Pokoknya tentang Geofisika

Seismogenic Shear Zones in the LithosphericMantle: Ultramafic Pseudotachylytes in theLanzo Peridotite (Western Alps, NW Italy)

GIOVANNI B. PICCARDO1*, GIORGIO RANALLI2 ANDLUISA GUARNIERI1

1DIPTERIS, UNIVERSITA' DI GENOVA, 16132 GENOVA, ITALY2DEPARTMENT OF EARTH SCIENCES AND OTTAWA^CARLETON GEOSCIENCE CENTRE, CARLETON UNIVERSITY,

OTTAWA, ONTARIO K1S 5B6, CANADA

RECEIVED DECEMBER 6, 2008; ACCEPTED SEPTEMBER 15, 2009ADVANCE ACCESS PUBLICATION NOVEMBER 6, 2009

At Mt. Moncuni (Lanzo Massif,Western Alps) plagioclase peri-

dotites and early mid-ocean ridge basalt (MORB) gabbroic dykes

are deformed by shear zones containing cataclastic bands and both

fault-vein and injection-vein pseudotachylytes, which are crosscut by

late MORB porphyritic dykes. Fault-vein pseudotachylytes have

thicknesses of the order of 1mm; injection-vein pseudotachylytes have

a typical thickness of 1^10 cm and contain spinifex textures.

Structural, petrological and geochemical data show that the pseudo-

tachylytes formed by near-complete melting of the host peridotite,

at ambient temperature^pressure conditions (T¼ 600�1008C,P50·5 GPa) close to the brittle^ductile transition of ultramafic

rocks, during exhumation of the lithospheric mantle in the early

stages of formation of the LigurianTethys oceanic basin. Estimates

of the average volume fraction of unmelted clasts and of the ambient

and liquidus temperature, together with thermophysical parameters,

allow the determination of the melting energy per unit volume.

Coseismic displacement is not observable at Mt. Moncuni, and con-

sequently the dynamic shear resistance cannot be inferred.We show

that commonly proposed relations between fault-vein thickness and

displacement are of limited value, given the difficulty in identifying

‘single-event’pseudotachylytes and the mobility of the melt. However,

we also show that dynamic shear resistance can be predicted to

decrease sharply if the melt coats the whole fault plane, partly as a

consequence of the nonlinear viscosity of silicate melts at high strain

rates. The Mt. Moncuni pseudotachylytes are the result of upper

mantle seismicity at shallow depth (z520 km) over a time period

of at most 5 Myr. Estimation of the total seismic energy release and

moment (caused by an unspecified number of small to moderate

earthquakes) requires an assessment of the total pseudotachylyte

volume. This is highly uncertain, with a probable qualitative error

margin of �1 order of magnitude.The inferred values of cumulative

seismic energy release and moment are of the order of 1015�1 J and

1019�1 N m, respectively, resulting in a seismic energy release

rate of approximately 108�1 J/a. This value is compatible with

present-day seismic rates at extensional plate margins.

KEY WORDS: Lanzo Massif; lithospheric mantle; ultramafic

pseudotachylytes

I NTRODUCTIONPseudotachylytes are the product of frictional meltingduring seismic slip (Sibson, 1975; Di Toro et al., 2005;Sibson & Toy, 2006). The great majority of pseudotachy-lytes are found in upper^middle crustal rocks, and occurin the brittle regime [see the classical example given bySibson (1975)], but in some cases they occur also in the duc-tile regime (T46008C;White, 1996; Lin et al., 2003; Uedaet al., 2008). Brittle ‘fault-hosted’ pseudotachylytes are usu-ally associated with cataclasites and sometimes myloniticshear zones, often in amphibolite-facies metamorphicassemblages. Sometimes, near the base of the crustal

*Corresponding author. Telephone: þ39 010 3538308. Fax: þ39 010352169. E-mail: [email protected]

� The Author 2009. Published by Oxford University Press. Allrights reserved. For Permissions, please e-mail: [email protected]

JOURNALOFPETROLOGY VOLUME 51 NUMBERS 1 & 2 PAGES 81^100 2010 doi:10.1093/petrology/egp067 at Padjadjaran U

niversity on March 8, 2012

http://petrology.oxfordjournals.org/D

ownloaded from

http://petrology.oxfordjournals.org/

seismogenic zone, rupturing and ductile shearing are pene-contemporaneous in mylonite belts (Sibson & Toy, 2006).Some pseudotachylytes have been found in subduction

accretionary complexes (e.g. Kodiak accretionary com-plex, Alaska, Rowe et al., 2005; Schistes Lustre¤ s of CapeCorse, Corsica, Austrheim & Andersen, 2004; Andersen& Austrheim, 2006; Shimanto accretionary complex,Japan, Ujiie et al., 2007). Pseudotachylytes in in situ

exhumed mantle rocks have been found in the Balmucciaperidotite, Ivrea^Verbano zone, Western Alps (Obata &Karato, 1995; Ueda et al., 2008) and in the Mt. Moncuniperidotite, Lanzo Massif, Western Alps (Piccardo et al.,2007a). At Balmuccia, the pseudotachylytes occur inspinel lherzolite. Compositions of host rock and pseudota-chylyte are similar, showing near-total melting of theformer, followed by rapid crystallization.The pseudotachy-lytes were probably formed during a late-stage, low-T(�6008C), high-stress (t�200^300 MPa) deformation epi-sode of the host rock at conditions near the brittle^ductile transition (Obata & Karato, 1995). However, somesamples from Balmuccia, overprinting a spinel lherzoliteamphibole-bearing protomylonite, were probably formedat higher temperatures (T� 700^8008C; Ueda et al.,2008).The pseudotachylytes in the Mt. Moncuni peridotitehave been described by Piccardo et al. (2007a). Here,impregnated plagioclase peridotites are cut by metre- todecametre-scale shear zones. Within the shear zones andin the host peridotite, both fault and injection pseudotachy-lyte veins (sensu Sibson, 1975) are present, postdating theductile deformation in the shear zones.Here we first review and update the stuctural, petrologi-

cal, and geochemical characteristics of the Mt. Moncunipseudotachylytes, which help to constrain the ambient con-ditions at the time of their formation. Then we discussways in which observational constraints can be used toinfer seismic parameters. In particular, we show thatfault-vein thickness is not a reliable proxy for coseismic dis-placement in cases where the latter cannot be determinedfrom observation, and use a combined linear^nonlinearrheology of silicate melts to predict a large drop from fric-tional to viscous shear resistance. Finally, on the basis ofreasonable estimates of thermophysical parameters, ambi-ent conditions and total pseudotachylyte volume, we esti-mate the cumulative seismic energy release and momentassociated with the upper mantle seismic activity that gen-erated the pseudotachylytes.

GEOLOGICAL BACKGROUND ANDCHARACTER I ST ICS OF THE MT.MONCUNI PSEUDOTACHYLYTESThe Lanzo Peridotite Massif (�150 km2; Fig. 1) is locatedin the Western Alps �30 km NW of the city of Turin(northwestern Italy). It contains a large proportion of

fresh peridotite, consisting predominantly of plagioclaselherzolite with minor amounts of spinel lherzolite, pyroxe-nite and dunite, surrounded and partially overprinted byserpentinite (see, e.g. Boudier, 1978). The Lanzo Massifwas originally interpreted as an upwelling asthenosphericdiapir that underwent low-pressure (plagioclase-facies)partial melting and melt redistribution while beingemplaced during the early stages of opening of theJurassic LigurianTethys (Nicolas, 1984, 1986). On the basisof Nd^Sr isotope studies, Bodinier et al. (1991) interpretedthe northern sector of the Lanzo body as a fragment ofsub-continental mantle that became accreted to the ther-mal lithosphere at �400^700 Ma. The Lanzo peridotitesare primarily overlain by a reduced thickness of metamor-phic oceanic crust (i.e. metabasites, Mn-rich metaquart-zites and calcschists; Lagabrielle et al., 1989; Pelletier &Mu« ntener, 2006), similar to the other Jurassic ophiolitesequences of the Western Alps and Northern Apennines.This indicates that they were emplaced at the sea-floorduring opening of the Jurassic oceanic basin.Recent investigations (see Piccardo et al., 2007b, 2009;

Piccardo, 2008, and references therein) have documentedthe complex tectonic and petrological evolution of theLanzo peridotites prior to their sea-floor exposure in theJurassic basin in the context of the extensional evolutionof the EuropeÂdria continental lithosphere and the open-ing of the Jurassic Ligurian Tethys basin. These rocks (1)were derived from the subcontinental lithospheric mantle,(2) underwent subsolidus exhumation towards shallowlevels during continental rifting, leading to Jurassic sea-floor spreading in the LigurianTethys basin, (3) underwentsignificant interaction with mid-ocean ridge basalt(MORB)-type melts via diffuse and focused porous flow,and were consequently transformed into pyroxene-depleted reactive spinel peridotites, plagioclase-enrichedimpregnated peridotites, and replacive spinel harzbur-gite^dunite channels, and (4) were later intruded byMORB gabbroic dykes and pods (e.g. Piccardo, 2009;Piccardo & Guarnieri, 2009, and references therein).The small ultramafic body at Mt. Moncuni is a satellite

of the Lanzo Massif and consists mostly of plagioclaseperidotite, with widespread early gabbroic and late subvol-canic porphyritic mafic dykes (Piccardo et al., 2007a).Both peridotites and gabbroic dykes are strongly deformedand show decametre-scale ductile shear zones containingcataclastic bands. Within the latter, millimetre-wide pseu-dotachylyte veins are present, mostly concordant with thetectonite^mylonite foliation of the shear zones and thehost rock (‘fault-vein type’ sensu Sibson, 1975; Figs 2a and3a, b), connected by ‘bridges’ of ‘injection-vein type’ sensuSibson, 1975; Fig. 3c and d). Larger (decimetre-wide,metre-long) veins of ‘injection-vein type’ propagate fromthe shear zones into the host rock, showing sharp discor-dant contacts with the peridotite (Fig. 2b^d).

