Pre-mineralization thermal evolution of the Palaeoproterozoic gold-rich Ashanti belt, Ghana

Over the past decade, the increasing volume ofhigh-precision geochronological data hasdemonstrated that gold deposition was not acontinuous process through geological time. Thepattern for orogenic gold deposits in thePrecambrian shows distinct peaks separated byseveral hundreds of millions of years (Goldfarbet al. 2001). Two main periods (at 2800–2500 and2100–1800 Ma) correlate well with episodes ofincreased continental growth. Geochrono-logical data on juvenile continental crust havebeen compiled and revised by Condie (2000),who suggested that the three major crustalformation episodes (estimated at 2700, 1900 and1200 Ma) were related to ‘superevents’ in themantle. For the Palaeoproterozoic peak(1900 Ma), a superplume event has beenproposed (e.g. Condie et al. 2000). Similarly, theassociated periods of orogenic gold formationare ‘most commonly explained by major mantleoverturning in the hotter early Earth, withassociated plumes causing extreme heating atthe base of the crust’ (Goldfarb et al. 2001).Indeed, it is now generally agreed that globalgeodynamics played a key role in the formationof several classes of ore deposits (e.g. Barley etal. 1998; de Boorder et al. 1998; Isley & Abbott1999; Pirajno 2001). If the occurrence of super-events in the mantle, the formation of super-

continents, and the genesis of world-classorogenic gold deposits (or ‘superdeposits’) areall related, then studies on the possible thermalconditions prevailing at the time of ore depositformation could provide useful results.

Deep thermal processes beneath continentsare well constrained below thermally stableareas for present times (e.g. Jaupart &Mareschal 1999; Lenardic et al. 2000). Heat flowstudies in the Canadian and Fennoscandianshields (e.g. Kukkonen 1998; Rolandone et al.2002) as well as petrological studies on mantlexenoliths (Kukkonen & Peltonen 1999; Russell& Kopylova 1999) have shown that mantle heatflow values at the base of the Precambrian crustrange from 10 mW m–2 to 16 mW m–2. However,these present estimates cannot be applied toPrecambrian times, and the possibility ofstudying thermal regimes billions of years agothus becomes a real challenge. A number ofdistinct arguments suggest that the thermalregime of the Archaean continental crust wasnot very different from that of the present-daycrust (e.g. Richter 1985). If this were indeed thecase, one would expect similar, low mantle heatflow values at the base of continents, unlesssome large-scale thermal perturbation, such asa superplume event, affected the base of thecrust. The formation of world class deposits

Pre-mineralization thermal evolution of the Palaeoproterozoicgold-rich Ashanti belt, Ghana

V. HARCOUËT1,2, L. GUILLOU-FROTTIER1, A. BONNEVILLE2 &J. L. FEYBESSE1

1Bureau de Recherches Géologiques et Minières, Orléans, France(e-mail: [email protected])

2Institut de Physique du Globe, Paris, France

Abstract: The region of the gold-rich Ashanti belt in southern Ghana was chosen as thesubject for a detailed regional thermal modelling study. Geological studies, in addition tolaboratory measurements of thermal properties and heat-production rates, allow us toconstrain a finite-element thermal modelling. Scenarios integrating variations of the struc-ture of the crust and various chronological settings were examined. We calculated thethermal regime before and after the thrust tectonism that affected the region during theEburnean orogeny (2130–2095 Ma), just before ore deposit formation. This gives a newinsight into the regional thermal state of the crust before the mineralizing events. To satisfythe thermobarometric observations, the most probable mantle heat flow must be60 mW m–2, which is at least three times greater than the present-day value. At shallowdepths, our results also indicate anomalies of lateral heat flow reaching 25 mW m–2, focusedon the margins of each lithological unit, including the Ashanti belt. These anomalies arerelated to the distortion of the isotherms in the first few kilometres that can be explainedmostly by lateral contrasts in thermal conductivity. Such anomalies could be of importancefor the mineralizing events, as they would favour fluid circulation locally.

