Decompression melting driving intraplate volcanism in ... · intraplate volcanism, some of which has been attributed to mechanisms such as shear-driven upwelling [Conrad et al., 2011]

Decompression melting driving intraplatevolcanism in Australia: Evidencefrom magnetotelluric soundingSahereh Aivazpourporgou1, Stephan Thiel2,3, Patrick C. Hayman1, Louis N. Moresi4, and Graham Heinson2

1School of Earth, Atmosphere and Environment, Monash University, Clayton, Victoria, Australia, 2School of Earth andEnvironmental Sciences, University of Adelaide, Adelaide, South Australia, Australia, 3Geological Survey of South Australia,Adelaide, South Australia, Australia, 4School of Earth Sciences, Melbourne University, Parkville, Victoria, Australia

Abstract A long-period magnetotelluric (MT) survey, with 39 sites covering an area of 270 by 150 km, hasidentified melt within the thinned lithosphere of Pleistocene-Holocene Newer Volcanics Province (NVP) insoutheast Australia, which has been variously attributed to mantle plume activity or edge-driven mantleconvection. Two-dimensional inversions from the MT array imaged a low-resistivity anomaly (10–30Ωm)beneath the NVP at ~40–80 km depth, which is consistent with the presence of ~1.5–4% partial melt inthe lithosphere, but inconsistent with elevated iron content, metasomatism products or a hot spot. Theconductive zone is located within thin juvenile oceanic mantle lithosphere, which was accreted onto thickerProterozoic continental mantle lithosphere. We propose that the NVP owes its origin to decompressionmelting within the asthenosphere, promoted by lithospheric thickness variations in conjunction with rapidshear, where asthenospheric material is drawn by shear flow at a “step” at the base of the lithosphere.

1. Introduction

The majority of the volcanoes around the globe occur at subduction zones or mid-ocean ridges, with originsconsistent with plate tectonic theory, while the remainder are in intraplate settings, and are generallyattributed to mantle plumes. However, plume theory does not explain the spatial pattern and timing of allintraplate volcanism, some of which has been attributed to mechanisms such as shear-driven upwelling[Conrad et al., 2011] and buoyant decompression melting [Raddick et al., 2002].

The Newer Volcanics Province (NVP) in south eastern Australia is an active intraplate volcanic field withenigmatic origins. It includes ~400 eruption points over an area of ~15,000 km2 [Hare and Cas, 2005],overlying Otway Basin sediments in the south and Palaeozoic rocks of the Lachlan and Delamerian orogens,which are separated by the Moyston Fault, in the north (Figure 1).

Teleseismic tomographic studies [Graeber et al., 2002; Rawlinson and Fishwick, 2012] imaged a low-velocityzone at 50–150 km depth beneath the NVP, which was interpreted as a hot spot caused by a mantle plume[Graeber et al., 2002] or edge-driven convection (EDC) [Rawlinson and Fishwick, 2012]. However, consideringthe Australian-Indian plate has a northerly drift [Sandiford et al., 2004], the spatial pattern of the NVP and lackof systematic age progression of the volcanoes is inconsistent with a simple mantle plume model. Theupwelling rate (~1.5 cm/yr) for the youngest basalts of Mount Gambier (~5000 years old) and Mount Schank(~4500 years old) is uncharacteristically low, even for a weak plume [Demidjuk et al., 2007]. EDC fails toexplain the absence of contemporaneous volcanism along the south eastern margin of Australia [Demidjuket al., 2007; Farrington et al., 2010], but recent work by Davies and Rawlinson [2014] suggests that EDC focusedinto a locally thinner region of the lithosphere can explain the NVP volcanism. In addition, the large agegap (~85–50 Ma) between rifting in Victoria and the eruptions of the NVP [Vogel and Keays, 1997] shows theNVP magma generation is not associated with thermal residuals of continental rifting.

Seismic studies [Cayley et al., 2011; Rawlinson and Fishwick, 2012] imaged thinner (~100 km) lithosphereunderneath the Bendigo and Stawell zones (bounded by the Moyston Fault and Heathcote Fault Zones),while thicker lithosphere underlies the Selwyn Block to the east, and Proterozoic crust underplates theGrampians-Stavely Zone of the Delamerian Orogen to the west (Figure 1). This magnetotelluric (MT) study findsa low-resistivity zone underneath the thinnest lithosphere, which we show is consistent with presence of partialmelt and argue decompression melting due to lithospheric thickness variations is the cause of NVP volcanism.

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 346

PUBLICATIONSGeophysical Research Letters

RESEARCH LETTER10.1002/2014GL060088

Key Points:• Presence of large conductor inthe lithosphere

• Presence of partial melt underneathan intraplate volcanic field in Australia

• Magma genesis caused bydecompression melting

Supporting Information:• Readme• Figure S1

Correspondence to:S. Aivazpourporgou,[email protected]

Citation:Aivazpourporgou, S., S. Thiel, P. C.Hayman, L. N. Moresi, and G. Heinson(2015), Decompression melting drivingintraplate volcanism in Australia:Evidence from magnetotelluric sounding,Geophys. Res. Lett., 42, 346–354,doi:10.1002/2014GL060088.