JOURNAL OF PETROLOGY VOLUME 51 NUMBERS 1 & 2 JANUARY & FEBRUARY 2010

82

at Padjadjaran University on M

arch 8, 2012http://petrology.oxfordjournals.org/

Dow

nloaded from


Fig. 1. Sketch map of the Lanzo peridotite massif, in theWestern Alps and the location of the Mt. Moncuni body.1, Gran Paradiso Orthogneiss;2, Sesia^Lanzo Zone; 3, Schistes Lustres; 4, Pennine Ophiolite; 5, Pennine and Lanzo Serpentinites; 6, Lanzo Peridotites; 7, Late CenozoicSediments.

PICCARDO et al. ULTRAMAFIC PSEUDOTACHYLYTES, LANZO

83



Dow

nloaded from


Pseudotachylyte veins are widespread, covering at least�5% of the outcrop area. The porphyritic mafic dykescut across both peridotites and shear zones.

TEXTURAL AND PETROGRAPHICFEATURESPlagioclase peridotites

Plagioclase peridotites exhibit porphyroclastic texturesconsisting of a deformed spinel-facies assemblage[olivine (Ol)þ orthopyroxene (Opx)þ clinopyroxene(Cpx)þ spinel (Sp)] surrounded by a micro-granularmagmatic aggregate with a plagioclase (Plg)-rich gabbroiccomposition. Plagioclase content is up to 20% by volume.These peridotites show peculiar micro-structural fea-

tures. A first group consists of (1) rims and coronas of newunstrained olivine surrounding and partly replacingdeformed and exsolved pyroxene porphyroclasts, and (2)millimetre-size unstrained euhedral olivine crystals growninside exsolved orthopyroxene porphyroclasts. These struc-tures indicate olivine formation at the expense of mantlepyroxenes, in association with melt^rock interactioninduced by a silica-undersaturated melt (Piccardo et al.,2007a). A second group consists of (1) millimetre-sizeorthopyroxene patches and veins replacing kinked mantle

olivine, (2) isolated unstrained plagioclase crystals cuttingdeformed and kinked mantle prophyroclasts, and (3) milli-metre-size plagioclase-rich gabbroic aggregates interstitialto the deformed mantle porphyroclasts. These structuresindicate reactive interaction and interstitial crystallizationof orthopyroxene(^silica)-saturated melts (Piccardo et al.,2007a).As already described for the South Lanzo plagioclase

peridotites (Piccardo et al., 2007b), these peridotites arederived from pristine spinel-facies peridotites that under-went melt^peridotite reactive interaction and melt referti-lization by porous flow percolation and interstitialcrystallization of basaltic melts.

Tectonite^mylonite peridotite shear zones

The shear zones deforming the peridotites exhibit stronglyfoliated tectonite^mylonite fabrics, consisting of a grano-blastic fine-grained, Plg-bearing mineral aggregate, sug-gesting that early stages of deformation andrecrystallization occurred, most probably, at plagioclase-facies conditions (P51 GPa). Mg-hornblende (Mg-horn)amphiboles are widespread and in equilibrium with thePlg-bearing assemblage. Amphibole development in theshear zones suggests that they became preferential path-ways for fluid migration. Along the tectonite^mylonite

Fig. 2. Field characteristics of rocks at Mt. Moncuni: (a) millimetre-wide fault-vein pseudotachylyte in a shear zone, running parallel to themain tectonite foliation; (b^d) centimetre-wide injection-vein pseudotachylytes cutting at variable angles the main foliation of the hostperidotite.


84



Dow

nloaded from


foliation, thin, millimetre-size, near-parallel cataclasticbands are present. Late tremolitic amphibole, replacingpyroxenes and Ol, is stable with albite (Ab)-rich Plg andchlorite in the cataclastic mineral assemblage, suggestinga transition from amphibolite- to greenschist-facies condi-tions. The sequence of metamorphic assemblages and theoccurrence of hydrous minerals are indicative of progres-sive exhumation towards lower temperatures and pres-sures, increasing fluid activity, and probably transitionfrom plastic to brittle behaviour.

Pseudotachylyte veins

Within the mylonitic and cataclastic bands thin,millimetre-wide concordant pseudotachylyte veins (fault-vein type, following the terminology of Sibson, 1975) arepresent. They consist of an ultra-fine-grained to crypto- ormicrocrystalline matrix (Fig. 3a and b), showing signifi-cant amounts (510^15% by volume) of clastic olivinegrains or aggregates and lithic mylonitic or cataclastic

clasts. The ultra-fine structure suggests that they passedthrough a melt stage, whereas their peridotitic composi-tion (see below) requires that the host peridotite under-went localized almost complete melting to form theparental peridotitic melts of the pseudotachylytes. SEMbackscattered electron (BSE) images and microanalyticalinvestigations reveal that they are composed of ultra-fineaggregates of Mg-rich, cryptocrystalline minerals, mostprobably derived by low-grade metamorphic recrystalliza-tion of an original Mg-rich glass that has been completelyreplaced by low-grade minerals (e.g. serpentine minerals,chlorite, tremolitic amphibole, etc.).Larger (decimetre-wide and metre-long) sinusoidal

injection-vein pseudotachylytes bands crosscut the shearzones and extend into the deformed host peridotite show-ing very sharp contacts (injection-vein type, following theterminology of Sibson, 1975) (see Fig. 2). They are poor inclastic olivine grains or aggregates and lithic mylonitic orcataclastic clasts, and consist of fine-grained granular

Fig. 3. Photomicrographs of the ultramafic pseudotachylytes at Mt. Moncuni. (a) Fault-vein type pseudotachylyte, running within a thin cata-clastic band parallel to the foliation in a Plg þ Mg-hornblende peridotite tectonite (plane-polarized light). (b) As in (a), crossed Nicols.(c) ‘Bridge’ of injection-vein pseudotachylyte connecting two fault-vein pseudotachylytes and cutting the foliation of a Plg þ Mg-hornblendeperidotite tectonite (plane-polarized light). (d) As in (c), crossed Nicols. (e, f) Spinifex textures, made of radial aggregates of orthopyroxenelaths, in decimetre-wide injection-vein type pseudotachylyte.


85



Dow

nloaded from


aggregates of OlþOpxþCpx� spinel (Sp), where Ol pre-dominates (450% by volume) over Opx (430% byvolume) and subordinate Cpx, with interstitial cryptocrys-talline material. Well-developed spinifex stuctures (sensuDonaldson, 1982) are observed (Fig. 3e and f), consistingof radial aggregates of elongated Opx crystals with Cpxrims, surrounded by an aggregate of euhedral to roundedOl grains enclosed in interstitial cryptocrystalline material(Fig. 3g and h). The spinifex structures and the peculiarcompositions of the component minerals (see below) indi-cate very rapid crystallization at very high temperaturesin the discordant pseudotachylyte veins.The larger, decimetre-wide pseudotachylyte veins have a

clear internal structure consisting of: (1) an outer border(about 1cm thick) made of a very fine-grained, microcrys-talline aggregate most probably of olivine; (2) an internallayer (about 1cm thick) mainly composed of olivine andabundant spinel grains; (3) the inner part of the pseudota-chylyte vein showing a spinifex-like texture composed ofopx laths and interstitial cryptocrystalline material(Fig. 4a).

Microtextures in injection pseudotachylyte veins

Accurate microstructural investigations using SEM BSEimages, coupled with SEM analyses (see below) showthat, notwithstanding the presumably very rapid crystalli-zation within the injection pseudotachylytes, almost allthe rock experienced rapid cooling and the various miner-als appeared on the liquidus following a precise crystalliza-tion sequence. In addition to spinel, which showscomplicated microstructures and has been not investigatedin this study, olivine is the first liquidus mineral and formsrounded, subhedral grains distributed in the whole-rocks(Fig. 4b^g). Olivine grains grew abundantly, reaching atleast 50% of the system volume. Later, orthopyroxenelaths crystallized, both radiating from a single point andgrowing across the olivine-bearing, ‘crystal mush’-likepseudotachylyte at this stage (Fig. 4b^d), forming at theexpense of the molten material and enclosing the alreadyformed olivine grains. The interstitial spaces between theorthopyroxene laths are occupied by abundant olivinegrains that are surrounded by extremely fine-grained tocryptocrystalline material. This interstitial material is fre-quently recrystallized to an ultra-fine-grained aggregate(Fig. 4f) of most probably low-grade metamorphicminerals.The last stage of crystallization is represented by forma-

tion of clinopyroxene both as rims surrounding the outerborder of the orthopyroxene laths (Fig. 4d) and as intersti-tial grains between the rounded olivine crystals outsidethe orthopyroxene laths (Fig. 4e). Detailed analyticalwork allows us to recognize the major element composi-tions of the different minerals in the different structuralsites and the variation of some chemical parameters forthe same mineral phase during crystallization (see below).

Similar microstructural investigation of fault-vein pseu-dotachylytes to unravel primary textural features failed togive useful information, largely because these pseudotachy-lytes are composed of extremely fine-grained material industy aggregates that gave unreliable chemical composi-tions. The pristine material in the fault-vein pseudotachy-lytes, whether originally glassy or cryptocrystalline, hasbeen completely replaced by ultra-fine-grained aggregatesof secondary metamorphic minerals.

COMPOSIT IONAL FEATURES OFTHE PSEUDOTACHYLYTESAnalytical methodsBulk-rock major and trace element analyses were per-formed at ActLabs (Toronto, Canada) using the4-lithoresearch schedule. Analytical techniques anderrors can be found at http://www.actlabs.com/gg_rock_litho_usa.htm.Major element mineral chemistry and BSE images were

obtained at DIPTERIS, University of Genova, using aPhilips SEM 515 electron microscope equipped with anenergy-dispersive spectrometer (EDS) (acceleratingpotential 15 kV; sample current 20 nA). Major elementmineral chemistry data were also obtained at the Istitutodi Geoscience e Georisorse^Consiglio Nazionale delleRicerche (IGG^CNR) of Padova, Italy, using a CamecaSX50 electron microprobe equipped with four wave-length-dispersive spectrometers and one EDS spectrometer(accelerating potential 15 kV; sample current 15 nA).

