imaging crustal structure variation across southeastern...
TRANSCRIPT
Tectonophysics 582 (2013) 112–125
Contents lists available at SciVerse ScienceDirect
Tectonophysics
j ourna l homepage: www.e lsev ie r .com/ locate / tecto
Imaging crustal structure variation across southeastern Australia
Fabrice R. Fontaine a,⁎, Hrvoje Tkalčić b, Brian L.N. Kennett b
a Laboratoire GéoSciences Réunion, Université de La Réunion, IPGP, UMR 7154, 15 avenue René Cassin, BP 7151, 97715 Saint Denis cedex 9, La Réunion Island, Franceb Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia
⁎ Corresponding author at: Laboratoire GéoSciences RéuIPGP, Sorbonne Paris Cité, UMR 7154, 15 avenue René Cascedex 9, La Réunion Island, France. Tel.: +33 2 62 93 82
E-mail addresses: [email protected] ([email protected] (H. Tkalčić), brian.kennett@a
0040-1951/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.tecto.2012.09.031
a b s t r a c t
a r t i c l e i n f oArticle history:Received 5 March 2012Received in revised form 23 September 2012Accepted 28 September 2012Available online 5 October 2012
Keywords:Crustal thicknessMohoCratonMount GambierMurray Basin
A broad-band seismic network of 28 three-component seismometers was deployed in southeastern Australia toexamine variations in crustal thickness across the transition between Precambrian and Phanerozoic lithosphere.Receiver function observations and modelling of P-to-S conversions at the Mohorovičić discontinuity (Moho)have been employed to investigate: (i) the variations in the Moho depth across southeastern Australia, and(ii) the nature of the transition between crust andmantle. Data from temporary deploymentswere used togetherwith data from the few permanent broad-band stations in the region. The extraction of P-receiver functions fromhigh-quality seismic data recorded on these stations has enabled the determination of the crustal thicknessacross the region. The crustal thicknesses lie in the range 28–48 km. TheMoho depth is generally well correlatedwith the Earth surface elevation in the southeastern Australia. TheMoho estimates from receiver functions are ingood agreementwith results from reflection profiling. The average crustal thickness is found to be around 39 kmbeneath the Precambrian area in the west and even thicker beneath the Lachlan Orogen in the east (~43 km).The average crustal thickness in between, beneath the Murray Basin is thinner ~32 km. Interestingly, the crustin the Mount Gambier volcanic area is rather thick ~41 km, suggesting that the limit between the Delamerianand western Lachlan orogens is located east of Mount Gambier. Our results favour a position for the TasmanLine generally consistent with the interpretation by Direen and Crawford (2003) and thus to the east of the lo-cation favoured by many authors. The broader crust–mantle transition and thicker crust beneath the LachlanOrogen suggest the presence of magmatic underplating at the base of the lower crust. The intermediate natureof the crust–mantle transition also suggests magmatic underplating beneath the Gawler Craton and theCurnamona Province.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
The crustal structure (crustal thickness and nature of the basement)of tectonic units is not well constrained in southeastern Australia due tothe presence of sedimentary rocks and of thick regolith across most ofthe surface. Although active seismology can provide the most reliableand detailed information on crustal architecture, such information isconfined to limited regions. A complementary source of informationcomes from passive seismic techniques such as receiver functionsutilised in this study.
The Gawler Craton contains less than 5% basement exposure in anarea of roughly 530,800 km2 (Payne et al., 2009). Unravelling the na-ture of the crustal structure and the lateral extent of crustal provincesis necessary to improve understanding of the evolution of southeasternAustralia. The primary aim of this study is therefore to investigate the
nion, Université de La Réunion,sin, BP 7151, 97715 Saint Denis07; fax: +33 2 62 93 82 66.F.R. Fontaine),nu.edu.au (B.L.N. Kennett).
rights reserved.
lateral variation of crustal structure and nature of the crust to mantletransition between the Precambrian and Phanerozoic belts of south-eastern Australia in a corridor extending from the Gawler Craton tothe east coast of Australia.
Since 2004 more than 7000 km of full-crustal reflection profileshas been collected across Australia providing information on the var-iation in Moho depth. However, the estimates of crustal thicknessesbased on reflection data rely on a simple assumption of an rms veloc-ity of 6 km/s at the base of the crust (e.g. Drummond et al., 2006;Kennett et al., 2011). The secondary goal of our study is therefore tocalibrate Moho depth estimates from recent seismic reflectionprofiles carried out in southeastern Australia (Cayley et al., 2011;Drummond et al., 2006; Fraser et al., 2010; Preiss et al., 2010) usingour receiver function measurements, which are particularly sensitiveto the gradients in seismic velocities.
1.1. Tectonic setting
The Lachlan Orogen in southeastern Australia (Fig. 1b) was accret-ed to the Precambrian core of the continent in a sequence of stages(e.g., Braun and Pauselli, 2004; Cayley, 2011; Collins, 2002; Direenand Crawford, 2003; Gray and Foster, 2004). We show in Fig. 1 a
Tasm
an L
ine ?
Curnamonaprovince
LachlanOrogen
Sydney Basin
NewEnglandOrogen
TasmanSea
NVP
134˚ 136˚ 138˚ 140˚ 142˚ 144˚ 146˚ 148˚ 150˚ 152˚ 154˚
−40˚
−38˚
−36˚
−34˚
−32˚
−30˚
−28˚
CAN
SoCPSKIPPYGA
a)
b)
GEOSCOPE
SO07
SO08
SO09 SO11
SO10
SO12
SO17SO13
SO14
SO15
SO16
ARPSMG01MG02
MG04
MG03
MG05MG06
STKA
SO06
MountGambier
Murray BasinBBOO
TOO
MUR1
MUR2
MUR3
MUR4
MUR5
SO01SO02
SO03
SO05
SO04
Melbourne
Adelaide
SydneyDelamerianOrogen
Precambrian Phanerozoic
GawlerCraton
Otway BasinGippslandBasin
0˚ 60˚ 120˚ 180˚ 240˚ 300˚
−60˚
−40˚
−20˚
0˚
20˚
40˚
60˚
CAN
A
A’
Adelaide
Geosyncline
Fig. 1. a) Location of the 323 events used for receiver function (RF) analysis at station CAN. b) Simplified geological map of southeastern Australia modified from Gray and Foster(2004) with the location of the SoCP seismic network and the permanent seismic stations from Geoscience Australia (GA) and GEOSCOPE networks. Key to marked feature: NVP,Newer Volcanic Province. The red dashed line shows the location of the Tasman Line based on the reinterpretation of Direen and Crawford (2003). The line AA′ is shown in Fig. 6.
113F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
simplified model of a geological map for southeastern Australia.The map area is separated in two parts: (i) the western part formedby Precambrian units: the Gawler Craton, the Curnamona Provinceand the Adelaide Geosyncline and (ii) the younger eastern part withPhanerozoic terranes. The Gawler Craton contains Archean–early
Palaeoproterozoic supracrustal and magmatic lithologies, whichare surrounded, overlain and intruded by Palaeoproterozoic (2000–1610 Ma) to Mesoproterozoic (1590–1490 Ma) units (e.g. Payneet al., 2009). The basement of the Curnamona Province is of latePalaeoproterozoic age. Palaeozoic deformation affected the marginal
114 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
zones of the Province, but the central core is cratonic. The AdelaideGeosyncline is Neoproterozoic to Middle Cambrian basin complexwith deep subsidence which experienced at least five major succes-sive rift cycles (Preiss, 2000).
The large-scale structural elements of the eastern Phanerozoicpart are mainly:
• sedimentary basins with Mesozoic and Cenozoic cover: MurrayBasin, Otway Basin, Gippsland Basin and Sydney Basin,
• three orogens: the early Palaeozoic DelamerianOrogen (550–470 Ma),the Middle Palaeozoic Lachlan Orogen (450–340 Ma) and the LatePalaeozoic to Early Mesozoic New England Orogen,
• and the Newer Volcanic Province, a wide volcanic field site (around15,000 km2) of over 400 monogenetic vents, which represents theproduct of the most recent volcanic activity in Australia, spanningfrom the Pliocene to the Holocene (e.g., Lesti et al., 2008). Themagma reservoirs were estimated to be subcrustal and below 32 kmbased on the spacing and distribution of the density maxima (Lestiet al., 2008). The Mount Gambier volcanic sub-province is theyoungest region of this intraplate continental plain basalt province.Mount Gambier has experienced recent volcanic activity: with signif-icant volcanism at about 4000 years B.P. (Blackburn et al., 1982).
The Phanerozoic part of Australia was formed after the break-up ofRodinia around 780 Ma (Wingate et al., 1998), followed by the growthof orogenic belts along the eastern margin of Gondwana (e.g., Gray andFoster, 2004). Plate convergence in an oceanic setting along the easternmargin of Gondwana from ca. 520 Ma to 340 Ma produced crustalthickening (e.g., Gray and Foster, 2004). In the Mesozoic, southeasternAustralia was the continental margin of the subducting Pacific plate.