Received 27 OCT 2014Accepted 27 DEC 2014Accepted article online 5 JAN 2015Published online 29 JAN 2015

2. MT Data

In 2010 and 2011, long-period MT data (10–10,000 s) were collected, using AuScope MT instruments, at thesame teleseismic locations of Graeber et al. [2002] in a rectangular grid (270 km×150 km), which consistsof 39 stations forming four parallel east-west trending profiles with 50 km spacing (Figure 1). The MT datawere processed applying the robust technique of Chave and Thomson [2004].

Phase tensors analysis [Caldwell et al., 2004] shows that the minimum phases in the Stawell and Bendigozones (Lachlan Orogen) exceed 45° for periods up to 300 s (Figure 1), suggesting conductivity increases atupper mantle depths. In contrast, on the western Delamerian Orogen, the minimum phases lack increasingconductivity trends over the same period range. Up to 300 s, the major axes of phase ellipses in theLachlan Orogen are aligned NW-SE, which is consistent with the average regional strike (N60°E) over thebandwidth of 20–300 s for the same area (rose plot (a) in Figure 1). For the rest of the survey, including line A,major axes of phase ellipses points toward the coastline, due to the coast effect, and are aligned NE-SW, whichis the average regional strike direction (N30°E) for the Delamerian side of survey (rose plot (b) in Figure 1).

3. Two-Dimensional Modeling

Two-dimensional models are developed using the Rodi and Mackie [2001] code for transverse electric (TE)and transverse magnetic (TM) polarization modes, subsequent to rotation of MT response functions to thedominant strike angle. Station D10 is removed from inversions since the majority of frequencies have 3-Dcharacteristics. The start model includes two layers, including materials of 100Ωm resistivity up to ~410 kmoverlying a homogenous half space of 10Ωm. The errors for TE apparent resistivity were set to 100% forthe first inversion run and were gradually reduced to 10%. In the case of TMmode, the error floors were set to10% for all inversions. The tipper error was set to 0.02, and for both TE and TM modes, the error floors ofphases were set to 5° for all inversions. Using these settings, the final root-mean-square (rms) misfit for lineA–D, respectively, are RMSA = 2.09, RMSB = 2.05, RMSC = 2.10, and RMSD = 2.06.

Figure 1. MT phase tensors for period of 300 s overlying on the NVP geological map. The rose plots (bottom right) are theaverage strike direction over the period band of 20–300 s for stations to the (a) east of the Moyston Fault (Lachlan),excluding the line A; and (b) west of the Moyston Fault (Delamerian), including line A. The right inset map shows theVictorian part of the Lachlan Orogen and the Delamerian Orogen (DEO). Faults include AF: Avoca Fault, HFZ: HeathcoteFault Zone, and YF: Yarramyljup Fault.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 347

4. Results and Interpretations4.1. Resistivity Structures

The most prominent feature identified is a major low-resistivity zone of 10–30Ωm underneath the Bendigoand Stawell zones at ~40–80 km depth (A1, B1, C1, and D1 in Figure 2a), referred to here as the NVP anomaly,which extends ~40–90 km E-W and ~150 km NW-SE. In addition, there is a good spatial agreementbetween the NVP anomaly and the low-velocity zones at 50–150 km depth identified by teleseismic studies[Graeber et al., 2002; Rawlinson and Fishwick, 2012].

The NVP anomaly is located underneath a V-shaped resistive area at the eastern side of lines A–D (features A3,B3, C3, and D3 in Figure 2a), which represents Cambrian volcanic rocks of the Stawell and Bendigo zones(Figures 2b and 2c), as inferred by seismic reflection studies [Cayley et al., 2011]. The eastern margin of theV-shaped resistive blocks underneath stations A10, B10, C10, and D09 is a west-dipping structure thatspatially coincides with the Heathcote Fault Zone, while the western margin of the V-shaped featureimages the east-dipping Moyston Fault. The resolution of MT models is significantly reduced below ~90 kmunderneath the NVP anomaly.

Additionally, a very resistive structure within the lower crust and upper mantle beneath the Stawell andGrampians-Stavely zones is resolved (features B4 and C4 in Figure 2a), which is spatially consistent withregional positive gravity anomalies [Murphy et al., 2006] and high seismic velocities [Rawlinson et al., 2011].This zone may represent Proterozoic continental basement of the Delamerian Orogen, as inferred fromseismic and gravity studies [Murphy et al., 2006; Rawlinson et al., 2011].

(b)

(c)

(a)

Figure 2. (a) MT 2-D inversionmodels located above simplified geological map (see Figure 1 for the geologymap legend). Lines 1 and 2 are seismic reflection profileswith the interpretations shown in (b) and red dots are MT stations; (b) interpretations of the seismic cross sections of lines 1 and 2 shown in (a) modified after[Cayley et al., 2011]; (c) top 40 km of MT model of line B along stations B05–B10. In (a) and (b), faults include AF: Avoca Fault, HFZ: Heathcote Fault Zone, MF: MoystonFault, and YF: Yarramyljup Fault.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 348

4.2. Interpretations of the NVP Anomaly

Laboratory and MT studies show that the bulk conductivity of the mantle depends on several parametersincluding composition, degree of partial melt, and temperature [e.g., Constable, 2006; Fullea et al., 2011;Roberts and Tyburczy, 1999; Yoshino, 2010]. MTmeasurements along with western Victorian geochemical dataare used to show that the NVP anomaly is a product of mantle partial melting and that composition andtemperature, although they may contribute to the conduction, cannot solely be the cause of the anomaly.

