1-s2.0-S089953629800058

Pergamon Journal of African Earth Sciences, Vol. 27, No. 2, pp. 223-240, 1998

o 1998 Elsevier Science Ltd B I l . C ~ A a a a . I ~ / a Q ~ t l t t t t l = Q . V All rights reserved. Printed in Great Britain r l i , g lk /Qu¢l g g v f = ~ u l v v v t / u Jrl 0899-5362/98 $19.00 + 0.00

Metamorphism of the Palaeoproterozoic Magondi mobile belt north of Karoi, Zimbabwe

HUBERT MUNYANYIWA 1 and PIETER MAASKANT 2 1University of Zimbabwe, Department of Geology, P© Box MP 167 Mount Pleasant,

Harare, Zimbabwe =Vrije Universiteit, Faculty of Earth Sciences, De Boelelaan 1085, 1081 HV Amsterdam,

the Netherlands

Abstract--The Palaeoproterozoic Magondi mobile belt flanks the Zimbabwe Archaean Craton to the northwest. The belt is composed of metamorphosed sedimentary, volcanic and volcaniclastic rocks associated with quartzofeldspathic gneisses intruded by granitoids, some charnockitic, in the high-grade part of the belt. The belt is metamorphosed from low- grade greenschist-facies in the south and middle to upper amphibolite-facies in the north. Granulite-facies rocks are developed in the extreme north and northwestern part of the belt. Garnet-biotite geothermometry in metapelites indicates that temperatures increase from 590-600°C in the mid-amphibolite-facies through 640-690°C in the upper amphibolite- facies terrain and up to 730°C in the granulite-facies areas. In the granulite-facies terrains, garnet-biotite temperatures are similar to temperatures calculated using garnet-cordierite, garnet-clinopyroxene and, to some extent, two-feldspar geothermometers. Pressures calculated with the GASP barometers are 6 ± 1 kbar for both upper amphibolite- and granulite- facies, suggesting that the granulite-amphibolite-facies transition is primarily isobaric. The calculated pressures for granulites do not support models which invoke the formation of granulites by continent-continent collision. Instead the P-T data suggest that the Magondi mobile belt granulites were formed in a region of high heat flow, with heat possibly being supplied by deep-seated plutons, o 1998 Elsevier Science Limi ted.

Rdsum6--La ceinture mobile pal6oprotdrozo'(que de Magondi flanque le craton archden du Zimbabwe au nord-ouest. La ceinture se compose de roches sddimentaires, volcaniques et volcano-clastiques mdtamorphisdes ass©c/des ~ des gneiss quartzo-feldspathiques recoupds par des granitd(des et quelques charnockites dans la partie de degr(~ dlevd de la ceinture. La ceinture est m6tamorphis6e depuis le facies des schistes verts au sud jusqu'aux facies moyen et supdrieur des amphibolites au nord. Les roches du facies des granulites se ddveloppent dans les parties extremes nord et nord-ouest de la ceinture. Le gdothermom(~tre grenat-biotite sur les mdtapdlites indique que les tempdratures augmentent de 590-600°C dans le facies moyen des amphibolites ~ 640-690°C dans le facies supdrieur des amphibolites et jusqu'~ 730°C dans les rdgions dans le faci6s des granulites. Dans les terrains granulitiques, les temp6ratures grenat-biotite sont semblables 8 celles calculdes par les thermom(~tres grenat-cordidrite, grenat-clinopyrox~ne et, jusqu'8 un certain point, deux feldspaths. Les pressions calcul6es par le barom~tre GASP sont de 6 + 1 kbar pour les facies des amphibolites et des granulites, suggdrant que la transition amphibolite- granulite est essentiellement is©bare. Les pressions calculdes ne favor/sent pas les mod(~les de formation des granulites par collision continent-continent. Au contraire, nos rdsultats P- T sugg6rent que les granulites de la ceinture mobile de Magondi se sont form6es dans une rdgion de flux de chaleur important, la chaleur ayant pu ~tre fournie par des plutons mis en place en profondeur. © 1998 Elsevier Science Limi ted.

(Received 10 February 1997: revised version received 10 February 1998)

Journal of African Earth Sciences 223

H. MUNYANYIWA and R MAASKANT

INTRODUCTION

Metamorphism of the Magondi mobile belt in northwestern Zimbabwe is largely known from petrographic data acquired during regional reconnaissance mapping by the Geological Survey of Z imbabwe , and only a few geothermobarometr ic studies have been a t tempted (Treloar, 1988; Treloar and Kramers, 1989; Munyanyiwa et al., 1993). A full appreciation of the geological evolution of a metamorphic terrain requires detailed, quantitative estimates of P and T. In particular, an unders tand ing of the nature of the amphibolite-granulite-facies transition may be achieved from the distribution of temperature and pressure cond i t ions ex is t ing during metamorphism. The aim of this paper is to present P-Tdata for the medium- to high-grade terra ins of the Magondi mobi le belt in northwestern Zimbabwe. The temperatures calculated wi th various geothermometers indicate a north to northwest increase in the grade of metamorphism wi th in the belt, whereas pressures calculated from garnet- sillimanite-plagioclase-quartz geobarometers remain relatively constant from amphibolite- facies to granulite-facies terrains.

GEOLOGICAL SETTING

The Palaeoproterozoic Magondi mobile belt flanks the Zimbabwe Archaean Craton to the northwest (Fig. 1 ). The belt is composed partly of a supracrustal sequence termed the Magondi Supergroup (Leyshon and Tennick, 1988), which has been subdivided into the basal Deweras Group and the unconformably over ly ing Lomagundi and Piriwiri Groups. The Deweras Group crops out in two main parts of the belt (Fig. 2) and consists of metamorphosed conglomerates, arkoses, feldspathic sandstones, g reywackes , maf ic tho le i i t i c lavas and subordinate quantities of evaporites (Master, 1991). The Lomagundi Group, exposed along the entire length (-250 km) of the belt, comprises mainly quartzites, dolomites and phyllites, together w i th high-grade metamorph ic equivalents of these rocks to the north of Karoi (Fig. 2). The Piriwiri Group, the westernmost part of the sequence, is composed of pyritiferous and graph i t i c phy l l i tes , s lates, cherts, greywackes and volcaniclastic rocks, including tuffs and agglomerates.

The Magondi Supergroup was deformed and metamorphosed during the Ubendian orogenic cycle, about 2.0-1.8 Ga ago, and is now part of

, I Figure 1. Simplified geological map of Zimbabwe showing the mobile belts and cover surrounding the Archaean craton (after Stagman, 1978). Box shows the Magondi mobile belt enlarged on Fig. 2.

224 Journal of African Earth- Sciences


~u

UJ

ZAMBEZI /5 \X~ VALLEY

'P Shaml

LAKE

KARIBA

\\\\\x~

, -., \ ,..-/"----

- \ \ \ \ \\

F~ \ \ \ \ ' . \ \ \ \

x \ \ ,, \ \ . . . \ \ \ \ \ \ ' . \ \ \ ~ \ \ j

X X v v x Knroi X X X x x.] X X

~ X X "

X X

X

0 0 o

T.~> _ _ -

Mhom

+ + # + +

+

+

0 30 Km I I I I a le V

L E 6 E N O 0

"~IYouNGER COVER ~

Z~ ,o.=u.o,... 7~"'°u;

x ~-'~ Urungwe granite 6neisses

Oueen

" C o p p e r King_

I T A + + + - ~:'i::"_F + + + J~ rig*a+ Jr -~j~ +++ +÷ ++÷ ++

+ + + + ~F t + + + +

~- + + + ÷ + + + + +

)+ + + +

4- 4- + + + +

+

-t- + -t-

v ~ . . V

G,..,,o,,. v v v v v v v q _ ~ i= p F f , v v v ~ + t v v v v ~ v

• ~ vv V V v V/-I- I V v 4- -I- t-

Figure 2. Geological map of northwest Zimbabwe showing the Magondi Supergroup and the basement gneisses and granitoids of the Archaean craton and the surrounding cover (after Stagman, 1978). RK: Rukomechi; NY: Nyaodza.


H. M U N Y A N Y I W A and R M A A S K A N T

a large east vergent fold and thrust belt along the western margin of the Zimbabwe Craton (e.g. Treloar, 1988; Hartnady et al., 1985). The grade of metamorphism within the Magondi mobile belt increases along s t r ike to the nor th and northwest, from greenschist-facies near and south of the Shackleton and Avondale Cu mines near the craton margin, through mid-amphibolite- facies around Karoi, to upper amphibolite- and granulite-facies north and northwest of Karoi (Fig. 2). Granulite-facies rocks are developed east of Makuti, in the Rukomechi area, and west of Karoi, in the Nyaodza area (Fig. 2). Granulite- facies rocks include charnockites, enderbites, quartzofeldspathic gneisses, garnet-sillimanite and cordier i te-si l l imanite gneisses and less abundant diopside-hornblende bearing calc- silicate rocks.

