Hydrothermal alteration, fluid flow and volume change in shear zones: the layered mafic–ultramafic Kettara intrusion (Jebilet Massif, Variscan belt, Morocco

Hydrothermal alteration, fluid flow and volume change in shearzones: the layered mafic–ultramafic Kettara intrusion (JebiletMassif, Variscan belt, Morocco)

A. ESSAIF I , 1 R. CAPDEVILA,2 S . FOURCADE,2 J . L . LAGARDE,3 M. BALLEVRE3 AND CH. MARIGNAC4

1Departement de Geologie, B.P. 2390, Faculte des Sciences Semlalia, 40000 Marrakech, Maroc ([email protected])2Geosciences Rennes, UMR 6118 CNRS, Universite de Rennes 1, 35042 Rennes Cedex, France3M2C (UMR 6143 CNRS), Universite de Caen, 14032 Caen Cedex, France4Ecole des Mines de Nancy, parc de Saurupt, 54042 Nancy Cedex, France

ABSTRACT During emplacement and cooling, the layered mafic–ultramafic Kettara intrusion (Jebilet, Morocco)underwent coeval effects of deformation and pervasive fluid infiltration at the scale of the intrusion. Inthe zones not affected by deformation, primary minerals (olivine, plagioclase, clinopyroxene) werepartially or totally altered into Ca-amphibole, Mg-chlorite and CaAl-silicates. In the zones of activedeformation (centimetre-scale shear zones), focused fluid flow transformed the metacumulates(peridotites and leucogabbros) into ultramylonites where insoluble primary minerals (ilmenite, spineland apatite) persist in a Ca-amphibole-rich matrix. Mass-balance calculations indicate that shearing wasaccompanied by up to 200% volume gain; the ultramylonites being enriched in Si, Ca, Mg, and Fe, anddepleted in Na and K. The gains in Ca and Mg and losses in Na and K are consistent with fluid flow inthe direction of increasing temperature.

When the intrusion had cooled to temperatures prevailing in the country rock (lower greenschistfacies), deformation was still active along the shear zones. Intense intragranular fracturing in the shearzone walls and subsequent fluid infiltration allowed shear zones to thicken to metre-scale shear zones withtime. The inner parts of the shear zones were transformed into chlorite-rich ultramylonites. In the shearzone walls, muscovite crystallized at the expense of Ca–Al silicates, while calcite and quartz weredeposited in �en echelon� veins. Mass-balance calculations indicate that formation of the chlorite-richshear zones was accompanied by up to 60% volume loss near the centre of the shear zones; the ultra-mylonites being enriched in Fe and depleted in Si, Ca, Mg, Na and K while the shear zones walls areenriched in K and depleted in Ca and Si. The alteration observed in, and adjacent to the chlorite shearzones is consistent with an upward migrating regional fluid which flows laterally into the shear zone walls.Isotopic (Sr, O) signatures inferred for the fluid indicate it was deeply equilibrated with host lithologies.

Key words: fluid flow; gabbros; Morocco, shear zones; volume change.

INTRODUCTION

During regional metamorphism and orogenesis, tem-perature gradients and deformation are proposed as themain driving forces for fluid flow in the crust (e.g.Etheridge et al., 1983; Yardley, 1986; Ferry, 1994;Oliver, 1996). Large-scale pervasive fluid flow is oftenin the direction of increasing temperature (McCaig &Knipe, 1990; Ferry & Dipple, 1991; Ferry, 1994)whereas focused fluid flow is most likely upwardlydirected (Beach & Fyfe, 1972; Kerrich et al., 1977;Ferry & Dipple, 1991). Fluid flow behaviours may bedifferent in the vicinity of hot intrusions that intrudedat the time of deformation and metamorphism. Thesesyntectonic and synmetamorphic intrusions recordboth regional metamorphic fluid flow and �local� fluidflow related to thermal effects of the intrusions. Theygive rise to thermal perturbations and develop con-vection cells that circulate fluids between the intrusions

and the host rocks (Taylor, 1974; Leger & Ferry, 1993;Ferry, 1994). Of particular interest are intrusions thatare emplaced at shallow crustal levels and give rise tothermal softening in the surrounding rocks (White &Knipe, 1978; Brun & Cobbold, 1980), induce rheo-logical heterogeneities and localize deformation, con-sequently they should also localize fluid flow.

This paper presents structural and chemical evidencefor large-scale fluid infiltration during regional meta-morphism, deformation, emplacement and cooling ofthe syntectonic layered mafic–ultramafic Kettaraintrusion (Jebilet massif, Morocco). We argue that thefluid flow can be related to two main episodes (i) anepisode of cooling of the intrusion to temperatureconditions prevailing in the country rock (lowergreenschist facies), during which an up-temperaturepervasive fluid flow developed at the scale of the intru-sion (ii) an episode following the thermal re-equilibra-tion of the intrusion, during which a regional upwardly

J. metamorphic Geol., 2004, 22, 25–43 doi:10.1111/j.1525-1314.2004.00495.x

� 2004 Blackwell Publishing Ltd 2 5

directed fluid flow is focused along interconnectedmetre-scale shear zones that crosscut the intrusion.

GEOLOGICAL SETTING

The Jebilet massif, located north of Marrakech, pre-sents an E–W section of the Variscan belt of Morocco

(Huvelin, 1977) (Fig. 1a,b). The Kettara intrusion islocated in the Central Jebilet unit which corresponds toa crustal block bounded by two major crustal shearzones (Fig. 1b). A NNE trending thrust-wrench faultseparates the Central and the Western Jebilet Units(Bouloton & Le Corre, 1985; Le Corre & Bouloton,1987; Mayol, 1987). This zone is a part of the west

(a)

(b)

(c)

Fig. 1. (a) Inset: The Jebilet massif in the framework of the Palaeozoic outcrops of North Africa (in grey). (b) Geological sketch mapof the Jebilet massif (modified from Huvelin, 1977). Box encloses area covered by Fig. 1(c). (c) Ductile and brittle shear zones ofCentral Jebilet. Ductile shear zones correspond to arrays of anastomozing shear zones, which seem restricted to zones of thermalinfluence of magmas. Outside these zones, ductile shear zones evolve into brittle domains. Note the location of the Kettara intrusion inan area where dextral shearing connects sinistral shear zones (MSZ, Marrakech shear zone). Large arrows: shortening (k3) andstretching (k1) regional directions. The framed intrusions correspond to the locations of rocks isotopically studied in Table 2: Kt.Kettara: MGTK and MTK felsic samples and all the gabbroic and mylonitic samples (GK2, DK13, MK5, GK3, DK25, MK4, MK3,DK12, DK19, DK26). Kt. Bouzlaf: TBZ sample and host schists (BF3, BF8, BF10) sampled around the intrusion. Kt. Hamra: KAZsample.

2 6 A . E S S A I F I E T A L .

� 2004 Blackwell Publishing Ltd

Moroccan shear zone that extends from Tichka in theSouth to Rabat in the North (Pique et al., 1980;Lagarde & Michard, 1986). A SSE transcurrent shearzone separates the Central and the Eastern JebiletUnits (Marrakech shear zone; Lagarde & Choukroune,1982). Between these major shear zones (WMSZ,MSZ), Variscan deformation in Central Jebilet ismainly accommodated by minor shear zones localizedin thermally softened aureoles around magmatic stocks(Essaifi, 1995; Essaifi et al., 2001b). These minor shearzones are organized into interconnected N–S sinistraland ENE dextral shear zones (Fig. 1c).

The central Jebilet unit consists of Upper Viseanmarine sedimentary rocks that have been folded anddeveloped a schistose fabric during very low to low-grade metamorphism. This unit is characterized by awidespread bimodal magmatism (Bordonaro, 1983),represented by numerous felsic, mafic or compositestocks which form small elongated intrusions a fewhundreds of metres in thickness and a few kilometresin length (Fig. 1b,c). The mafic and felsic intrusionstrend N–S to NE–SW, parallel to the Hercynianstructures and are spatially associated with poly-metallic sulphide mineralizations (Bernard et al., 1988;

Essaifi et al., 1995). The magmatic bodies wereemplaced during the post-Visean, Hercynian shorten-ing (Essaifi et al., 2001b).

The largest mafic intrusion in the Central JebiletUnit is the Kettara massif (Aarab, 1984; Jadid, 1989;Essaifi, 1995), located 30 km north-west of Marrakech(Fig. 1c). The Kettara massif is a stratified intrusionoutcropping over about 2.5 km2. It consists of med-ium- to coarse-grained ultramafic cumulates (plagio-clase-bearing wherlites, troctolites and olivine-bearinggabbros) and mafic cumulates (massive and laye-red leucogabbros) (Fig. 2a). The grains, commonly0.5–1 cm in diameter, contain numerous intragranularfractures filled with secondary minerals. Magmatictextures are preserved in spite of an incipient to mod-erate recrystallization. The magmatic minerals includeolivine, clinopyroxene, plagioclase, spinel, ilmenite andapatite. Variations in modal composition within theintrusion define an ultramafic unit, that occurs alongthe western part of the intrusion, and a mafic unitwhich is dominant in its eastern part (Fig. 2). Thecoarse-grained ultramafic and mafic rocks are sur-rounded by a narrow zone of fringing microgabbros.Numerous felsic and mafic dykes crosscut both the

Fig. 2. (a) Geological sketch map of the Kettara intrusionshowing the distribution of major rock types and the dykecomplex (other minor intrusions are omitted). (b) Map-scaleschistosity trajectories (S1) within the layered Kettaraintrusion and host rocks. Note (i) the sigmoidal pattern of S1

along ENE dextral shear zones (large arrows) (ii) the con-jugate metre-scale ductile shear zones (little arrows) withinthe intrusion (only representative shear zones are illustra-ted). Framed numbers correspond to locations of samplesin Table 1 (1: DK17-18–19, 2: DK20-21–22, 3: DK27-28–29,4: MK5-4-3, 5: DK30-31–32, 6: DK25-26, 7: DK13-MK2, 8:DK3-4-5, 9: DK10-11–12, 10: DK14-15–16, 11: DK7-8-9).(c) Vertical cross section through the Kettara intrusion. Thediagrammatic section (horizontal plane) illustrates hetero-geneity of deformation in the intrusion (metre-scale shearzones (in grey) bounding weakly deformed lenticulardomains) and development of quartz-calcite veins in theunfoliated domains.

F L U I D F L O W I N M AF I C– UL TR AM A F I C S H E A R Z O N E S 2 7


intrusion and the adjacent country-rocks (Fig. 2a).The bottom of the intrusion can be observed in a smallwindow. Whole-rock chemical analyses show that theultramafic and mafic cumulates derive from a parentaldepleted tholeiitic magma (Essaifi, 1995).

The Kettara region exemplifies the style of defor-mation in Central Jebilet. Ductile deformation ischaracterized by a single and widespread subverticalschistosity bearing a stretching lineation subparallel tolarge-scale fold hinges. This regional schistosity strikesNNE–SSW and its trajectories display a progressivecurvature in the vicinity of the intrusion where it be-comes oriented ENE–WSW, indicating that the intru-sion was emplaced in an area of ENE dextral shearing(Fig. 2b). The NE–SW orientation of the intrusion isin accordance with the regional strain field and con-sistent with the NW–SE shortening recognized at thescale of the Jebilet massif (Essaifi et al., 2001b).