Throughout the discussion, the pressure and temperature ranges derived from western Victorian xenolithstudies [O’Reilly and Griffin, 1985; Sutherland et al., 2004] are used to constraint P-T condition, which representthe upper boundary of the NVP anomaly at Moho depth or ~40 km (~900°C, 0.9 GPa) and the lower boundaryof the NVP anomaly at ~80 km (~1200°C, 2.2 GPa).4.2.1. CompositionFor most minerals conduction can be attributed to the presence of impurities such as ferric iron or hydrogensince a typical silicate or oxide mineral, with no transition metals in the crystal lattice and all electronsstrongly bound to the atomic nuclei, behaves as an insulator [Karato and Wang, 2013]. For these minerals,conduction occurs when thermally activated electrons move from filled valance band to empty conductorband, referred to as intrinsic conduction. In general, for most minerals the gap between valance andconduction bands is large, the activation energy is high, and the contribution of intrinsic charge carriers toconduction is small. For this reason, the conductivity of typical minerals can be significantly affected byimpurities such iron and hydrogen [Karato and Wang, 2013]. Here we discuss whether the impurities,including iron, hydrogen, and metasomatism products, could be the cause of the NVP anomaly.4.2.1.1. IronAlthough increased iron content is well known to enhance conductivity [Romano et al., 2006], iron contentvariations in the western Victorian mantle can only have a small contribution to the NVP anomaly. Laboratorystudies on iron-bearing minerals show that at temperatures less than 1200°C, electrical currents areconducted either by electrons hopping between Fe2+ and Fe3+ or moving electron holes in the valance band[Karato and Wang, 2013]. The magnesium number (Mg#=Mg2+/(Mg2++Fe2+), where Mg and Fe are in molarconcentrations) generally ranges from 0.8 to 0.9 in mantle olivine, pyroxene, and garnet [Pearson et al., 2003],which is consistent with average values for NVP mantle minerals [e.g., Griffin and O’Reilly, 1988]. At 1000°Cand 10–19 GPa the resistivity of these minerals is measured 100Ωm for Mg# 0.8 and 10,000Ωm for Mg# 0.9[Romano et al., 2006]. Thus, Fe content alone is unlikely to be the cause of the NVP anomaly. However,laboratory measurements on garnet conductivity are performed under significantly higher pressures relativeto the NVP anomaly conditions (0.9–2.2 GPa) and the pressure effect on conductivity of iron-bearing mineralsremains controversial [Karato and Wang, 2013].4.2.1.2. HydrogenDissolution of hydrogen (water) in the mantle can enhance the conductivity by several orders ofmagnitude in an isotropic fashion, referred to as proton conduction [e.g., Karato, 1990; Poe et al., 2010;Wang et al., 2008; Yoshino et al., 2009]. However, the mechanism(s) for, and the magnitude of, protonconduction remains controversial [Jones, 2014; Jones et al., 2012]. Geochemical analysis of volcanic glassfrom the NVP show the water content of the glass range from 0.05 to 0.1 wt % [Van Otterloo, 2012], whichwe used as an estimate for mantle water contents. The resistivity of solid lherzolite can be calculatedfrom formula of Wang et al. [2008] for varying water and iron content (see Appendix A for formula). Usingthe NVP lherzolite xenolith data, resistivity is calculated to be ~2,000Ωm at 900°C for water content of0.1 wt % and XFe = 0.1 [e.g., Griffin and O’Reilly, 1988], which is significantly higher than that of the NVPanomaly (~10–30Ωm). In addition, teleseismic studies provide support for this interpretation. Remotesensing studies of hydrogen in the mantle show that if there is less than 0.1 wt % water in the mantlethen the reduction of P wave velocity is less than ~0.5% [Karato, 2006], while the measured reduction ofP waves velocities for the western Victorian mantle (coinciding with the NVP anomaly) is ~3.8% [Graeberet al., 2002; Rawlinson and Fishwick, 2012]. Although water enhances the mantle conductivity, the NVPanomaly cannot be solely explained by the current estimates for water in western Victorian mantle.4.2.1.3. MetasomatismMetasomatism can introduce and remove minerals and fluids to the mantle and can affect overall bulkconductivity, depending on the mobility of the impurities relative to the mobility of intrinsic charge carriers[e.g., ten Grotenhuis et al., 2004]. Studies of western Victorian peridotite xenoliths find abundant conductive

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 349

minerals, including amphibole and sulfides, as well as H2O-CO2 fluid inclusions [e.g., Griffin and O’Reilly, 1988],which indicates the effects of metasomatism on bulk conductivity need to be considered.

At lower crustal conditions amphibole conductivity is controlled by dehydration, which releases hydrogenand oxygen and transforms Fe2+ to Fe3+, and consequently increases conductivity [Wang et al., 2012].However, amphibole conductivities at mantle conditions are poorly constrained. In addition, amphiboles inwestern Victorian xenoliths are rare, have low Fe content (<4.5 wt % FeO) [Griffin and O’Reilly, 1988] and formlarge isolated clusters along foliation planes, rather than forming an interconnected network. Thus,amphibole is unlikely to be the cause of the NVP anomaly.

