Elsevier

Tectonophysics

Volume 753, 20 February 2019, Pages 1-14
Tectonophysics

Seismic structure beneath the Reykjanes Peninsula, southwest Iceland, inferred from array-derived Rayleigh wave dispersion

https://doi.org/10.1016/j.tecto.2018.12.020Get rights and content

Highlights

  • A site-specific S-wave structure beneath the Reykjanes Peninsula, SW Iceland is determined down to 60 km.

  • Rayleigh-wave phase velocity dispersion in a relatively wide range of periods, from 3 to 50 s, is computed.

  • Two novel array techniques are applied: the zero-crossing point method and the phase-plane method.

  • Low S-wave velocity channels were found in the upper crust and the uppermost mantle.

Abstract

The aim is to obtain a site-specific S-wave structure beneath the Reykjanes Peninsula, southwest Iceland. Nine broadband stations of the Reykjanet network are used to find Rayleigh-wave phase velocity dispersion in a relatively wide range of periods (from 3 to 50 s). The records analyzed were made in the years 2013 to 2015 and concern fourteen selected earthquakes whose epicentral distances range from tens of kilometers to almost ten thousand kilometers. Our approach to retrieving Rayleigh phase velocity dispersion involves two partly independent methods allowing for array apertures larger than a wavelength: 1) the zero-crossing point method, and 2) the phase-plane method. The two methods used here work with seismograms decomposed into quasi-harmonic components and implicitly assume a single plane wave propagation. The good match between the dispersion curves obtained by means of the two methods indicates that the assumption has been reasonably fulfilled. The Rayleigh wave phase velocity dispersion data are inverted into a horizontally layered isotropic S-wave velocity model of the Earth's crust and uppermost mantle by a modified method of the single-parameter variation. At shallow depths, the derived model is rather similar to some previous models that were derived predominantly from the arrival times of body waves. At depths exceeding about 20 km, the dispersion data require low S-wave velocities, indicating a noticeable low-velocity zone.

Introduction

The Reykjanes Peninsula, southwest Iceland, is a geodynamically active region characterized by recent volcanism, occurrence of geothermal fields and high seismicity. It is situated on the spreading plate boundary separating the North American and Eurasian plates. The peninsula geology is controlled by two interacting features (Fig. 1). First, it is an on-land continuation of the submarine Reykjanes Ridge (called also Reykjanes Volcanic Belt), a part of the Mid-Atlantic Ridge leading south of Iceland. The plate boundary bends where it passes onshore and follows the azimuth 70–80° being parallel to the axis of the peninsula. Second, it is the upwelling Iceland Mantle Plume and the associated hot spot, presently situated underneath central Iceland, with a presumed deep root in the mantle (Einarsson, 1991). The peninsula forms the transition between the Reykjanes Ridge in the west and the Western Volcanic Zone and the South Iceland Seismic Zone (SISZ) in the east. A major structural feature on the peninsula is the series of en echelon fissure swarms, each associated with a specific volcanic complex (Fig. 1, inset). The fissure swarms are oblique to the plate boundary and extend a few tens of kilometers into the plates on either side (Einarsson, 2008). The volcanic activity seems to recur approximately every thousand years. Pleistocene basaltic lavas, remains of subglacial eruptions and postglacial lava flows are typical for the peninsula surface geology (Thordarson and Larsen, 2007).

The crustal structure in Iceland and the surrounding area is very specific, neither typically continental nor typically oceanic. It has been studied for decades. Various methods have been used, namely seismic refraction (Darbyshire et al., 1998, Gebrande et al., 1980, Staples et al., 1997), seismic surveys (Smallwood et al., 1995), gravity modelling (Darbyshire et al., 2000a), magnetotelluric methods (Björnsson et al., 2005, Eysteinsson and Hermance, 1985), profile-based surface wave dispersion (Du and Foulger, 1999, Li and Detrick, 2006), receiver function (Darbyshire et al., 2000b, Du et al., 2002), and body wave tomography (Yang and Shen, 2005). Recent studies propose a crust with laterally varying thickness reaching as far as 40 km. The thickest crust is located in central and southeastern Iceland, above the center of the mantle plume (Foulger et al., 2003; Gudmundsson, 2003; Karson, 2016). In the southwest and along the ridge, the crustal thickness does not exceed 25 km but even in coastal regions it is not thinner than 15 km (Kumar et al., 2007). It is thus still much thicker than normal oceanic crust.

Regardless of the laterally varying structure, most of the studies confirm the crust that can be divided into two parts, the upper and the lower crust. The upper crust is composed of porous lava characterized by a rapid increase of seismic velocities with depth, which is interpreted as a result of the closure of pores and fractures with depth (Stefánsson et al., 1993). The bottom of the upper crust is marked with the vP isovelocity surface of 6.5 km/s and is found at depths of 3–10 km (Kaban et al., 2002). The lower crust displays a small velocity gradient and it is assumed to consist of low-porosity basalts with intrusions and high content of epidote (Flóvenz and Gunnarsson, 1991).