In the gabbroic rocks of the Kettara intrusion,deformation is very heterogeneous. Narrow andstrongly deformed zones bound lenticular metre to100 m-scale domains of weakly deformed to unde-formed gabbros (Fig. 3a,b). Deformation is concen-trated within two sets of interconnected, metre tocentimetre-scale shear zones. The dominant set strikesENE and the sense of shear inferred from finite straintrajectories (Gapais, 1989), sigmoidal veins (Ramsay,1967), and multiscale S ⁄C shear bands (Berthe et al.,1979) indicate dextral displacements. The minor set issinistral and strikes NNE. The principal finite strainshortening direction bisects the acute angle made bythe two sets of shear zones and is NW–SE and hori-zontal. This shortening direction corresponds to thatinferred from foliation trajectories, major fold trends,and syn-folding conjugate wrench faults at theMoroccan Meseta scale (Lagarde et al., 1990).

SAMPLING AND ANALYTICAL TECHNIQUES

Whole rock and isotope analyses were undertaken on selected sam-ples of variably deformed and altered mafic-ultramafic rocks of theintrusion as a means of investigating the effects of deformation andfluid–rock interaction.

Several margin-to-centre horizontal sampling traverses were madethrough the Kettara shear zones, including the weakly deformed toundeformed gabbros in the lenticular domains through the highlydeformed rocks in the inner parts of shear zones (Table 1). In eachtraverse, 3–5 kg samples were taken 0.5–2 m apart.

In leucogabbros, two traverses (DK17-18–19 & DK20-21–22)correspond to samples taken in two interconnected m-scale shearzones; the two traverses being located �30 m apart. Two othertraverses (DK27-28–29 & DK30-31–32), spaced approximately 3 mapart, were taken on the two sides of a metre-scale shear zone.Another traverse (MK5-4-3) was taken in another m-scale shear zonecrosscutting leucogabbros (Fig. 2b).

In peridotites, four traverses (DK3-4-5, DK7-8-9, DK10-11–12 &DK14-15–16) were taken in different interconnected metre-scaleshear zones. Two other traverses that include two samples in eachsection were taken. The DK25-26 section corresponds to a centi-metre-scale shear zone where the two collected samples are spaced< 0.5 m apart. Connection between this centimetre-scale shear zoneand other shear zones was not observed. The DK13-MK2 sectioncorresponds to a metre-scale chlorite-rich shear zone. The MK2

sample, collected in the centre of the shear zone, corresponds to anamphibole-rich schist preserved within chlorite-schists. The DK13sample was taken in the adjacent weakly deformed metaperidotitesapproximately 3 m from the MK2 sample.

Analytical techniques

Mineral compositions were determined by microprobe analysis witha Cameca SX 50 (Microsonde Ouest, Brest, France). Whole-rockchemical analyses were performed by X-ray fluorescence at theUniversity of Rennes (France). The precision for major elements isestimated at 1–5%, except for MnO and P2O5 (10%). For traceelements, precisions are of the order of 10% for concentrations lowerthan 30 p.p.m. For concentrations higher than 30 p.p.m., the preci-sion is 3%. Specific gravities were measured in water using apycnometer. Before measurement, the rock sample was degassed inwater during 4 h. Each measure was repeated at least three times, andthe deduced uncertainty is ± 0.01 g cm)3.

Selected samples have been analysed for rare earth elements byICP MS at the C.R.P.G. (Nancy, France). Precisions are of the orderof 10% for concentrations lower than one p.p.m. For concentrationshigher than one p.p.m., the precision is 5%.

Sr isotopic analyses were carried out at the University of Rennesusing the Faraday cups of a five-collector Finnigan MAT 262mass spectrometer. All measured 87Sr ⁄ 86Sr ratios were normalized to86Sr ⁄ 88Sr ¼ 0.1194, and were measured relative to NBS 987 SrStandard ¼ 0.71020. The error of 87Sr ⁄ 86Sr ratios reported inTable 2, including the statistical error obtained during the massspectrometer run and other error sources such as instrumentalreproducibility, is estimated to be ± 3.

Chlorite, amphibole and clinopyroxene fractions were separ-ated from selected samples and were analyzed for O or Sr isotopes(Table 2). Oxygen was extracted using BrF5 (Clayton & Mayeda,1963), converted to CO2 gas and analysed on a VG SIRA 10 massspectrometer at the university of Rennes. Isotopic compositions,quoted in the d notation relative to V-SMOW were normalizedusing internal standards (basaltic glass Circe 93) and internationalreference materials (carbonate NBS 19, quartz NBS 28). Totaluncertainties with respect to the SMOW scale are estimated to be0.1–0.15 d unit.

MINERAL ALTERATION WITHIN THE KETTARAINTRUSION

Mineral alteration within the undeformed cumulates

The ultramafic unit

The ultramafic unit consists of peridotites (Fig. 3c)corresponding to orthocumulates with 55–70% olivineand 1–2% chrome spinel as the cumulus phases;endiopside (Mg-rich augite), plagioclase and ilmeniteas the intercumulus minerals (20–40%). From thebottom to the top of the ultramafic unit, modal vari-ations correspond, respectively, to plagioclase-bearingwherlites, troctolites, and olivine-bearing gabbros(Jadid, 1989). Olivine (Fo 87–88) exists as residualgrains (1–5 mm) partially replaced by tremolite (Mg#¼ 0.81–0.96) or serpentine (Mg# ¼ 0.90), and containschrome spinel (Cr# ¼ 0.5–0.9) surrounded by horn-blende. Clinopyroxene (Mg# ¼ 0.86, Al2O3 ¼ 3%) ispresent as residual patches within an alteration matrixmade of chlorite (XMg ¼ 0,9), amphibole ranging incomposition from hornblende to actinolite ⁄ tremolite,and rarely epidote (Ps 8–18%) between olivine ghosts.

2 8 A . E S S A I F I E T A L .


Clinopyroxene has a clear appearance and containsinclusions of spinel and ilmenite. Plagioclase (An82%)is rarely preserved in the ultramafic rocks; it is usuallytransformed into microcrystalline aggregates ofchlorite, fine-grained muscovite (sericite), prehnite andepidote.

The mafic unit

The mafic unit consists of massive and layered leuco-gabbros. The massive leucogabbbros crosscut andoverly the ultramafic unit; they also form a few maficdykes which crosscut the whole intrusion (Fig. 2a). The

Fig. 3. Microstructures and metamorphic assemblages: (a) metre-scale shear zones crosscutting ultramafic metacumulates;(b) centimetre-scale interconnected shear zones bounding metre-scale lenticular domains of unfoliated leucogabbros; (c) undeformedultramafic cumulates (wherlite) made of olivine (Ol) (cumulus phase) partially altered into amphibole, and an intercumulus phase(plagiocase + clinopyroxene) completely altered into chlorite (Chl) (d) Undeformed leucogabbros made of olivine completelytransformed into amphibole (Amp) and plagioclase (PLG) (partially altered into prehnite, epidote and sericite) as a cumulus phase,and clinopyroxene (CPX) as the intercumulus phase (e) foliated ultramafic cumulates (spotted schists) made of stretched olivine alteredinto amphibole and a chlorite + amphibole matrix (f) Augen chlorite schist made of albitized plagioclase clasts and a chlorite +amphibole matrix.



Table

1.

Geo

chem

ical

data

of

mylo

nit

esan

dth

eir

pro

tolith

sfr

om

the

Ket

tara

shea

rzo

nes

.S

am

ple

sex

ten

dfr

om

the

wea

kly

def

orm

edto

un

def

orm

edgab

bro

s(*

)in

the

len

ticu

lar

do

main

sto

the

hig

hly

def

orm

edro

cks

(***)

inth

ein

ner

part

so

fsh

ear

zon

esth

rou

gh

the

mo

der

ate

lyd

efo

rmed

gab

bro

s(*

*).

Dep

end

ing

on

the

wid

tho

fth

esh

ear

zon

esa

mp

led

,ea

chp

rofi

leco

mp

rise

stw

oo

rth

ree

sam

ple

s(f

or

loca

tio

ns,

see

Fig

.2b

).S

pec

ific

gra

vit

y(i

ng

cm)

3)

has

bee

nm

easu

red

inw

ate

r.P

rd:

un

def

orm

edp

erid

oti

te;

F-P

rd:

fract

ure

dp

erid

oti

te;

My-P

rd:

mylo

nit

ized

per

ido

tite

;L

gb

:u

nd

efo

rmed

leu

cogab

bro

;F

-Lgb

:fr

act

ure

dle

uco

gab

bro

;M

y-L

gb

:m

ylo

nit

ized

leu

cogab

bro

;T

r-sc

h:

trem

olite

sch

ist;

Ch

l-sc

ht:

chlo

rite

sch

ist;

Tr+

Ch

l-sc

ht:

trem

olite

an

dch

lori

teb

eari

ng

sch

ist.