Bulk-rock compositionsPlagioclase peridotites (Table 1) have relatively highAl and Ca, high Mg and low Si contents (Al2O3¼2·87^3·20wt %, CaO¼ 2·83^2·99wt %, SiO2¼43·6^45·0wt %, MgO¼ 39·6^40·5wt %). Bulk-rock chondrite-normalized rare earth element (REE) patterns are almostflat in the middle to high REE (MREE^HREE) region(Er up to 3� chondrite) and variably light REE(LREE)-depleted (CeN/SmN¼ 0·17^0·21).These character-istics are very similar to those observed in the SouthLanzo impregnated plagioclase peridotites (Piccardoet al., 2007b).Some samples from the decimetre-wide injection veins,

selected to avoid the presence of lithic clasts, also havebulk-rock peridotitic compositions (SiO2¼42·9^44·3wt %, Al2O3¼2·4^3·8wt %, CaO¼ 2·3^3·1wt %,MgO¼ 39·4^41·9wt %) (Table 1). Bulk-rock chondrite-normalized REE patterns are almost flat in the MREE^HREE region (Er up to 2� chondrite) and variablyLREE-depleted (CeN/SmN¼ 0·22^0·68). They plot withinthe compositional range of the impregnated plagioclaseperidotites of South Lanzo (Piccardo et al., 2007b).


86



Dow

nloaded from

http://www.actlabs.com/


Fig. 4. SEM BSE images of a decimetre-wide injection-vein pseudotachylyte. (a) Traverse (from lower left to top right) from the outer border tothe inner zone of the vein. The vein is concentrically zoned, and is formed by: (1) an outer, extremely fine-grained, 1cm wide zone, very richin Mg, most probably composed of cryptocrystalline olivine; (2) an intermediate, very fine-grained zone composed of Mg-olivine and abundantinterstitial grains of spinel; (3) an inner zone where the mineral grain size is larger and spinifex textures are present. This evidence points tothe following crystallization order: olivine! olivineþ spinel! orthopyroxene! clinopyroxene. (b) Details of the inner zone of the pseudo-tachylyte vein, showing the main textural setting of the spinifex-bearing area, mostly composed of orthopyroxene laths and interstitial aggre-gates of rounded olivine grains surrounded by an extremely fine-grained cryptocrystalline matrix. (c) Detail of a spinifex texture, where themain component minerals can be recognized: (1) elongated orthopyroxene laths; (2) rounded olivine grains in an interstitial cryptocrystalline(black) matrix. (d) Detail of an orthopyroxene lath: (1) enclosed euhedral olivine grains; (2) orthopyroxene; (3) clinopyroxene forming at theouter rim of the orthopyroxene laths and interstitially between olivine grains and orthopyroxene. (e) Detail of an olivine aggregate (1) sur-rounded by interstitial cryptocrystalline matrix (black) (3). It should be noted that sporadically some interstitial clinopyroxene is present (2).(f) Close-up of the cryptocrystalline material (3) between rounded subhedral olivine grains (1), with some interstitial clinopyroxene (2). Thecryptocrystalline material appears to have been significantly altered to an ultra-fine-grained aggregate of metamorphic minerals. Small areasare rich in Mg and Al, other areas are rich in Ca and Al, suggesting the possible association of chloriteâmphibole and plagioclase.


87



Dow

nloaded from


Mineral chemistryPlagioclase peridotites

Plagioclase peridotites are composed of magnesian olivine(Fo 89) and pyroxenes (orthopyroxene Mg-number 90,clinopyroxene Mg-number 92) and anorthite-rich plagio-clase (An 81^88) (Tables 2 and 4). Spinels are significantlyenriched in Ti (TiO2 0·47^0·99wt %) with respect tothose in common mantle peridotites, as is usual in spinels

from impregnated plagioclase peridotites (Piccardo et al.,2007b, and references therein). Orthopyroxene grains inthe host deformed peridotite very close to the pseudotachy-lyte veins sometimes have very high Ca contents (CaO2·79^3·34wt %).Mylonitic peridotites in the shear zones (Table 3)

are composed of magnesian olivine (Fo 89^90) and pyrox-enes (orthopyroxene Mg-number 90, clinopyroxene

Table 1: Bulk-rock major and trace element compositions of representative ultramafic rocks from Mt. Moncuni

Sample: MM1 MM15 MM21 MM13 MM16A MM17 MM18

Rock type: Plg perid. Plg perid. Shear perid. Pseudot. Pseudot. Pseudot. Pseudot.

SiO2 44·98 43·62 44·99 43·74 44·11 42·93 44·31

Al2O3 2·87 3·20 4·79 4·84 2·44 2·91 3·77

Fe2O3(T) 9·26 9·26 7·97 8·49 9·29 9·14 8·88

MnO 0·13 0·13 0·11 0·13 0·12 0·13 0·13

MgO 39·59 40·46 38·15 38·55 41·32 41·88 39·42

CaO 2·83 2·99 3·33 3·68 2·34 2·56 3·14

Na2O 0·23 0·22 0·57 0·35 0·31 0·28 0·24

K2O 0·01 0·12 0·08

TiO2 0·11 0·11 0·08 0·10 0·08 0·09 0·11

P2O5 0·01

Total 100·00 100·00 100·00 100·00 100·00 100·00 100·00

Recalculated LOI free (ppm)

La 50·05 50·05 0·07 0·13 0·22 0·13 50·05

Ce 0·23 0·2 0·27 0·44 0·5 0·38 0·2

Pb 55 55 55 55 55 55 55

Sr 3 2 41 32 5 4 12

Pr 0·07 0·07 0·06 0·09 0·07 0·07 0·05

Nd 0·55 0·54 0·43 0·61 0·45 0·44 0·43

Zr 6 54 54 4 54 54 54

Hf 0·2 0·3 0·1 0·2 0·1 0·2 0·1

Sm 0·27 0·28 0·19 0·24 0·18 0·18 0·22

Eu 0·103 0·127 0·104 0·092 0·065 0·069 0·09

Gd 0·38 0·45 0·25 0·35 0·25 0·27 0·32

Tb 0·08 0·1 0·05 0·07 0·05 0·06 0·07

Tl 50·05 50·05 50·05 50·05 50·05 50·05 50·05

Dy 0·58 0·71 0·36 0·49 0·37 0·41 0·46

Ho 0·13 0·15 0·08 0·11 0·08 0·08 0·1

Y 3·7 4·3 2·5 2·9 2·2 2·5 2·9

Er 0·4 0·48 0·25 0·33 0·27 0·27 0·32

Tm 0·061 0·073 0·042 0·05 0·04 0·045 0·049

Yb 0·39 0·47 0·28 0·32 0·27 0·29 0·33

Lu 0·06 0·069 0·043 0·046 0·044 0·042 0·051

Sc 14 15 13 13 13 13 16

V 86 104 47 50 46 49 61

Cr 2480 2950 2500 2530 2980 2790 2780

Plg perid., host plagioclase peridotite; shear perid., tectonite–mylonite peridotite in shear zones; pseudot., decimetre-wideinjection-vein pseudotachylyte; LOI, loss on ignition.


88



Dow

nloaded from


Mg-number 88^92) and anorthite-rich plagioclase (An 75^82). Amphiboles are also developed within the shearzones: (1) Mg-hornblende formed in equilibrium withplagioclase-bearing assemblages in the tectonitic^mylonitic bands; (2) tremolite replacing clinopyroxeneand olivine, formed in the cataclastic bands coexistingwith rather sodic plagioclase.

Injection pseudotachylyte veins

As previously described, the discordant decimetre-widepseudotachylyte veins consist of dominant spinifex-textureolivine grains enclosed in a matrix of orthopyroxene laths

with minor clinopyroxene rims. The spinifex structure,modal mineralogy and the compositions of the constituentminerals indicate that the injection veins formed by veryrapid crystallization at the very high temperature of anultramafic melt.The constituent minerals of the veins have unusual com-

positions (Tables 5^7). In general, the euhedral olivinegrains are highly magnesian (Mg-number 91·7^87·5); theorthopyroxene laths and clinopyroxene rims and intersti-tial grains are also magnesian (opx Mg-number 89·8^92·8and cpx Mg-number 76·3^89·9, respectively). Olivine hasunusually high contents of Cr (Cr2O3 in the range

Table 2: Mineral major element compositions of representative Plg peridotites

Sample: MM1�

Mineral: CPX PC CPX PR CPX GR OPX PC OPX PR OPX GR SP PLG

SiO2 51·52 51·73 52·03 56·59 56·86 56·29 0·12 48·27

TiO2 0·57 0·74 0·68 0·21 0·23 0·26 0·47 0·01

Cr2O3 1·36 1·12 1·27 0·68 0·64 0·69 39·32 0·00

Al2O3 4·81 4·63 3·99 2·74 2·56 2·76 25·69 32·97

Fe2O3 2·70 0·05

FeO 2·46 2·16 2·36 6·73 6·58 6·64

MnO 0·09 0·09 0·05 0·15 0·11 0·13 0·26 0·04

MgO 15·61 16·30 15·93 32·84 32·97 32·51 10·54 0·00

CaO 23·31 23·20 23·43 0·87 0·64 1·39 0·01 16·46

Na2O 0·44 0·47 0·46 0·00 0·00 0·30 0·01 2·17

K2O 0·01 0·00 0·00 0·03 0·04 0·01 0·00 0·02

Mg-no. 91·9 93·1 92·3 89·7 89·9 89·7

An 80·7

Sample: MM20

Mineral: CPX PC CPX PR CPX GR OPX PC OPX PR OPX GR PLG

SiO2 52·33 51·92 53·27 56·45 56·70 56·43 50·30

TiO2 0·48 0·51 0·60 0·14 0·16 0·18 0·02

Cr2O3 1·32 1·46 1·54 0·85 0·75 0·81 0·01

Al2O3 4·73 5·44 3·66 3·10 2·82 3·09 31·57

Fe2O3 0·03

FeO 2·80 2·74 2·64 6·69 6·41 6·24

MnO 0·08 0·07 0·07 0·14 0·12 0·08 0·01

MgO 15·81 15·55 16·17 32·47 32·89 32·66 0·02

CaO 22·75 22·82 22·62 1·36 1·43 1·56 15·30

Na2O 0·50 0·46 0·46 0·00 0·03 0·00 3·02

K2O 0·01 0·00 0·00 0·00 0·00 0·00 0·01

Mg-no. 91·0 91·0 91·6 89·6 90·1 90·3

An 73·7

PC, porphyroclast core; PR, porphyroclast rim; GR, interstitial grain.�From Piccardo et al. (2007a).