The q u a r t z o f e l d s p a t h i c gne isses are predominantly medium- to coarse-grained biotite gneisses, locally migmatised during partial melt ing. They are composed of alkali and plagioclase feldspars with biotite as the dominant mafic mineral associated with less abundant hornblende and locally, in some samples, garnet, clino- and orthopyroxene. These gneisses were mapped as paragneisses partly on the basis of their layered structure and high modal biotite content (e.g. Stagman, 1962). However, the gneisses have been severely affected by intense ductile deformation and metamorphism under medium- to high-grade conditions during the Magondi Orogeny and are locally migmatised. The nature of the protoliths in most places has therefore been obliterated. Their widespread occurrence north and nor thwes t of Karoi, however, suggests that, originally, they could have been granitoids rather than an extensive thick pile of arkoses and feldspathic sandstones. Preliminary Rb-Sr and U-Pb radiometric data suggest that the gneisses are Palaeoproterozoic to Neoarchaean in age (Loney, 1969; Kr6ner, pers. c o m m . , 1996). More specifically, KrSner, (pers. c o m m . , 1996) obtained Pb-Pb zircon ages of 1920+0 .3 Ma and 1963+0 .2 Ma for two samples of the Kariba gneiss unit, which relate to the present study.

To the northwest, the Magondi mobile belt is unconformably over la in by a supracrusta l sequence termed the Makuti Group (Fig. 2). This cons is ts of q u a r t z o f e l d s p a t h i c rocks, metabas i tes, metape l i tes , quar tz i tes and surbodinate intercalations of marbles and calc- s i l icate rocks (Broder ick, 1976; Fey and Broderick, 1990; Munyanyiwa et al. , 1997). Deformation and metamorphism of the Makuti

Group, together with the underlying basement, the latter also previously deformed during the Palaeoproterozoic Magondi Orogeny, took place during Pan-African tectonothermal events at ca 800 Ma (Loney, 1969).

Temperatures w i th in the medium-grade Magondi terrain were calculated by Treloar (1988) to be < 5 0 0 ° C for the garnet zone, 550- 600°C for the staurolite zone and up to 630°C for the kyanite zone, using the biotite-garnet geothermometer. Pressures of at least 6 kbar were tentatively estimated for the kyanite zone from the intersection of the staurolite + quartz = kyanite+garnet equilibrium of Yardley (1981). Granulite-facies conditions were estimated from the enderbites by Munyanyiwa et al. (1993) to have been 700-800°C and 5-7 kbar, using the garnet-orthopyroxene geothermometer of Lee and Ganguly (1988) and the plagioclase- orthopyroxene-garnet-quartz geobarometer of Newton and Perkins (1982) . Simi lar P - T conditions for granulite-facies metamorphism were also calculated by Treloar and Kramers (1989) , based on both cord ier i te -bear ing metapelit ic assemblages and orthopyroxene- bearing charnockitic assemblages.

The tectonic setting of the Magondi mobile belt has been discussed by Leyshon and Tennick (1988), Stowe (1989) and Master (1991). Leyshon and Tennick (1988) suggested, on the basis of a lack of a collision suture, ophiolites or m~langes within the Magondi mobile belt, that the belt developed in an ensialic geosyncline along the western margin of the Zimbabwe Craton. Stowe (1989), however, pointed out that the Magondi supracrustal sequence shows a general depositional polarity from the proximal Lomagundi Group carbonate-quartz i te shelf sequence in the east to distal Piriwiri Group continental-slope type sediments and possible trench facies in the west. The distribution of the Magondi supracrustals was interpreted to record the transition of a passive-margin setting into geosynclinal f lysch-type deposits (Stowe, 1989). Stowe (1989) further suggested that the 1.8-2.0 Ga Proterozoic terrains in southern Africa, including the Magondi mobile belt, may be interpreted in plate-tectonic terms. Master (1991) on the other hand, interpreted the Magondi Supergroup as having been deposited in a back-arc continental basin developed in response to an easterly directed subduction zone.

In the present study, geothermobarometric results are presented of a number of metapelite samples obtained from the middle amphibolite- fac ies ter ra in no r theas t of Karoi around

226 Journal of African Earth Sciences

Metamorphism of the Palaeoproterozoic Magondi mobile belt north o f Karoi, Zimbabwe

Shamrocke mine, through the upper amphibolite- facies terrain around and west of Karoi, and from granulite-facies rocks east of Makuti and west of Karoi (Fig. 2). P-T data from one granulite- facies quartzofeldspathic gneiss sample from Kariba (Fig. 2) are also presented. The tectonic implications of these results are discussed in the light of prevailing models of granulite-facies metamorphism.

FIELD RELATIONS AND PETROGRAPHY Amphibolite-facies pelitic and semipelitic schists These rocks crop out mostly within the mid- to upper amphibolite-facies terrain between Karoi and Shamrocke mine (Fig. 2). They are included in the Lomagundi and Piriwiri Groups on the 1:1 000 000 geological map of Zimbabwe (Stagman, 1978), although it is not possible in this study to distinguish between the Lomagundi and Pi r iwi r i metape l i tes because they are petrographically similar.

The rocks are typically medium- to coarse- grained, comprising mainly bioti te, quartz, muscovite and plagioclase. Garnet is found in minor amounts in all the examined specimens, whereas sillimanite occurs in high-grade, upper amphibolite-facies rocks (e.g. sample hm-O03). Zircon, rutile and apatite are present in trace quant i t ies. The metapel i tes have a well- developed fol iat ion defined by biot i te and muscov i te . In upper amph ibo l i te - fac ies metapelites, sillimanite also helps define the fol iat ion together wi th biot i te, but minor muscovite is present as a secondary mineral.

Granulite-facies metapelites Granulite-facies metapelites, the high-grade metamorphic equivalents of the Lomagundi and Piriwiri Groups, crop out around Rukomechi and Nyaodza (RK and NY, Fig. 2). In the Nyaodza area, garnet-sillimanite gneisses are the dominant pelitic granulites and are associated with minor cordier i te-si l l imanite gneisses. The pelit ic granulites are interleaved with quartzofeldspathic migmatitic rocks of the Chipisa and Urungwe Gneisses (Broderick, 1976; Fey and Broderick, 1990). Structural analysis of the pelitic granulites and quartzofeldspathic gneisses reveals that the two units are structurally concordant and were deformed together during the Magondi Orogeny.

The garnet-sillimanite gneisses are medium- to coarse-grained rocks. They are characterised by a diffuse to well-developed migmati t ic compositional layering defined by felsic bands up to 5 mm thick with alternations of mafic layers

2-4 mm thick. Quartz, alkali feldspar and plagioclase are the main minerals in the felsic layers, whereas sillimanite, garnet and biotite are the dominant minerals in the mafic bands. Zircon is a common accessory mineral and occurs as inclusions in biotites or as small and well-rounded discrete crystals. Magnetite is present in minor amounts in some gneisses where it is intergrown with tiny ilmenite crystals, and in places contains inclusions of dark green hercynitic spinel. Traces of spinel are found in the majority of the garnet-sillimanite gneisses as anhedral cyrstals. As observed in many other granulite-facies metapelites (e.g. Bohlen et al., 1986), hercynitic spinel shows complicated textural relations with other minerals. For instance, the mineral is not in contact with quartz but is commonly surrounded by rims of sillimanite and in some cases by rims of garnet and magnetite. Sillimanite is in direct contact with garnet and quartz. This texture indicates a retrograde reaction in which a high temperature assemblage (hercynitic spinel + quartz) was replaced by a low temperature assemblage (sillimanite +garnet), possibly during isobaric cooling as suggested by Munyanyiwa et al. (1993).

Locally, the garnet-sillimanite gneisses are conspicuously porphyroblastic, with sub- to euhedral garnet porphyroblasts up to 3 cm in size. The porphyroblasts contain numerous inclusions of sillimanite, biotite and quartz. Garnet porphyroblasts in these samples have overgrown earlier formed brown biotite, which suggests that the porphyroblasts are late- to post-tectonic with respect to the main Magondi fabric-forming event. Perthite occurs in minor amounts in most of the garnet-si l l imanite gneisses. Orthoclase is locally present as porphyroblasts up to 1 cm across in a few samples (e.g. sample hm-384). Sillimanite forms two textural types in a few samples; the first type is aligned parallel to the foliation or is strongly kinked. In places these sillimanite grains are recrystallised into finer acicular sillimanite crystals of -0.2 mm in size, associated with fine- grained retrograde green biotite. The second sillimanite type is sub- to euhedral, crosscuts the foliation and is also overgrown on the first generation of sillimanite, which implies that it is late- to post-tectonic.