Sam

ple

Lit

ho

logy

DK

20*

Lgb

DK

21**

F–L

gb

DK

22***

Ch

l–sc

h

DK

17*

Lgb

DK

18**

F–L

gb

DK

19***

Ch

l–sc

h

DK

27*

Lgb

DK

28**

F–L

gb

DK

29***

Ch

l+T

r–sc

h

MK

5*

F–L

gb

MK

4**

My–L

gb

MK

3***

Ch

l–sc

h

DK

30*

F–L

gb

DK

31**

My–L

gb

DK

32***

Ch

l–sc

h

DK

25*

F–P

rd

DK

26***

Tr–

sch

DK

13*

Prd

MK

2***

Tr–

sch

DK

3*

F–P

rd

DK

4**

My–P

rd

DK

5***

Ch

l–sc

h

DK

10*

F–P

rd

DK

11**

My–P

rd

DK

12***

Ch

l–sc

h

DK

14*

F–P

rd

DK

15**

My–P

rd

DK

16***

Tr+

Ch

l–sc

h

DK

7*

Prd

DK

8**

My–P

rd

DK

9***

Ch

l–sc

h

SiO

2(%

)48.0

847.6

425.1

948.1

946.4

725.2

146.5

546.1

429.9

644.4

233.1

825.5

644.9

833.8

425.2

644.1

750.3

840.0

752.6

140.1

138.5

327.9

640.5

539.6

425.8

640.7

841.5

39.1

242.7

235.3

228.1

4

Al 2

O3

17.4

216.6

220.0

916.9

517.6

219.8

419.8

616.8

418.7

418.0

921.7

120.0

621.3

922.8

720.6

97.9

54.1

37.2

82.9

97.5

411.2

17.8

88.4

911.2

520.5

37.6

69.3

311.4

514.3

216.3

818.1

3

Fe 2

O3

8.5

9.7

230.5

79.9

413.3

229.7

17.1

37.4

928.7

67.8

121.8

428.3

17.5

918.5

631.3

88.5

58.2

69.8

87.8

310.1

115.6

22.2

48.8

812.8

326.2

19.5

9.2

514.2

87.8

222.3

26.3

1

Mn

O0.1

60.1

70.3

30.2

20.2

20.3

20.1

50.1

50.2

80.1

30.2

70.3

20.1

10.2

40.3

80.1

10.1

80.1

70.2

0.1

50.2

20.2

50.1

30.1

90.3

50.1

50.1

40.2

10.1

40.1

70.1

7

MgO

9.2

79.2

412.9

49.2

98.4

313.2

89.9

613.3

612.9

13.3

512.1

15.2

98.3

38.7

613.4

526.9

223

28.2

722.5

627.4

121.3

420.3

26.7

522.8

916.8

728.3

525.7

121.5

318.1

313.1

316.3

6

CaO

12.1

510.9

80.8

410.2

57.1

20.9

511.2

68.5

50.5

7.9

70.7

50.2

311.2

14.5

40.3

3.7

28.8

45.2

210.2

24.9

34.9

30.5

55.5

45.3

30.3

54.2

55.6

65.5

99.0

84.0

10.7

5

Na

2O

1.6

1.7

40

1.5

41.1

80

1.4

21

00.5

71.0

20

1.3

40.6

70

00

00

00

00

00

00

0.1

20.2

90.7

70

K2O

0.1

60.2

70.1

0.6

30.6

40.0

30.2

91.2

60.8

21.0

10.7

70.0

20.8

21.4

70.0

20.0

20.0

20.0

60.0

10.0

40.0

10

0.0

30.0

20.1

50.0

40.0

20.0

20.6

0.2

10.0

5

TiO

20.9

1.0

81.7

40.9

61.0

81.6

10.6

60.5

10.7

60.4

40.6

40.5

20.5

50.7

10.4

90.4

90.2

40.4

0.2

30.4

30.5

60.9

80.4

20.4

61.4

60.4

90.5

10.4

70.4

70.4

90.6

7

P2O

50.0

70.0

90.1

30.0

90.0

80.1

10.0

60.0

50.0

60.0

30.0

40.0

40.0

40.0

60.0

40.0

40.0

20.0

30.0

10.0

40.0

70.0

80.0

40.0

40.0

90.0

40.0

40.0

40.0

40.0

40.0

6

LO

I2.0

12.7

18.4

22.3

33.5

88.5

52.7

4.4

78.0

35.6

8.2

88.9

13.4

77.1

68.7

47.6

15.0

88.4

93.8

48.8

96.9

39.6

48.2

77.2

9.0

28.4

47.5

46.8

5.7

66.8

99.0

7

To

tal

100.3

2100.2

6100.3

5100.3

999.7

499.6

1100.0

499.8

2100.8

199.4

2100.6

99.2

699.8

398.8

8100.7

599.5

8100.1

599.8

7100.5

99.6

599.3

999.8

899.1

99.8

5100.8

999.7

99.7

99.6

399.3

799.7

199.7

1

Zr

(p.p

.m.)

57

59

53

55

58

51

47

37

32

35

31

27

37

36

23

31

23

27

35

29

34

40

29

29

53

31

32

28

35

27

29

Y21

23

923

26

21

14

15

18

10

83

12

21

67

26

77

74

78

17

78

811

74

Sr

158

143

4137

103

9154

118

298

84

3187

118

26

11

15

10

13

53

14

411

11

11

959

28

6

Rb

917

13

49

53

321

109

101

85

61

465

132

0.5

10

61

50.5

0.5

41

56

10.5

42

13

3

Co

42

43

146

37

42

138

39

52

121

52

87

130

46

64

144

72

49

97

85

100

179

200

80

77

146

97

76

86

62

90

130

V189

227

513

215

249

483

127

121

234

101

146

152

113

171

203

95

57

89

48

95

138

188

88

103

334

101

117

102

89

139

175

Ni

171

169

195

172

160

203

285

402

364

448

424

738

223

233

321

1326

821

1132

393

1005

1241

1406

1192

981

642

1386

999

1048

640

616

679

Cr

309

282

214

330

302

369

413

311

275

1110

1250

870

264

463

482

2163

1008

2002

765

1491

1911

3170

1971

1710

869

2015

1911

2083

1580

1663

2259

Ba

34

49

91

76

87

58

34

73

83

57

74

26

36

70

28

21

12

13

0.5

19

19

41

92

22

138

19

17

10

30

49

57

Ga

15

16

56

15

18

48

14

12

47

11

39

59

14

33

53

98

75

734

65

79

12

79

12

10

36

57

Cu

33

36

411

75

86

42

44

66

65

74

319

63

49

87

23

22

565

624

Zn

50

55

128

58

63

126

51

55

123

58

101

126

43

97

151

48

42

61

37

55

90

128

45

81

138

62

59

102

60

84

98

La

0.8

11.0

50.5

70.3

60.6

60.2

40.7

60.3

9

Ce

2.8

22.9

71.9

41.1

42.1

41.3

11.7

71.7

5

Pr

0.4

50.4

80.3

90.3

80.2

60.4

40.4

2

Nd

2.6

62.7

92.3

11.0

72.0

92.0

62.3

02.3

3

Sm

0.8

90.6

70.8

70.4

30.8

11.0

00.8

01.0

0

Eu

0.4

10.3

30.2

10.1

70.3

50.1

70.3

40.4

4

Gd

1.1

60.8

91.1

70.5

81.0

51.3

11.3

81.2

9

Tb

0.2

10.1

50.2

10.1

00.1

80.2

30.2

00.2

4

Dy

1.3

60.7

61.4

41.2

01.5

51.2

81.5

4

Ho

0.3

0.1

70.3

00.1

60.2

70.3

50.3

10.3

2

Er

0.8

10.5

20.8

50.4

50.7

11.0

10.8

50.8

4

Yb

0.8

50.5

50.8

10.4

70.7

51.1

90.8

10.8

9

Lu

0.1

20.1

10.1

20.0

70.1

10.1

70.1

50.1

4

Sp

ecifi

c

gra

vit

y

2.8

62.9

53.0

22.9

32.9

13.0

32.9

2.8

83.0

32.9

2.9

22.9

72.8

72.9

33.0

32.8

72.9

72.9

2.9

92.8

82.9

42.9

22.9

2.9

42.9

82.8

42.8

82.9

92.9

42.9

83.0

0

3 0 A . E S S A I F I E T A L .


leucogabbros evolve towards layered leucogabbros inthe eastern part of the intrusion, and to zircon-bearingleucogabbros with Na-rich plagioclase (An55–70%) atthe top of the intrusion (northern side). Ferrogabbros,with up to 8% of ilmenite, also occur within theintrusion (Jadid, 1989). Leucogabbros, which consti-tute the most widespread rock type in the Kettaraintrusion (Fig. 2), have ortho- to mesocumulate tex-tures with 5–20% olivine (Fo 81), 45–65% plagioclase(An61–86%) and about 1% chrome spinel as thecumulus phases; 15–30% endiopside (Mg# ¼ 0.85,Al2O3 ¼ 2%) and ilmenite as the intercumulus phases.Relict plagioclase is present as tabular (8–10 mm)phenocrysts (An61–85%) and plagioclase is also en-closed as small crystals (An70–86%) in clinopyroxene(Fig. 3d). While the latter are generally preserved fromalteration, the former are commonly transformed intoaluminous prehnite (XFe < 0.025), epidote (Ps < 4%)and weakly phengitic muscovite. The development ofthese alteration products is accompanied by a decreaseof the Ca content of the plagioclase phenocrysts. Rarerelics of olivine form elongated (8 mm) or subrounded(4 mm) grains enclosing spinel but are almost entirelyreplaced by tremolite (Mg# ¼ 0.81–0.86). In somecases Mg-hornblende develops around spinel. Clin-opyroxene has a clear appearance and forms anhedralpatches with inclusions of plagioclase, ilmenite andspinel. Its alteration product is amphibole (ranging incomposition from hornblende to tremolite). Chloriteand epidote develop at the contact between clinopyr-oxene and plagioclase. Ilmenite is altered into anataseand titanite while chrome spinel is altered into alumi-nous chromite, ferrochromite and chromiferous mag-netite. Where ilmenite grains are preserved fromsecondary alteration, their composition shows an

increase in MnO (2–7.25% MnO) and a decrease inMgO (MgO < 0.7%) relatively to compositions re-ported from unaltered tholeiitic intrusions (Haggerty,1976). Such chemical changes characterize the incipientstages of the alteration process and result from diffu-sional processes under low to medium-grade meta-morphism (Cassidy & Groves, 1988).

Mineral alteration in the outer parts of the shear zones

In the outer parts of the shear zones, the gabbroiccumulates are not foliated. These unfoliated gabbrosof the shear zone walls are however, characterizedby an intense intragranular fracturing of plagioclaseand clinopyroxene grains and by an increase in themodal percentage of secondary minerals (amphibole,chlorite, muscovite, prehnite, epidote, anatase) thatreplace the primary minerals as well as filling theintragranular fractures. In these zones olivine isentirely replaced by tremolite, the percentage ofplagioclase decreases, and clinopyroxene totally van-ishes near the transition bet ween the unfoliated andthe foliated domains (Figs 4 & 5). Chlorite andamphibole modal proportions increase and becomeprevalent near the shear zones. In contrast, themaximum development of prehnite, muscovite andclinozoisite is observed at the beginning of the shearzone walls.

Mineral alteration in the inner parts of the shear zones

Rocks within Kettara shear zones have a well-developed foliation formed by chlorite and amphibole.With increasing strain, the unfoliated metacumulatesare converted into mylonites whose protoliths are still

Table 2. Rb and Sr concentrations (ID or XRF (*)), Sr and O isotopic ratios of various rocks in the Kettara intrusion andnearby rocks (for locations, see Fig. 1c). Suffixes m and f refer to the isotopic compositions that are thought to be those ofrelated magmas and metasomatizing fluid (altered samples), respectively (see text for explanations). Abbreviations: Cpx: clino-pyroxene; Gb: gabbro; Prd: peridotite; F-Lgb: fractured leucogabbro; My-Prd: mylonitic peridotite; Chl-sch: chlorite schist; Tr-sch:tremolite schist; mg: microgranite; mtdh: microtrondhjemite; sch: schist.

Sample Location Rb (ID) Sr (ID) 87Rb ⁄ 86Sr 87Sr ⁄ 86Sr ± 87Sr ⁄ 86Sr (330) 87Sr ⁄ 86Sr (300) d18O ⁄ SMOW

Amphibole

d18O ⁄ SMOW

Chlorite

GK 2 cpx Cpx from Gb Kt. Kettara m 0.37 15.88 0.067 0.704787 14 0.7045

DK 13 Prd – m 6.56 12.98 1.46 0.711337 8 0.7045

MK 5 F-Lgb – f 83.5 94.8 2.55 0.717165 10 0.7052 0.7063

GK 3 F-Lgb – f 77.3 332.5 0.673 0.711756 8 0.7086 0.7089 5.69 ± 0.06

DK 25 My-Prd – f 1.52 6.44 0.685 0.709952 9 0.7067 0.7070

MK 4 augen Chl-scht – f 58 72.8 2.31 0.717721 7 0.7069 0.7079

MK3 Chl-sch – f 6.01 ± 0.09

DK 12 Chl-sch – f 6.65 10.87 1.77 0.720173 7 0.7119 0.7126

DK 19 Chl-sch – f 3.28 7.86 1.21 0.714948 7 0.7093 0.7098

DK 26 Tr-sch – f 0.65 10.4 0.18 0.708478 10 0.7076 0.7077 5.71 ± 0.01

KAZ mg Kt. Hamra m 103* 170* 1.75 0.71863 8 0.7104 0.7112

TBZ mg Kt. Bouzlaf m 86* 126* 1.97 0.72053 0.7113 0.7121

MGTK mtdh Kt. Kettara f 3* 348* 0.025 0.711856 9 0.7117 0.7117

MTK mtdh – f 15* 379* 0.114 0.71238 7 0.7118 0.7119

BF 3 sch Kt. Bouzlaf 7.93 400.2 0.057 0.712301 7 0.7120 0.7121

(dupl.) 8.42 402.8 0.06 0.712317 7 0.7120 0.7121

BF 8 sch – f 175.7 66.7 7.635 0.743536 7 0.7077 0.7109

BF 10 sch – f 122.4 103.7 3.42 0.728122 7 0.7121 0.7135



recognizable. The peridotites become spotted schists(Fig. 3e); the spots are elliptical and correspond totremolite-bearing ghosts after olivine in a chlorite-richmatrix. The leucogabbros become augen schists(Fig. 3f); the eyes corresponding to scarce plagioclaseclasts that are albitized, sericitized or prehnitized.Plagioclase deforms as a result of pressure solution andpervasive microfracturing; the plagioclase clasts arereoriented in a ductile matrix that is composed mainlyof chlorite (Mg# ¼ 0.3–0.4) and tremolite (Mg# ¼ 0.6–0.9) whose total abundances reach 90% of the rockvolume (Fig. 5).