89



Dow

nloaded from


0·21^0·40wt %) and Ca (CaO in the range 0·29^0·39wt %) and, in some cases Al (Al2O3 up to 0·15^0·17wt %). Detailed observations reveal that the higherCr and Ca concentrations and the highest Mg-numbervalues (in the range 91·7^90·6) are frequently recorded byolivine grains enclosed in the orthopyroxene laths. Core^rim variations are variable and frequently contrasting, pos-sibly as a result of the very small grain size and analyticallimits. High concentrations of these elements in olivinefrom basaltic^peridotitic compositions are generallyconsidered indicative of high equilibration temperatures.

It is possible that those olivine grains enclosed in the ortho-pyroxene laths retain compositions attained during thehighest temperatures of incipient crystallization of theultramafic melt, as the formation of the orthopyroxenelaths caused isolation of the olivine from the melt and theclosure of elemental exchange. Olivine grains in the inter-stitial cryptocrystalline material have slightly lower Mg-numbers (down to 87·5).Orthopyroxene has variable contents of Ti, Al, Cr and

Ca, and is particularly rich in Cr and Ca. The more sys-tematic variations are related to the micro-textural site

Table 3: Mineral major element compositions of a deformed Plg peridotite in a shear zone

Sample: MM22

Mineral: CPX GR CPX GR CPX GR OPX GR OPX GR OPX GR OPX GR

SiO2 52·22 53·11 51·61 56·02 56·27 56·69 56·17

TiO2 0·68 0·38 0·50 0·21 0·25 0·25 0·16

Cr2O3 1·40 1·27 1·47 0·92 0·74 0·63 0·64

Al2O3 4·35 3·74 4·98 3·52 3·05 2·62 2·59

Fe2O3

FeOT 2·60 2·55 2·84 6·55 6·50 6·54 6·34

MnO 0·07 0·08 0·08 0·17 0·12 0·20 0·14

MgO 15·69 16·31 15·44 32·89 32·63 32·73 33·02

CaO 23·33 23·18 22·82 0·99 1·09 1·21 0·72

Na2O 0·48 0·47 0·52 0·06 0·03 0·01 0·00

K2O 0·00 0·00 0·02 0·00 0·04 0·00 0·00

Mg-no. 91·5 91·9 90·6 89·9 89·9 89·9 90·3

Sample: MM22

Mineral: OL GR PLG� PLG�� AMPH� AMPH��

SiO2 42·02 48·25 65·30 47·41 54·52

TiO2 0·03 0·03 0·00 0·25 0·49

Cr2O3 0·00 0·01 0·00 0·24 1·21

Al2O3 0·00 32·58 21·50 10·60 3·49

Fe2O3 0·07 1·02 0·09

FeOT 9·21 0·10 3·38 2·38

MnO 0·23 0·02 0·00 0·15 0·15

MgO 48·42 0·02 0·10 19·64 21·97

CaO 0·06 16·60 2·50 12·30 12·71

Na2O 0·00 2·13 10·40 2·51 0·81

K2O 0·00 0·03 0·06 0·03 0·07

Mg-no. 90·4

An 88·1 11·7

GR, granoblastic grain; AMPH�, hornblende amphibole; AMPH��, tremolitic amphibole; PLG�, An-rich plagioclase inmylonitic aggregates; PLG��, Ab-rich plagioclase in cataclastic matrix.


90



Dow

nloaded from


(i.e. lath core^rim, vicinity of cpx rim); the more Cr- andCa-rich compositions are confined to the cores of thelaths, presumably corresponding to the earlier crystallizedportions. In some cases the cores have very high Ca con-tents (CaO in the range 1·15^2·03 wt %), indicating veryhigh equilibration temperatures (see below).Clinopyroxene has variable contents of most of the com-

ponent elements. With respect to the two main texturalsites [i.e. cpx rims on orthopyroxene laths (Site 1) andinterstitial grains between olivine in the cryptocrystalline

matrix (Site 2)] Al decreases, whereas Mg-number andCr increase from Site 1 to Site 2. In Site 1 Al is high tovery high (Al2O3 in the range 10·0^14·51wt %),Mg-number is relatively low (in the range 76·3^86·0) andCr is low (Cr2O3 in the range 0·23^0·49wt %). Mg-number and Cr variations show a broad inverse correla-tion with similar variations in olivine. The extremely highAl contents in Cpx from basaltic^peridotitic systems arewidely recognized as records of very high equilibrationtemperatures.

Table 4: Mineral major element compositions in a Plg peridotite hosting a pseudotachylyte

Sample: MM16B

Mineral: CPX PC CPX PR CPX GR OPX PC OPX GR� OPX GR� OPX GR�

SiO2 51·18 51·21 51·88 57·00 56·84 54·43 55·77

TiO2 0·63 0·58 0·58 0·21 0·30 0·35 0·26

Cr2O3 1·19 1·57 1·26 0·69 0·80 0·69 0·72

Al2O3 5·33 4·93 4·07 2·02 2·76 3·54 2·74

Fe2O3

FeOT 3·91 3·06 3·18 6·30 6·26 6·19 6·44

MnO 0·09 0·11 0·13 0·13 0·10 0·21 0·17

MgO 16·73 15·67 16·64 32·97 32·06 30·37 30·92

CaO 20·64 23·00 21·80 1·13 2·35 2·79 3·34

Na2O 0·39 0·44 0·36 0·02 0·01 0·09 0·09

K2O 0·04 0·05 0·00 0·01 0·00 0·00 0·00

Mg-no. 88·4 90·1 90·3 90·3 90·1 89·7 89·5

An

Sample: MM16B

Mineral: OL PLG SP

SiO2 41·41 46·57 0·10

TiO2 0·01 0·14 0·99

Cr2O3 0·01 0·00 29·99

Al2O3 0·02 33·92 33·67

Fe2O3 0·42 5·96

FeOT 10·10 0·00 16·00

MnO 0·11 0·00 0·00

MgO 47·98 0·02 14·32

CaO 0·02 16·59 0·18

Na2O 0·01 2·23 0·00

K2O 0·02 0·00 0·00

Mg-no. 89·4

An 80·4y

OPX GR�, Opx crystal very close (51mm) to the contact with the injection-vein pseudotachylyte; PC, porphyroclast core;PR, porphyroclast rim; GR, interstitial grain.yFrom Piccardo et al. (2007a).


91



Dow

nloaded from


Geothermometry estimatesEstimates of equilibration temperatures have beenobtained by applying the method of Wells (1977), based oncoexisting clinopyroxeneôrthopyroxene solid solutionsand the method of Brey & Kohler (1990), based on the Cacontent in Opx. Both methods were used to estimate nom-inal equilibrium temperatures in the Plg-peridotites, Plg-bearing peridotite tectonite^mylonites in shear zones, thehost Plg-peridotites of the pseudotachylytes and the pseu-dotachylytes, using the mineral compositions reported inTables 2^7. The method of Holland & Blundy (1994) wasused, in addition, to explore amphibole^plagioclase equili-bration conditions in mylonitic and cataclastic bands from

shear zones using the mineral compositions reported inTable 3.The Wells (1977) and Brey & Kohler (1990) methods

give discrepant nominal temperatures, even when thesame opx composition is used. The Wells (1977) methodgives nominal temperature values that are systematicallylower by some 1008C than those given by the Brey &Kohler (1990) method. This most probably reflects the dif-ferent closure temperatures of elemental exchangesbetween the considered mineral phases.The application of the Wells (1977) method to cpxôpx

pairs considered to be in textural equilibrium gives thefollowing nominal temperatures: (1) 873^9018C for

Table 5: Olivine major element composition in an injection-vein pseudotachylyte

Sample: MM16B

Mineral: Ol grain� Ol grain� Ol grain� Ol grain� Ol grain� Ol grain�� Ol grain�� Ol grain��

SiO2 40·76 40·47 40·81 40·50 40·84 40·69 40·28 40·17

Cr2O3 0·40 0·31 0·09 0·25 0·07 0·21 0·24 0·29

Al2O3 0·00 0·00 0·17 0·15 0·07 0·11 0·00 0·00

FeO 8·12 8·17 8·95 9·12 9·19 9·46 9·88 12·08

MnO 0·21 0·17 0·19 0·21 0·23 0·17 0·18 0·23

MgO 50·46 50·58 50·00 49·23 49·91 49·11 49·24 47·45

NiO 0·45 0·34 0·40 0·30 0·30 0·39 0·37 0·40

CaO 0·39 0·34 0·21 0·29 0·22 0·33 0·34 0·25

Mg-no. 91·7 91·7 90·9 90·6 90·6 90·2 89·9 87·5

Ol grain�, olivine grain enclosed in an orthopyroxene lath of the spinifex texture; Ol grain��, olivine grain in the interstitialcryptocrystalline matrix surrounding opx laths.

Table 6: Orthopyroxene major element composition in a injection-vein pseudotachylyte

Sample MM16B

Mineral Opx lath�y Opx lath� Opx lath� Opx lath� Opx lath�� Opx lath�� Opx lath�� Opx lath��

SiO2 55·13 56·22 56·32 56·27 54·92 55·46 57·54 55·19

TiO2 0·19 0·07 0·08 0·08 0·26 0·26 0·12 0·26

Cr2O3 1·74 1·38 1·31 1·35 0·85 0·74 0·81 0·80

Al2O3 2·79 1·73 1·37 1·55 2·35 2·27 0·69 2·31

FeO 6·54 5·79 5·71 5·75 7·26 7·17 4·89 7·22

MnO 0·20 0·20 0·22 0·21 0·22 0·29 0·14 0·26

MgO 32·25 33·58 33·67 33·63 33·02 32·08 35·51 32·55

CaO 2·03 1·18 1·15 1·17 1·07 1·06 0·54 1·07

Mg-no. 89·8 91·2 91·3 91·2 89·0 88·9 92·8 88·9

Opx lath�, core of the orthopyroxene lath; Opx lath��, rim of the orthopyroxene lath.yFrom Piccardo et al. (2007a).