The garne t -s i l l iman i te gneisses from Rukomechi (RK, Fig. 2) are coarse-grained, massive to weakly foliated rocks; the foliation is defined by lepidoblastic, brown biotite. Mineralogically, the rocks resemble the garnet- si l l imanite gneisses from Nyaodza. Trace



A (a)

Grt Bt

A (b)

M"

Grt Bt

A

Bt

A

9 F

Gd Bt

Figure 3. AFM (AI203-FeO-MgO) diagrams depicting the Fe-Mg distr ibution between garnet and biot i te in the pefit ic granufites (hm-302, hm-346, hm-384 and hm-396) and a pefit ic schist (hm-O06). The Fe-Mg distr ibution between garnet, biot i te and cordierite is shown on (d). Sample hm-O06 does not contain sil l imanite.

amounts of late muscovite replace plagioclase. Calcite, where present, is also a secondary mineral after plagioclase. Quartz exhibits a mortar texture in which quartz porphyroclasts with undulose ext inct ion are surrounded by recrystallised fine-grained quartz with a well- developed granoblastic polygonal texture.

The co rd ie r i t e - s i l l iman i t e gne isses are intimately associated with the garnet-sillimanite gneisses west of Karoi in Nyaodza, but none were found in Rukomechi . The rocks are medium- to coarse-grained and contain garnet, biotite and quartz in lesser quantities than in the garnet-sillimanite gneisses. Cordierite, in



stable coexistence with garnet based on the textural evidence, occurs as turbid crystals up to 2 cm in size and in some outcrops forms porphyroblasts up to 4 cm in diameter. In thin section, cordierite shows penetration and sector twinning and is pinitised locally. It contains numerous inclusions of biotite, quartz, sillimanite and zircon. Sub- to euhedral magnetite is common in the cordierite-sillimanite gneisses and conta ins inc lus ions of hercyn i t i c spinel. Sillirnanite occurs along the margins of the spinel, as in the garnet-sillimanite gneisses. Garnet occurs in minor amounts in the cordierite- sillimanite gneisses as subhedral crystals and contains numerous inclusions of graphite and hercynitic spinel.

The mineral assemblages in both the garnet- sillimanite and cordierite-sillimanite gneisses are muscovite-free, granulite-facies assemblages. Textural and mineralogical evidence suggest that the migmatitic layering present in these rocks represents a vapour-absent melt reaction during granulite-facies metamorphism.

Quartzofeldspathic gneisses As pointed out above, quartzofeldspathic gneisses are abundant in the high-grade part of the Magondi mobile belt. The sample hm-300, P-Tdata from which are presented in this study together with the metapelite data, is part of the strongly deformed granulite-facies Kariba Gneiss. The Pb-Pb zircon ages of two Kariba Gneiss samples of 1920_+0.3 Ma and 1963_+0.2 Ma (KrOner, pers. comm. , 1996) indicate that the Kariba Gneiss is syntectonic with respect to the Magondi Orogeny. It is coarse-grained and contains plagioclase, quartz and biotite as the main mineral const i tuents. Alkali feldspar, garnet, hornblende and clinopyroxene are found in minor amounts, whereas zircon and apatite occur in trace quantities. Orthopyroxene is rarely found in specimens of the Kariba gneisses, but is common in enderbit ic and charnocki t ic gneisses in Rukomechi and Nyaodza.

MINERAL CHEMISTRY Analytical technique Mineral analyses were obtained at the Vrije Un ive rs i te i t , Amsterdam. A Cambridge Microscan 9 ® electron microprobe with two automated wavelength-dispersive spectrometers and an on-line ZAF correction program was used. Analyses were performed at an operating voltage of 15 kV with a probe current of 30 nA. Well- calibrated natural silicates and oxides were used

Gts

Amphebohte faczu 6rs p r ~ ~ s

Prp Aim ÷ Sps Figure 4. Compositions of garnets of the Magondi metapelites plotted on a grossular (Grs)-almandine (AIm) + spessartine (Sps)-pyrope (Prp) diagram.

as standards. The feldspars, biot i tes and cordierites were analysed with a rastered broad electron beam of -600 I~m 2. For analysis of minerals that are stable under the probe, a focused electron beam of -1 i~m in diameter was used. P-T values were calculated with P- TVU, an internal program housed at the Vrije Universiteit, Amsterdam. Partial AFM (AI203-FeO- MgO) diagrams depicting the Fe-Mg distribution between garnet and biotite, and between garnet, biotite and cordierite, are given in Fig. 3a-d.

Garnets Garnets in the metapelites are almandine-rich (Table 1 and Fig. 4, cf. Deer et al., 1982). The Xu0 [Mg/(Mg + Fe)] is 0.1-0.12 in garnets from amphibolite-facies pelitic schists and 0.3-0.35 in garnets from granulite-facies metapelites. Grossular components are genrally low (2-3 mol.%). Garnets in pel i t ic granul i tes are homogenous in composition except, in some samples (e.g. hm-310), for the extreme outer rims (outer 10 pm), which are relatively Fe-rich.

A number of studies have shown that, due to high di f fusion rates at high temperature, chemical zonation is rarely preserved in minerals during high-grade metamorphism (e.g. Harley, 1989; Essene, 1989). The presence of Fe-rich garnet rims adjacent to Mg-rich phases, such as biotite in granulite-facies rocks, is probably due to re-equi l ibrat ion dur ing re t rogradat ion (Fitzsimons and Harley, 1994).


Table 1. Representative microprobe analyses of garnets

hm-310c hm-310r hm-075c hm-O75r hm-396r hm-396c SiO 2 38.78 38.85 37.81 36.87 38.52 38.79

AI203 2 2 . t 2 22.24 21.44 21.89 21.71 21.46 FeO* 29.18 29.06 34.41 35.91 33.79 32.23 MnO 0.81 0.72 0.66 0.89 1.31 1.29 MgO 8.72 8.65 4.78 3.34 5.22 5.98 CaO 1.09 1.14 0.74 0.75 0.94 0.80 Total 100.70 100.66 99.85 99.04 101.51 100.55 Si 2,985 2.988 3.009 2,991 3.009 3.036 AI 2 .006 2.005 2.010 2,010 1.999 1.979 Fe 1.878 1.869 2,290 2.436 2.208 2.109 Mn 0,053 0.047 0 .044 0.061 0.087 0.086 Mg 1.000 0.992 0,567 0.404 0.608 0.698 Ca 0 .090 0.094 0,063 0.065 0.079 0,067

XFe 0.622 0.623 0.773 0.821 0 .740 0.712

XMn 0.018 0.016 0.015 0.021 0.029 0.029

XMg 0.331 0.330 0.191 0.136 0 .204 0.236

Xca 0 .030 0.031 0.021 0.022 0 .026 0.023 0 12.00 12.00 12.00 12.00 12.00 12.00

Total Fe as FeO; c: core; r: rim; av: mean of f ive uniform point analyses.

hm-346r hm-346c 39.46 39.01

22.10 21.78 28.45 28.63

3.04 3.32 7.54 7.24 0.94 0.90

101.53 100.88 3.024 3.018 1.996 1.986 1.823 1.852 0.197 0.218 0.861 0.835 0.077 0.075

0.616 0.621

0.067 O.O73

0.291 0 .280

0.026 0.025

12.00 12.00

hm-3OOr hm-300c hm-OO3r hm-O03c m-384av m-348av m-O73av 37.86 37.64 37.58 37.19 38.54 37.97 37.59

21.11 20.98 21.07 21.08 21.83 21.56 21.44

29.52 30.28 33.44 33.76 30.35 31.43 32.96 1.43 1.52 5.06 4.51 1.26 0.87 1.11 2.56 2.25 2.25 2.62 7.05 6.08 4.19 7.82 7.18 1.09 1.22 0.97 1.12 2.10

100.30 99.85 100.70 100.38 100.01 99.02 99.39 3.005 3.011 3.023 2.997 3.009 3.011 3.006 1.974 1.978 1.997 2.002 2.009 2.015 2.021 1.959 2.025 2.249 2.275 1.982 2.085 2.204 0.096 0.103 0.345 0.308 0 .084 0.058 0 .075 0.303 0.268 0 .270 0.375 0 .820 0.719 0 .499 0.665 0.615 0.095 0.105 0.081 0.095 0 .179

