolution wavefield imaging

Abstract

We processed two independent data sets: 2,458,973 three‐component seismograms from all EarthScope Transportable Array (USArray) stations and 141,080 pairs of high‐quality receiver functions from the EarthScope Automated Receiver Survey using the generalized iterative deconvolution method and shaped the output with a 0.1 Hz Ricker wavelet. We used these data as input to our 3‐D plane wave migration method to produce an image volume of P to S scattering surfaces under all of the lower 48 states. Model simulations show that compared to the widely used common conversion point stacking method, our method is capable of resolving not only dipping discontinuities but also subtler details of amplitude and topography variations at scales near the diffraction limit. Maps of attributes derived from these discontinuities reveal several previously unobserved features of the 410 and 660 discontinuities (d410 and d660). Topography with many tens of kilometers is resolved at a range of scales. Additionally, we observed large variations of amplitude on the radial component and in the radial to transverse amplitude ratio that correlate with inferred variations in discontinuity topography. We argue that this combination of observations can be explained by roughness at a range of scales. The structural fabric we imaged is preferentially aligned orthogonal to the Farallon‐North American relative motion. Two regions we infer to have exceptional roughness show spatial correlation with locations where tomography models indicate that the subducted slab is passing through the transition zone. We suggest that these rough regions are indicators of vertical mantle flows through the transition zone.

Imaging Upper‐Mantle Structure Under USArray Using Long‐Period Reflection Seismology

“…Noteworthy transition‐zone features seen in Figures and include the following: The apparent western deepening of the 410 occurs near −100° longitude, whereas the deepening of the 660 occurs along a NNW/SSE trending boundary that approximately coincides with the Rocky Mountain Front. A pronounced depression is seen in the 660 centered on Utah with maximum depths in southern Utah, which roughly agrees in position with the sharp, localized 660 depression imaged in the receiver‐function analysis of Cao and Levander () and a broader depression seen in Wang and Pavlis (). However, other USArray receiver‐function studies (Gao & Liu, ; Schmandt et al, ) have not resolved this feature. The 410‐km reflector nearly disappears near 40° latitude under Nevada, Utah, and western Colorado.…”

“…2. A pronounced depression is seen in the 660 centered on Utah with maximum depths in southern Utah, which roughly agrees in position with the sharp, localized 660 depression imaged in the receiver-function analysis ofCao and Levander (2010) and a broader depression seen inWang and Pavlis (2016). However, Depths to a negative polarity reflection observed between 40 and 100 km under much of USArray.…”

“…These values are very close to the average discontinuity depths estimated from global SS precursors (e.g., 418 and 660 km from Flanagan & Shearer, ) and the USArray receiver function analysis of Gao and Liu (), who obtained average depths across the United States of 416 and 667 km. Wang and Pavlis () obtained slightly different USArray receiver‐function average depths of 415 and 660 km; they attributed their shallower d 660 to their use of the AK135 reference 1‐D model rather than iasp91 (their application of timing corrections based on the 3‐D P wave model of Burdick et al, , may also have been a contributing factor).…”

Complete agreement of the post-spinel transition with the 660-km seismic discontinuity

“…Open and solid circles with error bars identify the stable phases as ringwoodite and bridgmanite + periclase, respectively. The black solid line is the expected condition of the 660-km discontinuity 7,8,16–21 . The red (T-3BM) and blue (T-Vinet) solid lines are phase boundaries evaluated using the fixed points obtained in this study and the Clapeyron slopes of ref.…”

“…Taking into account the uncertainty of the Clapeyron slope between −0.003 GPa/K and −0.001 GPa/K 1–5,13–15 , the transition pressure would be 22.7–23.3 GPa and 23.4–23.8 GPa, respectively, according to these scales. Meanwhile, the global average depth of the discontinuity is 660 ± 10 km, a variance in distance that corresponds to a pressure variance of 23.4 ± 0.4 GPa 16–21 . Given these facts, the present findings are entirely consistent with the depth of the 660-km discontinuity, therefore corroborating the compositional models in which the Earth’s mantle is composed of ferromagnesian silicates.…”