From: MCDONALD, I., BOYCE, A. J., BUTLER, I. B., HERRINGTON, R. J. & POLYA, D. A. (eds) 2005. MineralDeposits and Earth Evolution. Geological Society, London, Special Publications, 248, 103–118. 0305-8719/$15.00© The Geological Society of London 2005.

05SP248_Harcouet (COL) (JB-S) 1/8/05 1:39 pm 03

probably requires anomalous thermal processes,located at shallow depth or at the crust–mantleinterface. The use of various geological data, andin particular, thermobarometric data, shouldhelp to quantify thermal regimes of the past.

In this study, we start from the apparentrelationship between Palaeoproterozoicorogenic gold deposits and the deep thermalprocesses prevailing at the time of their forma-tion. Since numerous geological data have beenacquired in mineralized areas, the modelling ofpast thermal regimes before and during miner-alizing events may help to constrain thermalconditions at depth.

Among the major orogenic gold deposits, theAshanti belt of Ghana possesses severalfeatures that are particularly well-suited tostudying deep thermal processes 2000 Ma ago.First, the Ashanti belt of Ghana was formedduring the ‘Palaeoproterozoic event’ of Condie(2000). Secondly, many thermobarometric dataare available and can provide strong constraintsto any thermal model. Moreover, numerousrock samples collected during several fieldcampaigns in the area (Milesi et al. 1991, 1992;Feybesse & Milesi 1994) can be used to measurethermal properties of the main lithological units,thus reducing the number of unknownparameters in the models. Thirdly, the volcanicand sedimentary belts in Ghana are sufficientlyparallel (over a distance reaching ten times thecrustal thickness) to consider the area as a two-dimensional structure, thus justifying the 2Dapproach to the modelling.

This study presents the preliminary results ofthermal modelling of the Ashanti and surround-ing belts, before and during the collisionaltectonics leading to the pre-mineralizing stage.After defining a geological model according tothe available thermobarometric and geochrono-logical data, various hypotheses on crustalstructure, erosion rates and sediment thicknesswere tested, corresponding to different geo-logical scenarios. As the model depends stronglyon thermal properties, measurements of thermalconductivities and heat-production rates wereperformed. In all cases, thermobarometric dataare the main constraints. As a consequence, ourresults argue for a suggested geological scheme,provide quantitative constraints on the deepthermal processes prevailing immediatelybefore this particular Palaeoproterozoic miner-alizing event, and suggest some shallow thermalprocesses that may lead to particular fluid–rockinteractions. This latter consequence is thesubject of an ongoing study.

Geological setting

Main stages of crustal evolution

The Palaeoproterozoic rocks covering a largeparts of SW Ghana (Fig. 1) belong to theBirimian Supergroup and the Tarkwaian Group.The Birimian Supergroup includes sedimentarybasins separated by subparallel NE-trendingvolcanic belts consisting of basalts and someinterflow sediments. The Tarkwaian conglomer-ates overlie the Birimian rocks and wereprobably derived from the erosion of theserocks.

A detailed chronology for the emplacementof these different groups is proposed. We firstconsider a phase of magmatic accretion between2250 and 2170 Ma, corresponding to a period ofvolcanoplutonism, leading to the emplacementof the greenstone belts constituting the so-calledvolcanic belts. The addition of extensive monzo-granitic plutonism, which occurred between2160 and 2150 Ma (Opare-Addo et al. 1993b;Feybesse et al. 2000), led to the formation of ajuvenile continental crust. After this crust-formation stage, basin opening and sedimenta-tion took place within the Birim, Afema andComoé basins between 2150 and 2100 Ma.