Seismicity on the Reykjanes Peninsula is concentrated in a relatively narrow zone along the plate boundary (Árnadóttir et al., 2004, Keiding et al., 2009, Tryggvason, 1973). Earthquakes are frequent but generally small. On the western part, earthquake swarms including only few earthquakes of magnitude 5 or larger occur. The typical mechanism is normal faulting. Towards the east, the mainshock-aftershock character of the activity gradually increases and strike-slip faulting becomes more prominent. Earthquakes of magnitudes exceeding 6 occur rarely.

When studying seismic swarms as well as the mainshock-aftershock sequences it is essential to locate the individual events as precisely as possible, for which it is necessary to know the correspondent velocity structure of the region. The main goal of this study was to derive a site-specific structure of the crust and uppermost mantle beneath the Reykjanes Peninsula. In view of the above-mentioned lateral heterogeneity, any averaged model or structure obtained for another (even not too remote) area is not applicable to this region. Up to now, the SIL model by Stefánsson et al. (1993) has been used for routine data processing and earthquake locations in SW Iceland. It has been, however, derived for the SISZ region, about 100 km to the east of our region of interest. Thus the need to obtain the site-specific Reykjanes Peninsula seismic structure becomes even more covetable as the SIL model does not seem to be entirely adequate. Among previous studies, only few are relevant for the Reykjanes Peninsula itself. Let us name the Reykjanes-Iceland Seismic Experiment (RISE) from 1996, with one of the profile lines going along the peninsula (Weir et al., 2001), and the South Iceland Seismic Tomography (SIST) refraction profile from early 1980s (Bjarnason et al., 1993), the middle part of which passed close to the east end of the peninsula. Both of those experiments resulted merely in a P-wave velocity structure of the crust above Moho. Further, the crustal structure along the Reykjanes Peninsula, both P- and S-wave velocities, has been investigated by tomographic studies but only down to relatively shallow depths. Tryggvason et al. (2002) present a 3-D tomographic model of the upper 15 km of the crust. Geoffroy and Dorbath (2008) applied the double-difference tomography method which allowed them to interpret the upper crust only down to 6 km.

In this study, broadband seismic records are used to derive the site-specific structure from Rayleigh wave phase velocity dispersion in a wide range of periods. In a surface wave array analysis, differential travel times (station-to-station time delays) are measured as a function of period, from which an optimal period-dependent slowness vector and, consequently, phase velocity dispersion and period-dependent true back azimuth are obtained. Many powerful array data analysis techniques have been developed in recent decades for this purpose. An overview of those methods can be found, e.g., in Rost and Thomas (2002). In most of the array methods, the wavefield is approximated as a plane wave propagating along a flat Earth's surface through the array, and beam forming is used to determine back azimuth, phase velocity, and to increase the signal-to-noise ratio by stacking the time-shifted seismograms. Seismogram cross correlation is traditionally used to compute the differential travel times (Menke and Levin, 2002, Pedersen et al., 2006 and others). Another possibility is to use wave gradiometry as the array data processing technique (Langston, 2007, Liang and Langston, 2009). The novelty of the approach used in this paper consists in decomposing the array seismograms into a set of quasi-harmonic components with the amplitude and phase only slowly changing in time. The beam forming approach is applied to those components, which enables us to identify more precisely the very small time differences between them at the individual array stations, i.e. much smaller than the time sampling. In our study, Rayleigh wave dispersion is determined using two novel beam forming methods, the zero-crossing point method (Brokešová and Málek, 2018) and the phase-plane method explained here. Both methods fall into the category of phase gradiometry since they deal predominantly with the differential phase among signals at different array stations. Although not fully independent (the same data are used in both of them), the two methods are procedurally so different and work with different types of information contained in the data that each of them can be used to verify the resulting dispersion curve obtained by the other one.

The dispersion curve is inverted for the 1D isotropic model of the crust and the uppermost mantle down to the depth of 60 km. When inverting the phase velocities, we apply the single-parameter variation method (Novotný et al., 2001). It is well known that Rayleigh wave dispersion is controlled predominantly by the distribution of S-wave velocities, but less by P-wave velocities and densities (Brune and Dorman, 1963, Li and Detrick, 2006). Therefore, we focused on retrieving the S-wave velocity structure. Note that the S-wave structure is, in general, much more difficult to obtain by other methods which are not based on surface waves. In the authors' experience, it is sometimes impossible to identify precise S-wave onsets in local-earthquake records because of the seismogram complexity typical for the region.

