Lg-wave attenuation in the Australian crust
Introduction
Lg waves form a prominent component of seismograms for continental sources at regional distances. Lg propagates in continental crust with a group velocity around 3.5 km/s, and was first described by Press and Ewing (1952). Lg is dominated by relatively high frequencies between 0.2 and 5 Hz for continental paths, and usually has an unclear onset but distinct later large amplitudes. A description of the nature of Lg can be made in terms of either the superposition of higher modes of surface waves, or the interference of multiply reflected S waves (e.g., Oliver and Ewing, 1957, Knopoff et al., 1973, Kennett, 1984, Kennett, 1985).
The Lg wave is sensitive to the structure of crust along its propagation path, and so we can hope to extract differences in crustal properties from the nature of the propagation. A detailed discussion of the interaction of Lg waves with structure is provided in Kennett (2002, Chapter 20). Variations in the characteristics of Lg propagation across Asia were recognized by, e.g., Ruzaikan et al., 1977, and similar studies have been made for many parts of the world. Gregersen (1984) showed how the efficiency of Lg-wave propagation (expressed as the relative amplitude of Lg to Sn) could be used to obtain variations in lateral crustal structure heterogeneity. The increasing volumes of high quality data have allowed detailed Lg tomography in many areas (e.g. Zhao et al., 2010).
The Fourier amplitude spectra of Lg waves can be exploited to study crustal attenuation. Typically, tectonically active regions are usually characterized by relatively strong attenuation while stable regions with lower heat flow, thinner sediment cover and a lack of fluids are related to relatively weaker Lg attenuation. Thus, Pasyanos et al. (2009) found higher loss of Lg-wave energy in tectonic regions in the Middle East compared with the shield regions, and Zhao and Xie (2016) found strong attenuation in the continental collision orogenic belt.
The amplitudes of Lg waves are affected by intrinsic attenuation and structural effects, but by building up extensive path coverage we can attempt to isolate the dominant influences in conjunction with information from other fields. For example, Sheehan et al. (2014) suggested that seismic attenuation measurements were able to provide information complementary to seismic velocity estimates and so help to distinguish between compositional and thermal mechanisms for observed anomalies.
There is a strong contrast in the surface geology of Australia with Precambrian rocks in the west and centre which cover two-thirds of the continent, and Phanerozoic fold belts in the east. There has been no mountain building in the ancient continent for more than 200 Myr. The complex development of Australian lithosphere over more than 2 Ga has exerted fundamental influences on the overall tectonic stability and the character of the landscape, the distribution of earthquakes and associated seismic risk, the evolution of sedimentary basins, as well as heat flow and other resource endowment (Kennett and Blewett, 2012).
The structure of the Australian crust has been extensively studied using a wide range of seismological techniques including refraction and reflection, ambient noise, and receiver functions (e.g. Clitheroe et al., 2000, Saygin and Kennett, 2012, Salmon et al., 2013, Korsch and Doublier, 2016). So far, the properties of Lg with their strong sensitivity to crustal variations have not been exploited, and have the potential to provide further understanding of the evolution of the continental lithosphere.
The Tasman Line indicating the transition zone between the Precambrian zone in the west and Phanerozoic region in the east of Australia was originally proposed as the eastern edge of Precambrian outcrop (Hill, 1951). Subsequently, attempts have been made to relate this transition to the break-up of the Rodinian supercontinent (e.g. Scheibner and Veevers, 2000). However, there still exists some uncertainty as to the nature of the transition since there are many conflicting interpretations and models based on different classes of geophysical data (Direen and Crawford, 2003, Kennett et al., 2004). In this work, we try to see if the variations in Lg Q provide further insight into the characteristics of the Tasman line.
For the Australian region, more attention has been paid to seismic attenuation for the entire lithosphere than within the crust. Gudmundsson et al. (1994) provided evidence for a strong change increase in attenuation in the upper mantle beneath the lithosphere in the northern part of the Australian continent. Cheng and Kennett (2002) used a spectral ratio method to estimate the differential attenuation between P and S waves for the paths through the upper-mantle beneath northern Australia. Kennett and Abdullah (2011) extended this work and found that the crust and lithospheric mantle in the eastern part of Australia has larger attenuation compared with that of western and central Australia. The characteristics of the Lg phase provide information on the structure of the Australian continental crust that has not previously been extensively exploited. Bolt (1957) provided a summary of the velocities of Lg waves for the whole Australian continent based on a rather limited set of observations. Bowman and Kennett (1991) investigated the nature of Lg in the northern part of Central Australia, demonstrating the complex influence of a strong velocity gradient at the base of the crust. Mitchell et al. (1998) mapped out variations in attenuation from a limited suite of Lg coda Q observations and showed a strong contrast between low crustal attenuation in central and western Australia and stronger attenuation in eastern Australia.