The emplacement of the Birimian rocks wasfollowed by a period of tectonic accretion, theEburnean orogeny, for which two main phasesof deformation (D1 and D2) can be described(Fig. 2). The main deformation is associatedwith the initial D1 thrust tectonism that tookplace between 2130 and 2105 Ma. This compres-sive deformation along a SE–NW directionresults in crustal thickening by stacking of thedifferent units. During this compressive stage,the Birimian rocks were deformed and subse-quently uplifted and eroded. The resulting Tark-waian detrital rocks were deposited between2133 ± 4 and 2097 ± 2 Ma (Oberthür et al. 1998;Pigois et al. 2003). They were not affected by D1deformation and they were probably depositedmainly at the end of D1 tectonism (Milesi et al.1991; Milesi pers. comm.). Regional metamor-phism associated with D1 ranges from lowergreenschist facies to lower amphibolite facies(Kleinschrot et al. 1993; John et al. 1999; Klemdet al. 2002; Pigois et al. 2003) and is dated ataround 2100 Ma (Oberthür et al. 1998). Asecond tectonic phase, D2, affecting both theTarkwaian and Birimian rocks, began at2095 Ma. The major epigenetic Birimian lode-gold event can be placed late in the Eburneanorogeny, after the peak of metamorphism wasreached. The age of hydrothermal alteration hasbeen determined by Oberthür et al. (1998), who

104 V. HARCOUËT ET AL.

05SP248_Harcouet (COL) (JB-S) 1/8/05 1:39 pm Page 104

THERMAL EVOLUTION OF THE ASHANTI AU BELT 105

Fig. 1. Simplified geological map of south Ghana, West Africa. Cross section A–A� is shown in Figure 3. Alsoshown are the locations of the major gold mines 1–14. Thermobarometric data used in this study are encircled(see Table 1).

Fig. 2. Succession of the different lithological units (top) and the main tectonic phases (bottom). Thissimplified scheme allows the reconstruction of a geological scenario prior to the mineralizing event (after theD1 phase). Time intervals are positioned according to geochronological data from several studies (see text forreferences).


give ages of 2092 ± 3 and 2086 ± 4 Ma. The mostrobust age for mineralization, determined at 2063± 9 Ma, was given by Pigois et al. (2003). Mostof the gold deposits are concentrated along thewestern flank of the Ashanti belt. They arelocated in the vicinity of major fault zones, alongNE-striking shear zones, at the contact betweenthe Birimian sedimentary basins, Tarkwaianconglomerates and volcanic belts (Fig. 1).

The Ashanti belt

The Ashanti belt in south Ghana hosts thelargest gold deposits known in West Africa. Itcontains potential resources of c. 5000 tons ofgold (Milesi et al. 1991; Pigois et al. 2003). Themineralization is temporally related to the latestage of the Eburnean orogeny, which tookplace between 2130 and 1980 Ma and affectedthe supracrustal rocks. It has the characteristicsof synorogenic, mesothermal gold deposits andoccurs mainly along shear zones on the westernmargin of the belt. Figure 2 shows the mainsuccessive events that affected the Ashanti belt.

There is still no general agreement concern-ing the nature of the basement in Ghana. Someauthors have proposed the presence of anArchaean basement, at least beneath the Birimbasin and SE Ghana (Milesi pers. comm.), butthe type of basement underlying the Ashantibelt is still controversial. The Ashanti belt mayrepresent the boundary between a continentaldomain and the western units possibly being ofoceanic type. Such hypotheses, as well as distinctscenarios for crustal evolution, could have beentested if there had been sufficient fieldconstraints. However, in this study we assume acontinental basement below the Ashanti beltand the western units.

Field constraints and thermal properties

Thermobarometric constraints

Pre-Eburnean. The pressure estimates from themigmatites yield pressures in excess of 5 kbar attemperatures between 600 and 1000 °C, corre-sponding to emplacement at a relatively deepcrustal level (Opare-Addo et al. 1993a). Theanatexis, indicated by the presence of deformedneosomes in these migmatites, began just before2130 Ma (pre-D1 stage) at a deep crustal level(c. 6 kbar). At this depth, the temperatureconditions required for partial melting to occurfall between 650–800 °C, depending on watercontent. In the following, a value of 700 °C isadopted.