The boundary between the Precambrian and Phanerozoic parts ofAustralia is often referred to as the Tasman Line. This was initiallydefined in Queensland as the boundary line between the craton andgeosyncline in early Palaeozoic times (Direen and Crawford, 2003;Hill, 1951). The definition was based on the limited outcrop of thePrecambrian rocks available at the surface. The recent reinterpretationby Direen and Crawford (2003) proposes that the Tasman Line is not asimple “line” associated to one single tectonic event: the Rodinia–Gondwana breakup, and they found no outcrop or geophysical evi-dence of rocks dating from the period 780 to 730 Ma at the locationsof the Tasman Line. The main locus of Rodinia breakup is possibly tothe east.
1.2. The use of receiver functions
The coda of teleseismic P waves contains considerable informationabout the structure of the Earth directly beneath a seismic station.The P-receiver function technique isolates P-to-S conversions generat-ed at crustal discontinuities beneath the recording site (Langston,1977, 1979) by source equalisation. Conversion primarily occurs atthe Moho, which represents one of the most significant interfaces(taking into account the possible presence of sedimentary layersbeneath some seismic stations) in terms of the variations in elasticproperties. The principle underlying the receiver function techniqueis to concentrate on conversions beneath the station by removal ofthe instrument response and the effect of passage through the bulk ofthe mantle through deconvolution of different components of groundmotion. In this study, we use the radial receiver function, obtained bydeconvolving the radial component (along the great-circle to thesource) by the vertical component, as a means of investigating the na-ture of the crust–mantle transition beneath southeastern Australia andcrustal thickness. Receiver functions are sensitive to shear velocity con-trasts of interfaces located beneath the seismic station, however, theyare only weakly sensitive to absolute velocities (e.g., Julià et al., 2000;Tkalčić et al., 2006). The S-velocity structure would be better resolvedwith a joint inversion of radial receiver function and surface wavedispersion observations (e.g. Julià et al., 2000; Tkalčić et al., 2006).
1.3. Seismic reflection data acquisition and processing
The set of seismic reflection profiles is fully described inDrummond et al. (2006), Fomin et al. (2006, 2010), and Cayley et al.(2011). The profiles have been recorded in a consistent way usingthe same equipment, with 18–20 s two-way time at a sample rateof 2 ms. Three in-line vibrators, each applying 250,000 N of peakforce, were used as the energy source. Three upsweeps of 12 s dura-tion using the Varisweep technique were used at each vibrationpoint. Vibroseis sweeps are typically in the range 6 to 96 Hz.Geophone groups of 12 phones were spaced at 40 m interval.
A typical processing sequence is (e.g. Drummond et al., 2006):
1) refraction statics corrections,2) velocity analysis,3) pre- or post-sort spectral equalisation,4) sort to common mid-point gathers,5) dip moveout corrections, stack,6) post-stack migration,7) band-pass filtering,8) and in some cases, semblance filtering before display.
1.4. Nature of the seismic boundary between the crust and the mantle:the Mohorovičić discontinuity
The Moho is commonly defined in seismology as the boundarythat marks an abrupt increase in P-wave velocity to values typical ofthe upper mantle (e.g., for the studies of the Australian region, seeClitheroe et al., 2000; Collins et al., 2003). However, different defini-tions have been used for the value of the upper mantle velocity:e.g. Vp>7.6 km/s (Clitheroe et al., 2000); Vp>7.8 km/s (Collins etal., 2003). The crust–mantle transition shows noticeable variationsin character across Australia, with both sharp and gradational transi-tions reported (e.g., Clitheroe et al., 2000; Kennett et al., 2011;Reading et al., 2007; Shibutani et al., 1996). The seismic discontinuitybetween the crust and the mantle may not always correspond to apetrologic boundary (e.g., O'Reilly and Griffin, 1985; and discussionin Clitheroe et al., 2000). Furthermore, the origin of the Moho couldbe multi-genetic for continental crust (e.g., Eaton, 2005). Reflectionand refraction profiles demonstrate that the crust–mantle boundarycan show considerable variation. Sometimes, there is no clear bound-ary at the base of the crust suggesting that the impedance contrastbetween crust and mantle is weak. The nature of the transition fromcrust to mantle is reflected in the nature of the reverberations andconversions occurring beneath the station. Thus, Zheng et al. (2008)have shown that a thick transition between the crust and mantle re-sults in diffused and weakened PpPms amplitude (the first Moho mul-tiple) in receiver functions, while a sharp crust–mantle boundarygenerates strong PpPms amplitude.
2. Data analysis
2.1. Information on station deployments
The seismic data are derived from a network of 28 portable threecomponent broad-band seismometers (Fig. 1) deployed for 9 monthsat a time for the SoCP (Southern Cratons to Palaeozoic) experimentand from 4 permanent broadband seismic stations from the Geosci-ence Australia network: BBOO, ARPS, TOO and STKA in the samearea. We also analyze data from one GEOSCOPE permanent seismicstation at Mount Stromlo (CAN) near Canberra. The data from thisstudy have been employed in the construction of the recent modelof Moho depth in Australia: AusMoho (Kennett et al., 2011). Herewe present a specific interpretation of the data in the light of theseismic measurements (receiver functions and reflection profiles)carried out in southeastern Australia.
115F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
2.2. Data and stacking
We used P waves from epicentral distances between 30 and 90°for events with mb≥5.5, hand-selected for good signal-to-noiseratio (SNR≥2) and cut the seismic traces at 5 s prior to and 30 safter the initial P-wave arrival. The seismograms were rotated on tothe azimuth between receiver and source. A Gaussian low pass filterwas applied to the waveforms before the deconvolution is performedto eliminate high frequencies (the Gaussian width is 2.5 rad/s) inorder to eliminate influence from small-scale crustal heterogeneities.The P radial receiver functions (RFs) were obtained by deconvolvingthe radial component, directed along the great-circle between sourceand receiver, with the vertical component (Fig. 2) in the time domain(Ligorria and Ammon, 1999). The RF waveform was sampled at a fre-quency of 10 Hz giving a total of 351 data samples. The radial RFs ateach seismic station were then stacked for a set of back-azimuths,with a narrow range of ray-parameters based on the followingprocedure:
(1) Select the quadrant (back-azimuths between N0° and N90°,N90° and N180°, N180° and N270°, N270° and N360°) withthe highest number of RFs.
(2) Compute pmedian: the median of the ray parameters of allseismic events in this interval.
(3) Select events with a ray parameter=pmedian±0.004 (s/km).Most data come from seismogenic belts surrounding Australiaand this narrows down the range of useful ray parameters.For example, the useful ray parameter range for station CANis between 0.067 and 0.075 s/km.
(4) Stack the RFs selected in the previous step. Only the best dataare used for stacking and we focused on obtaining the mostbasic information assuming a horizontally layered structure.
−0.20
0.20.40.6
SO
−0.20
0.20.4
−0.20
0.20.4
−0.20
0.20.4
−0.20
0.20.4
−0.20
0.20.4
−5 0 5 10−0.2
00.20.4
time
Fig. 2. Observed radial RFs and stacked RF for station SO07. Dashed lines show±1 standard dshown at the bottom.
We show in Fig. 2 an example of individual RF traces we usedfor the stack. Before each stack we examined the coherencyof individual RFs using the cross-correlation matrix approachfrom Tkalčić et al. (2011): we compute the cross-correlationcoefficients for each pair of RFs. Then RFs are selected orrejected using empirical values for the minimum coefficientof cross-correlation χ and the percentage of coherent RFs τ.More specifically, a single RF is deemed suitable for a stack ifcross-correlation coefficient is higher or equal to χ with atleast τ per cent of other RFs. In this study, for most stations,the small number of RFs (b50) results in χ=0.70 and τ around10% or less. In the case of station SO07, χ=0.85 (Fig. 2) and wefound insignificant difference of crustal thickness derived fromthe NA inversion of a single RF and the inversion of the stackedRF at the same station.
For station CAN, for which we have a larger number of RFs, weconsidered a higher value of χ: 0.90. The stacking process may inducea loss of some features of the crust–mantle boundary particularly as-sociated with the detection of dipping and anisotropic structures thatdepend on the back-azimuthal variation of the RF amplitude. Whenthe stacking bounds are tight in back-azimuth (≤10°) and epicentraldistance then the signal should not be degraded (e.g. Cassidy, 1992).At CAN we have such tight bounds, but for all other stations we couldmiss information relevant to a dipping Moho or the presence of crust-al anisotropy, as we are stacking over all RFs in one quadrant (90°variation of back-azimuth). Investigating the transverse RF variationwith back-azimuth can help to determine complexity beneath aseismic station (e.g., Frederiksen et al., 2003). However, unfortunatelywith a temporary seismic deployment of 9 months at a timefor the SoCP experiment it was not possible to obtain a broadback-azimuthal coverage in southeastern Australia. Due to this
07
15 20 25 30
stack
synthetic
±1σ
(s)
Event 1
Event 2
Event 3
Event 4
Event 5
eviation bounds around the linear stack. Synthetic RF resulting from the NA inversion is
116 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
limitation we only consider isotropic and planar discontinuity tocompare the results from all seismic stations across southeasternAustralia.