The highly strained zones are ultramylonites whoseprotoliths cannot be identified. Whatever the compo-sition of the adjacent undeformed rock, the ultramyl-onites are mainly composed of chlorite and amphibole(Fig. 4). The relative abundances of these latter min-erals vary largely along the shear zones, from � 0% to� 100%. The widest (m-scale) and the most abundantshear zones are chlorite-rich, while the amphibole-richultramylonites are present as relict lenses within thechlorite-rich shear zones or as scarce narrow cm-scaleshear zones. From the primary minerals present in theundeformed rocks adjacent to shear zones, only spinel,ilmenite and apatite (Figs 4 & 5) persist in the ultra-mylonites where they show pressure shadows and rollingstructures. Chlorite and tremolite from the ultramyl-onites have the same XMg as in the mylonites.

Minerals in veins

Different types of veins are present in the Kettaraintrusion where they occur in both ultramafic andmafic cumulates. Calcite and dolomite microscopicveins crosscut microfractures filled with chlorite,amphibole and sericite. Mesoscopic veins are (a) up to8 cm wide shear fractures filled with prehnite, pump-ellyite, muscovite and potassic feldspar (b) up to 30 cmwide quartz-calcite bearing �en echelon� veins, and(c) up to 10 cm wide quartz veins present occasionallyat the centre of the chlorite shear zones. �En echelon�veins trend at 45� relative to the direction of the shearzones; however, they are progressively reoriented andfolded in the vicinity of the highly deformed zones. Theabove geometric relationships indicate that formationof calcite ⁄ quartz veins was contemporaneous withshear zone development (Essaifi et al., 1995). As theshear zones widen, the veins become progressivelyinvolved in the ductile deformation. Within thequartz–calcite veins, quartz is located at the marginswhile calcite is present in the centre. The shear frac-tures are scarcer than the �en echelon veins� and theirassociation with shear zones is not clear. However theiroccurrence is restricted to the unfoliated gabbrosadjacent to shear zones and the sense and direction ofshear inferred from these fractures is in accordancewith that inferred from the shear zones, indicating thatthey were probably coeval with shear zone develop-ment. The quartz veins are stretched concurrently to

20

40

60

80

100

20

40

60

80

100

Pl

Cpx

Chl

10

20

30

10

20

30

Amp

Prh

EpMs

0

1

2

3

0

1

2

3

Cal

Spl+Ilm

An

lenticulardomains

shearzone walls ultramylonites mylonites

chlorite schistaugenchlorite schist

less fracturedleucogabbro

fracturedleucogabbros

Increasing strain

Mo

dal

pro

po

rtio

ns

(%)

Chl- rich shear zones

Fig. 5. Modal variations through chlorite-rich shear zone. Notethe maximum development of muscovite, prehnite and epidote atthe boundary between the weakly deformed to undeformedlenticular domains and the shear zones walls while the maximumof amphibole is observed at the boundary between the shear zonewalls and the inner parts of shear zones. Note also the increase ofthe proportions of the insoluble minerals toward the highlydeformed zones consequently to volume loss and mass leaching.

Lenticular Shear zone Inner parts of shear zonesdomains walls

Less fractured Fractured Mylonitized UltramylonitizedMinerals Protolith cumulates cumulates cumulates cumulatesOlSpl

prim

ary

PlCpxIlmApHbl*Act-HblAct / TrMg-Chl

seco

ndar

y

Fe-ChlMsCzoPrhAbSrpAnCal

Increasing strain

Fig. 4. Primary and secondary minerals in the Kettara shearzones and their protoliths. * part of hornblende is probablymagmatic. Among the primary minerals, only the insoluble ones(Ilm, Spl, Ap) persist in the inner parts of the shear zones. Mg-Chlin the unfoliated domains becomes Fe-Chl in the foliated rocks.

3 2 A . E S S A I F I E T A L .


shear zones and show evidence of low temperaturerecrystallization as indicated by quartz subgrain rota-tion (Fitz Gerald & Stunitz, 1993).

DEFORMATION MECHANISMS

Brittle deformation dominates in the outer part of theshear zones, as shown by fracturing of plagioclase andclinopyroxene grains, producing a progressive grainsize reduction and allowing fluids to penetrate therock, which enhances formation of hydrated minerals(Fitz Gerald & Stunitz, 1993). In the inner parts ofshear zones, evidence for pressure solution is, forexample, indicated by pressure shadows around pla-gioclase clasts in augen schists and around relativelyinsoluble minerals (ilmenite, spinel, apatite) in ultra-mylonites. Since hydrated minerals are less resistant todeformation than anhydrous minerals (Brodie & Rut-ter, 1985; White & Knipe, 1978), hydration has assistedand localized ductile deformation in the shear zones.Reaction softening assisted by advection of hydrousfluids is a mechanism often invoked for promotinglocalized deformation in granitoids at the brittle–ductile transition (e.g. Bos & Spiers, 2001 and refer-ences therein) in contrast to the closed system fluidconditions which may allow the existence of dis-tributed large strains within large volumes even at low-temperature conditions (Le Hebel et al., 2002).

No evidence of plagioclase or clinopyroxene recry-stallization has been identified in the Kettara shearzones; such a mechanism has however, been describedin gabbroic rocks deformed at high temperatures in�dry� ductile shear zones (Bonatti et al., 1975; Helms-taedt & Allen, 1976; Mevel, 1984). The absence of hightemperature recrystallization in the Kettara shearzones could be due to the lack of significant defor-mation at relatively high temperature, but it mostlikely results from a fluid-assisted deformation. In fact,when high temperature deformation of the oceaniccrust is fluid-assisted, neocrystallization of high-temperature hydrated minerals occurs instead ofrecrystallization of anhydrous minerals (Mevel, 1984).

TEMPERATURES OF ALTERATION

Alteration in the Kettara intrusion started at hightemperatures (amphibolite facies) as indicated by thecrystallization of secondary hornblende, which differsfrom the primary hornblende by its high Cl content(Mevel, 1984; Stakes & Vanko, 1986; Coogan et al.,2001) (Fig. 6a) and by its temperature of formation.Using the empirical geothermometer of Otten (1984),secondary hornblende crystallized between 850 and600 �C while primary hornblende crystallized between950 and 1300 �C. Crystallization of secondary horn-blende occurred in two sites: in the olivine grains, andat the clinopyroxene–spinel contact.

Alteration of the nonfoliated metacumulates con-tinued at progressively lower temperatures (greenschist

facies) as recorded by (i) a wide spectrum of amphibolecomposition (Fig. 6b), ranging from Al-rich andTi-rich to Al-poor and Ti-poor compositions, sug-gesting continuous growth of amphibole at decreas-ing temperatures (ii) tremolite ⁄ actinolite formationat the expense of olivine, pyroxene and hornblende(iii) transformation of plagioclase into clinozoisite,prehnite and sericite (iv) crystallization of chloriteat the expense of clinopyroxene, plagioclase andamphibole, and (v) crystallization of serpentine at theexpense of olivine in the ultramafic rocks. The associ-ation Act ⁄Tr + Chl + Ms + Prh + An indicatestemperatures around 300–350 �C (Liou et al., 1985)and chlorite compositions indicate temperatures be-tween 275 and 350 �C (Cathelineau, 1988; Xie et al.,1997). The lowest temperatures recorded in the intru-sion correspond to the low greenschist facies, i.e. to theregional metamorphism in the country rocks (Essaifiet al., 2001a). We infer that alteration in the intrusionoccurred during magma emplacement and cooling.

In contrast to the nonfoliated metacumulates, theinner parts of shear zones record essentially thelow temperature alteration. Even where rare relicsof amphibolite facies shear zones are observed,

0

0.1

0.2

0.3

0 0.1 0.2 0.3 0.4 0.5

Cl%

Ti (atomic proportions)

0

0.5

1

6.006.507.007.508.00

Tremolite

Actinolite

Ferro-Actinolite

Tr-Hbl

Ac-HblMagnesio-Hornblende

Ferro-Hornblende

Alumino-

Tschermakite

Tsch-Hbl

Fe-Tsch

-Hbl

Ferro-

TschermaliteM

g/(M

g+F

e2+

)

Si

(a)

(b)

brown

greenuncoloured

Fig. 6. Amphibole developed in the non foliated leucogabbros.Filled symbols (amphibole alteration after olivine); Opensymbols (amphibole alteration after clinopyroxene).(a) Ti vs.Cl showing the high Ti–low Cl brown hornblende (probablymagmatic) and high Cl–medium Ti hydrothermal brown horn-blende.(b) Nomenclature (after Leake, 1978) showing the widerange of the amphibole, ranging from hornblende to actinolite.



greenschist facies minerals (chlorite, tremolite, albite,prehnite, sericite) prevail in the Kettara shear zoneswhere the mineral associations are Chl + Ab + Act-Hbl + Act ⁄Tr + Ms + Prh + An in the myloniticrocks and Chl + Act-Hbl + Act ⁄Tr + An in theultramylonitic rocks. Chlorite compositions in thefoliated leucogabbros indicate temperatures between345 and 430 �C (Cathelineau, 1988; Xie et al., 1997).The low temperature alteration recorded insidethe shear zones is thus slightly higher than the lowtemperature alteration recorded in the unfoliateddomains, indicating that probably focused fluid flowhad increased the temperature of rocks in the shearzones.

WHOLE ROCK CHEMISTRY AS A RECORD OFFLUID–ROCK INTERACTIONS

In the undeformed rocks, in spite of hydrothermalalteration, element concentrations are compatible witha process of magmatic differentiation by means ofcrystal accumulation (Jadid, 1989; Essaifi, 1995). Incontrast, element concentrations in the highly defor-med inner parts of the shear zones show larger varia-tions than the undeformed counterparts (Table 1). To

test if some elements could have been immobile duringhydrothermal alteration, bulk rock compositions ofthe highly deformed rocks in the shear zones werecompared to their undeformed counterparts on anelement-by-element basis (Fig. 7). In these diagrams,the least altered and deformed metacumulates shoulddefine the magmatic trends while the highly altered andfoliated rocks should define the metasomatic trends.The magmatic and the metasomatic trends are distinctfor the majority of the elements, except for theincompatible elements P, Ti and V. For these threeelements, the magmatic and the metasomatic trendsare superposed indicating that the ratios Ti ⁄P, Ti ⁄Vand V ⁄P are the same in the foliated and the nonfoli-ated rocks. This fact can be explained by similar mob-ilities of Ti, P and V during the metasomatic process(Gresens, 1967), but this requires these elements to bedissolved into or precipitated from the metasomaticfluid without any relative fractionation, which is geo-logically very unlikely. The alternative is that Ti, P andV were immobile, a common behaviour in hydrother-mal environments (e.g. Grant, 1986; Dipple et al.,1990; Marquer & Burkhard, 1992). The large concen-tration changes in shear zones could then be easilyexplained through volume changes. These elements

0

0.04

0.08

0.12

0 0.5 1 1.5

%P2O5

0

10

20

30

0 0.5 1 1.5

%MgO

0

10

20

0 0.5 1 1.5

%Fe2O3

0

10

20

0 0.5 1 1.5

%Al2O3

0

20

40

60

80

100

0 0.5 1 1.5

Zr (ppm)

0

10

20

30

0 0.5 1 1.5

Y (ppm)

0

4

8

12

0 0.5 1 1.5

%CaO

%TiO2

20

30

40

50

0 0.5 1 1.5

%SiO2

0

100

200

300

0 0.5 1 1.5

V (ppm)

%TiO2

Fig. 7. Magmatic and metasomatic trends in the Kettara intrusion (whole rock element concentrations vs. whole rock TiO2

concentrations). The magmatic trends (curves) are drawn from data points representing the undeformed metacumulates while themetasomatic trends are defined by data points corresponding to the mylonitic and ultramylonitic rocks in the shear zones (opencircles). Only V-TiO2 and P-TiO2 magmatic trends remain unchanged regardless of the alteration and deformation states of the rocks.