Syn-D1 Eburnean tectonism. The minimumpeak-metamorphism conditions reported forthe Eburnean orogeny are 500–610 °C atc. 4.5–6 kbar (Table 1) in the southern Kibi.They were measured from samples ofmetapelites and amphibolites (Klemd et al.2002). Pigois et al. (2003) suggest peak meta-morphism conditions of 550 °C at minimumpressures of 5 kbar in the southern part of theAshanti belt, which is consistent with previousestimates of 490–650 °C at 4–6 kbar (John et al.1999; Klemd et al. 2002). These thermobaromet-ric data are shown in Table 1 and were takeninto account during construction of the geolog-ical model (section 4).

Thermal properties

In order to introduce petrophysical constraintsinto the model, measurements of thermalconductivity and heat-production rates weremade on representative rock samples collected


Table 1. Thermobarometric data (pressure and temperature estimates) for peak-metamorphism conditions atvarious locations in the Ashanti area. The spatial distribution of these estimates is shown on Figure 1

Location Rock type Pressure Temperature Reference(kbar) (°C)

Nsuta Mine Greenstone belts and <5 500 Kleinschrot et al. (1993)metasediments

Southern Ashanti Belt Amphibolites and 5–6 500–650 John et al. (1999)metagranitoid rocks

Southern Kibi–Winneba belt Metapelites and 4.5–6 500–610 Klemd et al. (2002)amphibolites

Southern Ashanti belt and Birimian rocks 4–6 490–580 Klemd et al. (2002)NW Sefwi belt

Damang Argillite <5 550 Pigois et al. (2003)


in Ghana (greenstones, monzogranites, sedi-ments and Tarkwaian conglomerates) andGuinea (basement and sediments).

Thermal conductivity. For the thermal conduc-tivity measurements, we used a QTM (QuickThermal conductivity Meter) based on amodified hot wire method and commercializedby Showa Denko K.K. The measurement prin-ciple is the following (Tavman & Tavman 1999):a thin wire fixed to an insulating material ofknown thermal properties receives a constantelectrical current. This process generates aconstant heat (Q) per unit length and per unittime. According to Galson et al. (1987), specificexperimental conditions yield measurements asprecise as those obtained by the divided-barapparatus method. These authors provide theexperimental procedure required to use theQTM method. In our study, optimal conditionsfor the number of measurements on eachsample were respected with at least six measure-ments per sample. In most cases, the size of therock sample exceeded the minimum require-ments (20 � 50 � 70 mm). With this procedure,the obtained reproducibility is ± 5% and uncer-tainty remains lower than ± 10–20% (for drysamples).

By supplying a constant power to the heaterelement, a temperature rise (∆T) of the wire isgenerated. The value of ∆T is then measured bya thermocouple as a function of time during ashort heating interval. This temperature versustime record is used to calculate the thermalconductivity (k) according to the followingequation (Carlsaw & Jaeger 1984):

k = F – H, (1)

where T1 and T2 are temperatures at time t1 andt2 respectively, Q is the heat flow per unit timeand per unit length of the heating wire, and Fand H are specific constants of the probe, deter-mined by calibration.

Measurements were made on 31 rock samplesrepresentative of the main lithological units(Table 2). The measured thermal conductivitiesfor greenstones and monzogranites fall withinthe range of classical values for these rock types.The apparently high thermal conductivity of theBirimian sediments and Tarkwaian conglomer-ates is a result of metamorphism and burial, ascompaction is the dominant controlling factor inthe variation of thermal conductivity in the firstfew kilometres (e.g. Clauser & Huenges 1995).However, since the basins were still opening anderosion and conglomerate deposition were stilloccurring during D1, we had to use a thermalconductivity value of sediments and conglomer-ates before metamorphism and burial. Inaccordance with the thermal conductivity valuesproposed by Jaupart & Provost (1985), anidentical value of 1.8 W m–1 K–1 was taken forboth lithologies during their deposition. Aftercompaction due to burial, the measured thermalconductivity of the sediments was adopted,namely 2.9 W m–1 K–1 (see a similar case inSerban et al. 2001). For the basement and themantle, mean values from the literature(Turcotte & Schubert 2002) were taken at2.6 W m–1 K–1 and 3.3 W m–1 K–1, respectively.