92



Dow

nloaded from


Plg-peridotites; (2) 881^9198C for Plg-bearing tectonites^mylonites in shear zones; (3) 10738C for host peridotites ofthe injection-vein pseudotachylyte; (4) 1032^11218C forpseudotachylyte (opx lath^cpx rim). Plg-peridotites andPlg-bearing peridotite tectonite^mylonites in shear zonesrecord relatively low-temperature conditions, presumablyrepresenting the ambient temperature in the Plg-faciesperidotites at the inception of shear zone formation. Thehost peridotites of the pseudotachylytes record somewhathigher temperature conditions that can provide evidenceof thermal input during intrusion in of the veins of ultra-mafic melt.The application of the Brey & Kohler (1990) method

gives more texture-related temperature information asOpx is almost ubiquitous in all rock types and texturalsites and, moreover, any specific textural site can be betterrepresented by a single Opx composition: (1) 947^11498Cfor Plg-peridotites; (2) 976^10758C for Plg-bearingtectonites^mylonites in shear zones; (3) 10858C for hostperidotites of the injection-vein pseudotachylyte; (4) 1315^14328C for opx grains very close (a few millimetres) tothe injection-vein pseudotachylyte; (5) 1092^12678C fororthopyroxene laths (core) in pseudotachylyte; (6) 910^10748C for orthopyroxene laths (rim) in pseudotachylyte.The Plg-peridotites record, in places, rather high-temperature conditions, most probably inherited from thethermal conditions of the melt impregnation event,whereas the Plg-bearing tectonite^mylonites in shearzones indicate lower temperatures that can be related toplastic deformation under Plg-facies conditions. The hostperidotites of the pseudotachylytes record slightly higher

temperature conditions that presumably reflect the ther-mal input during the ultramafic melt intrusion in veins.Orthopyroxene grains in the host peridotite that are veryclose (51mm) to the contact with a decimetre-size injec-tion pseudotachylyte vein, lacking clinopyroxene exsolu-tion lamellae and showing rather uniform compositions,frequently have very high Ca contents (CaO in the range2·35^3·34wt %) yielding very high equilibration tempera-ture estimates in the range 1315^14328C. These tempera-tures are probably related to the thermal regime locallyinduced in the host-rock at the time of injection of theultramafic melt. Orthopyroxene laths in the pseudotachy-lytes give decreasing temperature estimates from core (upto 12678C) to rim (9108C). The former can be interpretedas the temperature approaching the Opx liquidus tempera-ture in the cooling melt, whereas the latter can be relatedto closure of elemental exchange in the already solidifiedpseudotachylyte.Some general inferences related to the injection pseudo-

tachylytes can be attempted, notwithstanding the differ-ences in the nominal temperatures obtained with the twogeothermometry methods. Orthopyroxene lath composi-tions in the injection pseudotachylytes record the veryhigh liquidus temperatures of opx in the ultramafic melt,and the rapidly decreasing temperature during very rapidcrystallization and quenching.The method of Holland & Blundy (1994) has been used

to explore amphibole^plagioclase equilibration conditionsin tectonite^mylonites and cataclastic bands from theshear zones. This method has been applied to pairsof hornblende and An-rich plagioclase grains from the

Table 7: Clinopyroxene major element composition in an injection-vein pseudotachylyte

Sample: MM16B

Mineral: Cpx rim�y Cpx rim� Cpx rim� Cpx rim� Cpx rim�y Cpx int�� Cpx int�� Cpx int��

SiO2 44·10 47·97 47·57 47·16 50·27 51·68 51·57 51·46

TiO2 1·86 0·92 1·43 1·93 0·62 0·69 0·68 0·67

Cr2O3 0·23 0·49 0·37 0·25 0·44 1·44 1·47 1·49

Al2O3 14·51 10·80 10·41 10·02 7·55 4·21 4·08 3·94

FeO 7·22 5·59 6·28 6·96 6·19 3·26 3·25 3·23

MnO 0·24 0·29 0·29 0·29 0·25 0·20 0·21 0·22

MgO 13·04 15·52 14·93 14·34 17·97 15·62 15·90 16·17

NiO 0·00 0·00 0·00 0·00 0·00 0·00 0·00 0·00

CaO 17·86 18·70 18·54 18·38 16·55 22·78 22·62 22·45

Na2O 0·74 0·26 0·47 0·68 0·59 0·46 0·43 0·39

K2O 0·07 0·00 0·00 0·00 0·05 0·07 0·07 0·06

Mg-no. 76·3 83·2 80·9 78·6 83·8 89·5 89·7 89·9

Cpx rim�, cpx rim surrounding orthopyroxene lath; Cpx int��, cpx interstitial between the olivine grains in the crypto-crystalline matrix.yFrom Piccardo et al. (2007a).


93



Dow

nloaded from


tectonite^mylonite bands; these yield mean temperaturesof 8308C. Pairs of tremolitic amphibole and Ab-rich plagi-oclase grains in the cataclastic bands yield mean tempera-tures of �6008C. The higher temperature could be relatedto incipient infiltration of water during the formation ofthe sheared tectonite^mylonite peridotites, whereas thelower temperature, measured using grain compositionsfrom the clastic matrix of a cataclastic zone, may indicatethe ambient temperature conditions in the shear zones pre-ceding seismogenetic events.

Timing of pseudotachylyte formationThe timing of pseudotachylyte formation is only partiallyconstrained. Shear zones and related pseudotachylytes areyounger than the early MORB gabbroic dykes, whereasthey predate intrusion of the late MORB porphyriticmafic dykes. The age of MORB dyke intrusion in theLanzo peridotite has recently been investigated byKaczmarek et al. (2008) based on U^Pb dating of zircons.Their results point to a small age difference (5�1 Myr)between the earlier intrusion of less fractionated gabbroicrocks, characterized by Plg accumulation (163�1 Ma)and the late intrusion of the more fractionated gabbroicdykes (158�2 Ma). Kaczmarek et al. (2008) inferred thatmagmatism spanned a period of a few million years andthat the time interval is possibly linked to the chemicalevolution of the magmatic system that produced the vari-ous types of gabbroic dyke. Accordingly, taking intoaccount the observation that the pseudotachylytes formedbetween the intrusion of the more primitive and the moredifferentiated MORB dykes, we can assume that the pseu-dotachylytes and their causative seismic activity were con-fined to a relatively short period in late Jurassic times andwere related to the exhumation of the mantle peridotitesduring extension and opening of the Ligurian Tethys(Piccardo et al., 2007a, 2007b).

RELAT IONS BETWEENPSEUDOTACHYLYTE FORMATIONAND SEI SMIC PARAMETERSPseudotachylytes provide a record of seismicity at the timeof their formation. The Mt. Moncuni pseudotachylytescontain information about the seismic activity in a regionof the shallow upper mantle undergoing extension duringthe formation of the Ligurian Tethys ocean. Here we usefield and petrological data from Mt. Moncuni to constrainsome parameters of this upper mantle seismicity. First webriefly review the basic equations, discuss the reliability ofsome of the correlations that are relevant to cases such asthis when the seismic displacement is not measurable, andanalyze the consequences to be expected from a switchfrom frictional to nonlinear viscous shear resistance. Thenwe try to estimate the rate of seismic energy release

generating the Mt. Moncuni pseudotachylytes and com-pare it with present-day rates at divergent plateboundaries.If most mechanical work during seismic slip is converted

to heat, the volume of pseudotachylyte melt can be relatedto the seismic energy release and thus to the seismicmoment and magnitude. In addition, it is theoreticallypossible to constrain earthquake source parameters suchas dynamic shear resistance, average dynamic friction,and slip-weakening distance from pseudotachylyte obser-vables (e.g. Sibson, 1975; Wenk et al., 2000; Caggianelliet al., 2005; Di Toro et al., 2005, 2006; Andersen &Austrheim, 2006). However, this is often difficult, becauseof limitations to the accuracy of field information even inthe best-exposed cases, and uncertainties as to the numberof seismic events that generated a given pseudotachylytepopulation.

Melt volume, seismic energy release andfrictional shear resistanceNeglecting surface energy and work against gravity, whichare usually negligible, the energy balance in seismic fault-ing is (see, e.g. Scholz, 2002)

W � Q þ ES ð1Þ

whereW is mechanical work, Q is the generated heat, andES is the released seismic energy. The last is relativelysmall (the seismic efficiency Z� ¼ES/W is usually �0·06;McGarr, 1999). Therefore, assuming that all mechanicalwork is converted to heat, the heat generated per unitarea of fault is

Q A¼�f vt¼�f d ð2Þ

where tf is the frictional shear resistance along the ruptureplane, v is the sliding velocity (which in seismic slip is ofthe order of 0·1^1m/s), t is time, and d¼ vt is the coseismicdisplacement.The resulting temperature increase, taking place under

essentially adiabatic conditions, can be estimated on thebasis of classical solutions of the heat conduction equation(Carslaw & Jaeger, 1959), both for constant (Sibson, 1975)and time-dependent (Caggianelli et al., 2005) sliding veloc-ity. Short slip times and displacements (of the order of sec-onds and centimetres, respectively) are sufficient toexceed the melting temperature, provided that the shearresistance is sufficiently large. The heating of the host-rockis very localized and decays rapidly with time. Therefore(1) even relatively small shocks have the potential to pro-duce melting, (2) injection veins must be intruded in theambient rock penecontemporaneously with seismic slip,and (3) solidification of both fault and injection veinstakes place at most within a few tens of seconds aftermelting.