3 4 A . E S S A I F I E T A L .


would be enriched in ultramylonites that have experi-enced a net volume loss and would be diluted in rocksthat have experienced a net volume gain. This is feas-ible, for example, if these elements were hosted in rel-atively insoluble minerals, the abundance of whichwould increase or decrease, respectively, in rocks thathave experienced a net volume loss or a net volumegain. In the undeformed rocks, Ti and V are mostlycontained in ilmenite and titanite while P is hosted byapatite. All these minerals are observed to persist in theultramylonites.

Volume changes

Volume changes related to the formation of shearzones can be calculated using the composition-volumerelation of Gresens (1967). According to this relationthe volume change, defined as the volume factor Fv orthe ratio of the volume of the altered rock to that ofthe original rock, is expressed by the equation Fv ¼CO

j dO ⁄CAj dA where CO

j and CAj are the concentra-

tions (in weight percent) of an immobile component�j�, dO and dA the densities with the subscripts O and

A referring to the original and to the altered rocks,respectively. Rock densities (Table 1), which weremeasured in water and determined with an uncertaintyof ± 0.01 g cm)3, are higher in the ultramylonites thanin the unfoliated metacumulates. Because the ratio oftwo immobile elements must be the same in protolithand product samples (Gresens, 1967), volume factorscalculated should be identical in principle (accordingto the previous equation) and similar in practice. Usingrock densities and assuming the elements P, Ti, and Vto be immobile, the calculated volume factors rangefrom 0.4 to 2 (Fig. 8). This indicates that the Kettarashear zones correspond to a spectrum of shear zonebehaviours, ranging from those that have lost 60% byvolume (Fv ¼ 0.4) to those that have increased involume by about 200% (Fv ¼ 2), through those thatwere volume conservative.

The volume loss shear zones are chlorite-rich ultra-mylonites while the volume-gain shear zones aretremolite-rich ultramylonites. The iso-volume shear

zones are ultramylonites where chlorite and tremoliteabundances are approximately equal.

Mass transfer estimates

Using Gresens’ approach, Grant (1986) presented asimplified graphical method to evaluate chemicalmobility during rock alteration. In this method, ele-ment concentrations in the altered rock are plotted vs.their concentrations in the original rock. The immobileelements form a straight line through the origin, andthe line is known as an �isocon�. The mobile elementsplot away from this line and their position with respectto the line reveals the direction of their mobility (lossor gain).

Four isocon plots are presented. One isocon diagramillustrates the evolution from the undeformed lenti-cular domains to the outer parts of the chlorite shearzones (Fig. 9). The three other isocon plots comparethe inner parts of shear zones to their protoliths andexemplify the volume loss; the volume gain and theconstant volume shear zones (Fig. 10).

0

0.5

1

1.5

2

2.5

3

0 0.5 1 1.5 2 2.5 3(Fv)Ti

(Fv)

P

0

0.5

1

1.5

2

2.5

3

0 0.5 1 1.5 2 2.5 3

(Fv)Ti

(Fv)

V

Fig. 8. Calculated volume factors assumingthe immobility of Ti, P and V. The Kettarashear zones correspond to a spectrum ofdilational changes in shear zones, rangingfrom those which have lost 60% by volumeto those which have increased in volume byabout two times, through those which werevolume conservative.

V/20100P2O5

15TiO2

Fe2O3

Al2O3

1/2SiO2

100MnO

2Na2O

CaO

100K2O

LOI

Sr/6Y

Rb

Co/3 Cr/20

Ba/2

Ni/25

1.2Ga

Cu/3

Zn/2

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Undeformed leucogabbro (DK20)

Fra

ctu

red

leu

cog

abb

ro (

DK

21)

Fig. 9. Isocon plot of fractured leucogabbro (shear zone wall)element concentrations vs. concentrations in protolith. Majorelements are plotted in weight percent; trace elements in p.p.m.Data are scaled as indicated. Error bars represent ± maxuncertainty of analyses. The P–Ti–V isocon is shown.



Protolith-shear zone walls evolution

A fractured leucogabbro in the outer part of a chlorite-rich shear zone is compared to a leucogabbro takensome metres away from the foliated rocks in the shearzone (Fig. 9). Most of the elements plot near the P–Ti–V isocon, indicating that element exchanges within theshear zone walls are minor comparatively to the innerparts of shear zones (see below). However K2O, Rband Ba plot above the isocon, and thus are interpretedas gained. SiO2, CaO, Al2O3, and Sr plot below theisocon and are interpreted as lost. The mass balanceequation for major elements in the shear zone wall is:

100 g leucogabbro þ 0:06 g K2O þ 0:20 g LOI

¼ 81:46 g fractured leucogabbro þ 9:22 g SiO2

þ 3:86 g Al2O3 þ 0:58 g Fe2O3 þ 1:73 g MgO

þ 0:02 g MnO þ 0:02 g TiO2 þ 3:19 g CaO

þ 0:18 g Na2O

The enrichment in K2O and Rb is consistent withsericitization at the vicinity of the shear zones. Thissericitization is accompanied by losses in CaO andSiO2, consistent with formation of quartz-calcite veinsaround the shear zones.

Volume loss shear zone

In the volume loss shear zone, an augen chlorite-richmylonite is compared to a fractured leucogabbro(Fig. 10a). Most of the elements plot above the con-stant-mass isocon, confirming that rock alteration wasnot mass conservative. With respect to the isocondefined by the immobile elements P, Ti and V, most ofthe elements plot below this line and thus interpretedas lost. They are principally Si, Ca, Sr, K, Rb and Mg.A few elements plot above the P–Ti–V isocon, especi-ally Fe, Mn, Zn and Ga and thus were gained. Themass-balance equation for the major elements in theshear zone is:

100 g leucogabbro þ 7:66 g Fe2O3 þ 0:06 g MnO

þ 0:15 g Na2O þ 0:01 g TiO2 þ 0:25 g LOI

¼ 71:49 g augen chlorite-schist þ 21:10 g SiO2

þ 2:77 g Al2O3 þ 7:48 g CaO þ 4:83 g MgO

þ 0:47 g K2O

The losses in Si, Ca, Sr, K, and Rb are consistent withdissolution of feldspar while the loss in Mg and theenrichment in Fe (+ Zn and Ga) is consistent withthe enrichment of chlorite in Fe from the unde-formed lenticular domains to the inner parts of shearzones. The formation of augen chlorite-schists is alsoaccompanied by a weak gain in Na that reflects albi-tization of feldspar clasts. However this gain occursonly in augen chlorite-schists; in the ultramylonites,Na is lost, reflecting the total dissolution of feldspar.

Ga/2Al2O3

1/2SiO2

Fe2O3

MnO

MgO

CaO

10Na2O

10K2O

30TiO2

100P2O5

LOI

Zr/2

2Y

V/5

Ba/3

Rb/3Sr/4

Cr/80Ni/30

Co/4

Cu

Zn/4

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Fractured leucogabbro (MK5)

Aug

en c

hlor

ite s

chis

t (M

K4)

P-Ti-V isocon

a)

1/2SiO2

Fe2O3

Al203

100MnO

MgO

CaO

100K2O

30TiO2

100P2O5

Zr/2

2Rb

2Y1.5Sr

Co/4

Cr/80

Ni/50

V/5Ga

2Ba

Cu

Zn/4

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Undeformed wherlite (DK13)

Trem

olite

sch

ist (

MK

2)

P-Ti-V isocon

b)

1/2SiO2

Fe2O3

Al2O3

100MnO MgO

CaO

10Na2O100K2O

30TiO2

100P2O5

LOI

Zr/3

2Y 2Sr

Rb

Ni/100

Co/3

Cr/80

V/5

Ga

Cu

Ba

Zn/4

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Fractured troctolite (DK14)

chlo

rite

+ tr

emol

ite s

chis

t (D

K16

) c)

Fig. 10. Isocon plots of mylonite to ultramylonite elementconcentrations vs. concentrations in least deformed andaltered metacumulates. Major element data are plotted inweight percent; trace elements in p.p.m. Data are scaledas indicated. Error bars represent ± max uncertainty ofanalyses. The P-Ti-V isocon is shown. (a) volume-lossshear zone; (b) volume-gain shear zone; (c) constant-volumeshear zone.

3 6 A . E S S A I F I E T A L .


Volume gain shear zone

In the volume-gain shear zone, an amphibole-richultramylonite is compared to an undeformed ultra-mafic metacumulate (Fig. 10b). Most of the elementsplot below the constant mass isocon, confirming hereagain, that the deformation ⁄ alteration process was notmass conservative. They are located above the P–Ti–Visocon, indicating that most of the elements wereenriched in the ultramylonite. In addition to gains inFe, Mn, Ga, and Zn as in the chlorite rich-mylonite,the tremolite-schist is also enriched in Si, Ca, Mg, Srand Zr. A few elements (Ba, Rb & K2O) were lost, andNa, which is completely leached from the ultramyl-onite as well as from its undeformed equivalent, wasalso lost. The mass-balance equation for the majorelements in the volume-gain shear zone is:

100 g ultramafic cumulate þ 75:38 g SiO2

þ 17:21 g CaO þ 21:22 g MgO þ 7:3 g Fe2O3

þ 0:27 g MnO þ 0:1 g TiO2

¼ 220:64 g tremolite-schist þ 0:73 g Al2O3

þ 0:04 g K2O þ 0:01 g P2O5 þ 0:07 g LOI

The enrichment in Si, Ca and Sr in the volume-gainshear zones reflects the crystallization of tremolite inspite of chlorite while the loss in Rb, Ba, Na and K2Oreflects dissolution of feldspar grains.

Constant volume shear zone

In the iso-volume type shear zone, a tremolite andchlorite-bearing ultramylonite is compared to an iso-tropic ultramafic rock (Fig. 10c). In contrast to the twoother shear zone types, most of the elements plot nearthe constant mass and the P–Ti–V isocons, indicatingthat here the alteration was probably mass conserva-tive. The mass-balance equation for the major elementsin this shear zone is:

100 g ultramafic cumulate þ 3:91 g Al2O3

þ 4:93 g Fe2O3 þ 0:06 g MnO þ 1:4 g CaO

þ 0:12 g Na2O ¼ 101:04 g chlorite-tremolite

bearing ultramylonite

þ 1:25 g SiO2 þ 6:60 g MgO þ 0:02 g K2O

þ 1:57 g LOI

Losses and gains are not as high as in the volume lossor volume-gain shear zones. The most significant lossesare in Rb and Cu, followed by K and Ba, while themost important gains are in Ga and Zn, Fe and Al,then Mn and Ca.