0.648 0.673 0 .760 0.743 0.668 0.705 0 .745

0.032 0 .034 0.117 0.101 0.028 0 .020 0.025

0.100 0.089 0.091 0.122 0.276 0.243 0 .169

0.220 0 .204 0.032 0 .034 0.027 0.032 0.061

12.00 12.00 12.00 12.00 12.00 12.00 12.00


Table 2. Representative microprobe analyses of biotites

hm-003 hm-073 hm-075 hm-006 hm-134 hm-396 hm-346 hm-384 hm-300 hm-348 SiO 2 35.12 35.48 36.00 35.41 35.79 35.65 36.92 35.59 34.61 35.37 TiO 2 2.85 2.53 2.37 2.21 3.57 3.52 3.84 4.03 4.93 4.61 AI203 18.91 17.75 18.82 19.45 14.77 17.01 16.63 17.90 13.40 17.40 FeO* 21.76 19.20 18.76 21.36 19.98 17.56 14.62 15.89 24.25 16.46 MnO 0.18 0.03 0.01 0.04 0.08 0.04 0.07 0.00 0.00 0.01 MgO 6.79 9.52 7.72 8.30 10.51 11.17 13.24 10.50 7.48 9.69 Na2 O 0.17 0.17 0.07 0.30 0.10 0.12 0.10 0.08 0.09 0.09 K20 9.36 9.18 9.94 8.99 9.12 9.69 9.95 9.37 8.36 9.64 Total 95.14 93.86 93.69 95.76 93.97 94.86 95.37 93.33 93.27 93.27 Si 5.420 5.480 5.560 5.375 5.556 5.432 5.506 5.444 5.539 5.443 AI jv 2.580 2.520 2.440 2.625 2.444 2.568 2.494 2.554 2.461 2.557 AI vl 0.859 0.711 0.986 0.855 0.258 0.487 0.429 0.673 0.060 0.599 Ti 0.331 0.294 0.275 0.252 0.417 0.403 0.431 0.460 0.593 0.534 Fe 2.808 2.480 2.423 2.704 2.594 2.238 1.823 2.033 3.246 2.119 Mn 0.024 0.004 0.001 0.005 0.011 0.005 0.009 0.000 0.000 0.001 Mg 1.562 2.192 1.777 1.878 2.432 2.537 2.944 2.394 1.785 2.222 Na 0.051 0.051 0.021 0.088 0.030 0.035 0.029 0.024 0.028 0.025 K 1.843 1.809 1.958 1.741 1.806 1.884 1.893 1.828 1.801 1.892 Xig 0.357 0.469 0.423 0.410 0.484 0.531 0.618 0.541 0.355 0.512 O 22.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00 22.00

Total Fe as FeO.

Biotites Biotites in amphibolite-facies pelitic schists have Ti values of 0.24-0.28 per formula unit (pfu) based on 22 O and XMg (=Mg/Mg +Fe) values of 0.40-0.41. In contrast, biotites in granulite- facies metapelites have Ti contents of 0.40-0.59 pfu and XMgvalues of 0.54-0.62. AI w varies from 0.05 to 1.2 pfu in biotites in amphibolite-facies pelitic schists and from 0.2 to 1.0 pfu in biotites in pelitic granulites (Table 2).

The high Ti contents in biotites in pelitic granulites are due to the presence of a Ti- saturating phase which is mainly ilmenite (cf. Guidotti, 1984).

The variations in XM0and AI w in biotites are cont ro l led by the t schermak subs t i t u t i on [MgSi(AIWAl~V) 1] t h rough the ne t - t rans fe r reaction (equation 1). This reaction accounts for the decrease in AI w and increase in X io in b i o t i t e s f r om p e l i t i c s c h i s t s to pe l i t i c granulites. However, Guidotti (1 984), Guidotti et al. (1988), Mohr and Newton (1983) and Nesbitt and Essene (1982) have shown that XMQ values in biot i tes may also increase wi th increasing O and S fugacit ies. The presence of i l m e n i t e and, l o c a l l y , g r a p h i t e and magnetite, and the absence of sulphides and

hematite in the rocks examined in this study indicate that changes in fO 2 and fS 2did not contr ibute signif icantly to changes in XMo in biotites.

Feldspars Plagioclase feldspars in individual metapelitic rock types are almost uniform in composition (Table 3). Composit ions vary from An~7_l 9 in amphibolite-facies pelitic schists to An21_23 in upper amph ibo l i t e - fac ies metape l i tes . In granulite-facies metapelites, the plagioclase anorthite content is An28.29 in garnet-sillimanite gneisses and is andesine (An4o) in the cordierite- s i l l imani te gneisses. Because plagioclase feldspars are in rocks of broadly similar pelitic composition, the compositional variations of the plagioclase are, to a large extent, controlled by metamorphic grade.

Alkali feldspar composit ions fall on the K- Na join with K/ (K+Na) generally ranging from 0.67 to 0 .96. In one cordier i te-s i l l imani te gneiss (sample hm-384), the K/(K + Na) value is as-low as 0.50. The alkali feldspars in this specimen are ext remely exsolved and it is possible that some of the Na in the analysis may be from adjacent plagioclase lamellae.

4KMg3AISi3Olo(OH) 2 (Bt) + 12SiO 2 (Qtz) = 3Mg3AI2SizO12 (Prp) + 4KAISi308 (Kfs) + 3MgSiAl_2(Tsch-Bt) + 4H20 (1)



Table 3. Representative microprobe analyses of feldspars

hm-396(pl) hm-396(pl)hm-396(kf)hm-396(kf) hm-384(pl)hm-384(kf) hm-075(pl) hm-075(pl) SiO 2 61.08 60.94 64.96 64.82 58.30 63.53 60.96 60.69

AI20 3 24.20 24.45 18.39 t 8.37 25.40 20.41 24.13 24.40 CaO 6.11 6.28 0.13 0.06 7.75 2.11 5.96 6.07 Na20 8.05 7.97 2.54 2.48 6.96 4.93 8.13 8.33

K20 0.18 0.17 13.02 13.35 0.10 8.34 0.31 0.31 Total 99.62 99.81 99.04 99.08 98.51 99.32 99.49 99.80 Si 2.723 2.835 3.000 2.996 2.640 2.896 2.724 2.708 AI 1.273 1.342 1.000 1.000 1.357 1.097 1.271 1.284 Ca 0.292 0.313 0.001 0.003 0.376 0.103 0.285 0.290 Na 0.696 0.726 0.227 0.222 0.611 0.436 0.704 0.721 K 0.010 0.010 0.767 0.787 0.006 0.485 0.020 0.020 0 8.00 8.00 8.00 8.00 8.00 8.00 8.00 8.00

pl: plagioclase; kf: alkali feldspar.

Table 4. Representative microprobe analyses of cordierites

hm-384c hm-384c hm-384r hm-310r hm-310c hm-384r

S iO 2 49.26 49.28 49.16 49.74 49.69 49.51

A I 2 0 3 32.82 32.62 32.80 32.65 32.35 32.19

FeO* 5.96 6.48 6.64 5.13 5.19 5.13 MnO 0.00 0.03 0.09 0.01 0.01 0.00 MgO 9.78 9.44 9.28 9.68 9.74 9.53 Total 97.72 97.85 97.97 97.21 96.98 96.38

Si 5.030 5.047 5.034 5.089 5.098 5 .108 AI 3.958 3.938 3.959 3.937 3.913 3.914 Fe 0.510 0.555 0.569 0.439 0.445 0.443 Mn 0.000 0.003 0.008 0.001 0.001 0.000 Mg 1.492 1.441 1.417 1.476 1.490 1.466

XMg 0.745 0.722 0.714 0.771 0.770 0.768

O 18.00 18.00 18.00 18.00 18.00 18.00

Total Fe as FeO; r: rim; c: core.

Cordierite Cordierites are Mg-rich (Mg/Mg +Fe = 0.71-0.77) and have a slight within-grain chemical zonation revealed by Mg-rich cores relative to the rims (Table 4). The cordierite contains inclusions of biotite and sillimanite and coexists with garnet and associated granitic leucosomes. These textural relationships suggest that the cordierite formed by the following vapour-absent melting reaction involving biotite, sillimanite and quartz (cf. Owen, 1991):

Bt + Sil + Qtz = Crd + Grt + melt. (2)

The occurrence of this reaction is also supported by low modal biotite, sillimanite and quartz in the cordierite-bearing metapelites relative to the cordierite-free, garnet-sillimanite gneisses.

Spinel The spinel in these rocks is a hercynite spinel solid solut ion (FeAI204-MgAI204) , wi th Xug varying between 0.30 and 0.35. The mineral contains up to 0.5 wt% ZnO and up to 0.6 wt% Cr203 (Table 5). Recalculation of spinel compositions according to stoichiometry with the method of Bohlen and Essene (1977) gives 2.0-6.4 wt% Fe203 (Table 5).

As no ted above , h e r c y n i t i c sp ine l is invariably rimmed by sil l imanite and is not in contact with quartz, whereas sill imanite is in textural equilibrium with garnet and quartz. This texture suggests the fol lowing retrograde reaction:

3Hc + 5Qtz = Aim + 2Sil, (3)



Table 5. Representative micropobe analyses of spinels

hm-384r hm-384c hm-384c hm-310c hm-310r hm-310r

AI203 58.69 58.63 59.52 56.96 58.64 58.03

Cr203 0.36 0.33 0.35 0.27 0,57 0.00

Fe203 2.07 2.54 1.90 5.65 4.70 6.37

FeO 31.61 31.28 30.92 29.44 28.04 28.17 MnO 0.07 0.00 0.05 0.34 0.14 0.09 MgO 6.06 6.23 6.63 7.21 8.63 8.69 ZnO 0.15 0.25 0.20 0.19 0.16 0.00 Total 99.02 99.26 99.57 100.07 100.87 101.35 AI 1.944 1.939 1.952 1.880 1.890 1.870 Cr 0.008 0.007 0.008 0.010 0.010 0.000

Fe 3+ 0.044 0.054 0.040 0.120 0.100 0.130

Fe 2+ 0.743 0.734 0.720 0.690 0.640 0.640 Mn 0.002 0.000 0.001 0.100 0.000 0.000 Mg 0.254 0.261 0.275 0.300 0.350 0.350 Zn 0.003 0.005 0.004 0.000 0.000 0.000 O 4.00 4.00 4.00 4.00 4.00 4.00

Fe203, FeO calculated from stoichiometry; r: rim, c: core.