Heat-production rate. The concentration inheat-producing elements is also an importantfactor controlling the geotherm. Heat-produc-tion rates were therefore calculated for eachunit, using the concentrations in the radiogenicelements U, Th and K (Table 3). U and Thconcentrations were determined by ICP-MS,and K concentration by X-ray fluorescencespectrometry.

Since the emplacement of the Ashanti belttook place at around 2100 Ma, heat-productionrates for this period had to be calculated usingpresent-day concentrations of radiogenicelements. Mean production rates in the past canbe calculated using the present concentrationsin radiogenic elements with the followingequation (Turcotte & Schubert 2002):

Q t tT Tln( / )2 1

2 1−


Table 2. Thermal conductivity measurements of Birimian and Tarkwaian rock samples from the Ashanti beltand Guinea. For each sample, six measurements were made and thermal conductivity is the mean value of all thesamples

Rock type Number of samples Thermal conductivity Standard deviation(W m–1 K–1) (W m–1 K–1)

Greenstones 7 2.93 0.74Monzogranites 6 3.10 0.59Sediments 12 2.89 0.90Conglomerates 6 3.12 0.62(Tarkwaian)


A(t) = ρ{(9.3919 10–5 et�0.155

+ 0.4399 10–5 et�0.9846)[U0]+ (2.64 10–5 et�0.04951)[Th0]+ (3.475 10–5 et�0.5545) [K0]}

(2)

where A(t) is the heat-production rate at timet (Ga), for a material of density ρ.[U0], [Th0] and[K0] are the present concentrations of the differ-ent isotopes measured in the laboratory. Table 3shows 104 present-day heat-production rates,measured on rock samples representative of thearea and their corresponding values 2100 Maago.

Geological model for crustal evolutionfrom the initial to the pre-D2 stages

Structural constraints

The detailed structure and composition of themodelled area is described below. Figure 3 illus-trates a geological cross-section inferred frommore than 15 years of various field campaigns inthe area. Figure 4 summarizes different stagesof crustal evolution from the initial to the pre-D2 stages. In order to build a realistic geologi-cal model, we first had to constrain the thicknessand lateral extent of the different rock units.


Table 3. Heat-production rates and their standard deviations (s.d.) for Birimian and Tarkwaian rock samplesfrom the Ashanti belt, at present (measurements) and 2100 Ma ago (see equation (2))

Rock type Number U K Th Heat-productionof samples (ppm) (%) (ppm) rate µW m–3

s.d. s.d. s.d. present s.d. t = 2100 Ma s.d.

Basement 48 0.98 1.03 2.00 1.29 8.47 9.60 1.05 0.90 1.71 1.28Greenstones 12 0.81 0.64 0.60 0.42 1.76 1.49 0.40 0.28 0.69 0.45Monzogranites 9 1.83 1.26 2.25 1.19 4.87 2.44 1.04 0.44 1.88 0.82Sediments 25 1.71 1.38 1.96 1.44 4.32 2.17 0.77 0.43 1.38 0.75Tarkwaian conglomerates 10 2.79 3.42 1.64 1.49 6.78 4.52 1.12 0.92 1.85 1.62

Fig. 3. Geological cross-section through A–A� (Fig. 1) as deduced from field campaigns. The NW direction ofthrusting events is outlined.

Fig. 4. Model for the Ashanti and surrounding belts, between 2130 and 2105 Ma. The initial stage (a) accountsfor thermobarometric data in the monzogranites. At 2105 Ma (b), the central unit has been overthrust by thosefrom the east, implying the burial of sediments and metamorphism in the greenschist facies, while the Ashantibelt was recovered by sediments and barren conglomerates (Kawere). The third stage (c) is the result oferosion and deposition.