The comparison of the mass transfer estimates bet-ween the three types of shear zones indicates that,independently from the volume variation, Fe, Mn, Gaand Zn are enriched while Na, K, Ba and Rb are lost inthe ultramylonites shown in Fig. 10. Si, Mg, Ca and Sr

are lost in the volume loss shear zone (Fig. 10a) butgained in the volume-gain shear zone (Fig. 10b). Alwas gained in the constant volume shear zone butlost in the volume loss as well as in the volume gainshear zones.

Mobility of the rare-earth elements

In the Kettara intrusion, in addition to verifying themobility of Rare Earth Elements (REE), chondrite-normalized patterns in the three shear zone types allowus to verify the homogeneity between the protolithsand their mylonitized equivalents (Fig. 11). In theconstant volume shear zone type (Fig. 11a), the REEpattern of the ultramylonite mimics that of the prot-olith confirming not only that the volume change is notsignificant but also that the protolith and the myloni-tized equivalent are homogeneous. This superpositionof REE patterns can also be interpreted as an indica-tion of the immobility of the rare earth elements. Thisinterpretation is however, misleading. If the REE wereimmobile, their concentration would increase in thevolume loss shear zones and decrease in the volume-gain shear zones. This is not the case; the behaviour ofthe Light REE (LREE) is opposite to that of theHeavy REE (HREE) (Fig. 11b,c). The LREE increase

(a) Iso-volume shear zone

1

10

100

(b) Volume-loss shear zone

1

10

100

Ana

lysi

s/C

hond

rite

(c) Volume-gain shear zone

0.1

1

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

ultramyloniteprotolith

Fig. 11. Chondrite normalized (Anders & Grevesse, 1989) rareearth elements (REE) patterns in iso-volume (a), volume-loss (b)and volume-gain (c) shear zones. REE are slightly mobile inthe Kettara shear zones and their behaviour cannot be relatedto volume changes.



in the volume loss shear zones and decrease in thevolume-gain shear zones while the HREE behave dif-ferently; they decrease in the volume loss shear zonesand increase in the volume-gain shear zones. Hence theREE are slightly mobile in the Kettara shear zones andtheir behaviour is related to the minerals that crystal-lized in the shear zones. The REE patterns in thevolume loss and volume-gain shear zones are,respectively, compatible with the REE patterns ofchlorite and amphibole. Chlorite concentrates LREE(Cullers et al., 1974), while amphibole concentratesHREE (Arth, 1976; Philpotts & Schnetzler, 1970). Infact, the composition of the ultramylonitic rocks variesbetween that of chlorite and that of tremolite (Fig. 12).These results indicate a complete redistribution of theREE from the magmatic minerals to the secondaryminerals during the alteration process.

Sr and O ISOTOPE GEOCHEMISTRY

The Sr whole-rock isotopic compositions of variablydeformed and altered rocks are given in Table 2together with those of a few nearby coeval acidicintrusions and those of a few representative schistssampled around these intrusions. Isotopic composi-tions are calculated for (i) 330 Ma, which is the ageobtained (U–Pb on zircon) by Essaifi et al. (2003) forthe emplacement of synkinematic acidic intrusions, anage also ascribed to the emplacement of the Kettarabody and major hydrothermal alteration because bothare also syntectonic; (ii) also at 300 Ma for the rocksaffected by hydrothermal alteration since an importantthermal event is detected in the Hajjar sulphide deposit

within the Marrakech area (Ar–Ar dating; Watanabe,2002). Initiation of hydrothermal alteration wasnecessarily synchronous with, or quickly followed themagmatic event because secondary minerals recordedboth high-T and low-T conditions corresponding tothe thermal re-equilibration of the intrusion with thecountry rock. The intrusion was of small size, empla-ced at high structural levels and therefore experiencedrapid cooling, but hydrothermal alteration continuedalong the shear zones where the recorded temperaturesare slightly higher than in the undeformed counter-parts. The present-day 87Sr ⁄ 86Sr ratio is positivelycorrelated with the 87Rb ⁄ 86Sr ratio, but the data do notyield an isochron relationship, which was expectedsince the shear zones were not closed systems withrespect to Rb and Sr (see above). Initial isotopic ratiosare 0.7045 in the peridotite as well as in the clino-pyroxene separate from a gabbro sampled in the centreof a lenticular domain (�30 m from the nearest shearzone), and this indicates the Sr isotopic signature of themantle derived magmatic suite. Chlorite- or tremolite-rich samples from shear zones have isotopic ratios inthe 0.707–0.7126 range (330–300 Ma), which indicatesthe rather radiogenic signature of the metasomaticfluid. This is consistent with the fact that the nearbyhighly metasomatized microgranites, now with tron-dhjemitic compositions (Essaifi, 1995; Essaifi et al.,2004) share the same Sr isotopic ratios, likely repre-senting those of the metasomatizing fluid (samplesMTK and MGTK, Table 2).

Oxygen isotopic compositions of amphibole fromthe fractured leucogabbro, from the tremolite-richultramylonite and of chlorite from the chlorite-richultramylonite are very close (d18O ¼ 5.7–6.0&) andsimilar to those of MORB-type material. Nevertheless,the fact that the Sr isotopic compositions of thehydrated rocks ⁄minerals are strongly modified byhydrothermal alteration demonstrates that these iso-topic signatures cannot be pristine but are controlledby the fluid. O isotopic similarity for chlorite andtremolite is predicted by the clinochlore-H2O andtremolite-H2O fractionation equations of Zheng (1993)in upper greenschist to lower amphibolite facies con-ditions. These equations yield a fluid d18O value of7–7.5& in the range 400–500 �C which, associated toSr isotopic data, indicate that the hydration fluid in theKettara intrusion was equilibrated with the hostlithologies under rather low fluid ⁄ rock ratios (rock-dominated fluid systems, greenschist facies conditions)prior to interaction with the intrusion. Indeed, the Srisotopic signatures of the fluids indicate that theirrecharge system could have been the host schists seriesor ⁄ and the coeval granitic intrusions since they sharecomparable Sr isotopic ratios (0.712–0.7135) at thetime of the hydrothermal activity. Hydration in Kett-ara was thus produced by advection of metamorphicfluids sensu lato, i.e. dewatering metamorphic fluids,magmatic-derived fluids, or surface-derived fluidsincluding formation waters which were re-equilibrated

Fe

Si

Mg-Chl

Fe-Chl

Tr

Al

protolithsshear zones[ inner parts

outer parts

Fig. 12. Projection of the Kettara shear zones in Si, Fe andAl ternary diagram (cation proportions). The whole rockcomposition of the Kettara shear zones evolves toward thechlorite-tremolite join.

3 8 A . E S S A I F I E T A L .


with host lithologies. In the discharge system (shearzones) fluid focusing induced high fluid ⁄ rock ratios asenlightened by the large chemical changes observed(see above). Variations of the Sr isotopic ratio amongthe group of chlorite-rich and tremolite- rich mylonitesmay be explained by variations of the fluid and rock Srcontents and of the fluid ⁄ rock ratio during interaction(e.g. Lecuyer et al., 1990) and ⁄ or by the degree of Srisotopic equilibration between the fluid circulating in agiven shear zone and its recharge system (host schists).

DISCUSSION

Mass balance calculations for the shear zones indicatecomplex patterns of chemical changes which, togetherwith d18O of minerals, can be used to infer masstransfer mechanisms, fluid sources, flow directions andfluid fluxes.

Amphibole-rich vs. chlorite-rich shear zones

Since the amphibole-rich ultramylonites are presentas narrow (cm-scale) shear zones or as preservedcentimetre-scale lenses of amphibole schists withinthe broad (m-scale) chlorite-rich shear zones, weinterpret the amphibole-rich ultramylonites as thefirst nucleated shear zones in the Kettara intrusion.As the shear zones evolved, the increased permeab-ility allowed more fluids to infiltrate the shear zonesand those that continued to undergo fluid infiltrationwere retrogressed into chlorite-rich ultramylonites.Retrogression of amphibole-rich to chlorite-rich shearzones was progressive as indicated by the wide rangeof the calculated volume changes, and by the com-position of the ultramylonitic rocks, which variesbetween that of chlorite and that of tremolite(Fig. 12). The Kettara shear zones were thus thick-ening with time.

Fluid flow directions

One of the striking features of the Kettara shearzones is that the mineralogical compositions of thehighly deformed rocks in the inner parts of shearzones do not depend on the mineralogical composi-tion of the protoliths. Hence, we infer that a largevolume of fluids circulated through these zones andthat the composition of the ultramylonites was fluid-controlled. Formation of amphibole-rich shear zoneswas accompanied by gains in Si, Ca, Mg and lossesin K and Na while formation of chlorite-rich shearzones was accompanied by losses in Si, Ca, Mg andgains of K in the shear zone walls. Thereforeamphibole-rich and chlorite-rich shear zones couldcorrespond to interaction of gabbros with two �dis-tinct� fluids. The first fluid, charged in Ca and Mg,transformed gabbros into amphibolite. The secondfluid, depleted in Ca and Mg, infiltrated the sameshear zones and transformed gabbros and former

amphibolite into a chlorite-rich rock. Taking intoaccount the above constraints, the variation in fluidcomposition cannot be related to lithological differ-ences, but rather to movements of fluids along gra-dients in pressure and temperature. Fluid flow alongpositive temperature gradients would favour fixationof Ca and ⁄ or Mg (Hemley et al., 1971; Giggenbach,1984; Rose & Bird, 1994). In contrast, losses in Caand gains in K and Na appear to be typical formetasomatized quartzo-feldspathic ductile fault zonesand are explained as the result of fluid flow in thedirection of decreasing temperature (Dipple & Ferry,1992; Streit & Cox, 1998).

Accordingly, chemistries of the Kettara first-stageand second-stage fluids are consistent with move-ments along increasing and decreasing temperaturegradients, respectively (Dipple & Ferry, 1992; Streit& Cox, 1998). Transformation of various lithologiesinto chlorite actually appears to be typical of upflowzones in ophiolites and oceanic crust (e.g. Kranidiotis& MacLean, 1987; Richards et al., 1989). Inthe Kettara intrusion, the upflow zones are accom-panied, as in quartzo feldspathic lithologies (Ferry &Dipple, 1991; Rumble, 1994), by formation of quartzveins.

We propose that the first metasomatizing fluids(leading to amphibole mylonites) originated as porefluids in the surrounding greenschist facies metapelitesand in the contact metamorphic aureole. These porefluids flowed pervasively through the intrusion, indu-cing diffuse distribution of hydrated minerals and werelocally entrained into the newly created shear zones.Likely, the whole fluid pattern evolved towards athermally driven convection system centred on theintrusion as the shear zones network developed(Fig. 13). When the intrusion had cooled to tempera-tures of the country rock, the intrusion-driven hydro-thermal system ceased. Nevertheless, fluid circulationwas maintained through the intrusion in the dilatantshear zones where the still active shearing maintained ahigh permeability. O ⁄Sr isotope data suggest that therecharge system of the fluid was more or less the same(i.e. host schists lithologies). Thus, this second type offluid could possibly originate from dewatering of thesedimentary pile at deeper levels of Central Jebilet.

