Distinct crustal structure of the North American Midcontinent Rift from P wave receiver functions

Abstract

Eighty‐two broadband seismic stations of the Superior Province Rifting Earthscope Experiment (SPREE) collected 2.5 years of continuous seismic data in the area of the high gravity anomaly associated with the Midcontinent Rift (MCR). Over 100 high‐quality teleseismic earthquakes were used for crustal P wave receiver function analysis. Our analysis reveals that the base of the sedimentary layer is shallow outside the MCR, thickens near the flanks where gravity anomalies are low, and shallows again in the MCR's center where the gravity anomalies peak. This pattern is similar to that found from local geophysical studies and is consistent with reverse faulting having accompanied the cessation of rifting at 1.1 Ga. Intermittent intracrustal boundaries imaged by our analysis might represent the bottom of the MCR's mostly buried dense volcanic layers. Outside the MCR, the Moho is strong, sharp, and relatively flat, both beneath the Archean Superior Province and the Proterozoic terranes to its south. Inside the MCR, two weaker candidate Mohos are found at depths up to 25 km apart in the rift's center. The intermediate layer between these discontinuities tapers toward the edges of the MCR. The presence of this transitional layer is remarkably consistent along the strike of the MCR, including beneath its jog in southern Minnesota, near the Belle Plaine Fault. We interpret these results as evidence for extensive underplating as a defining characteristic of the rift, which remains continuous along the Minnesota jog, where due to its orientation, it is minimally affected by the reverse faulting that characterizes the NNE striking parts of the rift.

P Wave Teleseismic Traveltime Tomography of the North American Midcontinent

“…The standard deviations of both groups of relative delay times are similar to the standard deviation of all relative delays of 0.4 s. The mean delay time of the on‐rift group is roughly 0.05 s later than that of the away‐from‐rift group (with a mean relative delay around 0 s), and the difference between the mean delay times of the two groups is approximately an order of magnitude smaller than the groups' standard deviations, which are similar to the standard deviation of the distribution of all relative delays in this study. This 0.05‐s difference in magnitude is consistent with what one would expect from an underplated layer found along the rift by Zhang et al (). The standard deviation of our relative delay times from equation is 0.4 s. This is less than the standard deviation of relative delay times of 0.5 s measured for the Kenya rift (Park & Nyblade, ) and the Ethiopian hot spot (Bastow et al, ), which are active segments of the East African rift.…”

“…The distance between these weaker discontinuities decreases with increasing distance from the rift axis, and the intermediate impedance material between the two discontinuities is interpreted as “underplated,” following Behrendt et al () and dozens of additional publications of crustal structure of the MCR beneath Lake Superior. Zhang et al () infer that the underplated “layer” is located at 30–50 km in depth and extends axially along the segment of the MCR between Lake Superior and Iowa and likely beyond (supporting information Figure S11). A second test scenario is an identically shaped low‐velocity anomaly but residing in the upper mantle lithosphere at depths of 100–120 km (supporting information Figure S12).…”

Seismic Imaging of the North American Midcontinent Rift Using S‐to‐P Receiver Functions

“…Each event is rotated into theoretical P and S components using a free‐surface transformation matrix (Bostock, ). We divided the region into areas that are thickly sedimented (approximately ≥ 1 km thick) and those that are not based on waveform fitting of Ps receiver functions beneath SPREE stations by Zhang et al (). In locations with thick sediment, we assumed surface velocities V p = 4.00 km/s and V s = 2.00 km/s; otherwise, V p = 5.90 km/s and V s = 3.41 km/s.…”

“…The one‐dimensional migration model for each receiver function is obtained by tracing the approximate Sp ray path through a three‐dimensional Earth model that is based upon previously determined rift and near‐rift Earth structure (Zhang et al, ) and US‐CrustVs‐2015 Moho depths (Schmandt et al, ) in the outer bounds of the grid. For the rifted region, Moho beneath the flanks, the base of an underplate layer beneath the rift axis, and the crustal V p / V s values are defined, which we base on H‐κ stacking and waveform fitting results beneath the SPREE stations (Zhang et al, ). We assumed these values extend both further north and also south along the rift axis to fill in the larger area considered in our study.…”

“…The final crustal thickness used for migration is determined by the Moho piercing point of each waveform in the above model. The crustal P wave velocity structure is designed to be a linear gradient centered on the average P wave velocity of the North American crust (Zhang et al, ): V p = 5.90 km/s at surface and V p = 6.90 km/s at the crustal thickness of the Earth model described above. In areas of sediment, we assume a linear gradient through the sediment and the crust with V p = 4.00 km/s at the surface, V p = 5.90 km/s at the base of the sediment, and the same linear gradient to V p = 6.90 km/s at the crustal thickness in the Earth model.…”