Recent laboratory and field studies have shown that small amount of sulfides, in the form of melt or well-connected crystallized networks, can enhance conductivity in the absence of water or other connected meltor fluid phases [e.g., Thiel et al., 2005;Watson et al., 2010]. For example, adding 1.4 wt % of sulfide impurities toolivine (Mg #= 0.91) can increase conductivity by several orders of magnitude [Watson et al., 2010].Sulfides content in the western Victorianmantle (0.1 wt %) [MacRae, 1979], however, is an order of magnitudelower than the laboratory studies cited above, and these low contents are typical of studies of metasomatismof mantle minerals [e.g., Andersen et al., 1987]. As a result, it is unlikely that sulfides alone, either as meltor solid, are responsible for the NVP high conductive zones.

In western Victoria, H2O-CO2 rich fluids are common products of metasomatism [e.g., O’Reilly and Griffin, 1988].Dissolution of carbon dioxide in water releases ions of H+ and HCO3, which can act as charge carriers andincrease the bulk conductivity [Unsworth and Rondenay, 2013]. However, these minor phases can increase thebulk conductivity if they form well interconnected networks along grain boundaries [Unsworth and Rondenay,2013]. Studies of volatile-rich peridotite xenoliths show that, in contrast to silicate and carbonate melts,H2O-CO2 rich fluids tend to fill isolated pores along grain boundaries rather than forming interconnected films[O’Reilly and Griffin, 2013]. As a result H2O-CO2 rich fluids do not significantly affect the bulkmantle conductivity.4.2.2. TemperatureAssuming the mantle beneath the western Victoria is dry and that the elevated temperatures are the origin ofthe NVP anomaly, the conductivity of the anomaly can be calculated if the geotherm is known. Laboratorystudies by Constable [2006] show that the resistivity of dry olivine at ~900°C to ~1200°C (the estimatedtemperature range for the NVP anomaly) varies from ~30,000Ωm to ~8,000Ωm, which is several orders ofmagnitude higher than the resistivity of the NVP anomaly (10–30Ωm). Since the estimated mantletemperatures are derived from geothermometry studies of western Victoria, one uncertainty with the mantlexenoliths is whether they record ambient mantle temperatures (preplume), or reequilibrated temperaturesafter a plume thermal perturbation. In general, the conduit of mantle plumes increases the ambient mantletemperatures by a maximum of 250–300°C [Davies, 1999]. This places an upper bound on mantletemperatures beneath western Victoria in the range of 1150–1200°C at ~40 km after equilibrium. In thistemperature range (1150–1200°C), the resistivity of dry olivine is estimated to be ~1000–20,000Ωm[Constable, 2006], which is still significantly higher than that of the NVP anomaly. It is concluded that the NVPanomaly cannot be exclusively produced by enhanced temperatures in dry peridotite or a hot spot.4.2.3. Partial MeltSince the NVP is an active volcanic province, it is necessary to consider whether the conductive mantleanomaly is a result of partial melt. The conductivity of melts has been extensively studied in laboratories andfrom field measurements [e.g., Constable and Heinson, 2004; Gaillard, 2004; Roberts and Tyburczy, 1999]. Herethe melt conductivities are estimated from geochemical data to investigate whether the NVP anomaly can beexplained by the presence of partial melt in the mantle.

Western Victorian P-T conditions (Figure 3a) were determined from Bullenmerri garnet-bearing xenolithdata [O’Reilly and Griffin, 1985] and using the MELTS code [Ghiorso and Sack, 1995], which facilitatesthermodynamic modeling of phase equilibrium in magmatic systems. In Figure 3a, it is clear that the westernVictorian mantle solidus is higher than the South East Australian (SEA) geotherm, indicating partial meltingwill only occur in the presence of volatiles or changes in P and/or T.

Resistivity of silicate melt with varying water content (Figure 3b) is calculated at different temperatures andpressures using the SIGMELTS web portal [Pommier and Le-Trong, 2011]. Figure 3b shows that the resistivity ofmelt phase containing 0.001–0.1 wt % water varies between ~1 and 30Ωm at temperatures of ~900–1150°Cand pressures of 1–2 GPa, consistent with the NVP anomaly.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 350

The melt fraction (ϕ) is determined by the Hashin-Shtrikman upper bound on the bulk conductivity formula[Shankland et al., 1981]:

ϕ ¼ 3σm σb � σsð Þσb þ 2σmð Þ σm � σsð Þ (1)

where σb is bulk conductivity (reciprocal of the resistivity), σm is melt conductivity, and σs is solid mantle rockconductivity. The melt fraction is calculated for temperatures of the upper boundary of the NVP anomaly at~40 km (~900°C) and the lower boundary of the NVP at ~80 km (~1200°C). The bulk resistivity of the NVPanomaly, derived from MT inversion models, is ~10–30Ωm, with the most resistive value of 30Ωm (equals to0.03 Sm�1) adopted for melt fraction calculations. The melt resistivity value is constrained from Figure 3b,which shows the minimum resistivity of melt with 1 wt % water is ~1Ωm (equals to 1 Sm�1) at 1100°C and1.7 GPa. For the melt fraction calculations, the minimum melt resistivity value of 1Ωm is adopted. Theresistivity of NVP solid lherzolite is calculated using the formula of Wang et al. [2008] and the NVP lherzolitegeochemistry, which is calculated to be ~2000Ωm (equals to 0.0005 Sm�1) at 900°C and 50Ωm (equals to0.02 Sm�1) at 1200°C (see Appendix A for formula). The reciprocal of bulk, melt, and solid rock resistivity valuescast into equation (1) yields a range of partial melt from 4% (T=900°C) to 1.5% (T=1200°C).