2.3. Neighbourhood Algorithm inversion for 5 crustal layer and 1 mantlelayer models
We characterise the crust and uppermost mantle structurewith a 6-layer 1-D seismic velocity model: sediment layer, basementlayer, upper crust, middle crust, lower crust, and uppermostmantle (Shibutani et al., 1996). A nonlinear inversion method, theNeighbourhood Algorithm (NA, Sambridge, 1999a, 1999b) wasemployed to match the observed radial RFs. During the inversion,the synthetic radial RF was calculated using the Thomson–Haskell ma-trix method (Haskell, 1953, 1962; Thomson, 1950). The full effects offree-surface reverberations and conversions were modelled. The NA isa direct-search method of inversion. It has the capacity to search effi-ciently by sampling simultaneously different regions of parameterspace. This inversion technique samples regions of a multidimensionalparameter space that have acceptable data fit. The NAmakes use of sim-ple geometrical concepts to search a parameter space. At each iteration,the entire parameter space is partitioned into a set of Voronoi cells(Voronoi, 1908) constructed about each previously sampled model. Inour case Voronoi cells are nearest neighbour regions defined by anL2-norm. The initial sets of samples are uniformly random, but as itera-tions proceed only a subset of chosen Voronoi cells are re-sampled(using a random walk within each cell). This allows the NA to concen-trate where data misfit is lowest. A further advantage of the NA overother direct-search methods is that only the rank of the misfit functionis used to compare models. This is of particular significance as it avoidsthe problems associated with scaling of the misfit function and allowsany type of user-defined misfit measure to be employed. The NArequires just two control parameters:
– ns, which is the number of models generated at each iteration– and nr which is the number of neighbourhoods re-sampled at each
iteration.
After several trials, we specify a maximum number of iterations:900 and ns=100 for the first iteration and ns=25 for all otheriterations and nr=25 for all iterations. The a priori distribution ofthe misfit function is assumed to be Gaussian (e.g. Sambridge,1999a, 2001). In this study, we used a chi-square metric similar tothat employed by Sambridge (2001) to compute the misfit function.The misfit function, an L2-norm, is defined as the sum of the squaresof the difference between the observed amplitude of the radial RFand the amplitude of the synthetic radial RF from a 6-layer model.
One of the advantages of the NA over other direct search methodsis that only the rank of the misfit function is used to compare modelsand this enhances the exploration of parameter space to find suitablemodels. Fig. 3a shows an example of the misfit function obtainedusing the NA at station SO07. The NA inversion provides a 1-D shearwavespeed model and an estimate of the Vp/Vs. However, neitherthe Vp/Vs ratios nor the absolute velocities are strongly constrainedin the NA inversion, which is most sensitive to the presence of discon-tinuities. During the inversion, as in the work of Shibutani et al.(1996) the model was parameterised in terms of 6 layers with inter-nal velocity gradients and the possibility of discontinuities at theboundaries. This leads to the 24 parameters, the Vs values at the topand bottom of the gradient zone, the thickness of the gradient zoneand the Vp/Vs ratio in each zone. We used similar bounds for the 24parameters to those of Shibutani et al. (1996) and Clitheroe et al.(2000) (Table 1). The NA method combines a Monte Carlo searchtechnique and the properties of the Voronoi geometry in parameterspace to find an ensemble of the best fitting models and performs aglobal optimization. We present density plots of the best 1000 datafitting S-velocity models generated by the Neighbourhood Algorithm
in Figs. 3b and 4. The model with the best fit to the data is plotted inred. Sambridge (1999b) has used a Bayesian approach to estimateresolution and confidence intervals on the 24 parameters in the RF in-version. The layer thickness distribution and the S-velocity distribu-tion are better constrained by the NA inversion than the Vp/Vs ratio,thus we place less reliance on the Vp/Vs ratio. Sambridge (1999b)showed using marginal probability density functions that the Vp/Vsratio from the NA inversion is better resolved in the first layer andpoorly resolved in the remaining layers. The inclusion of the Vp/Vsratio serves primarily to allow for some of the effects of the sedimen-tary layer beneath the stations with no a priori information(Bannister et al., 2003).
3. Results
3.1. Crustal thickness and nature of the Moho
3.1.1. Crustal thicknessWe assume that the base of the transition to mantle velocities
defines the Moho depth, in order to be in accordance with previousRF studies (e.g., Clitheroe et al., 2000) which produced Moho depthsclose to estimates from seismic refraction studies (Collins, 1991;Collins et al., 2003). Here, we take the upper mantle velocity to beVp≥7.6 km/s following Giese (2005), which means that Vs≥4.3–4.4 km/s for Vp/Vs ratio in the range 1.73–1.77 at the base of thegradient. Fig. 4 presents the shear wave velocity models from theNA inversion and data fits at SO01, CAN, MUR5 and MG01. TheS-velocity models from the inversion at the other seismic stationsare available as an electronic supplement. The results of the RF anal-ysis from the NA inversion are summarised in Table 2 and Fig. 5aand b. The uncertainty in Moho depth is ±2 km. To assess the accu-racy of our results compared to other styles of RF analysis, we alsoperform the same analysis at one seismic station in Hawaii: KIP andobtained a small difference of 1 km of crustal thickness compared toLeahy et al. (2010) who used an improved H–K stacking approach.We obtain a Moho depth to be around 18 km from the RF inversionwhereas Leahy et al. (2010) obtained a Moho at around 19 kmdepth beneath this station.
3.1.1.1. Lachlan Fold Belt or Lachlan Orogen. The Moho depths show agenerally thick crust ranging from 34 up to 51 km. These values areclose to those obtained by Clitheroe et al. (2000) and Collins et al.(2003) at several localities. Taking into account crustal thickness esti-mates from previous studies (Clitheroe et al., 2000; Collins, 1991;Collins et al., 2003; Shibutani et al., 1996), the average crustal thick-ness is thicker beneath the Lachlan Orogen ~43 km than beneaththe cratons. The lower crustal structure obtained at CAN correspondsto a broad velocity transition zone at the Moho and a crustal thickness~48 km. This model is fully consistent with previous refraction and RFstudies (Collins et al., 2003).
The Moho depth obtained from RFs at station ARPS (34 km) isclose to the value (around 35 km) estimated by Korsch et al. (2002)using a seismic-reflection transect close to this seismic site.
3.1.1.2. Murray Basin. The crust is thinner than elsewhere in thestudied area. The average crustal thickness is ~32 km. The crustalthickness varies between 28 km (for MUR3) and 35 km (for MUR5).The highest crustal thickness is observed at MUR5 and this may bedue to the proximity of the Lachlan Orogen.
3.1.1.3. Gawler Craton and Curnamona Province. Our study suggests acrustal thickness ranging from 28 to 45 km. The crust is thickerthan east of the cratons in the Delamerian Orogen and in the MurrayBasin. The Gawler Craton and the Curnamona Province show a thickcrust (≥40 km) at stations: SO07, BBOO, SO08, SO10, SO11, SO12,SO13 and SO17. The values ≥40 km observed at SO07, BBOO and
0
1
2
3
4
5
chi−
squa
re d
ata
mis
fit
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000
Number of solutions
SO07a)
NO. OF MODELS OF THE BEST 1000
1 10 100 1000
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0.0 1.0 2.0 3.0 4.0 5.0
Dep
th (
km)
H=42 km
Model: SO07S velocity (km/s)
b)
Fig. 3. a) Misfit function against number of models for which the forward problem has been solved using the NA approach. b) The 1-D shear wave velocity models obtained by theinversion of teleseismic RFs at SO07. All 22,600 models searched in the NA inversion are shown as gray shaded area. The best 1000 models are shown in yellow and green, with thecolour being logarithmically proportional to model number. The colour scale shows the increase in data fit from yellow to green. A solid red line represents the best fitting model.0 km depth corresponds to the station elevation. H is the Moho depth value.
117F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
SO08 are located on the Gawler Craton whereas the estimates forSO12 and maybe SO17 are associated with the Curnamona block.The crustal thicknesses at SO07, BBOO and SO08 are greater than40 km and could be related to their location in the Gawler Craton. In-terestingly, the crustal thickness decreases at SO08 (Fig. 5a) and thenincreases at SO12 and SO17 (Fig. 6). We also note that the shear wave
velocity variations are similar beneath BBOO and SO08 (Fig. 6) anddifferent from the structure obtained at SO09. The 1-D shear wavevelocity models show similar trends beneath SO17 and SO13. The ve-locity structure of the crust at SO17 is more complex than at SO12.This difference may be due to the fact that the Curnamona Provincedoes not extend to SO17. The average crustal thickness is found to
Table 1We have modelled the crust and uppermost mantle using 6 layers down to 70 km.Model parameter space bounds used in the NA receiver function inversion. Vsupper de-notes the S velocity at the top of a layer and Vslower that at the bottom of a layer (km/s),Vp/Vs denotes the velocity ratio in a layer.