94



Dow

nloaded from


The energy required to melt a volume V of rock(neglecting superheating, and considering that unmeltedclasts do not exchange heat) is (Sibson, 1975)

EV¼r½cPðTm�ToÞ þ ð1�cÞHV ¼ E�V ð3Þ

where r, cP, and H are density, specific heat, and latentheat of melting, respectively, Tm and To are melting andambient temperature, c is the volume fraction of unmeltedclasts in the pseudotachylyte, and E� is the melting energyper unit volume. As the released seismic energy is only asmall fraction of the released energy (say, ES¼ 0·01EV),the magnitude and moment of the causative earthquakescan be estimated from the empirical relations (see Scholz,2002)

MS¼ ðlog ES�48Þ=15 ð4Þ

log Mo¼ 15MSþ91 ð5Þ

Furthermore, assuming V¼Ah, where A and h are thearea and thickness of an ideal fault-vein pseudotachylyte,direct comparison of equations (2) and (3) gives

�f¼r½cPðTm�ToÞ þ ð1�cÞH ðh=dÞ ð6Þ

which shows that shear resistance is proportional to h/d. Itsdependence on displacement depends on the relation, ifany, between thickness and displacement.There is a qualitative difference between equations (3)

and (6). Equation (3) holds for both single and multipleevents (the total pseudotachylyte volume generated by asequence of seismic shocks is related to the total seismicenergy release), whereas equation (6) is theoretically validonly for ‘single-event’ pseudotachylytes, and requires thedetermination of displacement and fault-vein thickness,which is often difficult or impossible.

Thickness^displacement relationshipTo infer shear resistance from equation (6) when the seis-mic displacement is not observable, several empirical rela-tions relating displacement to fault-vein thickness havebeen proposed. The variety of these proposed relationsper se is an indication of the wide scatter of the data. Themost complete data set available to date comes from pseu-dotachylytes in tonalites of the Gole Larghe Fault Zone,Adamello batholith (Di Toro et al., 2005), where more than60 measured h/d ratios range between 0·001 and 0·006.Whereas only a weak correlation is present between h andd (Di Toro et al., 2005, fig. 5), a better correlation existsbetween tf and d (Di Toro et al., 2005, fig. 7). This correla-tion is approximately of the type tf � d^1, which is that pre-dicted by equation (6) if h is taken as constant (i.e. inpractice, randomly variable about a central value;Fig. 5a). Because determination of h and d on a well-exposed population of pseudotachylytes is not possible atMt. Moncuni, we have checked this hypothesis by plottingtf vs h for the Gole Larghe Fault Zone data (Fig. 5b).Frictional shear resistance shows no clear correlation withpseudotachylyte fault-vein thickness.It appears, therefore, as previously pointed out by

Andersen & Austrheim (2006), that empirical relationshipsbetween thickness and displacement should be used withgreat caution, with the possible exception of areas of excep-tionally good exposure (see, e.g. Pittarello et al., 2008).Even if a connection theoretically exists, observationaluncertainties are such as to cast considerable doubt on anyproposed empirical correlation between the two variables,for at least three reasons.(1) The original thickness of melt formed during seismic

slip is not necessarily uniformly distributed along the faultplane and may vary with position according to the

Fig. 5. (a) Variations of frictional shear resistance tf with coseismicdisplacement d predicted by equation (6) assuming constant fault-vein thickness h (0·3 cm in this case). This hyperbolic dependence canbe approximately verified when sufficient data are available (see dis-cussion in the text). (b) Frictional shear resistance vs thickness h forpseudotachylytes in tonalites from the Gole Larghe Fault Zone,Adamello batholith. No clear dependence is observable. (Data fromDi Toro et al., 2005; symbols refer to rock type and method of deter-mining h; see Di Toro et al. for details.)


95



Dow

nloaded from


distribution of asperities and stress concentrations.Furthermore, some melt (sometimes most melt) migratesaway from the fault plane to form dilational injectionveins in the surrounding rock (see, e.g. Sibson, 1975; Obata& Karato, 1995). It follows that the thickness of fault-veinpseudotachylytes may in many cases have only a weak con-nection with the melt volume, and consequently withearthquake size and coseismic displacement.(2) In cases where displacement can be measured, it is

often difficult to ascertain if it is the result of a single epi-sode or multiple episodes of slip. This is particularly truefor larger slips, of the order of several metres, sometimesinferred from kinematic indicators in the host-rock.(3) In addition to the above, both thickness and displace-

ment are very difficult to measure in the field and are

subject to large uncertainies that reduce the predictivevalue of any empirically derived relation.

Frictional vs viscous shear resistanceAs soon as a continuous layer of melt is formed along therupture plane, resistance to slip is no longer controlled bysolid friction, but by the viscosity of the melt. Viscosityvaries considerably according to composition, from �106^107 Pa s for felsic melts to 1^10 Pa s for mafic and ultrama-fic melts (Clark, 1966; Dingwell, 1995). In the case of pseu-dotachylytes, melt viscosity must be corrected for theunmelted clast volume fraction, but this effect is not verylarge.Volume fractions between 0·10 and 0·40 increase vis-cosity by a factor between 1·6 and 25 (Kitano et al., 1981;Ujiie et al., 2007).Melt viscosity in pseudotachylytes is usually assumed to

be linear (strain-rate independent; e.g. Ujiie et al., 2007).However, the viscosity of silicate melts becomes nonlinearat high strain rates (�10^3 s^1; Spray, 1993; Dingwell, 1995),which are easily exceeded during seismic slip. Formally,the problem is analogous to the superposition of linearand nonlinear creep mechanisms in the mantle (seeRanalli, 1995; Dal Forno & Gasperini, 2007). The changein rheology with strain rate can be described by theequation

g¼ ð1=2ZoÞ½1þ ðg=goÞðn�1Þ=n

�m ð7Þ

where g¼ v/h is the shear strain rate (�1^103 s^1 forv¼ 0·1^1m/s, h¼ 0·1^10 cm), tm is the correspondingshear stress, Zo is the linear viscosity (g � tm for g5go), nthe stress exponent (g� tm

n for g4go, usually n� 3), andgo¼10

^3 s^1 is the transition strain rate. Therefore, interms of the shear resistance to flow,

�m¼ 2Zo½1þ ðg=goÞðn�1Þ=n

�1g¼ 2Zg¼ 2Zðv=hÞ ð8Þ

where Z¼Zo[1þ (g/go)(n^1)/n]^1 is the effective melt vis-

cosity as a function of strain rate.Effective melt viscosities are shown in Fig. 6a for three

initial (low-strain rate) viscosities ranging from 1 to 104 Pas, the lower range of which is representative of ultramaficmelts even if they contain a sizeable proportion ofunmelted clasts (up to 40%). Predicted values of viscousshear resistance for ultramafic melts (Fig. 6b) are lowerthan the frictional shear resistance for all realistic valuesof displacement and strain rates relevant to pseudotachy-lyte formation. Therefore, as pointed out by Obata &Karato (1995), frictional melting considerably decreasesslip resistance, which may result in large seismic stressdrops and moments in upper mantle earthquakes. Thishas been confirmed by high-velocity slip experiments (DiToro et al., 2006; Nielsen et al., 2008), which show a decreasein dynamic friction with increasing slip rates.The above considerations, however, hold only if melt

lubrication occurs along a large part of the fault plane,

Fig. 6. (a) Effective viscosities Z of silicate melts as functions of strainrate g for different low-strain rate (linear) viscosities Zo [see equations(7) and (8)]. Curves A, B and C illustrate the results for Zo¼10

4, 102

and 1Pa s, respectively. Melt viscosities are nonlinear (shear-thinning) in the seismic range (g�1 s^1). (b) Corresponding valuesof viscous shear resistance tm in terms of melt (fault-vein) thickness hfor seismic slip speed v¼1m/s [equation (8)]. Curves A, B and C areas in (a). In all cases, tm� tf (compare, for instance, with Fig. 5b).


96



Dow

nloaded from


with the additional condition that the absolute value of thestiffness of the ambient rock be less than the absolutevalue of the weakening. The occurrence of a thresholdbelow which melting has no sizeable effect, and the depen-dence of weakening on the fraction of the fault plane thatis frictionally melted, cannot be assessed by field studiesalone.

UPPER MANTLE SEI SMIC ITYDUR ING EXTENSION ANDOPENING OF THE L IGUR IANTETHYSIn this section, estimates of the total volume of the Mt.Moncuni ultramafic pseudotachylytes are used to assessthe approximate level of upper mantle seismicity beneaththe incipient Ligurian Tethys in late Jurassic times.Geological and petrological constraints show that the seis-mic activity that produced the pseudotachylytes was con-fined to a relatively short time span (approximately from163 to 158 Ma) and occurred at P� 0·5 GPa; that is, inthe lithospheric mantle at depths z�15^20 km, dependingon the thickness and density of the overlying thinnedcrust. The widespread occurrence of pseudotachylytes andthe lack of a clear association with (brittle or ductile)fault planes render impossible the identification of any‘single-event’melting episode. Consequently, in the follow-ing discussion we reason primarily in terms of total (cumu-lative) seismic energy release.To infer the total energy required to produce a given

volume of melt [equation (3)], estimates of thermophysicalparameters, ambient and liquidus temperatures, and meltand clast volumes are needed. Adopted values of density,specific heat and latent heat of melting are those appropri-ate for peridotite (3320 kg/m3, 1150 J/kg per 8C, and 860kJ/kg, respectively; see e.g. Bradley, 1962; Andersen &Austrheim, 2006). The ambient temperature at the time ofpseudotachylyte formation is constrained to be well below9008C (the temperature of formation of the ductile shearzones that clearly predate the pseudotachylytes; see the dis-cussion on ‘Geothermometry estimates’above). Taking intoaccount that formation of the pseudotachylytes occurredduring continuing exhumation of the host-rock, and thatfault-vein pseudotachylytes are often associated with cata-clastic bands, a valueTo¼ 600�1008C is probably applica-ble. This temperature is close to the brittle^ductiletransition of ultramafic rocks (see Ranalli, 1995) and isconsistent with geothermal gradients to be expected inregions of lithospheric extension. Therefore, the seismicactivity was of the usual ‘velocity weakening’ brittle type(see Scholz, 2002), as inferred for similar ultramafic pseu-dotachylytes in the Balmuccia peridotite (Obata &Karato,1995), and not a result of high-temperature instabil-ities, as observed in some ultramafic pseudotachylytes