It should be mentioned that the melt fraction calculations from MT measurements is affected by ourknowledge of temperature, bulk chemical composition, and the melt geometry [Pommier and Garnero, 2014].For example, for these calculations, the melt resistivity is assumed to be temperature independent andchosen to be the minimum value (1Ωm), which results in the lowest calculated melt fraction. Figure 3bshows that the resistivity of melt containing 0.1 wt % water at 900°C (the upper boundary of the NVPanomaly) is ~9Ωm. Adopting melt resistivity value of 9Ωm yields the melt fraction to be ~35% for the bulkconductivity of 30Ωm and solid conductivity of 2000Ωm, which is clearly an overestimate for an intraplatevolcanism [Pommier and Garnero, 2014]. However, the Hashin-Shtrikman upper bound model used for meltfraction calculations estimates higher melt fractions relative to the parallel conductive model and is asuitable model to calculate the melt fractions above 0.025 [ten Grotenhuis et al., 2005].

In conclusion, to constraint the melt fraction calculations, the result of geochemical studies on westernVictorian basalts are considered. These calculations are in agreement with geochemical studies of NVPvolcanoes, which show that the most primitive magma compositions are produced by 4%–5% of partial melt

1.0

1.5

2.0

Pres

sure

(G

Pa)

Temperature (ºC)

Dry solidus of w

estern Victoria

liquidus

SEA

geotherm

Res

istiv

ity (

m)

Water content (wt%)

100

10-3 10-2 10-1 100800 1000 1200 1400 1600 1800

(b) The resistivity of the NVP basaltic melt for varying water content (wt%), temperature (ºC) and pressure (GPa)

P-T conditions of western Victorian mantle(a)

1150 °C, 2 GPa1100 °C, 1.7 GPa

1050 °C, 1.5 GPa1000 °C, 1.3 GPa

40

950 °C, 1.15 GPa

900 °C, 1 GPa

850 °C, 0.85 GPa

Figure 3. (a) Solidus (dashed black line to the left) of the NVP Bullenmerri xenoliths produced using MELTS code [Ghiorsoand Sack, 1995]. (b) The melt resistivity calculated using SIGMELTS web portal [Pommier and Le-Trong, 2011]. The inputparameters are water content varying between 0.001 and 1.0 wt %, 49.09 wt % SiO2 and 3.64 wt % Na2O [Van Otterloo,2012], temperature ranges between 850 and 1150°C and pressure ranges between 0.85 and 2 GPa. There is a slight increasein melt resistivity for increasing pressure from 1.7 GPa to 2 GPa, supported by laboratory measurements on silicates meltconductivity [Gaillard, 2004].

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 351

[Demidjuk et al., 2007; Van Otterloo et al., 2014]. Another constraint is provided by the primitive magmacomposition and low water contents (0.1 wt %), which indicates that partial melting does not exceed 10%[Irving and Green, 1976].

4.3. Mechanism for the Genesis of the NVP Magmas

Melting in the lithospheric mantle can be initiated by increasing temperature, reducing pressure, or addingvolatiles. Increasing temperature from a mantle plume is not supported by our MT models (see section 4.2.2)and is inconsistent with the timing and spatial association of volcanoes in the NVP [Hare and Cas, 2005].Decompression melting in the lithosphere is also not tenable because Victoria is in a state of compression,rather than extension [Sandiford et al., 2004]. The addition of volatiles, discussed below, is a possiblemechanism for initiating partial melt within the lithosphere.

An important constraint provided by our MT models (Figure 2) is the clear spatial association with the NVPanomaly and thinned lithosphere. Figure 2 illustrates that the NVP anomaly is located underneath Cambrianvolcanic rocks of the Bendigo and Stawell zones. This part of the Lachlan Orogen forms the upper part ofthinner, juvenile lithospheric mantle that was trapped between converging older and thicker subcontinentallithospheric mantle of the Selwyn Block in the east [Cayley, 2011] and the Delamerian Orogen to the west[Rawlinson and Fishwick, 2012; Robertson et al., 2014]. Lithosphere thickness beneath the ProterozoicDelamerian Orogen may be 150–200 km thick, while it is significantly thinner beneath the Lachlan Orogen(~100 km), inferred from teleseismic studies [Rawlinson and Fishwick, 2012].