Layer Thickness (km) Vsupper (km/s) Vslower (km/s) Vp/Vs
Sediment 0–2 1.0–3.0 1.0–3.0 2.00–3.00Basement 0–3 1.5–3.5 1.5–3.5 1.65–2.00Upper crust 1–15 2.6–3.6 2.8–4.0 1.65–1.80Middle crust 5–20 3.2–4.5 3.2–4.5 1.65–1.80Lower crust 5–20 3.2–4.5 3.2–4.5 1.65–1.80Mantle 5–30 2.8–5.0 2.8–5.0 1.70–1.90
118 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
be around 39 km beneath the Precambrian Gawler and Curnamonaunits. The average Moho depth obtained in this study for theCurnamona Province: 39.5 km (for SO12 and SO16) is similar to theresults from seismic reflection experiments in the same region(across the Broken Hill and Olary domains) around 39 km (Gibsonet al., 1998; Goleby et al., 2006). Recent reflection profiles in theCurnamona region indicate little changes in the depth to the crust–mantle boundary despite contrasts within the crust.
3.1.1.4. Mount Gambier region. The average crustal thickness is 44 km.Significant crustal thickness changes (from 37 to 47 km) are observedin Mount Gambier region over relatively short distances — around70 km. Stations MG01 and MG06 show a thick crust ≥40 km. Thesmaller crustal thickness estimated at MG02 could be due to theproximity to the coast and thus the zone affected by the breakup ofAustralia and Antarctica. We note a relatively good correlationbetween Moho depth and the surface topography (Fig. 5a).
The complexity of the RF waveforms for 6 of seismic stations(MG03, MG04, MG05, MUR1, MUR4 and SO14) prevented us fromusing the NA inversion to obtain accurate information about veloci-ties and Moho depths. Large-amplitude ringing can be generated bya low-velocity layer such as a sedimentary layer underlain by a strongcontrast in wave speed (e.g., Clitheroe et al., 2000). The ringing maybe due to near surface limestones with surface sediments. Weobserved a jump in Vp/Vs in the topmost layer of the best models atstations MUR5 and MG06 with ratios higher than 2. This jump isprobably due to a sedimentary layer below the sensor and was al-ready observed by Shibutani et al. (1996) in Eastern Australia fortwo stations located on sedimentary basins (SB01 and SA07). Similarhigh Vp/Vs values (higher than 2) related to sediments were alreadydescribed in the literature (e.g. Morozov et al., 2002; Yang et al.,2012; Zelt and Ellis, 1999). For some seismic stations a sedimentarylayer was detected, but we were also able to detect the Mohodepth (see for instance at MUR5). At MUR3 seismic station (a lessconstrained location), a high Poisson's ratio in the top layer is respon-sible for strong reverberations, which mask phases such as multiples.The velocity model resolution is reduced in comparison to stations onbedrock. Depth to major crustal velocity discontinuities and the Mohoare the best-constrained features of this kind of model.
3.1.2. Nature of crust–mantle transitionExcept for the stations installed in the Murray Basin, most sites
show a crustal thickness greater than 36 km. Using the character ofthe crust–mantle transition in the results of the modelling (Fig. 4e,f, g and h) we classify the Moho transition zone as sharp if its thick-ness is ≤2 km, intermediate for 2–10 km, and broad if it is ≥10 km,as suggested by Shibutani et al. (1996). Our Moho estimate willlie at the base of a gradient (in conformity with earlier work(e.g., Clitheroe et al., 2000)).
Fig. 4. a, b, c and d) Comparison between the observed average and the predicted radial RFsvelocity models obtained by the inversion of teleseismic RFs at SO01, CAN, MUR5 and MG0
Except for two seismic stations SO15 and STKA, which are close tothe limit between the Precambrian and Phanerozoic terranes, mostseismic stations located in the Precambrian region have an intermedi-ate crust–mantle transition zone. All seismic sites in the Eyre Peninsula(Fig. 5) in the Gawler Craton show an intermediate nature (SO07,BBOO, and SO08).
The crust–mantle transition at MUR3 (in the Murray Basin) issharp whereas it is intermediate at MUR2 and broad at MUR5. Thestations MG01 and MG06 show a thick crust ≥40 km with a broadcrust–mantle transition, whereas the crust is thinner at MG02(37 km) with a sharp Moho transition zone. The crust–mantle bound-ary is deep and mostly intermediate in character beneath the LachlanFold Belt.
3.2. Calibration of reflection profiles
In recent years a number of reflection profiles have been carriedout in southeastern Australia that provide valuable continuity be-tween localised crustal properties that we can extract from RFs. How-ever, there is limited velocity control on these full-crustal reflectionprofiles. As in the work of Kennett et al. (2011), we estimate Mohodepth from the reflection results using an average crustal velocity of6000 m/s, which gives a depth conversion of 3 km per 1 s two-waytime (TWT). Kennett et al. (2011) showed that such a simple approx-imation provides good agreement between seismic reflection and RFobservations in western Australia (Reading et al., 2007, 2011) withsmall discrepancy (around 1 km) and also between seismic refractionand RF measurements in southern Queensland. The influence of dif-ferent stacking velocities is described by Jones et al. (2005). Theyfound that for full crustal reflection images the nature of the recordsshows little dependence on the stacking velocities. We compare theresults from the reflection work with the RFs from the stations closestto the profiles.
A west–east seismic profile was carried out near the MountGambier region in 2009: GA-09-SD1 (Figs. 7 and 8a). A significantthickening of the crust is observed towards the Mount Gambier areaat the west of the line where the Moho reflection is at ~15 s (Cayleyet al., 2011). This deep Moho is very similar to that found from theRF at MG01 station (47 km). There is a relatively rapid change inMoho depth to the east, shallowing to 33 km which is also consistentwith the depth of 34 km obtained at ARPS.
A group of reflection profiles have been carried out to investigatethe crustal structure of the Gawler Craton and the CurnamonaProvince and the linkages between them. In 2003, a north–south re-flection profile was acquired in the Gawler Craton, near the majorore-deposit at Olympic Dam. The seismic data from the north–southseismic profile, 03GA-OD1, are represented in Fig. 8b (Drummond etal., 2006). The bottom of the reflection Moho is around 14 s TWT(~42 km) in the northern zone and 13 s TWT in the southern zonesuggesting a crustal thickness of ~39–40 km. In 2008, an east–westreflection transect (08GA-G1) was performed in the Eyre Peninsula,Gawler Craton in order to investigate the crustal architecture of theeastern part of the Gawler Craton (Figs. 7 and 8c). The central andwestern parts of the profile show a Moho at ~14 s TWT: 42 km(Fraser et al., 2010) whereas the eastern part indicates a Moho at~41 km (Fig. 8c). In 2009, an east–west transect (09GA-CG1) was ac-quired across the Flinders Ranges, South Australia. The primary goalof this survey was to elucidate the relationship, at depth, betweenthe Curnamona Province and the Gawler Craton (Preiss et al., 2010).The Moho is poorly imaged (Fig. 8d). At the eastern end of the linethe lower crust is mostly nonreflective. Weak reflections suggest aMoho depth ~13 s TWT (39 km). At the western end of the line
from the NA inversion at SO01, CAN, MUR5 and MG01; e, f, g and h) the 1-D shear wave1. Description as in Fig. 3b.
b)
datamodel
Rad
ial R
ecei
ver
Func
tion
Time (s)
CAN
Dep
th (
km)
Dep
th (
km)
a)
datamodel
SO01
S velocity (km/s)e)
d)
H=47 km
H=48 km−0.2
0.0
0.2
0.4
0.6
0.8
−4 −2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 NO. OF MODELS OF THE BEST 1000
1 10 100 1000
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0.0 1.0
Pms
Pms
Rad
ial R
ecei
ver
Func
tion
−0.2
0.0
0.2
0.4
0.6
0.8
−4 −2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Time (s)
c)MUR5
Dep
th (
km)
H=35 km
1
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0.0 1.0
Pms
−0.2
0.0
0.2
0.4
0.6
0.8
−4 −2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 NO. OF MODELS OF THE BEST 1000
1 10 100 1000
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0.0 1.0
Rad
ial R
ecei
ver
Func
tion
Dep
th (
km)
H=47 km
Pms
Time (s)
MG01
datamodel
datamodel
f)
g)
h)
−0.2
0.0
0.2
0.4
0.6
0.8
−4 −2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Time (s)
Rad
ial R
ecei
ver
Func
tion
NO. OF MODELS OF THE BEST 1000
1 10 100 1000
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
0.0 1.0
Model: CANS velocity (km/s)
Model: SO01
2.0 3.0 4.0 5.0
Model: MUR5S velocity (km/s)
NO. OF MODELS OF THE BEST 1000
10 100
2.0 3.0 4.0 5.0
2.0 3.0 4.0 5.0
Model: MG01S velocity (km/s)
2.0 3.0 4.0 5.0
119F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
Table 2Station location and results of the RF modelling with the NA inversion. Inversion qual-ity (Quality) is assessed visually and with the misfit value. SO02 seismic station provid-ed no clear RFs. Note that SO06 and CAN station were both at Mount Stromlo. H is thecrustal thickness and N is the number of RFs used for the stacking at each station.Nature represents the character of the Moho from the shear-wave velocity modelobtained at each station: thin ≤2 km, intermediate 2–10 km and broad >10 km.