SiO2

A QUARTZ

3Hc + 5Qtz = Aim + 2Sil

SILLIMANITE

GARNET ~.~F I

(FeO + MgO) SPINEL AI203

Figure 5. Plot of Si02-(FeO + MgO)-AI2Ozshowing the relationship between quartz, spinel, sillimanite and garnet in some metapelitic granulites. The phase relations and textural evidence suggest the reaction indicated on the figure (see text for discussion/.


H. MUNYANYIWA and R M A A S K A N T

Table 6. Garnet-biotite temperatures and garnet-aluminosilicate-plagioclase-quartz pressures (in °C and kbar)

4 kbar F-S H-L P-L P-A I-Ma I-Mb K-R Da Tho Bha ave K-R hm-396 783 707 675 675 677 655 674 698 703 648 680 4.7 hm-346 785 708 662 676 690 682 673 735 727 652 680 5.2 hm-073 719 669 647 646 662 643 654 642 682 643 650 7.2 hm-348 850 745 706 704 687 652 695 699 771 685 700 hm-075 875 761 721 725 778 727 689 680 789 699 720 6.0 hm-003 706 660 623 640 657 642 633 573 672 609 640 5.6 hm-384 891 767 719 720 748 714 705 741 797 696 730 4.9 hm-O06 623 608 597 599 578 551 598 527 612 585 600 hm-300 774 703 670 681 712 764 794 789 723 700 720 hm-270 754 689 661 694 676 680 701 722 707 668 690

F-S: Ferry and Spear (1978); H-L: Holdaway and Lee (1977); P-L: Perchuk and Lavrent'eva (1983); P-A: Perchuk and Aranovich (1986); I-Ma and IMb: Indares and Martignole (1985); K-R: Kleemann and Reinhardt (1994); Da: Dasgupta et al. (1991 ); Tho: Thompson (1976); Bha: Bhattacharya et al. (1992). Overall garnet-biotite temperature uncertainities may be estimated at _+50°C.

which is supported by phase relations in the (FeO + MgO)-AI203-SiO 2 (FMAS) system (Fig. 5). Because the right hand side of the reaction is a low temperature side relative to the left side, this reaction is a retrograde reaction taking place during isobaric cooling.

MINERAL THERMOBAROMETRY Garnet-biotite The garnet-biotite pair is the most common assemblage in metapelites used for temperature estimates. Many authors have calibrated this thermometer and at this moment there are more than twenty calibrations available (see references in Table 6). Most calibrations take Ca and Mn contents in garnet into account, while AI and Ti contents in biotite also are incorporated in the latest versions. Evaluations of the garnet-biotite thermometers are given by Chipera and Perkins (1988), Fonarev et al. (1990a) and Kleemann and Reinhardt (1994).

Temperature estimates were done on the combination matrix biotite-core garnet, in an at tempt to constrain peak metamorphic conditions as closely as possible. The cores of large garnets may differ in composition, as only two-dimensional sections are available in thin section. A distinct trend is usually visible towards higher Mg values for core compositions, which facilitates the proper choice of mineral analysis.

Table 6 shows temperature estimates at 4 kbar for various calibrations of the garnet-biotite geothermometer. The geothermometer shows only slight pressure dependence: 1 kbar difference results in a temperature increase of

about 5°C. Ferry and Spear (1978) values are added to demonstrate that for higher grade metamorphic rocks this thermometer, and also the ones closely related or derived from it (Hodges and Spear, 1982; Hoinkes, 1986; Pigage and Greenwood, 1982), give unrealistically high values for temperatures above 600°C. The remaining nine calibrations have been averaged (1~ values of about 30°C), values deviating more than 1~ have been subtracted, and new averages have been calculated and rounded off. The results are shown in Table 6. The obtained values closely correspond to the recommended values of Fonarev et al. (1990b), who suggested that the averages of temperatures obtained by the Holdaway and Lee (1977) and Perchuk and Lavrent'eva (1983) calibrations should be used for T> 550-600°C.

The mean temperatures (Table 6) reveal a west and northwest increase in temperatures from 590-600°C in mid-amphibolite-facies through 640-690°C in upper amphibolite-facies and up to 730°C in granulite-facies terrains (Fig. 6). These northwest and westward increasing temperatures are consistent with changes in mineral assemblages that reveal an increasing grade of metamorphism in these directions. The temperatures do not necessarily represent peak metamorphic temperatures, particularly in the granulites, which show some mineral chemical and petrographic evidence for retrogression, such as Fe-rich rims in garnet, sillimanite and garnet rims around her¢yn i t ic spinel, recrystallised and kinked second generation sillimanite and green biotite (see also Fitzsimons



0 -) Q3 :

=/. "]

N o"

/

0 30 K ilornctrcs I t

• MAKUTI

LAKE

KARlaA

/ ~ ~.f-/" . . . . . . . . . " / ~ I " / j o . . . . . . . . . . . .

.~.~. . . . . . . . . . . . . . . . . .

÷ ~ ÷ ÷ ÷ 4. ÷ ÷

/ ~1, ÷ 4. ÷ ÷ 4. ÷ ÷ 4. ! : : : : : : : : : . .

£hl- 8t zone

I~rt zone

t - Staur zone

- ~ Urungve granites

134 Sample number

710 Tempera~re in o£

u-Staur zone

Sit zone

6ranuUtes

Figure 6. Metamorphic map of the Magondi mobile belt (slightly modif ied after Treloar, 1988) and the plotted garnet- biotite temperatures indicating that the temperatures increase to the west and northwest (see text for discussion). The average temperature for samples hm-O09 and hm-134 have been adapted from Munyanyiwa (1995).

and Harley, 1994). Furthermore, as pointed out above, the "peak" temperatures in granulite- facies metapel i tes were calculated from compositions of high Mg garnet cores and matrix biotites (cf. Indares and Martignole, 1985)• Temperatures calculated for pelitic granulites using some of the garnet rim compositions are lower (660-550°C). This reflects varying stages of re-equilibration of garnet-biotite Fe2*-Mg exchange during cooling.

GASP (garnet-aluminosilicate-quartz-plagioclase) In metapelites containing the four minerals in apparent textural equil ibrium (i.e. garnet,

sillimanite, plagioclase and quartz), pressures have been calculated with the GASP equilibrium:

Anorthite = grossular + sillimanite + quartz (4) 3CaAI2Si208 = Ca3AI2Si3012 + 2AI2SiO s +SiO 2,

which has been calibrated with thermodynamic (e.g• Ghent, 1976; Newton and Haselton, 1981, 1984) and experimental data (Koziol and Newton, 1988).

Accuracy of the GASP barometer heavily depends on the chosen garnet and plagioclase activity models and the ~V of the reaction, Dif ferences of several kbar are obtained



between the barometers of Ghent (1976), Newton and Haselton (1981, 1984), Ganguly and Saxena (1984), Aranovich and Podlesskii (1983, 1989) and Koziol and Newton (1988, 1989). The Kleemann and Reinhardt (1994) calibration is chosen here because it is the most recent and is based upon the Koziol and Newton (1988) thermodynamic data and the garnet solution model of Berman (1990). With this calibration method, uncertainties in the pressure estimates (see Table 6) are +0.5 kbar. Pressures calculated for three granulite samples from Nyaodza (hm-346, hm-384 and hm-396) are 5.2, 4.9 and 5.2 kbar, respectively. Pelitic g ranu l i te samples f rom Rukomech i g ive pressures of 6.0 and 7.2 kbar for samples hm- 75 and hm-73, respectively. A pressure of 5.6 kbar is obtained for an upper amphibol i te- facies sample (hm-003) northeast of Karoi. The calculated pressure for the upper amphibolite- fac ies sample is s imi lar to pressures in granulite-facies rocks in the Nyaodza area and only sl ightly lower than 7.2 kbar, an upper value calculated for a pel i t ic granul i te in Rukomechi (Table 6).