Lateral variations of lithospheric thickness can cause regional asthenospheric upwelling and decompressionmelting if mantle rocks are near their solidus. The upwelling is enhanced if the viscosity is reduced in theregion with thinner lithosphere, as would be expected if partial melting occurs [Conrad et al., 2010]. Davies andRawlinson [2014] showed that this effect is enhanced if the thin lithosphere is of limited lateral extent. Thisleads to the possibility of a long-lived and self-sustaining region of partially molten asthenosphere beneath asmall region of anomalously thin lithosphere in a fast-moving continent. Anomalously hot asthenosphericmaterial coming from depth may contribute to raise the temperature at the base of lithosphere, which caninitiate partial melt within the lithosphere primed by volatiles. The presence of metasomatic fluids in thewestern Victorian lithosphere is supported by geochemical studies that show western Victorian mantle hasbeen through at least three distinct episodes of metasomatism, with the second episode overlapping the NVPvolcanism [O’Reilly and Griffin, 2013]. We propose that asthenospheric circulation driven by shear flow causesupwelling and decompression melting at the base of thinner lithosphere of the Bendigo and Stawell zones.Pooling of the melt at the base of the lithospheric mantle both raises the temperature of, and adds volatiles to,lithospheric mantle and initiates partial melting, which has been identified by our MT models.

5. Conclusion

A major conductive zone (10–30Ωm) is identified at ~40–80 km depth, underneath the resistive Cambrianvolcanic rocks of the Bendigo and Stawell zones, referred to here as the NVP anomaly. This anomaly isassociated with the presence of ~1.5–4% partial melt in the mantle lithosphere, supported by independentgeochemistry studies of the NVP basalts and xenoliths, and is not attributed to either mantle iron content,metasomatism products or enhanced temperatures (e.g., a mantle plume) in dry peridotite. The NVP anomalyis embedded within a thinner lithospheric mantle of the Bendigo and Stawell zones, which is trappedbetween older continental lithospheric mantle of the Selwyn Block and the Delamerian Orogen. We suggestthe NVP origin is due to decompression melting of the shear-driven flow of buoyant asthenospheric materialwithin a “step” at the lithospheric base of the Stawell and Bendigo zones.

Appendix A

The conductivity of the lherzolite for varying iron (XFe) and water (Cw) contents (wt %) can be calculated fromformula of Wang et al. [2008]:

σ ¼ 105:2±0:4C0:67±0:07w exp � 183 kJ=molð Þ

RT

� ��exp XFe

79±3 kJ=molð ÞRT

� �(A1)

where R is gas constant in Jmol K and T is temperature in K.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 352

ReferencesAndersen, T., W. L. Griffin, and S. Y. O’Reilly (1987), Primary sulphide melt inclusions in mantle-derived megacrysts and pyroxenites, Lithos,

20(4), 279–294.Caldwell, T. G., H. M. Bibby, and C. Brown (2004), The magnetotelluric phase tensor, Geophys. J. Int., 158(2), 457–469.Cayley, R. A. (2011), Exotic crustal block accretion to the eastern Gondwanaland margin in the Late Cambrian–Tasmania, the Selwyn

Block, and implications for the Cambrian–Silurian evolution of the Ross, Delamerian, and Lachlan orogens, Gondwana Res., 19(3),628–649.

Cayley, R. A., R. J. Korsch, D. H. Moore, R. D. Costelloe, A. Nakamura, C. E. Willman, T. J. Rawling, V. J. Morand, P. B. Skladzien, and P. J. O’Shea(2011), Crustal architecture of central Victoria: Results from the 2006 deep crustal reflection seismic survey, Aust. J. Earth Sci., 58(2),113–156.

Chave, A. D., and D. J. Thomson (2004), Bounded influence magnetotelluric response function estimation, Geophys. J. Int., 157(3), 988–1006.Conrad, C. P., B. Wu, E. I. Smith, T. A. Bianco, and A. Tibbetts (2010), Shear-driven upwelling induced by lateral viscosity variations and

asthenospheric shear: A mechanism for intraplate volcanism, Phys. Earth Planet. Inter., 178(3–4), 162–175.Conrad, C. P., T. A. Bianco, E. I. Smith, and P. Wessel (2011), Patterns of intraplate volcanism controlled by asthenospheric shear, Nat. Geosci.,

4(5), 317–321.Constable, S. (2006), SEO3: A new model of olivine electrical conductivity, Geophys. J. Int., 166(1), 435–437.Constable, S., and G. Heinson (2004), Hawaiian hot-spot swell structure from seafloor MT sounding, Tectonophysics, 389(1–2), 111–124.Davies, D. R., and N. Rawlinson (2014), On the origin of recent intraplate volcanism in Australia, Geology, doi:10.1130/G36093.1.Davies, G. F. (1999), The plumemode, in Dynamic Earth: Plates, Plumes and Mantle Convection, edited by G. F. Davies, pp. 293–323, Cambridge

Univ. Press, Cambridge, U. K.Demidjuk, Z., S. Turner, M. Sandiford, R. George, J. Foden, and M. Etheridge (2007), U-series isotope and geodynamic constraints on mantle

melting processes beneath the Newer Volcanic Province in South Australia, Earth Planet. Sci. Lett., 261(3–4), 517–533.Farrington, R. J., D. R. Stegman, L. N. Moresi, M. Sandiford, and D. A. May (2010), Interactions of 3D mantle flow and continental lithosphere

near passive margins, Tectonophysics, 483(1–2), 20–28.Fullea, J., M. R. Muller, and A. G. Jones (2011), Electrical conductivity of continental lithospheric mantle from integrated geophysical and

petrological modeling: Application to the Kaapvaal Craton and Rehoboth Terrane, southern Africa, J. Geophys. Res., 116, B10202,doi:10.1029/2011JB008544.