Station Latitude Longitude H (km) N Quality Nature
ARPS −36.7699 141.8383 34 17 1 IntermediateCAN −35.3210 148.9990 48 29 1 BroadMG01 −36.6541 140.4311 47 12 0.5 BroadMG02 −37.0663 139.8564 37 2 0.5 ThinMG06 −37.8647 140.8707 40 8 1 BroadSTKA −31.8769 141.5952 42 2 0.5 ThinTOO −37.5714 145.4906 34 2 0.5 IntermediateBBOO −32.8096 136.0583 45 11 1 IntermediateMUR2 −32.9316 145.5985 32 5 1 IntermediateMUR3 −33.9768 145.0808 28 13 0.25 ThinMUR5 −34.3642 146.2668 35 10 1 BroadSO01 −33.4357 146.4332 47 9 1 IntermediateSO03 −33.2548 148.8480 35 8 1 BroadSO04 −34.2847 149.0345 36 4 1 IntermediateSO05 −34.8280 147.7299 43 3 0.5 BroadSO06 −35.3188 148.9963 45 8 1 BroadSO07 −33.7839 135.8373 42 5 1 IntermediateSO08 −32.5177 136.5128 40 7 1 IntermediateSO09 −31.6692 137.0613 42 5 1 IntermediateSO10 −29.7119 137.0625 40 7 1 IntermediateSO11 −31.9705 138.0309 46 8 0.5 BroadSO12 −31.0911 138.6775 43 5 1 IntermediateSO13 −29.8957 139.7331 40 5 1 IntermediateSO15 −32.3197 139.8651 36 3 1 ThinSO16 −31.5931 140.7935 36 15 1 BroadSO17 −30.2730 139.3408 45 3 1 Intermediate
120 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
moderate reflections occurred in the lower crust at ~14 s TWT(36 km).
The discrepancies between the Moho depth determined from re-flection profile and crustal thickness obtained from RFs at nearbyseismic stations are in general less than 2 km (Fig. 8). Thus, thegood agreement between the two techniques suggests that using a6 km/s P wave velocity at the base of the crust provides reliable esti-mates of Moho depth.
4. Discussion
4.1. Tectonic interpretations of crustal thickness variations and nature ofthe Moho for the ensemble of stations
The average crustal thickness obtained in the present study fromthe broad-band station in southeastern Australia is 39 km (and themedian value is also 39 km). This value is similar to: (i) the 38 kmaverage thickness of continental crust in the CRUST 5.1 model ofMooney et al. (1998), a model of global crustal thickness based onprevious seismic refraction data, (ii) the value of 38.8 km computedby Clitheroe et al. (2000) from earlier RF analyses across the wholeAustralian continent, and (iii) the value of ca. 38 km for continentalcrust in Australia deduced by Collins et al. (2003) using a compilationof both RF studies and seismic refraction. These crustal thickness esti-mates have been incorporated into the compilation of Kennett et al.(2011) that provides a significant improvement over the resolutionof previous Moho maps (e.g., Clitheroe et al., 2000; Collins et al.,2003; Mooney et al., 1998). The individual values add considerabledetail for Southern Australia.
We assess the uncertainty of our estimates of Moho depths by usingthe density plots of the best 1000 data fitting shear-wave velocitymodels (Fig. 4e, f, g and h) and the nature of crust–mantle transition(if it is sharp, intermediate or broad). The estimated error in Mohodepths at all seismic stations is less than ±2 km. The crust–mantleboundary is the deepest beneath the Lachlan Fold Belt suggesting the
presence of underplating materials at the base of the crust. Interesting-ly, even though the Gawler Craton contains very old material before2560 Ma (Swain et al., 2005), we also note a thick crust ≥40 kmbeneath all seismic stations located in the Gawler Craton. This thickcrust may be a characteristic of the Gawler Craton crust.
4.1.1. Nature of crust–mantle transitionWe could not observe clear multiples on the RFs at most of our
seismic stations. Chevrot and van der Hilst (2000) also pointed outthe absence of clear multiples in our region of study. This feature sug-gests the absence of sharp seismic interface at the crust–mantleboundary beneath this region.
4.1.2. Gawler Craton and Curnamona ProvinceThe intermediate nature observed beneath the Gawler Craton is
consistent with previous result from Clitheroe et al. (2000) for thestation BBOO. Such intermediate nature of the crust–mantle transi-tion suggests magmatic underplating at the base of the crust. Thepresence of magmatic underplating is also suggested beneath othercratons.
The mafic underplating could result from two different processes:
i) a deep crust delamination with a change of the nature of crust–mantle transition due to lithospheric mantle partly or entirelyadded from the ambient convecting mantle.
ii) passage of the continent overtop a mantle upwelling resulting inthe thickening of the crust–mantle transition.
4.1.3. Evidence of intermediate or broad transition beneath other cratonsZandt and Ammon (1995) using measurements of crustal Poisson's
ratio at all continent except Antarctica suggested that magmaticunderplating could be involved in the formation of all cratons. Alower-crustal 10–25 km thick high velocity layerwas observed beneaththe Archean Wyoming Craton from seismic refraction profile (Eaton,2005; Gorman et al., 2002) at the south end of the profile. This layerwas interpreted as thick magmatic underplating. An intermediatenature of crust–mantle transition was also observed beneath theProterozoic North Australian Craton and interpreted as a result of thepresence of underplating in the lower crust possibly associated withsubduction (Clitheroe et al., 2000). A thick crust–mantle transition ofabout 5–10 km was also observed beneath the Taihangshan region inthe northern North China Craton (Zheng et al., 2006, 2008). The NorthChina Craton is one of the world's oldest Archean cratons. This transi-tion was interpreted as the result of slow secular interaction betweenthe crust and the underlying mantle, and underplating at the base ofthe lower crust. On the other hand, a sharp crust–mantle boundary isimaged beneath the Yanshan region in the northern North ChinaCraton (Zheng et al., 2006, 2008) and it is interpreted as regional delam-ination of the deep crust after orogenic thickened and/or magmaticunderplating.
4.1.4. Mount GambierRecent reflection profiles near MG01 (Fig. 8a) show a complex
structure beneath this region with a rapid and substantial thickeningof the crust towards the location of MG01 and a possible imbricatestructure (with an apparent duplication) at the base of the crust.This class of structure may explain the complexity in the observedRFs for MG01 and could be the reason for only a moderate fit(Fig. 4d) with 1-D synthetics. The thick crust in the Mount Gambierregion suggests the extension of the Gawler Craton beneath MountGambier and the imbricate structure may be linked to convergenceduring the Cambrian age Delamerian Orogeny.
4.1.5. Murray BasinAt MUR2 and MUR5, no clear PpPms arrivals are visible on the RFs,
which is consistent with a thick Moho transition zone beneath these
136˚ 140˚ 144˚ 148˚ 152˚
−40˚
−38˚
−36˚
−34˚
−32˚
−30˚
−28˚
36
32
36
3544
38
41
44 36
39
28
36
31
37
36
25
30
50
454251
52
39
44
43
34
35
35
36
36
34
4747
474542
39
29 27
31 31
36
1218
30
3541
32
39
16
15
39
37
42&36
28
43
40
36
34
34
45&40 3246
3642
40
40
45
48&4947
40
35
35
43
4240 36
−400 −200 0 100 200 400 600 800 1000 1200 1400 1600 1800 2200
crustal thickness (km)
b)
a) elevation (m)
this study previous studies
refraction
sharpintermediatebroad
sharpintermediatebroad
EyrePeninsula
342942
10
15
20
25
30
35
40
45
50
55
136˚ 140˚ 144˚ 148˚ 152˚
−40˚
−38˚
−36˚
−34˚
−32˚
−30˚
−28˚
FlindersRange
SnowyMountains
this study: previous RFs:
Fig. 5. a) The Moho values are superimposed on the surface topography (e.g., Smith and Sandwell, 1997). The Moho depths from this study are shown in red. Previous crustal thick-ness estimates from seismic refraction and reflection (Collins, 1991; Collins et al., 2003) and RF studies (Clitheroe et al., 2000; Shibutani et al., 1996) are also presented in black forcomparison. When a previous estimate was available at one of the SoCP seismic stations the value is shown in italics. The nature of the Moho from the shear-wave velocity modelobtained at each station is indicated by different symbols: a circle for thin Moho transition zone, a diamond for intermediate zone and a square for broad zone. The estimate of thewidth of the Moho transition zone is not made at the sample points of Collins (1991) and Collins et al. (2003). Refraction and reflection values are plotted as empty triangles.b) Location map of the depth of the crust–mantle seismic discontinuity beneath southeastern Australia. Two different symbols are used: stars represent location and crustalthicknesses from this study and octagons show results from previous studies (Clitheroe et al., 2000; Collins, 1991; Collins et al., 2003; Saygin, 2007; Shibutani et al., 1996).