Mass transfer mechanisms

The Kettara shear zones are interconnected fracturesallowing channelled fluid-flow at the regional-scale,but evidence for pervasive fluid input in the Kettaraintrusion is given by the widespread occurrence ofamphibole, Mg-chlorite and Ca–Al silicates in theunfoliated gabbros, minerals that are related tointeraction with solutions subject to increasing tem-peratures (Giggenbach, 1984; Rose & Bird, 1994).Fluid access to individual grains probably occurredduring the initial alteration event via grain bound-aries and networks of grain-scale fractures. In fact,



all the gabbroic rocks of the lenticular unfoliateddomains of the Kettara intrusion are fractured tosome degree.

In the second stage of alteration (thermal equili-brium between intrusion and country rock), focusingof fluid flow was maintained in the inner part of theshear zones because deformation, which can raisefluid pressure, was likely initiating hydraulic fractur-ing (Sibson et al., 1975; Vrolijk, 1987). On the otherhand, mass balance calculations indicate that theshear zone walls were enriched in K relatively to theless fractured gabbros. We relate this K enrichmentaround the shear zones (Essaifi et al., 1995) to eithera redistribution of the K leached in the inner parts ofthe shear zones or to the fact that the infiltratingfluid was originally K-bearing, or both. In any case,extensive formation of K–Al–OH silicates wasexpected to be typical in ⁄ or around major upflowzones (Giggenbach, 1984), allowing the ductility ofsilicate rocks to increase (Dipple & Ferry, 1992).Now, the formation of the quartz and ⁄ or calcite �enechelon veins� which are also found in the walls of

the chlorite shear zones appears to be complementaryto potassic alteration. Indeed, the reactivity of CO2

with respect to hydrogen metasomatism increaseswith decreasing temperature, and it is expected toincrease with the horizontal distance from majorupflow zones, leading to formation of K–Al-enrichedminerals, calcite and quartz (Giggenbach, 1984). Thisinterpretation is supported by the fact that the tem-peratures calculated from chlorite compositions arehigher than in the sericite-bearing undeformedcounterparts, which is consistent with heat advectionrelated to large fluid fluxes within the shear zones(Brady, 1988).

Compared to what is known in the other felsic andmafic pods of Central Jebilet (Essaifi, 1995; Essaifiet al., 2004), the protracted character of fluid flowin the Kettara shear zones seems remarkable. Thispeculiarity is related to the specific structural positionof the Kettara intrusion (Fig. 1c), immediately at thesouth of a horse tail termination of the Mesret Faultmarked by a subdivision into a number of 130–140 N-trending synthetic dextral faults.

Fig. 13. Model for fluid assisted deformation in the Kettara intrusion. Volume-gain, amphibole-rich shear zones nucleated while theintrusion was still cooling to temperatures of the country rock (lower greenschist facies). They progressively evolved into volume-loss,chlorite-rich shear zones that continued to undergo fluid infiltration after the rapid cooling of the intrusion. Quartz-calcite veinsdeveloped in the shear zone walls while deformation was evolving from ductile to brittle-ductile conditions. Formation of amphiboleshear zones was accompanied by gains of Si, Ca, and Mg (white arrows) and losses of K and Na (black arrows) during anup-temperature pervasive fluid flow at the scale of the intrusion. Formation of chlorite shear zones occurred during an upwardfocused fluid flow and was accompanied by extensive leaching of Si, Ca and Mg (black arrows) in the inner parts of the shearzones while K was deposited in the shear zone walls (little white arrows). Fe was gained while Na was leached from the shearzones during the entire hydrothermal system activity.

4 0 A . E S S A I F I E T A L .


Comparison with granitoid shear zones

Typical major element mobilities associated with feld-spar breakdown in granitoid shear zones are losses ofNa, Si, K and Ca in volume loss shear zones while thegain in Si is typical in constant volume and volumegain shear zones (Bailey et al., 1994; Condie & Sinha,1996). Comparison of our results on gabbroic rocks towhat is known about the behaviour of elements duringmylonitization of granitoid rocks indicate that exceptfor Ti, P, and to a lower extent Al, which can beconsidered as immobile elements, the behaviour ofelements in the Kettara hydrothermal system is con-trolled by (i) the chemical composition of the infiltra-ting fluids: Fe was gained in all the shear zone typesindicating interaction of gabbros with evolved, Fe-richfluids (ii) the direction of fluid flow: Ca and Mg werefixed in up-temperature flow zones and leached indown-temperature upflow zones (iii) the CO2 reactivitywith respect to hydrogen metasomatism: K wasdeposited in the shear zone walls (iv) volume changes:Si, the principal component of rocks, increase in vol-ume-gain and decrease in volume loss shear zones, and(v) fluid fluxes along the shear zones: Na was leachedin all the shear zones, probably as a result of the highfluid fluxes.

CONCLUSIONS

The layered mafic–ultramafic Kettara intrusion recor-ded both regional fluid flow and �local� fluid flowrelated to the thermal effects of the intrusion.

(a) The thermally driven local fluid flow, initiatedduring cooling of the intrusion, was contemporaneouswith a structurally driven fluid flow that operates alongdilatant, cm-scale shear zones nucleated in the intru-sion. Fluid infiltration, pervasive throughout the wholeintrusion, was focused along the shear zones wheregabbros were transformed into amphibole-rich ultra-mylonites. Fluid infiltration to the intrusion occurredvia grain boundaries and networks of grain-scalefractures, the fluid being entrained into the shear zoneswhere gains in Ca and Mg are related to up tempera-ture fluid flow. This kilometre-scale pervasive fluidflow induced by the thermal gradient between theintrusion and the host rocks was favoured by (i) thepermeability enhancement accompanying reaction anddilatational deformation (ii) the high fluid fluxes alongthe shear zones, and (iii) the synmetamorphic andrapid cooling of the intrusion emplaced at 330 Mainto still reactive, wet, monotonous upper Visean(350–333 Ma) related metapelites.

(b) The regional fluid flow was recorded becausethe thermal re-equilibration of the intrusion to tem-peratures of the country rock occurred while defor-mation was still active and the shear zones were stillevolving into metre-scale shear zones. As the shearzones thickened, the increased permeability allowedmore fluids to infiltrate the fault zones and those that

continued to undergo fluid infiltration were retro-gressed into chlorite-rich shear zones. The alterationpatterns indicate that this focused fluid flow waspervasive throughout the interconnected shear zonesand implies fluid infiltration at the regional scale. Incontrast to the amphibole-rich ultramylonites, thechlorite-rich fault zones are volume loss shear zoneswhere losses in Ca, Si, Mg and Na are related to fluidtransfer in retrograde temperature conditions. Largefluid fluxes along the shear zones were probablyaccompanied by a rise in temperature that led theCO2 reactivity to increase away from the majorupflow zones, and hence to the deposition of Karound the shear zones.

Despite the complex alteration patterns, the d18Ovalues of amphibole and chlorite are similar, indicatinggabbro interaction with fluids persistently equilibratedwith the metamorphic pile. As a whole, the complexKettara hydrothermal system illustrates the competi-tion between the structurally driven and thermallydriven fluid flow, and indicates that synmetamorphicintrusions can be used as markers of fluid flow duringregional metamorphism and orogenesis.

ACKNOWLEDGEMENTS

We thank D. Gapais from �Geosciences Rennes�for helpful discussions and L. Coogan from CardiffUniversity for comments on an earlier draft of thismanuscript. Thanks are due to F. Martineau andG. Gruau (Rennes University) for providing O andSr isotope analyses. The constructive reviews ofJ. E. Streit and an anonymous referee were greatlyappreciated, as well as editorial comments byD. Robinson. This work was partially financed by thegrant PICS n� 45.

REFERENCES

Aarab, E. M., 1984. Mise En Evidence Du CaractereCo-Genetique Des Roches Magmatiques Basiques et AcidesDans la Serie Volcano-Sedimentaire de Sarhlef (Jebilet, MarocHercynien). 3rd cycle Thesis, Nancy University, Nancy.

Anders, E. & Grevesse, N., 1989. Abundances of the elements:Meteoric and solar. Geochimica et Cosmochimica Acta, 53,197–214.

Arth, J. G., 1976. Behavior of trace elements during magmaticprocesses: a summary of theoretical models and their appli-cations. Journal of Research of the U.S. Geological Survey, 4,41–47.

Bailey, C. M., Simpson, C. & De Paor, D. G., 1994. Volume lossand tectonic flattening strain in granitic mylonites from theBlue Ridge province, central Appalachians. Journal of Struc-tural Geology, 16, 1403–1416.

Beach, A. & Fyfe, G., 1972. Fluid transport and shear zones atScourie Sutherland: evidence of overthrusting? Contributionsto Mineralogy and Petrology, 36, 175–180.

Bernard, A. J., Maier, O. W. & Mellal, A., 1988. Apercu sur lesamas sulfures des hercynides marocaines, Mineralium Depos-ita, 23, 104–114.

Berthe, D., Choukroune, P. & Jegouzo, P., 1979. Orthogneiss,mylonite, and non-coaxial deformation of granite: the



example of South Armorican shear zone. Journal of StructuralGeology, 1, 31–42.

Bonatti, E., Honorez, J., Kirst, P. & Radicati, F., 1975. Meta-gabbros from the mid-Atlantic ridge at 06�N: contact-hydro-thermal-dynamic metamorphism beneath the axial valley.Journal of Geology, 83, 61–78.

Bordonaro, M., 1983. Tectonique et petrographie du districta pyrrhotine de Kettara (Paleozoique des Jebilet, Maroc).3rd cycle Thesis, Strasbourg University, Strasbourg.

Bos, B. & Spiers, C. J., 2001. Experimental investigation into themicrostructural and mechanical evolution of phyllosilicates-bearing fault rocks under conditions favouring pressuresolution. Journal of Structural Geology, 23, 1187–1202.

Bouloton, J. & Le Corre, C., 1985. Le probleme de la tectoniquetangentielle dans les Jebilet (Maroc hercynien): Donnees ethypotheses. Hercynica, 2, 121–129.

Brady, J. B., 1988. The role of volatiles in the thermal history ofmetamorphic terranes. Journal of Petrology, 29, 1187–1213.

Brodie, K. H. & Rutter, E. H., 1985. On the relationship betweendeformation and metamorphism, with special reference to thebehavior of basic rocks. In: Metamorphic Reactions: Kinetics,Textures, and Deformation (eds Thompson, A. B. & Rubie,D. C.), pp. 138–179. Springer-verlag, New York.

Brun, J. P. & Cobbold, P. R., 1980. Strain heating in continentalshear zones: a review. Journal of Structural Geology, 2, 149–158.

Cassidy, K. F. & Groves, D. I., 1988. Manganoan ilmeniteformed during regional metamorphism of Archean mafic andultramafic rocks from Western Australia. Canadian Miner-alogist, 26, 999–1012.

Cathelineau, M., 1988. Cation site occupancy in chlorites andillites as a function of temperature. Clay Minerals, 23, 471–485.

Clayton, R. N. & Mayeda, R. K., 1963. The use of brominepentafluoride in the extraction of oxygen from oxides and si-licates for isotopic analysis. Geochemica et CosmochimicaActa, 27, 43–52.

Condie, K. C. & Sinha, K., 1996. Rare earth and other traceelement mobility during mylonitization: a comparison of theBrevard and Hope Valley shear zones in the AppalachianMountains, USA. Journal of Metamorphic Geology, 14, 213–226.

Coogan, L. A., Wilson, R. N., Gillis, K. M. & Macleod, Ch,2001. Near-solidus evolution of oceanic gabbros: Insightsfrom amphibole geochemistry. Geochimica et CosmochimicaActa, 256, 4339–4357.

Cullers, R. L., Yeh, L.-T., Chaudhuri, S. & Guidotti, C. V.,1974. Rare earth elements in Silurian pelitic schists from N.WMaine. Geochimica et Cosmochimica Acta, 38, 389–400.