(Ueda et al., 2008) and modelled, for example, byKelemen & Hirth (2007) andTimm et al. (2009).As petrological evidence shows near-complete melting of

the host peridotite, the relevant value ofTm is the liquidustemperature. The liquidus of peridotite depends on itshydration state. On the basis of the relative scarcity ofhydrous minerals, we infer that the confining conditionswere water-deficient, with the liquidus somewhere between�15008C (water-deficient conditions) and �17008C (dryconditions); we adopt Tm¼16008C. The volume fractionof clasts in the injection veins varies with position, but isusually in the 5^15% range; we take c¼ 0·10.With the above values of the parameters, the resulting

energy of melting per unit volume [equation (3)] isE� ¼ 6·39�109 J/m3. Uncertainties in the parameters areunlikely to vary this value by more than �1�109 J/m3,which is neglibible when compared with uncertainties inthe estimation of total pseudotachylyte volume. Theexpected relations between pseudotachylyte volume, totalenergy of melting, and seismic moment are shown in Fig.7.As mentioned in previous sections, pseudotachylytes

cover approximately 5% of the outcrops at Mt. Moncuni.Typical thicknesses of fault veins and injection veins are0·1^0·3 cm and 10^20 cm, respectively, with typical lengthsof about 3m. Mt. Moncuni has an approximately conicalshape with a basal radius of �1km and height� 300mabove the surrounding topography, resulting in a volumeof �3�108 m3. Assuming that the same area fractionextends to parts that are not visible, and making the sim-plest stereological extrapolation that the volume fractionis equal to the area fraction (see, e.g. Russ, 1986), the esti-mated total pseudotachylyte volume, contained in the pre-sent topography alone, is 1·5�107 m3. For an average 1%seismic efficiency, the associated cumulative seismicenergy and moment [equations (3) and (5)] are�9·6�1014 J and �1·9�1019 N m, respectively.The estimate of total pseudotachylyte volume is subject

to large and unquantifiable errors. Any extrapolation todepth is precluded by lack of observations, as is an estimateof the eroded volume. An error of one order of magnitude(e.g. if the pseudotachylyte volume were only 0·5% of thevolume of Mt. Moncuni, or if the 5% value applied to arock volume 10 times larger) would translate directly intoa similar error in the seismic energy and moment estimates(see Fig. 7). Only order-of-magnitude arguments are usedin the following discussion.If the estimated seismic energy had been released in a

single shock, the magnitude would have been �6·8 [equa-tion (4)]. However, this is clearly not the present case, asshown by the ubiquitous occurrence of pseudotachylytesand the lack of a clear slip plane. There are no observa-tional constraints at Mt. Moncuni on the average numberof ‘single-event’ pseudotachylytes associated with a seismicshock (and this number itself is in all likelihood highly


97



Dow

nloaded from


variable among seismic events), but hundreds of fault andinjection veins can be observed.A single seismic event can produce hundreds of pseudo-

tachylytes (see, e.g. Wenk et al., 2000). If, however, a‘single-event’ pseudotachylyte at Mt. Moncuni is assumedto consist of one fault vein and two injection veins (surelya lower bound assumption), and to have an approximatelysquare shape of which the observed length is a side (toaccount for the randomness of exposure), its typicalvolume would be �2^3 m3, which corresponds to MS �

2·2^2·3 and Mo � (2·5^3·8) � 1012 N m. Applying Brune’s(1970) model for small earthquakes and using the definitionof seismic moment, this hypothetical ‘single-event’ pseudo-tachylyte would be associated with a seismic shock withrupture radius and coseismic displacement in the 50^120m and 0·1^0·5 cm ranges, respectively.Although the evidence at Mt. Moncuni does not provide

any constraints on the number of seismic shocks, theabove estimate of total seismic moment is compatible withboth the physics of the earthquake source and global seis-motectonics. Most estimates of the energy per unit area offault released by earthquakes [QA, equation (2)] are in the107^108 J/m2 range (e.g. Andersen & Austrheim, 2006;Pittarello et al., 2008). The thickness of the resulting pseu-dotachylyte in ultramafic rocks is roughly QA/E� � 0·16^1·6 cm. Therefore, the estimated total volume could havebeen produced by a large number of events with limitedrupture areas (e.g. 100 events each with average rupturearea of 15 km2).On a global scale, the level of seismicity inferred from

the Mt. Moncuni pseudotachylytes is compatible withpresent-day seismicity, if stochastic uniformity of seismicactivity over time is assumed. The present-day global rate

of sesmic energy release is of the order of 1018 J/a (see, e.g.Stacey & Davis, 2008), of which510% comes from conti-nental and oceanic extension zones. Taking an estimate of105 km for the total length of extension zones, the seismicenergy release rate per unit length in these zones is51012

J/a per km. In the Mt. Moncuni area, the maximum dura-tion of seismic activity is bracketed by the age differencebetween the two generations of dykes (�5 Myr; see the dis-cussion in the section on ‘Timing of pseudotachylyte for-mation’), resulting in a minimum energy release rate of�2�108 J/a. The seismic activity could have lasted ashorter time, but further controls are lacking. However,even assuming a duration of 1Myr (and therefore �109 J/a for the seismic energy release rate) and a 10 km lengthof the local extensional boundary, the resulting energyrelease rate per unit length represented by the Mt.Moncuni pseudotachylytes would be �108 J/a per km; thatis, four orders of magnitude less than the present-day aver-age. These estimates are not to be taken at face value, butthey do show that the inferred seismic energy release atMt. Moncuni is not large when compared with present-day activity, and is robust in the sense that uncertaintiesof 1^2 orders of magnitude do not invalidate the conclu-sions.What is peculiar in this area is not the level of seismi-city, but the fact that it occurred in a region of exhumingupper mantle.

CONCLUSIONSThe occurrence of ultramafic pseudotachylytes at Mt.Moncuni shows that some segments of the upper mantlewere seismically active during lithospheric extension lead-ing to the formation of the Ligurian Tethys ocean basin.Field and petrographic observations, theoretical considera-tions on frictional and viscous shear resistance, and empir-ical seismological relations lead to the followingconclusions.(1) The pseudotachylytes formed by near-total melting of

the host peridotite at ambient temperature^pressure condi-tions T¼ 600�1008C, P50·5 GPa during progressiveexhumation of a region of lithospheric mantle that wasclose to the brittle^ductile transition in ultramafic rocks.Their age (and the maximum duration of the related seis-mic activity) is constrained to be approximately between163 and 158 Ma, on the basis of the time span of MORBgabbroic intrusion in the Lanzo peridotites.(2) Coseismic displacement is not observable at Mt.

Moncuni. The use of fault-vein thickness as a proxy for dis-placement is subject to wide and unquantifiable uncertain-ties, as shown in cases where both are measured withreasonable accuracy. Therefore, estimates of dynamicshear resistance should be limited to cases where bothquantities are measurable with an acceptable degree ofprecision.

Fig. 7. Total energy of melting EV and seismic moment Mo as func-tions of melt volume V, using the melting energy per unit volumeinferred from the conditions at Mt. Moncuni (E� ¼ 6·39�109 J/m3).Blue segments on the curves and red segments on the axes showthe estimates (within �1 order of magnitude) of these quantitiesadopted in the discussion. The total seismic energy ES is taken as1% of EV.


98



Dow

nloaded from


(3) Once melting occurs, the local shear resistancedepends on the melt viscosity, which is nonlinear (strain-rate softening) at high strain rates. The viscous shear resis-tance is orders of magnitude less than the frictional shearresistance under pseudotachylyte-forming conditions.However, any corresponding decrease in overall fault resis-tance depends on the distribution of melt on the faultplane.(4) The total seismic energy release and moment related

to the formation of the Mt. Moncuni pseudotachylytes,inferred from estimates of the energy of melting and pseu-dotachylyte volume, are of the order of 1015 J and 1019 Nm, respectively, with probable error margins of �1 orderof magnitude. These values do not require very largeearthquakes. For instance, 100 shocks with a rupture areaof 15 km2, or 1000 shocks with rupture area of 1·5 km2,would generate the required volume of melt. The inferredseismic energy release rates (108^109 J/a) are well withinpresent-day rates at extensional plate boundaries.Among the above four points, the first and the last are

based on evidence from Mt. Moncuni. The other two,although not verifiable from specific observations in thepresent case, are in our opinion of general validity. Thebasic question that remains to be answered is whetherupper mantle seismicity at relatively shallow levels inextending lithosphere, under conditions of decreasing tem-perature and pressure, is a general feature of this tectonicenvironment, or is confined to a limited number of occur-rences by some as yet unidentified factors.

ACKNOWLEDGEMENTSG.B.P.’s and L.G.’s work was supported by a grant from theItalian MIUR (Ministero dell’Istruzione, dell’Universita' edella Ricerca) (PRIN2007: ‘Mantle heterogeneity and geo-dynamic evolution of the lithosphereâsthenospheresystem’) and the University of Genova. G.R.’s work is sup-ported by a grant from NSERC (Natural Sciences andEngineering Research Council of Canada). The authorsare grateful to the Editor, Marjorie Wilson, two anon-ymous reviewers and Masaaki Obata for their criticism ofan earlier version of the manuscript and for many sugges-tions for improving the final version of the paper.

REFERENCESAndersen, T. B. & Austrheim, H. (2006). Fossil earthquakes recorded

by pseudotachylytes in mantle peridotite from the Alpine subduc-tion complex of Corsica. Earth and Planetary Science Letters 242,58^72.

Austrheim, H. & Andersen, T. B. (2004). Pseudotachylytes fromCorsica: fossil earthquakes from a subduction complex.Terra Nova

16, 193^197.Bodinier, J. L., Menzies, M. A. & Thirlwall, M. F. (1991). Continental

to oceanic mantle transitionçREE and Sr^Nd isotopic geochemis-try of the Lanzo lherzolite massif. In: Menzies, M. A., Dupuy, C.