121F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
two stations. The intermediate and broad nature of the crust–mantletransition beneath MUR2 and MUR5 could be due to the proximity ofthese two stations with the Lachlan Orogen and of underplating. Thesharp nature of the crust–mantle transition observed at MUR3suggests direct contact of the crust on top of the upper mantle withno underplating materials.
4.1.6. Lachlan OrogenOur results are consistent with previous observations (Clitheroe et
al., 2000; Collins et al., 2003; Shibutani et al., 1996). The variations inthe crustal thickness and the intermediate and broad transition be-tween crust and mantle beneath the Lachlan Orogen may be relatedto the presence of magmatic underplating at the base of the crust
SO07 BBOO SO08 SO09 SO12 SO17 SO13
−70
−60
−50
−40
−30
−20
−10
0
Dep
th (
km)
surface
A A'
Precambrian cratonsSW NE
Moho
S wavevelocity
H=42 km H=45 km H=40 km H=42 km H=43 km H=45 km H=40 km
Fig. 6. Synthesis of 1-D shear wave velocity models derived from teleseismic earthquakes across the Precambrian region lithosphere. The red lines are the average models of the best100 fitting models derived for each station in the NA inversion. The Moho is interpreted to be the base of the high-velocity gradient zone, shown in black shading. See Fig. 1 for thelocations of the interpreted sections A–A′. A dashed line connects seismic discontinuities that are similar beneath two adjacent seismic stations.
crustal thickness (km)
this study
previous RFsestimatesseismic reflection
TL17
D D’
C
C’
E’ E
B B’
136˚ 140˚ 144˚ 148˚ 152˚
−40˚
−38˚
−36˚
−34˚
−32˚
−30˚
−28˚
10
15
20
25
30
35
40
45
50
55
Fig. 7. Comparison between Moho depth from reflection profiles and estimates from RFs. Letters refer to the segments of reflection profile shown in Fig. 8.
122 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
(e.g., Clitheroe et al., 2000; Drummond and Collins, 1986; Shibutani etal., 1996). Interestingly, the tomographic model from Rawlinson et al.(2010) shows an increase of P-wavespeed at the SO01 location andthe authors interpret the high velocity zone as a result of the presenceof magmatic underplating. This mafic magmatic underplating mayhave occurred in a backarc region of a subduction zone as a result ofhigh degrees of adiabatic decompression melting of the asthenosphere(Collins, 2002).
4.2. Moho topography and other evidence for complex structure
The profile of average 1-D shear wave velocity models across thePrecambrian region shows a dipping structure (Fig. 6) with a dippingMoho towards the south of the Gawler Craton.
The trend of the mountain belt in the Lachlan Orogen (SnowyMountains) is approximately N–S (Fig. 5a) east of longitude 148°E.The crustal thickness tends to increase as the surface topography
Fig. 8. Four examples of reflection profiles showing the position of the Moho (Hr). Estimates owith the seismic reflection measurements. See Fig. 7 for the locations of the reflection trand) 03GA-CU1: W–E profile.
increases east of longitude 138°E (Fig. 5). This direction is close tothe elongation direction of Moho depth surface for most of the Lach-lan Orogen (Fig. 9 of Clitheroe et al., 2000). This dipping Moho couldbe related to a crustal thickening to the west due to the slab conver-gence or to the crustal thinning towards the east induced by the con-tinent–ocean transition.
4.3. Interpretation of receiver function measurements and reflectionprofile
The small difference in Moho depth estimates between RF and re-flection surveys is likely to arise from the different way that they sam-ple the crust. RFs exploit conversion of P-to-S wave generated intransmission at crustal discontinuities beneath the recorder. Whereasthe reflection method uses P wave that are reflected from seismicdiscontinuity. RFs are most sensitive to sharp velocity contrast andsensitivity to smooth velocity gradients comes from the longer
f crustal thickness (H) from RFs at nearby seismic stations are also shown for comparisonsects. a) 09GA_SD1: W–E profile. b) 03GA-OD1: N–S profile. c) 08GA-G1: W–E profile.
B B’
C C’
a)
b)
0
2
4
6
8
10
12
14
16
18
20
Moho
0
2
4
6
8
10
12
14
16
18
20
03GA-OD1: N-S profile
09GA_SD1: W-E profileMG01:H=47 km ARPS:H=34 km
SO10:H=40 km SO09:H=42 km
31 km 71 km
26 km 54 km
51 km 99 km
41 km 27 km
Hr~44 km
Hr~33 km
Hr~42 km Hr~39 km
Hr~41 km
Hr~36 km
Hr~39 km
Hr~42 km
Moho
Moho
Moho
Moho
Tw
o-w
ay ti
me
(s)
Tw
o-w
ay ti
me
(s)
0
20 km0
20 km
D D’c)
EE’
d)
024
6
8
1012141618
20
Tw
o-w
ay ti
me
(s)
0
2
4
6
8
10
12
14
16
18
20
Tw
o-w
ay ti
me
(s)
Moho
Moho
08GA-G1: W-E profile
03GA-CU1: W-E profile
BBOO:H=45 km SO08:H=40 km
SO16:H=36 kmTL17:H=42 km
19 km0
20 km0
1
16
Moho
123F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
124 F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
wavelength components of the incident wave. Reflection surveys arecarried out with much higher frequencies and pick up minorcontrasts in acoustic impedance (ρV, where ρ is the density and Vis the P-wave velocity). The reflection studies have better verticalresolution than our RF study because the dominant frequency of theseismic signal from the RF is around 0.4 Hz, whereas it is around40 Hz for the reflection profiles (e.g. Drummond et al., 2006). Forseismic reflection studies, the vertical resolution attainable at Mohodepth would be about 100 m. For our RF study, considering a domi-nant frequency of ca. 0.4 Hz a transition over 2 km will be indistin-guishable from an abrupt reflection. In general the interpretation ofthe reflection Moho is made at the base of a crustal reflectivity, andthis need not have direct correspondence to the features seen intransmission. Indeed the Fresnel zone, a measure of the likely resolv-ing length, is much different from the two methods at the base of thecrust. With a crustal thickness of 50 km, the Fresnel zone for RFs(~13 km) is much larger than for reflection measurements (lessthan 3–4 km) (Fomin et al., 2006).
4.4. Phanerozoic versus Precambrian regions
The Moho depth in the Mount Gambier volcanic area is thick~41 km suggesting that the limit between the Delamerian and westernLachlan orogens could be located east of Mount Gambier. Reflectionprofiles also indicate a very thick crust beneath Mount Gambier.Cayley et al. (2011) suggest the presence of Palaeoproterozoic–Cambrian ultramafic granulites at the base of the crust from interpreta-tion of these recent reflection profiles. A southern extension of GawlerCraton material may reach as far as the Mount Gambier.
Drummond and Collins (1986) suggested from seismic refractionstudies a correlation between increasing Moho depths and provinceage (excluding the Archean provinces). Clitheroe et al. (2000) ob-served a weak correlation using a RF analysis. The Lachlan Fold Beltor Lachlan Orogen was formed in the middle Palaeozoic (from 450to 340 Ma) and the age of the western part of the Precambrian regionis dominantly Proterozoic with an Archean core to the Gawler Craton.The crustal thickness of the Lachlan Fold Belt is thicker than beneaththe Proterozoic province (Fig. 5). On the west of the Lachlan Fold Beltlies the Delamerian Orogen from early Palaeozoic (550 to 470 Ma).On the east side is the New England Orogen, which was formedfrom late Palaeozoic to Early Mesozoic (310 to 210 Ma). Even in theLachlan Fold Belt no clear trend is observed between surface agesand Moho topography. As pointed out by Clitheroe et al. (2000)such correlations are difficult to justify because they rely on the as-sumption that the Earth surface age is correlated to the age of thelast crustal growth event at the base of the crust.
