Seismic body waves from distant earthquakes, which propagate near vertically beneath recording stations, provide tools for imaging shallow Earth structures with high vertical resolution. The most commonly used techniques such as P and S wave receiver functions utilize mode conversions from P to S waves or vice versa to retrieve information on the gradients of elastic properties in the crust and upper mantle. Here we demonstrate the feasibility and advantage of utilizing reflection signals through an improved method of teleseismic P wave coda autocorrelation. We recover clear reflections independently on vertical and radial components, which provide complementary constraints on the subsurface structures. Field data from two stations from different geological settings are analyzed, one of which is an ice station in Antarctica and the other is a bedrock station on the Kaapvaal craton in South Africa. The results from both analyses show the feasibility of the method to unveil P and S wave reflection signals from the ice‐rock interface and the Moho discontinuity. Extensive synthetic experiments are set up to corroborate our results.

Cross correlation of seismograms provides new information on the Earth both through the exploitation of ambient noise and specific components of earthquake records. Here we cross‐correlate recordings of large earthquakes on a planetary scale and identify a range of hitherto unobserved seismic phases in Earth's correlation wavefield. We show that both arrivals with the timing expected for the regular seismic wavefield and previously unexplained phases are produced by interference between seismic paths having the same ray parameter but with only a subset of propagation legs in common. This insight explains the origin and generation mechanism of the features of Earth's correlation wavefield and opens up new ways of addressing issues in global seismology. Strong similarity between observed and synthesized correlation wavefields indicates that the Earth's radial structure is remarkably well constrained in the intermediate period range.

Antarctica is largely covered by an ice cap of a variable thickness characterized by relatively low density and seismic velocities. Passive seismological deployments have a limited use in imaging a thin ice layer because of the dominance of a relatively low‐frequency content in the teleseismic wavefield. Here we use passive seismological data and an improved autocorrelation method utilizing P wave coda to image the ice cover. The resulting autocorrelograms are interpreted as reflectivity records from a virtual source on the surface and reflection pulses at the ice base. We convert the reflection delay of P waves to the ice thickness measurements using a homogeneous P wave speed compatible with previous studies. Apart from P wave reflectivity, we obtain S wave reflectivity from the autocorrelation of radial component. The ratio of S wave and P wave reflection times represents a measurement of the P over S wave speed ratio (and Poisson's ratio). The successful application to unveil the Antarctic ice sheet properties presented here opens a way for future studies to measure properties of the ice cover in Antarctica, other continents, and icy planets in future space missions.

No abstract

This paper reviews the concepts underlying the well-documented receiver functions (RFs) method, and places it in the conceptual framework of seismic interferometry. We first propose a simple and efficient approach for isolating the receiver-side seismic response (i.e. the record of reflections and conversions in the stratification beneath receivers): this method makes use of the P-wave coda recorded on the radial and vertical components of three-component stations, applies spectral whitening, which is followed by auto-and cross-correlation. The interferometric principle underpinning RFs analysis is shown theoretically and illustrated in practice using earthquake records and synthetic waveforms computed from simple structures. We point out to a major limitation, which is the contamination of the receiver-side response by propagation effects in the source-side structure. We then apply our approach to teleseismic earthquake data recorded in California. We show that the reconstructed vertical and horizontal seismic responses can be back-projected to illuminate the crustal and mantle structure. We build comparable ∼300-km-long seismic reflectivity profiles from pure P-wave reverberations and from the converted wavefield across the forearc and arc of the southern Cascadia subduction zone. Then, we show a case of processing data from narrow bandpass, short-period and single-component sensors, usually unsuitable for RFs analysis. Finally, through the same interferometric principle, we attempt to demonstrate a link between event-and noise-based seismic interferometry. We demonstrate that it is possible to extract approximate responses from the records of low-magnitude-down to 4.5-teleseismic earthquakes. We make a comparison of these estimates with the result from the autocorrelation of the continuous ambient noise seismic wavefield. While the amplitudes of the extracted receiver-side responses are mutually different, their phases are in a relative agreement. This development opens a way to the use of small magnitude teleseismic earthquakes to characterize the receiver-side structure.