Gaillard, F. (2004), Laboratory measurements of electrical conductivity of hydrous and dry silicic melts under pressure, Earth Planet. Sci. Lett.,218(1–2), 215–228.

Ghiorso, M. S., and R. O. Sack (1995), Chemical mass transfer in magmatic processes. IV. A revised and internally consistent thermodynamicmodel for the interpolation and extrapolation of liquid-solid equilibria in magmatic systems at elevated temperatures and pressures,Contrib. Mineral. Petrol., 119, 197–212.

Graeber, F. M., G. A. Houseman, and S. A. Greenhalgh (2002), Regional teleseismic tomography of the western Lachlan Orogen and the NewerVolcanic Province, southeast Australia, Geophys. J. Int., 149(2), 249–266.

Griffin, W. L., and S. Y. O’Reilly (1988), Mantle metasomatism beneath western Victoria, Australia: II. Isotopic geochemistry of Crdiopsidelherzolites and Al-augite pyroxenites, Geochim. Cosmochim. Acta, 52(2), 449–459.

Hare, A., and R. Cas (2005), Volcanology and evolution of the Werribee Plains intraplate, basaltic lava flow-field, Newer Volcanics Province,southeast Australia, Aust. J. Earth Sci., 52(1), 59–78.

Irving, A. J., and D. H. Green (1976), Geochemistry and petrogenesis of the newer basalts of Victoria and South Australia, J. Geol. Soc. Aust.,23(1), 45–66.

Jones, A. G. (2014), Reconciling different equations for proton conduction using the Meyer-Neldel compensation rule, Geochem. Geophys.Geosyst., 15, 337–349, doi:10.1002/2013GC004911.

Jones, A. G., J. Fullea, R. L. Evans, andM. R. Muller (2012), Water in cratonic lithosphere: Calibrating laboratory-determinedmodels of electricalconductivity of mantle minerals using geophysical and petrological observations, Geochem. Geophys. Geosyst., 13, Q06010, doi:10.1029/2012GC004055.

Karato, S. I. (1990), The role of hydrogen in the electrical conductivity of the upper mantle, Nature, 347(6290), 272–273.Karato, S. I. (2006), Remote sensing of hydrogen in Earth’s mantle, in Reviews in Mineralogy and Geochemistry, edited by H. Keppler and J. R.

Smith, pp. 343–375, Mineral. Soc. of Am., Chantilly, Va.Karato, S. I., and D. Wang (2013), Electrical conductivity of minerals and rocks, in Physics and Chemistry of the Deep Earth, pp. 145–182,

Wiley-Blackwell, New York.MacRae, N. D. (1979), Silicate glass and sulfides in ultramafic xenoliths, Newer Basalts, Victoria, Australia, Contrib. Mineral. Petrol., 68(3), 275–280.Murphy, F. C., T. J. Rawling, C. J. L. Wilson, L. J. Dugdale, and J. M. Miller (2006), 3D structural modelling and implications for targeting gold

mineralisation in western Victoria, Aust. J. Earth Sci., 53(5), 875–889.O’Reilly, S. Y., and W. L. Griffin (1985), A xenolith-derived geotherm from southeastern Australia and its geophysical implications,

Tectonophysics, 111, 41–63.O’Reilly, S. Y., and W. L. Griffin (1988), Mantle metasomatism beneath western Victoria, Australia: I. Isotopic geochemistry of Crdiopside

lherzolites and Al-augite pyroxenites, Geochim. Cosmochim. Acta, 52(2), 433–477.O’Reilly, S. Y., and W. L. Griffin (2013), Mantle metasomatism, inMetasomatism and the Chemical Transformation of Rock, edited by D. E. Harlov

and H. Austrheim, pp. 471–533, Springer, Berlin.Pearson, D. G., D. Canil, and S. B. Shirey (2003), Mantle samples included in volcanic rocks: Xenoliths and diamonds, in Treatise on

Geochemistry, edited by D. H. Heinrich and K. K. Turekian, pp. 171–275, Pergamon, Oxford, U. K.Poe, B. T., C. Romano, F. Nestola, and J. R. Smyth (2010), Electrical conductivity anisotropy of dry and hydrous olivine at 8GPa, Phys. Earth

Planet. Inter., 181(3–4), 103–111.Pommier, A., and E. J. Garnero (2014), Petrology-based modeling of mantle melt electrical conductivity and joint interpretation of

electromagnetic and seismic results, J. Geophys. Res. Solid Earth, 119, 4001–4016, doi:10.1002/2013JB010449.Pommier, A., and E. Le-Trong (2011), “SIGMELTS”: A web portal for electrical conductivity calculations in geosciences, Comput. Geosci., 37(9),

1450–1459.Raddick, M. J., E. M. Parmentier, and D. S. Scheirer (2002), Buoyant decompression melting: A possible mechanism for intraplate volcanism,

J. Geophys. Res., 107(B10), 2228, doi:10.1029/2001JB000617.Rawlinson, N., and S. Fishwick (2012), Seismic structure of the southeast Australian lithosphere from surface and body wave tomography,

Tectonophysics, 572–573, 111–122.

AcknowledgmentsInstrumentation and funding for thisproject were provided by AuScope,Monash University, and University ofAdelaide. Data can be obtained bycontacting Geoscience Australia (http://www.ga.gov.au/scientific-topics/disci-plines/geophysics/magnetotellurics).Thanks to Goran Boran for equipment,to the field crew, Hamish Adam,Leonhard Kocijan, David Willis,Chris Mays, and Jackson Van DenHoven, and to Ross Cayley for helpfulgeological discussions.