5. Conclusion
Receiver function modelling of broad-band seismic waveformsrecorded in southeastern Australia has been employed to constrainthe way in which crustal structure changes from the Precambrian cra-tons to the Palaeozoic fold belt. Our estimates of Moho depths rangebetween 28 and 48 km, with a mean value equal to 39 km, similarto Moho depths for global average of continental crust (38 km). Wesee a transition from a 40 km depth to Moho in the Precambrian,thinning beneath the Murray Basin and then thickening again in theexposed Lachlan Fold Belt. This leads to a general correlation betweenthe surface topography and the Moho depths. The intermediatenature of crust–mantle transition beneath the Gawler Craton andthe Lachlan Orogen suggests the presence of magmatic underplatingat the base of the lower crust.
We have shown that our RF estimates for Moho depth are consis-tent with recent reflection profiles and provide a detailed tie andcalibration of the methods. These detailed results provide strong
support for the approach used by Kennett et al. (2011) to produce acontinent wide map of the Moho.
The very thick crust in the Mount Gambier region obtained fromboth RF studies and reflection data suggests that the limit betweenthe Delamerian and western Lachlan orogens is located east ofMount Gambier. Our results therefore support the interpretation ofDireen and Crawford (2003) and place the Tasman Line well to theeast of the location favoured by many authors (e.g., Scheibner andVeevers, 2000).
Acknowledgements
We thank M.S. Sambridge for his inversion program and A.Arcidiaco and Q. Li for their help with the data. We thank G. Barruol,M.N. Vitry, A. Reading and R. Rambeloson for fruitful discussions. Weare grateful to C.D.N. Collins for access to his data. We would also liketo thank members of the RSES group (M. Salmon, P. Arroucau, N.Rawlinson, M. Heintz, T. Percival, C. Sinadinovski) for help in thefieldwork deployments and to the landowners for their support. Theconstructive comments of an anonymous reviewer and G. Leahyhave been appreciated. We are grateful to the developers of GMT soft-ware (Wessel and Smith, 1991). CAN station is maintained by bothGEOSCOPE (IPGP) and the Australian National University. BBOO,ARPS, TOO and STKA are maintained by Geoscience Australia. Wethank the Commonwealth of Australia and Geoscience Australia forthe reflection profiles.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.tecto.2012.09.031.
References
Bannister, S., Yu, J., Leitner, B., Kennett, B.L.N., 2003. Variations in crustal structureacross the transition from West to East Antarctica, Southern Victoria Land.Geophysical Journal International 155, 870–884.
Blackburn, G., Allison, G.B., Leaney, F.W.J., 1982. Further evidence on the age of tuff atMt Gambier, South Australia. Transactions of the Royal Society of South Australia106, 163–167.
Braun, J., Pauselli, C., 2004. Tectonic evolution of the Lachlan Fold Belt, southeasternAustralia: constraints from coupled numerical models of crustal deformation andsurface erosion driven by subduction of the underlying mantle. Physics of theEarth and Planetary Interiors 141, 281–301.
Cassidy, J.F., 1992. Numerical experiments in broadband receiver function analysis.Bulletin of the Seismological Society of America 82, 1453–1474.
Cayley, R.A., 2011. Exotic crustal block accretion to the eastern Gondwanaland marginin the Late Cambrian–Tasmania, the Selwyn Block, and implications for theCambrian–Silurian evolution of the Ross, Delamerian, and Lachlan orogens.Gondwana Research 19, 628–649.
Cayley, R., Korsch, R., Kennett, B., Skladzien, P., Jones, L., Morand, V., Gibson, G.,Rawling, T., Betts, P., 2011. Results of deep seismic reflection imaging of the easternDelamerian Orogen, South Australia and western Victoria, Australia, Data CDversion: 4 March, 2011. GeoScienceDepartment of Primary Industries, Victoria.
Chevrot, S., van der Hilst, R.D., 2000. The Poisson ratio of the Australian crust: geolog-ical and geophysical implications. Earth and Planetary Science Letters 183,121–132.
Clitheroe, G.M., Gudmundsson, O., Kennett, B.L.N., 2000. The crustal thickness ofAustralia. Journal of Geophysical Research 105, 13697–13713.
Collins, C.D.N., 1991. The nature of the crust–mantle boundary under Australia fromseismic evidence. In: Drummond, B.J. (Ed.), The Australian Lithosphere: GeologicalSociety of Australia Special Publication, 17, pp. 67–80.
Collins, W.J., 2002. Hot orogens, tectonic switching, and creation of continental crust.Geology 30, 535–538.
Collins, C.D.N., Drummond, B.J., Nicoll, M.G., 2003. Crustal thickness patterns in theAustralian continent. Geol. Soc. Australia Spec. Publ. 22, and Geol. Soc. AmericaSpec. Pap. 372, 121–128.
Direen, N.G., Crawford, A.J., 2003. The Tasman Line: where is it, what is it, is itAustralia's Rodinian breakup boundary? Australian Journal of Earth Sciences 50,491–502 http://dx.doi.org/10.1046/j.1440-0952.2003.01005.x.
Drummond, B.J., Collins, C.D.N., 1986. Seismic evidence for underplating of the lowercontinental crust of Australia. Earth and Planetary Science Letters 79, 361–372.
Drummond, B., Lyons, P., Goleby, B., Jones, L., 2006. Constraining models of the tectonicsetting of the giant Olympic Dam iron oxide–copper–gold deposit, South Australia,using deep seismic reflection data. Tectonophysics 420, 91–103.
125F.R. Fontaine et al. / Tectonophysics 582 (2013) 112–125
Eaton, D.W., 2005. Multi-genetic origin of the continental Moho: insights fromLITHOPROBE. Terra Nova 18, 34–43.
Fomin, T., Johnstone, D.W., Goleby, B.R., 2006. Seismic acquisition and processing:2003–2004 Curnamona Province seismic reflection survey (L164). In: Goleby,B.R., Korsch, R.J., Fomin, T., Conor, C.H.H., Preiss, W.V., Robertson, R.S., Burtt, A.C.(Eds.), The 2003–2004 Curnamona Province Seismic Survey, Geoscience Australia,Record, 2006/12, pp. 22–33.
Fomin, T., Holzschuh, L., Nakamura, A., Maher, J., Duan, J., Saygin, E., 2010. 2008Gawler–Curnamona–Arrowie (L189) and 2009 Curnamona–Gawler Link (L191)seismic surveys—acquisition and processing. In: Korsch, R.J., Kositcin, N. (Eds.),South Australian Seismic and MT Workshop 2010, Geoscience Australia, Record,2010/10, pp. 1–10.
Fraser, G.L., Blewett, R.S., Reid, A.J., Korsch, R.J., Dutch, R., Neumann, N.L., Meixner, A.J.,Skirrow, R.G., Cowley, W.M., Szpunar, M., Preiss, W.V., Fomin, T., Holzschuh, J.,Thiel, S., Milligan, P.R., Bendall, B.R., 2010. Geological interpretation of deep seismicreflection and magnetotelluric line 08GA-G1: Eyre Peninsula, Gawler Craton, SouthAustralia, Geoscience Australia, Record, 2010/10. , pp. 81–95.
Frederiksen, A.W., Folsom, H., Zandt, G., 2003. Neighbourhood inversion of teleseismicPs conversions for anisotropy and layer dip. Geophysical Journal International 155,200–212.
Gibson, G.M., Drummond, B., Fomin, T., Owen, A., Maidment, D.W., Gibson, D.L., Peljo,M., Wake-Dyster, K., 1998. Re-evaluation of crustal structures of the Broken HillInlier through structural mapping and seismic profiling. Australian GeologicalSurvey Organisation Record 1998/11, Canberra, ACT, Australia.
Giese, P., 2005. Moho discontinuity. In: Selly, R.C., Cocks, L.R.M., Plimer, I.R. (Eds.),Encyclopedia of Geology, 3. Elsevier Academic Press, Oxford, pp. 645–659.
Goleby, B.R., Korsch, R.J., Fomin, T., Conor, C.H.H., Preiss, W.V., Robertson, R.S., Burtt, A.,2006. The 2003–2004 Curnamona Province seismic survey: workshop notes.Geoscience Australia Record 2006/12, Canberra, ACT, Australia.
Gorman, A., Clowes, R.M., Ellis, R.M., Henstock, T.J., Spence, G.D., Keller, G.R., Levander,A., Snelson, C.M., Burianyk, M.J.A., Kanasewich, E.R., Asudeh, I., Hajnal, Z., Miller,K.C., 2002. Deep probe: imaging the roots of western North America. CanadianJournal of Earth Sciences 39, 375–398.
Gray, D.R., Foster, D.A., 2004. Tectonic evolution of the Lachlan Orogen, southeastAustralia: historical review, data synthesis and modern perspectives. AustralianJournal of Earth Sciences 51, 773–817.
Haskell, N.A., 1953. The dispersion of surface waves in multilayered media. Bulletin ofthe Seismological Society of America 43, 17–34.
Haskell, N.A., 1962. Crustal reflection of plane P and SV waves. Journal of GeophysicalResearch 67, 4751–4767.