Garnet-cordierite-sillimanite-(__+ quartz) and cordierite-biotite The var ious express ions for the garnet - cordierite geothermometer (Thompson, 1976; Ho ldaway and Lee, 1977; Wel ls, 1979; Perchuk and Lavrent'eva, 1983; Ellis, 1986; Aranovich and Podlesskii, 1989; Holland and Powell, 1989; Bhattacharya et al., 1988) show only sl ight var iat ion. The geothermometer shows only slight pressure dependence; 1 kbar difference gives a temperature increase of about 5°C. Mean temperature estimates for the garnet-cordierite-bearing samples hm-310 and hm-384 are 740°C ( 1 ~ = 2 0 ° C , at 7 kbar) and 735°C (1~= 16°C, at 4.9 kbar), similar to garnet-biotite mean temperatures (Table 6). The Perchuk and Lavrent 'eva (1983) cordierite- b io t i te the rmomete r g ives a tempera ture estimate of 735°C for sample hm-384. Pressure may be evaluated with the garnet-cordierite- si l l imanite-quartz barometer, which may be used on the basis of the Fe and Mg end- members. The amount of water and/or CO 2 in cordier i te great ly in f luences the pressure e s t i m a t e s ; here a h igh w a t e r c o n t e n t (XH2o=0.6-0.8) has ben assumed, consistent wi th most natural cordier i tes (Deer et a l . , 1992). The Aranovich and Podlesskii (1989) calibration gives a pressure range of 6.5-7.5 kbar for sample hm-310.

Two-feldspar thermometry This thermometer is very easily susceptible to resett ing, giving meaningless results. Only sample hm-396 gives values which can be compared with the garnet-biotite temperature estimates. This sample yields temperatures of 685, 690, 694 and 689°C, using, respectively, the calibrations of Price (1985), Haselton et al. (1983) and Green and Usdansky (1986).

Metabasite thermometry The garnet-clinopyroxene-hornblende-plagioclase assemblage in sample hm-300 of fers the possibility to investigate the use of the garnet- cl inopyroxene, garnet-hornblende and clinopyroxene-hornblende thermometers, and the garnet-clinopyroxene-quartz barometer.

The calibrations of Wells (1979), Ellis and Green (1979), Dahl (1980) and Powell (1985) give, respectively, temperature estimates of 701, 695, 719, 668, 674 and 705°C, at 6 kbar. One kbar difference results in a maximum temperature increase of 10°C. The garnet- hornblende pair gives, according to Wells (1979), Graham and Powell (1984), Powell (1985) and Perchuk and Lavrent 'eva (1990), temperatures of 675, 682, 662 and 701°C, respect ively, independent of the pressure applied. The garnet-cl inopyroxene-hornblende ca l ib ra t ions of Nasir and Abu-A l ja rayesh (1992), using the combinat ion of Powel l 's g a r n e t - c l i n o p y r o x e n e and Graham and Powel l 's garnet-hornblende thermometers, yield temperatures of 723 and 675°C for the f i rst and second equat ions, respect ive ly . C o n s i d e r i n g the va lues of t hese th ree thermometers, an average temperature of about 700°C seems reasonable and is close to the garnet-b iot i te and garnet-cordier i te temperatures obtained in the granulite-facies terrains to the east.

The Kohn and Spear (1989, 1990) calibrations for the garnet-hornblende-quartz assemblage, based on four equations (Kohn and Spear, 1989) and two equations (Kohn and Spear, 1990), give a pressure estimate of 6.3 kbar (1G=0.7 kbar). Similar pressures (6.5 kbar) are given by the garnet-cl inopyroxene-quartz geobarometer of Moecher et al. (1988).

The Pb-Pb zircon ages of 1920-1960 Ma for two samples of the Kariba Gneiss unit (KrSner, p e r s . c o m m . , 1996) i nd i ca te tha t the metamorphism of this unit is synchronous with the Magond i Orogeny and' t hus the P - T condit ions calculated for the Kariba Gneiss sample are associated with this orogeny.



DISCUSSION

The results of geothermobaromtery presented above indicate temperatures of up to 730°C and pressures of 6 + 1 kbar for the granulite-facies metamorphism in the Magondi mobile belt. The calculated temperatures and pressures are similar to the range 710-745°C and 5-7 kbar found by Treloar and Kramers (1989) using a garnet- cordierite geothermometer and a garnet-cordierite- biotite-sillimanite-orthoclase assemblage. The temperatures also fall wi th in the range of 750+50°C calculated by Munyanyiwa et al. (1993) for the Magondi enderbites using Lee and Ganguly's (1988) garnet-orthopyroxene geothermometer. The calculated granulite-facies conditions also fall within the sillimanite stability field, which is consistent with the presence of sillimanite in both the Nyaodza and Rukomechi pelitic granulites. The pressure of 6 .0+ 1 kbar recorded in one sample from the upper amphibolite-facies terrain is similar to pressures calculated for the granulite-facies terrains, whereas temperature estimates for the upper amphibolite-facies rocks are lower (640-690°C), as expected. This implies that the amphibolite- granulite-facies transition is primarily isobaric. However, this conclusion is based on only one amphibolite-facies sample, making it somewhat tenuous. Further work in the future should attempt to document pressure ranges in this part of the Magondi Belt in more detail.

Tectonic implications Bohlen and Mezger (1989) and Bohlen (1991) compiled a set of published P-Tdata indicating that regional granulite-facies terrains were formed at relatively low pressures (6-8 kbar), corresponding to mid-crustal depths of around 20-30 kin. On the other hand, mafic granulite- facies xenoliths, which are common in volcanic pipes and kimberlites, were found to have crystallised at higher pressures of 10-15 kbar, reflecting an origin from deeper crustal levels (35-50 km) or from the upper mantle. They further pointed out that in regional metamorphic granulites, the amphiboli te-granuli te-facies transition is essentially isobaric and is controlled by increasing temperatures from amphibolite- to granulite-facies. This led them to conclude that the regional metamorphic granulites were formed by crustal thickening involving magmatic underp la t ing in an is land-arc or ac t ive- continental margin setting. This mechanism of granulite formation is similar to the hot-spot model, where mafic-ultramafic magmatic bodies pond in the lower crust and provide the

necessary heat for metamorphism (Newton, 1987).

In contrast, Carswell and O'Brien (1993) suggested that some regional granulite terrains were formed at much higher pressures in excess of 9 kbar, a maximum value preferred by Bohlen and Mezger (1989) for regional granulites. The higher pressures implied crustal thickening during format ion of the granul i tes, possibly by cont inent-cont inent collision (Carswell and O'Brien, 1993). Carswell and O'Brien (1993) based their argument on the granulites from the Moldanubian Massif in Lower Austria, which show evidence of initial equilibration at 16 kbar and 1000°C. Carswell and O'Brien (1993) therefore questioned the emphasis placed by Bohlen and Mezger (1989) and Bohlen (1991) on regional granulite formation related to magmatic underplating rather than tectonic activity involving continent-continent collision.

Data presented above for the Magondi granulites show that there is no evidence for high pressures ( > 1 0 kbar) characteristic of thickened continental crust formed during continent-continent collision. The P-T values for the Magondi mobile belt granulites fall within the range calculated for regional granulite terrains by Bohlen and Mezger (1989) and Bohlen (1991). Such granulites are interpreted to be from a region of high heat flow caused by magmatic heating. Evidence for magmatic heating in the Magondi mobile belt high-grade terrain may be prov ided by assoc ia ted enderb i tes and charnockites which, on the basis of field evidence and U-Pb and Pb-Pb zircon ages, have been interpreted as synmetamorphic intrusives (Munyanyiwa et al., 1995). Significantly, some of the enderbites contain CO 2 fluid inclusions that show textures indicating trapping at or near the peak of granulite-facies metamorphism (Munyanyiwa et al., 1993). The rocks also bear petrographic evidence interpreted to reflect a primary magmatic origin for charnocki t ic / enderbitic rocks (cf. Ridley, 1992). It is likely that the Magondi mobile belt enderbites crystallised directly from CO2-saturated silicic magma emplaced into the mid-crustal levels, as implied in the model presented by Frost and Frost (1987) and Frost et al. (1989). The Magondi mobile belt granulites were formed in a region of high heat f low, with heat possibly being provided by underlying, deep-seated mafic magmatic bodies. The carbonic fluids could also have been the mechanism of heat transfer for granulite-facies metamorphism (cf. Frost and Frost, 1987; Frost et al., 1989). Also, precise



U-Pb and Pb-Pb single zircon ages of 1.9-1.96 Ga for the enderbites (Munyanyiwa et al., 1995) show that they are part of a widespread magmatic and metamorphic event in central and sou thern Af r ica . A ca lc -a lka l ine a f f i n i t y character is t ic of magmat ic arcs has been established for some of these magmatic rocks, for instance the Bangweulu Block granites, granodiorites and their extrusive equivalents in northern Zambia (Andersen and Unrug, 1984). Regionally, the Magondi mobile belt granulites are probably part of an active continental margin located along the northwestern margin of the Zimbabwe Craton during the Palaeoproterozoic (cf. Stowe, 1989).

In closing, temperatures determined with the garnet-biotite geothermometers are -600°C for the mid-amphibolite-facies, 640-690°C for the upper amphibolite-facies and around 730°C for the granulite-facies terrains. In pelitic granulites, these tempera tu res are s imi lar to those ca lcu la ted f rom the g a r n e t - c o r d i e r i t e geothermometer and, in some cases, from the two- fe ldspa r geo the rmomete r . Pressures calculated with the GASP geobarometers yield about 6.0 kbar for the upper amphibolite-facies terrains. The granul i te-facies terrains were deve loped at 5-7 kbar, w h i c h impl ies metamorphism at mid-crustal levels at similar depths to those upper amphibo l i te - fac ies metapelites.