Section snippets

Seismic data

In the summer of 2013, the Reykjanet seismic network, consisting of fifteen three-component stations, was installed in the Reykjanes Peninsula in order to monitor the local seismicity. The network is operated jointly by the Institute of Rock Structure and Mechanics and the Institute of Geophysics, both belonging to the Czech Academy of Sciences. Until 2016, nine of the Reykjanet stations were equipped with the Guralp CMG-40T instruments. They constituted a sub-array of broadband stations shown

Methods to determine array-derived phase velocity dispersion

In this pilot study, we prefer using Rayleigh waves instead of Love waves for several reasons. First, Rayleigh waves are less sensitive to azimuthal anisotropy (Maupin, 2007) which we average out. Second, Rayleigh waves can be observed on the vertical component alone. Analyzing the vertical component has many advantages compared to the horizontal ones. The vertical component does not depend on the true back azimuth (usually unknown prior to the analysis). Rayleigh waves in the

Broadband Rayleigh-wave phase-velocity dispersion in the Reykjanes Peninsula and its inversion for structure

To find a reliable structure of the crust and uppermost mantle, it is necessary to invert the dispersion curve in a wide interval of periods. Each of the selected earthquakes provides dispersion points only in an interval much narrower than required. Thus, we have to combine dispersion points from the various earthquakes to obtain the compound broadband dispersion curve to be inverted.

Such approach is justified by two assumptions adopted in our study. First, the array derived dispersion methods

Reykjanes Peninsula seismic structure

The final velocity-depth profile is specified in Table 1 (together with the starting model). The table also provides the vP/vS ratio down to the depth of 15 km. Below that depth, the vP structure has only a formal meaning and it is not supported by observations. The vP/vS ratio is not given there for its irrelevance.

The subsurface structure in the Reykjanes Peninsula seems to be characterized by a relatively high vP/vS ratio, except for the shallowest 1 km. The small vP/vS ratio in the first

Discussion

By applying two novel approaches to the data from the Reykjanet array we received the phase-velocity dispersion curve for the fundamental mode of Rayleigh waves in a broad period range involving both short periods (starting at 3 s) and long periods up to almost 50 s. For the given region, the corresponding dispersion curve has not been published yet in such a period range.

The main result of our present study is the S-wave velocity structure (Fig. 11). It shows three distinctive crust sections.

Conclusions

We have constrained the shear-wave structure of the crust and uppermost mantle beneath the Reykjanes Peninsula by analyzing fundamental mode Rayleigh waves recorded at nine broadband seismic stations of the Reykjanet network. The seismograms of fourteen selected earthquakes from various epicentral distances were used. Two recently developed techniques of dispersion analysis were applied to them, the ZCP method and the PP method. These approaches made it possible to determine Rayleigh wave phase

Acknowledgments

This work was supported by the Czech Science Foundation, under two projects, No P210/18-05053S and No P210/15-02363S. Locations and basic characteristics of the selected earthquakes were obtained from USGS NEIC through the open-source website http://earthquake.usgs.gov/earthquakes/. The authors also thank the editor and three anonymous reviewers for critical comments and suggestions that helped to improve the manuscript significantly.

References (52)

  • ThordarsonT. et al.

    Volcanism in Iceland in historical time: volcano types, eruption styles and eruptive history

    J. Geodyn.

    (2007)
  • YangT. et al.

    P-wave velocity structure of the crust and uppermost mantle beneath Iceland from local earthquake tomography

    Earth Planet. Sci. Lett.

    (2005)
  • AllenR.M. et al.

    Plume driven plumbing and crustal formation in Iceland

    J. Geophys. Res.

    (2002)
  • ÁrnadóttirT. et al.

    Coseismic stress changes and crustal deformation on the Reykjanes Peninsula due to triggered earthquakes on June 17, 2000

    J. Geophys. Res.

    (2004)
  • BjarnasonI.T.h. et al.

    Tomographic image of the Mid-Atlantic Plate Boundary in southwest Iceland

    J. Geophys. Res.

    (1993)
  • BjörnssonA. et al.

    Crustal formation and magma genesis beneath Iceland: magnetotelluric constraints

    Geol. Soc. Am. Spec. Pap.

    (2005)
  • BruneJ. et al.

    Seismic waves and Earth structure in the Canadian shield

    Bull. Seismol. Soc. Am.

    (1963)
  • DarbyshireF.A. et al.

    Crustal structure above the Iceland mantle plume imaged by the ICEMELT refraction profile

    Geophys. J. Int.

    (1998)
  • DarbyshireF.A. et al.

    A first detailed look at the Greenland lithosphere and upper mantle, using Rayleigh wave tomography

    Geophys. J. Int.

    (2004)
  • DarbyshireF.A. et al.

    Crustal strucure of central and northern Iceland from analysis of teleseismic receiver functions

    Geophys. J. Int.

    (2000)
  • DuZ. et al.

    Crustal structure beneath western and eastern Iceland from surface waves and receiver functions

    Geophys. J. Int.

    (2002)
  • DuZ.J. et al.

    The crustal structure beneath the northwest Fjords, iceland, from receiver functions and surface waves

    Geophys. J. Int.

    (1999)
  • EinarssonP.

    Plate boundaries, rifts and transforms in Iceland

    Jökull

    (2008)
  • EysteinssonH. et al.

    Magnetotelluric measurements across the eastern neovolcanic zone in South Iceland

    J. Geophys. Res.

    (1985)
  • FoulgerG.R. et al.

    Iceland-type crust

    Geophys. J. Int.

    (2003)
  • FunckT. et al.

    Moho and basement depth in the NE Atlantic Ocean based on seismic refraction data and receiver functions

  • View full text