Hill, D., 1951. Geology. In: Mack, G. (Ed.), Handbook of Queensland. AustralianAssociation for the Advancement of Science, Brisbane, pp. 13–24.
Jones, L.E.A., Johnstone, D.W., Goleby, B.R., 2005. Seismic acquisition and processing — 2003Gawler Craton seismic reflection survey (L163). In: Lyons, P., Goleby, B.R. (Eds.), The2003 Gawler Craton Seismic Survey: Notes from the Seismic Workshop Held at GawlerCraton State of Play 2004, Geoscience Australia, Record, 2005/19, pp. 35–47.
Julià, J., Ammon, C.J., Herrmann, R.B., Correig, A.M., 2000. Joint inversion of receiverfunction and surface wave dispersion observations. Geophysical Journal Interna-tional 143, 99–112.
Kennett, B.L.N., Salmon, M., Saygin, E., Group, AusMoho, 2011. AusMoho: the variationof Moho depth in Australia. Geophysical Journal International http://dx.doi.org/10.1111/j.1365.246X.2011.05194.x.
Korsch, R.J., Barton, T.J., Gray, D.R., Owen, A.J., Foster, D.A., 2002. Geological interpreta-tion of a deep seismic-reflection transect across the boundary between theDelamerian and Lachlan orogens, in the vicinity of the Grampians, westernVictoria. Australian Journal of Earth Sciences 49, 1057–1075.
Langston, C.A., 1977. Corvallis, Oregon, crustal and upper mantle structure from teleseismicbody waves. Bulletin of the Seismological Society of America 67, 713–724.
Langston, C.A., 1979. Structure under Mount Rainier, Washington, inferred fromteleseismic body waves. Journal of Geophysical Research 84, 4749–4762.
Leahy, G., Collins, J.A., Wolfe, C.J., Laske, G., Solomon, S.C., 2010. Underplating of theHawaiian Swell: evidence from teleseismic receiver functions. Geophysical JournalInternational 183, 313–329.
Lesti, C., Giordano, G., Salvini, F., Cas, R., 2008. Volcano tectonic setting of the intraplate,Pliocene–Holocene, Newer Volcanic Province (southeast Australia): role of crustalfracture zones. Journal of Geophysical Research 113 http://dx.doi.org/10.1029/2007JB005110.
Ligorria, J.P., Ammon, C.J., 1999. Iterative deconvolution and receiver functionestimation. Bulletin of the Seismological Society of America 89, 1395–1400.
Mooney, W.D., Laske, G., Masters, T.G., 1998. CRUST 5.1: a global crustal model at5°×5°. Journal of Geophysical Research 103, 727–747.
Morozov, I.B., Smithson, S.B., Solodilov, L.N., 2002. Imaging crustal structure along re-fraction profiles using multicomponent recordings of first-arrival coda. Bulletinof the Seismological Society of America 92 (8), 3080–3086.
O'Reilly, S.Y., Griffin, W.L., 1985. A xenolith-derived geotherm for southeasternAustralia and its geophysical implications. Tectonophysics 111, 41–63.
Payne, J.L., Hand, M., Barovich, K.M., Reid, A., Evans, D.A.D., 2009. Correlations and re-construction models for the 2500–1500 Ma evolution of the Mawson Continent.In: Reddy, S.M., Mazumder, R., Evans, D.A.D., Collins, A.S. (Eds.), PalaeoproterozoicSupercontinents and Global Evolution: Geological Society of London SpecialPublication, v. 323, pp. 319–355.
Preiss, W.V., 2000. The Adelaide Geosyncline of South Australia and its significance inNeoproterozoic continental reconstruction. Precambrian Research 100, 21–63.
Preiss, W.V., Korsch, R.J., Blewett, R.S., Fomin, T., Cowley, W.M., Neumann, N.L.,Meixner, A.J., 2010. Geological interpretation of deep seismic reflection line09GA-CG1: the Curnamona Province–Gawler Craton Link Line, South Australia.Geoscience Australia, Record, 2010/10, pp. 66–76.
Rawlinson, N., Kennett, B.L.N., Vanacore, E., Glen, R.A., Fishwick, S., 2010. The structureof the upper mantle beneath the Delamerian and Lachlan orogens from simulta-neous inversion of multiple teleseismic datasets. Gondwana Research 19, 788–799.
Reading, A.M., Kennett, B.L.N., Goleby, B., 2007. New constraints on the seismic struc-ture of West Australia: evidence for terrane stabilization prior to the assembly ofan ancient continent? Geology 35, 379–382.
Reading, A.M., Tkalčić, H., Kennett, B.L.N., Johnson, S.P., Sheppard, S., 2011. Seismicstructure of the crust and uppermost mantle of the Capricorn and PatersonOrogens and adjacent cratons, Western Australia, from passive seismic transects.Precambrian Research 196–197, 295–308.
Sambridge, M.S., 1999a. Geophysical inversion with a neighbourhood algorithm. I.Searching a parameter space. Geophysical Journal International 138, 479–494.
Sambridge, M.S., 1999b. Geophysical inversion with a neighbourhood algorithm — II.Appraising the ensemble. Geophysical Journal International 138, 727–746.
Sambridge, M.S., 2001. Finding acceptable models in nonlinear inverse problems usinga neighbourhood algorithm. Inverse Problems 17, 387–403.
Saygin, E., 2007. Seismic receiver and noise correlation based studies in Australia. Ph.D.thesis, The Australian National University, 175 pp.
Scheibner, E., Veevers, J.J., 2000. 19. Tasman fold belt system. In: Veevers, J.J. (Ed.),Billion-year Earth History of Australia and Neighbours in Gondwanaland. GEMOCPress, Sydney, pp. 154–234.
Shibutani, T., Sambridge, M.S., Kennett, B.L.N., 1996. Genetic algorithm inversion for re-ceiver functions with application to crust and uppermost mantle structure beneathEastern Australia. Geophysical Research Letters 23, 1826–1832.
Smith, W.H.F., Sandwell, D.T., 1997. Global sea floor topography from satellite altimetryand ship depth soundings. Science 277, 1956–1963.
Swain, G., Woodhouse, A., Hand, M., Barovich, K., Schwarz, M., Fanning, C.M., 2005.Provenance and tectonic development of the late Archean Gawler Craton, Australia:U–Pb zircon, geochemical and Sm–Nd isotopic implications. Precambrian Research141, 106–136.
Thomson, W.T., 1950. Transmission of elastic waves through a stratified solid. Journalof Applied Physics 21, 89–93.
Tkalčić, H., Pasyanos, M.E., Rodgers, A.J., Gök, R., Walter, W.R., Al-Amri, A., 2006. Amultistep approach for joint modeling of surface wave dispersion and teleseismicreceiver functions: implications for lithospheric structure of the Arabian Peninsula.Journal of Geophysical Research 111, B11311 http://dx.doi.org/10.1029/2005JB004130.
Tkalčić, H., Chen, Y., Liu, R., Huang, H., Sun, L., Chan, W., 2011. Multi-step modelling ofteleseismic receiver functions combined with constraints from seismic tomogra-phy: crustal structure beneath southeast China. Geophysical Journal International187 http://dx.doi.org/10.1111/j.1365-246X.2011.05132.x, 303-326.
Voronoi, M.G., 1908. Nouvelles applications des paramètres continus à la théorie desformes quadratiques. Journal für die reine und angewandte Mathematik 134,198–287.
Wessel, P., Smith, W.H., 1991. Free software helps map and display data. Eos, Transactions,American Geophysical Union 72, 441.
Wingate, M.T.D., Campbell, I.H., Compston, W., Gibson, G.M., 1998. Ion microprobeU–Pb ages for Neoproterozoic basaltic magmatism in south-central Australia andimplications for the break-up of Rodinia. Precambrian Research 87, 135–159.
Yang, Z., Sheehan, A.F., Yeck, W.L., Miller, K.C., Erslev, E.A., Worthington, L.L., Harder,S.H., 2012. Imaging basin structure with teleseismic virtual source reflectionprofiles. Geophysical Research Letters 39, L02303 http://dx.doi.org/10.1029/2011GL050035.
Zandt, G., Ammon, C.J., 1995. Continental crust composition constrained by measure-ments of crustal Poisson's ratio. Nature 374, 152–154.
Zelt, B.C., Ellis, R.M., 1999. Receiver–function studies in the Trans-Hudson Orogen,Saskatchewan. Canadian Journal of Earth Sciences 36, 585–603.
Zheng, T.Y., Chen, L., Zhao, L., Xu, W.W., Zhu, R.X., 2006. Crust–mantle structure differ-ence across the gravity gradient zone in North China Craton: seismic image of thethinned continental crust. Physics of the Earth and Planetary Interiors 159, 43–58.
Zheng, T.Y., Zhao, L., Zhu, R.X., 2008. Insight into the geodynamics of cratonicreactivation from seismic analysis of the crust–mantle boundary. GeophysicalResearch Letters 35 http://dx.doi.org/10.1029/2008GL033439.