The high-grade pelitic rocks (upper amphibolite- and granulite-facies) were retrogressed to a minor degree. Retrograde minerals include second gene ra t i on s i l l iman i t e , ch lo r i t e , muscovite, biotite and calcite. The hydrous/ carbonate minerals indicate an influx of CO 2- H20 fluids either during cooling, following the peak of metamorphism, or during a separate re t rograde me tamorph i c event , poss ib ly associated with the Pan-African Orogeny.

ACKNOWLEDGEMENTS

The authors are grateful to T. G. Blenkinsop, P. Dirks, H. A. Jelsma and P. J. Treloar for their helpful discussions. Thanks are due to Prof. R. E. Hanson and Dr P. J. Treloar for their careful reviews and const ruc t ive cr i t ic ism. H. M. acknowledges the Univers i ty of Zimbabwe Research Board grant 3249 from which the field work was funded. Thanks are due to W. J. Lustenhouwer who assisted with the microprobe analyses. H. A. Jelsma assisted with part of the field work described herein. This work was carried out under the framework of the MINREST

program, a scientific link between the Geology Department, University of Zimbabwe, Harare and the Vrije Universiteit, Amsterdam. The MINREST program is f i nanc ia l l y suppor ted by the Directorate General for International Cooperation (DGIS), Min is t ry of Foreign Af fa i rs of the Netherlands, via the Netherlands University Foundat ion for In te rna t iona l Cooperat ion (NUFFIC), which also sponsored the visit of H. M. to Amsterdam. This is a contribution to IGCP project 363 "Lower Proterozoic of the Sub- Equatorial Africa". Editor ial Handl ing - G. J. H. Ofiver

REFERENCES Andersen, L. S. and Unrug, R. 1984. Geodynamic evolution

of the Bangweulu Block, northern Zambia. Precambnan Research 25, 187-212.

Aranovich, L. Y. and Podlesskii, K. K. 1983. The cordierite- garnet-sil l imanite-quartz equilibrium: experiments and applications. In: Kinetics and equilibriumn mineral reactions (Edited by Saxena, S. K.) pp173-198. Springer-Verlag, Berlin.

Aranovich, L. Y. and Podlesskii, K. K. 1989. Geothermo- barometry of high grade metapelites: simultaneously operating reactions. In: Evolution of metamorphic belts (Edited by Daly, J. S., Cliff, R. S. and Yardley, B. W. D.) Geological Society Special Publication 43, 45-61.

Berman, R. G. 1990. Mixing properties of Ca-Mg-Fe-Mn garnets. American Mineralogist 75, 328-344.

Bhattacbarya, A., Mazmudar, A. C. and Sen, S. K. 1988. Fe-Mg mixing in cordierite: constraints from natural data and implications for cordierite-garnet geothermometry in granulites. American Mineralogist 73, 338-334.

Bhattacharya, A., Mohanty, L., Mail, A., Sen, S. K. and Raith. M. 1992. Non-ideal mixing in the phlogopite-annite binary: constraints from experimental data on Mg-Fe partitioning and reformulation of the garnet-biotite geotbermometer. Contributions Mineralogy Petrology 111, 87-93.

Bohlen, S. R. 1991. On the formation of granulites. Journal Metamorphic Geology 9, 223-229.

Bohlen, S. R. and Essene, E. J. 1977. Feldspar and oxide thermometry of granulites in the Adrinodack Highlands. Contribuitions Mineralogy Petrology 62, 951-964.

Bohlen, S. R. and Mezger, K. 1989. Origin of granulites and the formation of the lowermost continental crust. Science 244, 326-329.

Bohlen, S. R., Dollase, W. A. and Wall, V. W. 1986. Calibration and Applications of Spinel Equilibria in the System FeO-AI203-SiO 2. JournalPetrology 27, 1143-1156.

Broderick, T. J. 1976. The Explanation of the Geological Map of the Country East of Kariba. Short Report, 43, 98p. Geological Survey, Rhodesia.

Carswell, D. A. and O'Brien, P. J. 1993. Thermobarometry and Geotectonic Significance of High-Pressure Granulites: Examples from the Moldanubian Zone of the Bohemian Massif in Lower Austria. Journal Petrology 34, 427-459.

Chipera, S. J. and Perkins, D. 1988. Evaluation of garnet biotite geothermometers: application to the English River subprovince, Ontario. Contributions Mineralogy Petrology 98, 40-48.

Dahl, P. S. 1980. The thermal compo§itional dependence of Fe 2÷ -Mg distributions between coexisting garnet and pyroxene: applications to geothermometry. American Mineralogist 65, 852-866.



Dasgupta, S., Sengupta, P., Guha, D. and Fukuoka, M. 1991. A refined garnet-biotite Fe-Mg exchange geothermometer and its applications in amphibolites and granulites. Contributions Mineralogy Petrology 109, 130-137.

Deer, W. A., Howie, R. A. and Zussman, J. 1982. Rock- Forming Minerals. Vol., IA: Orthosilicates. 2nd edition. 919p. Longman, London.

Deer, W. A., Howie, R. A. and Zussman, J. 1992. Rock- Forming Minerals. 2rid edition. 696p. Longman, London.

Ellis, D. J. 1986. Garnet-liquid Fe2÷-Mg equilibria and implications for the beginning of melting in the crust and subduction zones. American Journal Science 286, 765- 791.

Ellis, D. J. and Green, D. H. 1979. An exprimental study of the effect of Ca upon garnet-cl inopyroxene Fe-Mg exchange equilibria. Contributions Mineralogy Petrology 71, 13-22.

Essene, E. J. 1989. The current status of thermobarometry in metamorphic rocks. In: Evolution of metamorphic belts (Edited by Daly, J. S., Cliff, R. A. and Yardley, B. W. D.) Geological Society Special Publication 43, 1-44.

Ferry, J. M. and Spear, F. S. 1978. Experimental calibration of the partitioning of Fe and Mg between biotite and garnet. Contributions Mineralogy Petrology 66, 113-117.

Fey, P. and Broderick, T. J. 1990. Explanation of the Geological Map of the Country East of Makuti, Hurungwe District. Short Report, 47, 84p. Geological Survey, Zimbabwe.

Fitzsimons, I. C. W. and Harley, S. L. 1994. The influence of retrograde cation exchange on granulite P-T estimates and the convergence technique for the recovery of peak metamorphic conditions. Journal Petrology 35, 543-576.

Fonarev, V. I., Grapchikov, A. A. and Konilov, A. N. 1990a. A consistent system of geothermometers for metamorphic complexes. International GeologicalReview 33, 743-783.

Fonarev, V. I., Konilov, A. N. and Grapchikov, A. A. 1990b. Geothermometry of metamorphic complexes from central- Kola Archaean granulite-gneiss region. Experiments, USSR Academy Sciences 89, 61-65.

Frost, B. R. and Frost, C. D. 1987. CO 2, melts, and granulite metamorphism. Nature 347, 503-506.

Frost, B. R., Frost, C. D. and Touret, J. R. L. 1989. Magma as a source of heat and f luids in granulite facies metamorphism. In: Fluid Movement, Element Transport and the Composition of the Deep Crust (Edited by Bridgwater, D.) pp1-18. NA TO ASI Series C281, Kluwer, Dordrecht.

Ganguly, J. and Saxena, S. 1984. Mixing properties of aluminosilicate garnets: constraints from natural and exper imenta l data and app l ica t ions to geothermobarometry. American Mineralogist 69, 88-97.

Ghent, E. D. 1976. Plagioclase-garnet-AI2SiOs-quartz: a potential geobarometer-geothermometer . American Mineralogist 61, 710-714.

Graham, C. M. and Powell, R. 1984. A garnet-hornblende geothermometer: calibration, testing, and application to the Pelona Schist , Southern Cal i forn ia. Journal Metamorphic Geology 2, 13-31.

Green, N. L. and Usdansky, S. I. 1986. Ternary-feldspar mixing relat ions and thermobaromet ry . American Mineralogist 71, 1100-1108.

Guidotti, C. V. 1984. Micas in metamorphic rocks. In: Micas (Edited by Bailey, S. W.) Reviews Mineralogy 13, 357-467.

Guidotti, C. V., Cheney, J. T. and Henry, D. J. 1988. Compositional variation of biotites as a function of metamorphic reactions and mineral assemblage in the pelitic schists of western Maine. American Journal Science 288A, 270-292.

Harley, S. L. 1989. The origin of granulites: a metamorphic perspective. Geological Magazine 126, 215-247.

Hartnady, C., Joubert, P. and Stowe, C. 1985. Proterozoic Crustal Evolution in Southwestern Africa. Episodes 8, 236-244.

