Imaging the full lithosphere with earthquake sources

Seismology Nov 24, 2020

Significant changes in physical properties produce conversions between compressional (P) and shear (S) waves. The Earth’s surface also returns energy to depth that can be reflected back to the surface. Analysis of this chain of conversions and reverberations yields information about the structure beneath the station.

The most common approach is the use of receiver functions that exploit the conversions and so can map out major interfaces. The receiver functions exploit the properties of the same time segment of different components of ground motion, suppressing the source contributions to enhance conversions.

A powerful alternative is to exploit the autocorrelation of a single component of ground motion, and thereby enhance consistent features equivalent to reflections from depth. The use of the vertical component of ground motion extracts an estimate of the reflectivity associated with P waves at higher frequencies than used for receiver functions.

With many closely spaced seismic stations it is therefore possible to build a reflection profiles for the entire lithosphere from the recordings of many different earthquakes (e.g., Sun & Kennett, 2020). The frequencies employed (1– 4 Hz) are much lower than in conventional reflection profiling using vibrator sources at the surface, but provide penetration of both the crust and the lithospheric mantle.

Figure 1: Tectonic blocks in southern Central Australia together with prior estimates of crustal thickness. The configuration of the dense MAL experiment is shown with purple symbols. The of triangles along an approximately north–south line mark the GOMA full-crustal reflection profile. The insert map shows the tectonic setting of Australia, with the main map area outlined in red. The approximate boundaries of the main cratonic blocks are shown: NAC – North Australian Craton, SAC – South Australian Craton, WAC – West Australian Craton

With support from AuScope, a line of 68 seismic recorders at approximately 3.5 km spacing was deployed in northern South Australia spanning a zone where previous results indicated a major change in crustal thickness from around 50 km to less than 30 km over a 200 km horizontal distance. This MAL experiment (Figure 1) extends from Marla on the sturt Highway across to the north of Oodnadatta and somewhat to the east. From nearly a year of recordings, in 2018–2019, Liang & Kennett (2020) have shown it is possible to construct an image of the P reflectivity through the entire lithosphere, as displayed in Figure 2.

Figure 2: Reflection imaging from teleseismic arrivals for the frequency band from 0.1–2.0 Hz, projected on to a profile along 27.3 S. The figure is approximately true scale. Major interpreted features are marked. The Moho from receiver function analysis is shown in light green, together with an inferred transition zone into the lithospheric mantle outlined with a dashed line. Bounds on the lithosphere–asthenosphere transition (LAT), from surface wave analysis, are indicated with light purple markers at the edge of the profile. The band of nearly continuous red reflectivity near 95 km depth is likely to be a mid-lithosphere discontinuity (MLD).

The change in crustal thickness art the eastern end of the profile is clearly evident and links to the estimates of the position of the edge of the Gawler craton from other geophysical evidence (particularly magnetic anomalies).  An intriguing feature is a transition zone at the base of the rust that wedges out to the east at the craton margin, this is indicated by the dashed light green line in Figure 2.

In conventional reflection profiles, with frequencies of 20 Hz and above, the base of the crust marks the end of P reflectivity. With the longer wavelengths from the illumination by distant earthquakes it becomes clear that the apparently bland lithospheric mantle has significant structure, but at larger scales than in the crust. Such behaviour had been inferred from a variety of other lines of evidence (Kennett et al., 2017) and now is seen in a direct image.

The MAL experiment has demonstrated that valuable information on lithospheric architecture can be extracted from distant earthquakes. Such reflection profiles exploiting distant earthquakes can be carried out in places where it is uneconomic or logistically difficult to use conventional equipment, though the timescale for data acquisition is much longer.

Background reading

Kennett, B.L.N., Yoshizawa, K., Furumura, T., 2017. Interactions of multi-scale heterogeneity: Australia. Tectonophysics,717, 193–213. doi: 10.1016/j.tecto.2017.07.009

Liang S. & Kennett B.L.N. (2020) Passive seismic imaging of a craton edge – Central Australia, Tectonophysics, 797, 228662. doi: 10.1016/j.tecto.2020.228662

Sun, W., Kennett, B.L.N., 2020. Common-reflection-point-based prestack depth migration for imaging lithosphere in Python: Application to the dense Warramunga array in northern Australia, Seismological Research Letters, 91, 2890–2899. doi: 10.1785/0220200078.


Brian Kennett

Brian Kennett AO FAA FRS is Emeritus Professor of Seismology at ANU. Ph.D Theoretical Seismology, Cambridge. Lindemann Fellow at UC San Diego and Lecturer at Cambridge. He moved to Australia in 1984.