The Editor thanks two anonymousreviewers for their assistance inevaluating this paper.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 353

Rawlinson, N., B. L. N. Kennett, E. Vanacore, R. A. Glen, and S. Fishwick (2011), The structure of the upper mantle beneath the Delamerian andLachlan orogens from simultaneous inversion of multiple teleseismic datasets, Gondwana Res., 19(3), 788–799.

Roberts, J. J., and J. A. Tyburczy (1999), Partial-melt electrical conductivity: Influence of melt composition, J. Geophys. Res., 104(B4), 7055–7065,doi:10.1029/1998JB900111.

Robertson, K., D. Taylor, S. Thiel, and G. Heinson (2014), Magnetotelluric evidence for serpentinisation in a Cambrian subduction zonebeneath the Delamerian Orogen, southeast Australia, Gondwana Res., doi:10.1016/j.gr.2014.07.013.

Rodi, W., and R. L. Mackie (2001), Nonlinear conjugate gradients algorithm for 2-D magnetotelluric inversion, Geophysics, 66(1), 174–187.Romano, C., B. T. Poe, N. Kreidie, and C. A. McCammon (2006), Electrical conductivities of pyrope-almandine garnets up to 19 GPa and 1700 °C,

Am. Mineral., 91(8–9), 1371–1377.Sandiford, M., M. Wallace, and D. Coblentz (2004), Origin of the in situ stress field in southeastern Australia, Basin Res., 16(3), 325–338.Shankland, T. J., R. J. Connell, and H. S. Waff (1981), Geophysical constraints on partial melt in the upper mantle, Rev. Geophys., 19(3), 394–406,

doi:10.1029/RG019i003p00394.Sutherland, F. L., J. D. Hollis, W. D. Birch, R. E. Pogson, and L. R. Raynor (2004), Cumulate-rich xenolith suite in Late Cenozoic basaltic eruptives,

Hepburn Lagoon, Newlyn, in relation to western Victorian lithosphere, Aust. J. Earth Sci., 51(3), 319–337.ten Grotenhuis, S. M., M. R. Drury, C. J. Peach, and C. J. Spiers (2004), Electrical properties of fine-grained olivine: Evidence for grain boundary

transport, J. Geophys. Res., 109, B06203, doi:10.1029/2003JB002799.ten Grotenhuis, S. M., M. R. Drury, C. J. Spiers, and C. J. Peach (2005), Melt distribution in olivine rocks based on electrical conductivity

measurements, J. Geophys. Res., 110, B12201, doi:10.1029/2004JB003462.Thiel, S., G. Heinson, and A. White (2005), Tectonic evolution of the southern Gawler Craton, South Australia, from electromagnetic sounding,

Aust. J. Earth Sci., 52(6), 887–896.Unsworth, M., and S. Rondenay (2013), Mapping the distribution of fluids in the crust and lithospheric mantle utilizing geophysical

methods, inMetasomatism and the Chemical Transformation of Rock, edited by D. E. Harlov and H. Austrheim, pp. 535–598, Springer, Berlin.Van Otterloo, J. (2012), Complexity in monogenetic volcanic systems: Factors influencing alternating magmatic and phreatomagmatic

eruption styles at the 5 ka Mt. Gambier Volcanic Complex, South Australia, Monash Univ., Melbourne, Australia.Van Otterloo, J., M. Raveggi, R. A. F. Cas, and R. Maas (2014), Polymagmatic activity at the monogenetic Mt Gambier Volcanic Complex in the

Newer Volcanics Province, SE Australia: New insights into the occurrence of intraplate volcanic activity in Australia, J. Petrol., 55(7),1317–1351.

Vogel, D. C., and R. R. Keays (1997), The petrogenesis and platinum-group element geochemistry of the Newer Volcanic Province, Victoria,Australia, Chem. Geol., 136(3–4), 181–204.

Wang, D., H. Li, L. Yi, and B. Shi (2008), The electrical conductivity of upper-mantle rocks: Water content in the upper mantle, Phys. Chem.Miner., 35(3), 157–162.

Wang, D., Y. Guo, Y. Yu, and S.-i. Karato (2012), Electrical conductivity of amphibole-bearing rocks: Influence of dehydration, Contrib. Mineral.Petrol., 164(1), 17–25.

Watson, H. C., J. J. Roberts, and J. A. Tyburczy (2010), Effect of conductive impurities on electrical conductivity in polycrystalline olivine,Geophys. Res. Lett., 37, L02302, doi:10.1029/2009GL041566.

Yoshino, T. (2010), Laboratory electrical conductivity measurement of mantle minerals, Surv. Geophys., 31(2), 163–206.Yoshino, T., T. Matsuzaki, A. Shatskiy, and T. Katsura (2009), The effect of water on the electrical conductivity of olivine aggregates and its

implications for the electrical structure of the upper mantle, Earth Planet. Sci. Lett., 288(1–2), 291–300.

Geophysical Research Letters 10.1002/2014GL060088

AIVAZPOURPORGOU ET AL. ©2015. American Geophysical Union. All Rights Reserved. 354