Haselton, H. T., Hovis, G. L., Hemmingway, B. S. and Robie, R. A. 1983. Calorimetric investigation of the excess entropy of mixing of albite-sanidine solid-solutions: Lack of evidence for Na, K short range order and implications for two-feldspar thermometry. American Mineralogist 68, 398-413.

Hodges, K. V. and Spear, F. S. 1982. Geothermometry, geobarometry and the AI2SiO S tr ipple point at Mt. Moosilauke, New Hampshire. American Mineralogist 67, 393-399.

Hoinkes, G. 1986. Effect of grossular-content in garnet on the partitioning of Fe amd Mg between garnet and biotite. Contributions Mineralogy Petrology 92, 393-399.

Holdaway, M. J. and Lee, S. M. 1977. Fe-Mg cordierite stability in high-grade pelitic rocks based on experimental, theoretical and natural observations. Contributions Mineralogy Petrology 63, 175-198.

Holland, T. J. B. and Powell, R. 1989. An enlarged and updated internally consistent thermodynamic data set with uncertainties and correlations; the system K20-Na20-CaO- MgO-MnO-FeO-Fe203-AI203-TiO2-SiO2-C-H2-O2. Journal Metamorphic Geology 8, 89-124.

Indares, A. and Mart ignole, J. 1985. Biot i te-garnet geothermometry in the granulite facies rocks: the influence of Ti and AI in biotite. American Mineralogist 70, 272-278.

Kleemann, U. and Reinhardt, J. 1994. garnet-biot i te thermometry revisited. The effects of AI vj and Ti in biotite. European Journal Mineralogy 6, 925-941

Kohn, M. J. and Spear, F. S. 1989. Emprical calibration of geobarometers for the assemblage garnet + hornblende + plagioclase + quartz. American Mineralogist 74, 77-84.

Kohn, M. J. and Spear, F. S. 1990. Two new geobarometers for garnet amphibolites, with applications to southeastern Vermont. American Mineralogist 75, 89-96.

Koziol, A. M. and Newton, R. C. 1988. Redetermination of the plagioclase-garnet-AI2SiOs-quartz geobarometer. American Mineralogist 73, 216- 223.

Koziol, A. M. and Newton, R. C. 1989. Grossular activity- composition relationships in ternary garnets determined by reversed d isp laced-equ i l ib r ium exper iments . Contributions Mineralogy Petrology 103, 423-433.

Lee, H. Y. and Ganguly, J. 1988. Equilibrium composition of coexist ing garnet and orthopyroxene: reversed experimental determination in the system FeO-MgO-AI203- SiO 2 and applications. Journal Petrology 29, 93-114.

Leyshon, P. R. and Tennick, F. P. 1988. The Proterozoic Magondi Mobile Belt in Zimbabwe-a review. South A frican Journal Geology 91, 114-131.

Loney, P. E. 1969. The geology of Kariba District, Rhodesia, with special reference to geochemistry and amphibolite petrochemistry. Ph. D, dissertation (unpubL) 110p. Leeds University, Leeds.

Master, S. 1991. Strat igraphy, tectonic sett ing and mineralization of the early Proterozoic Magondi Supergroup, Zimbabwe: a review. Information Circular 238, 75p. University of the Witwatersrand, Economic Geology Research Unit, Johannesburg.

Moecher, D. P., Essene, E. J. and Anovitz, L. M. 1988. Calculation and application of clinopyroxene-garnet- p lag ioc lase-quar tz geobarometers . Contributions Mineralogy Petrology 100, 92-106.

Mohr, D. W. and Newton, R. C. 1983. Kyanite-staurolite metamorphism in sulphidic schists of the Anakeesta Formation, Great Smoky Mountains, North Carolina. American Journal Science 283, 97-134.



Munyanyiwa, H. 1995. The metamorphic and structural geology and geochronology of the Magondi and Zambezi mobile belts, northwestern Zimbabwe. D. Phil. dissertation 416p. University of Zimbabwe, Harare, Zimbabwe.

Munyanyiwa, H., Touret, J. R. L. and J~lsma, H. A. 1993. Thermobarometry and fluid evolution of enderbites within the Magondi mobile belt, northern Zimbabwe. Lithos 29, 163-176.

Munyanyiwa, H., KrSner, A. and Jeackel, P. 1995. U-Pb and Pb-Pb single zircon ages for charno-enderbites from the Magondi mobile belt, northwest Zimbabwe. South African Journal Geology 98, 52-57.

Munyanyiwa, H. Hanson, R. E., Blenkinsop, T. G. and Treloar, P. J. 1997. Geochemis t r y of amph ibo l i t es and quartzofeldspathic gneisses in the Pan-African Zambezi belt, no r thwes t Z imbabwe: ev idence for b imodal magmatism in a continental rift setting. Precambrian Research 81, 179-196.

Nasir, S. and Abu-Aljarayesh, I. 1992. Emprical calibration of a geothermometer for garnet amphibolites and granulites based on the assemblage garnet-clinopyroxene-hornblende. Chemie Erde 52, 205-210.

Nesbitt, B. E. and Essene, E. J. 1982. Metamorphic thermometry and barometry of a portion of the Southern Blue Ridge Province. American Journal Science 282, 701 729.

Newton, R. C. 1987. Petrological aspects of Precambrian granulite facies terrains bearing on their origin. In: Proterozoic lithospheric evolution (Edited by Kr6ner, A.) American Geophysical Union, Geodynamics Series 17, 11-26.

Newton, R. C. and Haselton, H. T. 1981. Thermodynamics of the garnet-plagioclase-AI2SiOs-geobarometer. In: Advances in physical geochemistry (Edited by Newton, R. C., Navrotsky, A. and Wood, B. J.) pp131-147. Springer-Verlag, New York.

Newton, R. C. and Perkins, D. 1982. Thermodynamic calibration of geobarometers based on the assemblages garnet-plagioclase-orthopyroxene (clinopyroxene)-quartz. American Mineralogist 67, 203-222.

Owen, J. V. 1991. Cordierites + spinel parageneses in pelitic gneiss from contact aureoles of the Mistastin batholith (Quebec) and Taylor Brook gabbro complex (Newfoundland). Canadian Journal Earth Sciences 28, 372-381.

Perchuk, L. L. and Aranovich, L., Ya. 1986. Improvement of the biotite-garnet thermometer: correction for fluorine content in biotite. Doklady Akademii Nauk USSR 277, 131-135.

Perchuk, L. L. and Lavrent'eva, I. V. 1983. Experimental invest igat ion of exchange equil ibria in the system cord ier i te-garnet -b io t i te . In: Advances in phys ica l geochemistry (Edited by Newton, R. C. and Wood, B. J.) pp199-239. Springer-Verlag, New York.

Perchuk, L. L. and Lavrent 'eva, I. V. 1990. Garnet- orthopyroxene and garnet-amphibole geothermobarometry. International Geological Review 32, 486-507.

Pigage, L. C. and Greenwood, H. J. 1982. Internally consistent estimates of pressure and temperature: the staurolite problem. American Journal Science 282, 943- 969.

Powell, R. 1985. Regression diagnostic and robust regression in g e o t h e r m o m e t e r / g e o b a r o m e t e r : the garnet c l i nopyroxene geo the rmomete r rev is i ted. Journa l Metamorphic Geology 3, 231-243.

Price, J. G. 1985. Ideal site mixing in solid solution, with an application to two-feldspar geothermometry. American Mineralogist 70, 696-701.

Ridley, J. 1992. On the origins and tectonic significance of the charnockite suite of the Archaean Limpopo Belt, North Marginal Zone, Zimbabwe. Precambrian Research 55, 407-427.

Stagman, J. G. 1962. The geology of the Southern Urungwe District, Southern Rhodesia. Southern Rhodesia Geological Survey Bulletin 52, 82p.

Stagman, J. G. 1978. An outline of the geology of Rhodesia. Rhodesia Geological Survey Bulletin 80, 126p.

Stowe, C. W. 1989. Discussion on "The Proterozoic Magondi Mobile Belt in Zimabwe-a review'. South African Journal Geology 92, 69-71.

Thompson, A. B. 1976. Mineral reactions in pelitic rocks: I1. Calculations of some P-T-X (Fe-Mg) phase relations. American Journal Science 276, 425-454.

Treloar, P. J. 1988. The geological evolution of the Magondi Mobile Belt, Zimbabwe. Precambrian Research 38, 55 73.

Treloar, P. J. and Kramers, J. D. 1989. Metamorphism and geochronology of granulites and migmatitic granulites from the Magondi Mobile Belt, Zimbabwe. Precambrian Research 45, 277-289.

Wells, P. R. A. 1979. Chemical and thermal evolution of Archaean sialic crust, southern West Greenland. Journal Petrology 20, 187-226.

Yardley, 8. W. D. 1981. A note on the compostion and stability of the Fe-staurolite. Neues Jahrbuch Mineralogie Monatschafte 81, 127-132.


1-s2.0-S089953629800058

Documents