Our study covers the whole Australian continent, exploiting the improved seismic data available in the last 10 years in Australia. We employ Lg-wave attenuation tomography over a broad range of frequencies with good ray path coverage and so achieve higher resolution than in previous studies. We also examine the frequency dependency of QLg in the main crustal blocks in Australia.
Section snippets
Data
We have employed data from the network of permanent stations across the Australian continent, and from temporary broadband deployments made by the Australian National University. Vertical-component seismograms were collected in the period from January 1993 to October 2016, from 469 events recorded at 203 stations with epicentral distances ranging from 2° to 20°. To guarantee suitable signal-to-noise ratios, and avoid influences from complexity in the rupture process of larger earthquakes, we
Lg wave Q
We use a procedure to extract Lg wave attenuation across Australia that has previously been used for other areas (e.g., Zhao et al., 2010, Zhao et al., 2013b). We assume that the Lg wave propagates along the great circle path between source and station with geometrical spreading proportional to the square root of epicentral distance. This spreading function is suitable for a crust with a sharp transition to the mantle, but may not give as effective a representation where the transition from
Lg attenuation in Australia
The inversion domain for Lg-wave attenuation covers the entire Australian continent with a longitude span of 40° and a latitude span of 35°. Lg-wave Q models have been constructed for a suite of 58 frequencies between 0.05 and 10.0 Hz. We focus on the attenuation measurements for frequencies above 0.5 Hz, due to the larger noise levels for lower frequencies. We endeavour to characterize the differences in Lg attenuation and velocities between regions and the spatial variation in the frequency
Discussion
In Fig. 9 we bring together the full range of information on Lg behaviour for Australia, linked to the main tectonic provinces. We display the average Lg Q in the frequency band 0.6–1.0 Hz in Fig. 9(a), and alongside this we show the frequency exponent η for the same frequency band (Fig. 9b). For this frequency-band Lg waves should sample the entire crust with some modest emphasis on the shallower regions. The third panel (Fig. 9c) displays the group velocity for Lg waves (see Supplementary
Conclusions
With the improved coverage of seismic datasets for Australia, we are able to provide Lg-wave attenuation models at different frequencies at good spatial resolution to provide useful constraints on crustal structure. The resulting tomographic images provide the supplementary information to map major tectonic boundaries, as well as further insights into the characteristics of different geology blocks for Australia.
- (1)
The study of the spatial decay of spectral amplitudes of Lg waves shows that the
Acknowledgements
This work has been carried out at the Research School of Earth Sciences, Australian National University. The research was supported in large part by the China Scholarship Council (grant 201604910731) and the National Natural Science Foundation of China (grants 41374065 and 41674060), with additional support from the AuScope AuSREM project. We thank the IRIS Data Management Center (http://ds.iris.edu/ds/nodes/dmc) from which we extracted the waveform data, including the SKIPPY and KIMBA portable
References (46)
- et al.
Broadband observations of upper-mantle seismic phases in northern Australia and the attenuation structure in the upper mantle
Phys. Earth Planet. Inter.
(1994) - et al.
Grainsize-sensitive viscoelastic relaxation in olivine: towards a robust laboratory-based model for seismological application
Phys. Earth Planet. Inter.
(2010) - et al.
Major crustal boundaries of Australia, and their significance in mineral systems targeting
Ore Geol. Rev.
(2016) - et al.
The Moho in Australia and New Zealand
Tectonophysics
(2013) - et al.
Strong Lg-wave attenuation in the Middle East continental collision orogenic belt
Tectonophysics
(2016) - et al.
Crustal flow pattern beneath the Tibetan plateau constrained by regional Lg-wave Q tomography
Earth Planet. Sci. Lett.
(2013) - et al.
The shape of ground motion attenuation curves in southeastern Canada
Bull. Seismol. Soc. Am.
(1992) - et al.
Regional Lg attenuation for the continental United States
Bull. Seismol. Soc. Am.
(1997) Velocity of the seismic waves Lg and Rg across Australia
Nature
(1957)- et al.
Propagation of Lg waves in the North Australian craton: influence of crustal velocity gradients
Bull. Seismol. Soc. Am.
(1991)