Dipple, G. M. & Ferry, J. M., 1992. Metasomatism and fluidflow in dutile fault zones. Contributions to Mineralogy andPetrology, 112, 149–164.

Dipple, G. M., Wintsch, R. P. & Andrews, M. S., 1990. Identi-fication of the scales of differential element mobility in aductile fault zone. Journal of Metamorphic Geology, 8, 645–661.

Essaifi, A., 1995. Relations Entre Magmatisme, Deformationet al. Teration Hydrothermale: L’exemple Des Jebilet Cent-rales (Hercynien, Maroc). Memoires Geosciences Rennes, 66.(in French).

Essaifi, A., Ballevre, M., Marignac, Ch & Capdevila, R., 2001a.Decouverte et signification d’une paragenese a ilmenite zinc-ifere dans les metapelites des Jebilet centrales (Maroc).Comptes Rendus de l’Academie Des Sciences, Paris, 333,381–388.

Essaifi, A., Capdevila, R. & Lagarde, J. L., 1995. Transforma-tion de leucogabbros en chloritoschistes sous l’effet de l’alte-ration hydrothermale et de la deformation dans l’intrusion deKettara (Jebilet, Maroc). Comptes Rendus de l’Academie DesSciences, Paris, 320, 189–196.

Essaifi, A., Capdevila, R. & Lagarde, J. L., 2004. MetasomaticTrondhjemites and Tonalites: Examples in Central Jebilet

(Hercynian, Morocco). Journal of African Earth Sciences,submitted.

Essaifi, A., Lagarde, J. L. & Capdevila, R., 2001b. Deformationand displacement from shear zone patterns in the Variscanupper crust (Jebilet, Morocco). Journal of African EarthSciences, 32, 335–350.

Essaifi, A., Potrel, A., Capdevila, R. & Lagarde, J. L., 2003.Datation U-Pb: a ge de mise en place du magmatisme bimodaldes Jebilet centrales (Chaıne Varisque, Maroc). Implicationsgeodynamiques. Comptes Rendus Geosciences, 335, 193–203.

Etheridge, M. A., Wall, V. J. & Vernon, R. H., 1983. The role ofthe fluid phase during regional metamorphism and deforma-tion. Journal of Metamorphic Geology, 1, 205–226.

Ferry, J. M., 1994. Overview of the petrologic record offluid flow during regional metamorphism in northern NewEngland. American Journal of Science, 294, 905–988.

Ferry, J. M. & Dipple, G. M., 1991. Fluid flow, mineral reac-tions, and metasomatism. Geology, 19, 211–214.

Fitz Gerald, J. D. & Stunitz, H., 1993. Deformation of grani-toids at low grade metamorphism, I: Reactions and grain sizereduction. Tectonophysics, 221, 269–297.

Gapais, D., 1989. Les Orthogneiss: Structures, Mecanismes deDeformation et Analyse Cinematique. Memoires et Documentsdu CAESS, Rennes, 28. (in French).

Giggenbach, W. F., 1984. Mass transfer in hydrothermal sys-tems-A conceptual approach. Geochimica et CosmochimicaActa, 48, 2693–2711.

Grant, J. A., 1986. The Isocon Diagram. A simple solution toGresens’ equation for metasomatic alteration. EconomicGeology, 81, 1976–1982.

Gresens, R. L., 1967. Composition-Volume relationships ofmetasomatism. Chemical Geology, 2, 47–65.

Haggerty, S. E., 1976. Opaque mineral oxides in terrestrialigneous rocks. In: Oxides Minerals, Reviews in Mineralogy, 3.(ed. Rumble, D.), pp. 101–300. Mineralogical Society ofAmerica, Washington DC

Helmstaedt, H. & Allen, J. M., 1976. Metagabbronorite fromDSDP hole 334: an example of high-temperature deformationand recrystallisation near the Mid-Atlantic Ridge. CanadianJournal of Earth Sciences, 14, 886–898.

Hemley, J. J., Montoya, J. W., Nigrini, A. & Vincent, H. A.,1971. Some alteration products in the system CaO-Al2O3-SiO2-H2O. Society of Mining Geology, Japan, Special Issue, 2,58–63.

Huvelin, P., 1977. Etude geologique et gıtologique du massifhercynien des Jebilet (Maroc occidental). Notes et Memoires.du Service Geologique du Maroc, 232.

Jadid, M., 1989. Etude des processus de differenciation desroches magmatiques pre-orogeniques des Jebilet centralessur l’exemple du massif stratiforme de Koudiat Kettara(Maroc Hercynien). 3rd cycle Thesis, Marrakech University,Marrakech.

Kerrich, R., Fyfe, W. S., Gorman, B. E. & Allison, I., 1977.Local modification of rock chemistry by deformation. Con-tributions to Mineralogy and Petrology, 65, 183–190.

Kranidiotis, P. & Maclean, W. H., 1987. Systematics of chloritealteration at Phelps Dodge massive sulfide deposit, Matagami,Quebec. Economic Geology, 82, 1898–1911.

Lagarde, J. L., Aıt Omar, S. & Roddaz, B., 1990. Structuralcharacteristics of syntectonic plutons with special reference tolate Carboniferous plutons from Morocco. Journal of Struc-tural Geology, 12, 805–821.

Lagarde, J. L. & Choukroune, P., 1982. Cisaillement ductile etgranitoıdes syntectoniques: l’exemple du massif hercynien desJebilet (Maroc). Bulletin de la Societe Geologique de France,XXIV, 2, 299–307.

Lagarde, J. L. & Michard, A., 1986. Stretching normal to theregional thrust displacement in a thrust-wrench shear zone,Rehamna massif, Morocco. Journal of Structural Geology, 8,483–492.

Le Corre, C. & Bouloton, J., 1987. Un modele de �structure enfleur� associant decrochement et convergence: Les Jebilet

4 2 A . E S S A I F I E T A L .


centro-occidentales (Maroc hercynien). Comptes Rendus del’Academie Des Sciences, Paris, 13, 751–755.

Le Hebel, F., Gapais, D., Fourcade, S. & Capdevila, R., 2002.Fluid-assisted large strains in a crustal-scale decollement(Hercynian belt of South Brittany, France). In: DeformationMechanisms, Rheology and Tectonics: Current Status andFuture Perspectives, Special Publications, 200, (eds de Meer, S.,Drury, M. R., de Bresser, J. H. P. & Pennock, G. M.), pp. 85–101. The Geological Society, London.

Leake, B. E., 1978. Nomenclature of amphiboles. MineralogicalMagazine, 42, 533–563.

Lecuyer, C., Brouxel, M. & Albarede, F., 1990. Elementalfluxes during hydrothermal alteration of the Trinity ophiolite(California, USA). Chemical Geology, 89, 87–115.

Leger, A. & Ferry, J. M., 1993. Fluid infiltration and regionalmetamorphism of the Waits River Formation, north-eastVermont, USA. Journal of Metamorphic Geology, 11, 3–30.

Liou, J. G., Maruyama, S. & Cho, M., 1985. Phase equilibriaand mineral parageneses of metabasites in low-grade meta-morphism. Mineralogical Magazine, 49, 321–333.

Marquer, D. & Burkhard, M., 1992. Fluid circulation, pro-gressive deformation and mass-transfer processes in theupper crust: the example of basement-cover relationships inthe External Crystalline Massifs, Switzerland. Journal ofStructural Geology, 8 ⁄ 9, 1047–1057.

Mayol, S., 1987. Geologie de la partie occidentale de la bout-onniere paleozoıque des Jebilet, Maroc. PhD Thesis, MarseilleUniversity, Marseille.

McCaig, M. & Knipe, R. J., 1990. Mass-transport mechanismsin deforming rocks: recognition using microstructural criteria.Geology, 18, 824–827.

Mevel, C., 1984. Le metamorphisme dans la croute oceanique.Apport de la petrologie a la comprehension des phenomenesde circulation hydrothermale et de deformation (exemplesdans l’Atlantique). PhD Thesis, Paris VI University, Paris.

Oliver, N. H. S., 1996. Review and classification of structuralcontrols on fluid flow during regional metamorphism. Journalof Metamorphic Geology, 14, 477–492.

Otten, M. T., 1984. The origin of brown hornblende in theArtfjallet gabbro and dolerites. Contributions to Mineralogyand Petrology, 86, 189–199.

Philpotts J. A. & Schnetzler C. C., 1970. Phenocryst-matrixpartition coefficients for K, Rb Sr and Ba with applications toanorthosite and basalt genesis. Geochimica et CosmochimicaActa, 34, 307–322.

Pique, A., Jeannette, D. & Michard, A., 1980. The WesternMeseta shear zone, a major and permanent feature of theHercynian belt in Morocco. Journal of Structural Geology, 2,397–410.

Ramsay, J. G., 1967. Folding and Fracturing of Rocks. Mc-Graw-Hill, New York.

Richards, H. G., Cann, J. R., Richards, H. G. & Cowan, J. G.,1989. Mineralogical zonation and metamorphism of thealteration pipes of Cyprus sulfide deposits. Economic Geology,84, 91–115.

Rose, N. M. & Bird, D. K., 1994. Hydrothermally altereddolerite dykes in East Greenland: Implications for Ca-meta-somatism of basaltic protoliths. Contributions to Mineralogyand Petrology, 116, 420–432.

Rumble, D., 1994. Water circulation in metamorphism. Journalof Geophysical Research, 99, 15499–15502.

Sibson, R. H., Moore, J. M. & Rankin, A. H., 1975. Seismicpumping-a hydrothermal fluid transport mechanism. Journalof the Geological Society of London, 131, 653–659.

Stakes, D. & Vanko, D. A., 1986. Multistage hydrothermalalteration of gabbroic rocks from the failed MathematicianRidge. Earth and Planetary Science Letters, 79, 75–92.

Streit, J. E. & Cox, S. F., 1998. Fluid infiltration and Volumechange during mid-crustal mylonitization of Proterozoicgranite, King Island, Tasmania. Journal of MetamorphicGeology, 16, 197–212.

Taylor, H. G., 1974. The application of oxygen and hydrogenisotope studies to problems of hydrothermal alteration andore deposition. Economic Geology, 69, 843–883.

Vrolijk, P., 1987. Tectonically driven fluid flow in the Kodiakaccretionary complex, Alaska. Geology, 15, 466–469.

Watanabe, Y., 2002. 40Ar ⁄ 39Ar geochronologic constraints onthe timing of massive sulfide and vein-type Pb-Zn minerali-zation in the Western Meseta of Morocco. Economic Geology,97, 145–157.

White, S. H. & Knipe, R. J., 1978. Transformation and reactionenhanced ductility in rocks. Journal of the Geological Societyof London, 135, 513–516.

Xie, X., Byerley, G. R. & Ferrel, R. E., 1997. IIB trioctahedralchlorite from the Barberton greenstone belt: crystal structureand rock composition constraints with implications to geo-thermometry. Contributions to Mineralogy and Petrology, 126,275–291.

Yardley, B. W. D., 1986. Fluid migration and veining in theConnemara Schists, Ireland. In: Fluid–Rock InteractionsDuring Metamorphism (eds Walther, J. V. & Wood, B. J.),pp. 109–131. Springer, New York.

Zheng, Y. F., 1993. Calculation of oxygen isotope fractionationsin hydroxyl-bearing silicates. Earth and Planetary ScienceLetters, 120, 247–263.

Received 23 September 2002; revision accepted 8 October 2003.