& Nicolas, A. (eds) Orogenic Lherzolites and Mantle Processes. Journal

of Petrology, Special Lherzolite Issue, pp. 191^210.Boudier, F. (1978). Structure and petrology of the Lanzo peridotite

massif (Piedmont Alps). Geological Society of America Bulletin 89,1574^1591.

Bradley, R. S. (1962).Thermodynamic calculations on phase equilibriainvolving fused salts, Part II: Solid solutions and application to theolivines. AmericanJournal of Science 260, 550^554.

Brey, G. P. & Ko« hler, T. (1990). Geothermobarometry in four-phaselherzolites, II: New thermobarometers, and practical assessment ofexisting thermobarometers. Journal of Petrology 31, 1353^1378.

Brune, J. (1970). Tectonic stress and the spectra of seismic shear wavesfrom earthquakes. Journal of Geophysical Research 75, 4997^5009.

Caggianelli, A., de Lorenzo, S. & Prosser, G. (2005). Modelling theheat pulses generated on a fault plane during coseismic slip:Inferences from the pseudotachylytes of the Copanello cliffs(Calabria, Italy).Tectonophysics 405, 99^119.

Carslaw, H. S. & Jaeger, J. C. (1959). Conduction of Heat in Solids, 2ndedn. Oxford: Clarendon Press.

Clark, S. P., Jr (1966). Viscosity. In: Clark, S. P., Jr (eds) Handbook ofPhysical ConstantsçRevised Edition. Geological Society of America,

Memoirs 97, 291^300.Dal Forno, G. & Gasperini, P. (2007). Modelling of mantle postglacial

relaxation in axisymmetric geometry with a composite rheologyand a glacial load interpolated by adjusted spherical harmonicanalysis. GeophysicalJournal International 169, 1301^1314.

Dingwell, D. B. (1995). Viscosity and anelasticity of melts. In:Ahrens, T. J. (ed.) Mineral Physics and CrystallographyçA Handbook

of Physical Constants. American Geophysical Union, Reference Shelf 2,209^217.

DiToro, G., Pennacchioni, G. & Teza, G. (2005). Can pseudotachylytesbe used to infer earthquakes source parameters? An example oflimitations in the study of exhumed faults.Tectonophysics 402, 3^20.

Di Toro, G., Hirose, T., Nielsen, S. & Shimamoto, T. (2006). Relatinghigh-velocity rock-friction experiments to coseismic slip in the pre-sence of melts. In: Abercrombie, R., McGarr, A., Di Toro, G. &Kanamori, H. (eds). Earthquakes: Radiated Energy and the Physics of

Faulting. American Geophysical Union, Geophysical Monograph 170,121^134.

Donaldson, C. H. (1982). Spinifex-textured komatiites: A review oftextures, compositions and layering. In: Arndt, N. T. &Nisbet, E. G. (eds) Komatiites. London: Allen & Unwin,pp. 211^244.

Holland,T. & Blundy, J. (1994). Non-ideal interaction in calcic amphi-boles and their bearing on amphibole^plagioclase thermometry.Contributions to Mineralogy and Petrology 116, 433^447.

Kaczmarek, M.-A., Mu« ntener, O. & Rubatto, D. (2008). Trace ele-ment chemistry and U^Pb dating of zircons from oceanic gabbrosand their relationship with whole rock composition (Lanzo,Italian Alps). Contributions to Mineralogy and Petrology 155, 295^312.

Kelemen, P. B. & Hirth, G. (2007). A periodic shear-heating mechan-ism for intermediate-depth earthquakes in the mantle. Nature 446,787^790.

Kitano,T., Kataoka,T. & Shirota, T. (1981). An empirical equation ofthe relative viscosity of polymer melts filled with various inorganicfillers. Rheologica Acta 20, 207^209.

Lagabrielle,Y., Fudral, S. & Kienast, J. R. (1989). La couverture oce¤ a-nique des ultrabasites de Lanzo (Alpes Occidentales): Argumentslithostratigraphiques et pe¤ trologiques. Geodinamica Acta 3, 43^55.

Lin, A., Sun, Z. & Yang, Z. (2003). Multiple generations of pseudota-chylyte in the brittle to ductile regimes, Qinling^Dabie Shan ultra-high-pressure metamorphic complex, central China. Island Arc 12,423^435.


99



Dow

nloaded from


McGarr, A. (1999). On relating apparent stress to the stress causingearthquake fault slip. Journal of Geophysical Research 104, 3003^3011.

Nicolas, A. (1984). Lherzolites of the Western Alps: A structuralreview. In: Kornprobst, J. (ed.) Fifth International Kimberlite

Conference Proceedings. Amsterdam: Elsevier, pp. 333^345.Nicolas, A. (1986). A melt extraction model based on structural studies

in mantle peridotites. Journal of Petrology 27, 999^1022.Nielsen, S., Di Toro, G., Hirose,T. & Shimamoto,T. (2008). Frictional

melt and seismic slip. Journal of Geophysical Research 113, B01308,doi:10.1029/2007JB005122.

Obata, M. & Karato, S.-I. (1995). Ultramafic pseudotachylyte fromthe Balmuccia peridotite, Ivrea^Verbano zone, northern Italy.Tectonophysics 242, 313^328.

Pelletier, L. & Mu« ntener, O. (2006). High pressure metamorphism ofthe Lanzo peridotite and its oceanic cover, and some consequencesfor the Sesia^Lanzo zone (northwestern Italian Alps). Lithos 90,111^130.

Piccardo, G. B. (2008). The Jurassic LigurianTethys, a fossil ultra-slowspreading ocean: the mantle perspective. In: Coltorti, M. &Gre¤ goire, M. (eds) Metasomatism in Oceanic and Continental

Lithospheric Mantle. Geological Society, London, Special Publications 293,11^33.

Piccardo, G. B. (2009). The Lanzo peridotite massif, Italian WesternAlps: Jurassic rifting of the Ligurian Tethys. In: Coltorti, M.,Gre¤ goire, M., Downes, H., O’Reilly, S. (eds), PetrologicalEvolution of the European Mantle.The Geological Society of London,

Special Publications (in press).Piccardo, G. B. & Guarnieri, L. (2009). Alpine peridotites from the

Ligurian Tethys: an updated critical review. International Geology

Review (in press).Piccardo, G. B., Ranalli, G., Marasco, M. & Padovano, M. (2007a).

Ultramafic pseudotachylytes in the Mt. Moncuni peridotite(Lanzo Massif, Western Alps): Tectonic evolution and uppermantle seismicity. Periodico di Mineralogia 76, 181^197.

Piccardo, G. B., Zanetti, A. & Mu« ntener, O. (2007b). Melt^peridotiteinteraction in the South Lanzo peridotite: field, textural and geo-chemical evidence. Lithos 94, 181^209.

Piccardo, G. B.,Vannucci, R. & Guarnieri, L. (2009). Evolution of thelithospheric mantle at an extensional setting: insights from ophioli-tic peridotites. Lithosphere 1(2), 81^87.

Pittarello, L., Di Toro, G., Bizzarri, A., Pennacchioni, G.,Hadizadeh, J. & Cocco, M. (2008). Energy partitioning during

seismic slip in pseudotachylyte-bearing faults (Gole Larghe Fault,Adamello, Italy). Earth and Planetary Science Letters 269, 131^139.

Ranalli, G. (1995). Rheology of the Earth, 2nd edn. London: Chapman &Hall.

Rowe, C. D., Moore, J. C., Meneghini, F. & McKierman, A. W.(2005). Large-scale pseudotachylytes and fluidized cataclasitesfrom an ancient subduction thrust fault. Geology 33, 937^940.

Russ, J. C. (1986). Practical Stereology. NewYork: Plenum.Scholz, C. H. (2002).The Mechanics of Earthquakes and Faulting, 2nd edn.

Cambridge: Cambridge University Press.Sibson, R. H. (1975). Generation of pseudotachylyte by ancient seismic

faulting. Geophysical Journal of the Royal Astronomical Society 43,775^794.

Sibson, R. H. & Toy, V. G. (2006). The habitat of fault-generatedpseudotachylyte: Presence vs. absence of friction-melt. In:Abercrombie, R., McGarr, A., Di Toro, G. & Kanamori, H. (eds)Earthquakes: Radiated Energy and the Physics of Faulting. American

Geophysical Union, Geophysical Monograph 170, 153^166.Spray, J. G. (1993).Viscosity determinations of some frictionally gener-

ated silicate melts: Implications for fault zone rheology at highstrain rates. Journal of Geophyical Research 98, 8053^8068.

Timm, J., Medvedev, S., Ru« pke, L. H., Andersen, T. B.,Podladchikov, Y. Y. & Austrheim, H. (2009). Generation of inter-mediate-depth earthquakes by self-localizing thermal runaway.Nature Geoscience 2, 137^140.

Stacey, F. D. & Davis, P. M. (2008). Physics of the Earth, 4th edn.Cambridge: Cambridge University Press.

Ueda, T., Obata, M., Di Toro, G., Kanagawa, K. & Ozawa, K.(2008). Mantle earthquakes frozen in mylonitized ultramaficpseudotachylytes of spinel-lherzolite facies. Geology 36, 607^610.

Ujiie, K., Yamaguchi, H., Sakaguchi, A. & Toh, S. (2007).Pseudotachylytes in an ancient accretionary complex and implica-tions for melt lubrication during subduction zone earthquakes.Journal of Structural Geology 29, 599^613.

Wells, P. R. A. (1977). Pyroxene thermometry in simple and complexsystems. Contributions to Mineralogy and Petrology 62, 129^139.

Wenk, H.-R., Johnson, L. R. & Ratschbacher, L. (2000).Pseudotachylytes in the Eastern Peninsular Ranges of California.Tectonophysics 321, 253^277.

White, J. C. (1996). Transient discontinuities revisited: pseudotachy-lyte, plastic instability and the influence of low pore fluid pressureon deformation processes in the mid-crust. Journal of Structural

Geology 18, 1471^1486.


100



Dow

nloaded from


Documents

Pokoknya tentang Geofisika