Lateral extent of the rock units. In the following,lateral extent refers to a NW–SE direction, sincea 2D model is considered. The volcanic belts aretypically 15–40 km wide and 60–90 km apart(Taylor et al. 1992). From field observations andprevious maps (Klemd et al. 2002; Pigois et al.2003), the width of the Ashanti belt ranges from15–50 km, and is surrounded to the west by theAfema basin (c. 70 km wide) and to the east bythe Birim basin (c. 20–60 km wide), the Kibi belt(c. 40 km) and the monzogranite unit (c. 40 kmwide). We chose a mean lateral extent of 40 kmfor all the different rock units, in order tosimplify the calculations and to take intoaccount a certain amount of crustal shorteningthat occurred during the Eburnean orogeny.

Thickness of the lithological units. The geophys-ical constraints on the region are poor and wedid not have direct access to the data related tocrustal thickness, for which a value of 30 km wasassigned. Thermobarometric data were used toestimate the thickness of the different units. Ifpartial melting started at the base of themonzogranites (at least 700 °C at c. 6 kbar) thenthese rocks were at least 23 km thick beforethrusting. For our modelling approach, thispartial melting condition is used solely toconstrain temperature and depth at this givenlocation.

Concerning the volcanic units, we chose athickness of 20 km for the Ashanti and Kibibelts, in accordance with the peak pressure of4–6 kbar that was recorded in these rock units.The minimum and maximum sediment thick-nesses proposed for the sedimentary basins are3 and 8 km respectively. Different values weretested and the results for 6 km are presented.Pigois et al. (2003) estimate the presentminimum thickness of the sedimentary basins at2 km (within the basin facies). This is the lowerboundary for the thickness of the final sedimen-tary unit, after thrusting, erosion and the D2stage (which is beyond the scope of this study).

The thickness of the present-day Tarkwaian unitranges from 1800–3000 m (see a detaileddescription in Pigois et al. 2003). This mega-sequence resulted from the stacking of differentlithostratigraphic formations: the Kawereconglomerate (250–700 m), the Banketconglomerate (150–600 m), the Tarkwa forma-tion (120–400 m) and the Huni formation(<1300 m). 1 km is taken for the Kawere, whichwas deposited just after the thrusting event;3 km of conglomerates was then added to thepile, leading to a final value of 4 km for the Tark-waian conglomerates.

Chronological constraints

The stage just before 2130 Ma is considered torepresent the initial condition for the numericalmodelling (Fig. 2). At that time, all the unitsexcept the Tarkwaian conglomerates wereemplaced. From west to east, the 40 km-wideunits are: the Afema basin, which covers thegreenstone belts, the Ashanti volcanic belt, theBirim basin, the Kibi belt and the monzogran-ites.

The initial stage of tectonic accretion, D1,began at 2130 Ma. It resulted in unit stackingthat can be evaluated using the thermobaromet-ric data at peak metamorphism conditions(Table 1). The pressure constraints of 4–6 kbarimply that the sediments of the Birim basin, theAshanti greenstone belts and rocks of the Kibibelt were buried to a depth of 15–22.5 km, whena density of 2700 kg m–3 is used in the model.This value appears to be representative for mostrocks of the area, except for fresh sediments(Table 4).

Deposition of the Kawere unit on the Ashantibelt was contemporaneous with D1. It containspolygenic conglomerates deposited after 2132 ±2 Ma (age of the youngest detrital zircon fromthe base of the series; Davis et al. 1994;Oberthür et al. 1998). Following D1 tectonism,auriferous Tarkwaian conglomerates from the


Table 4. Main lithological units with associated density values and thermal properties. Density measurements arefrom Turcotte & Schubert (2002) except (1) from Barritt & Kuma (1998). Thermal properties are from this studyexcept (2) from Jaupart and Provost (1985)

Rock type Mean density Mean thermal conductivity Mean heat-production rate(kg m–3) (W m–1 K–1) (µW m–3)

Basement 2750 2.6 1.7Greenstones 28701 2.9 0.7Sediments 2200 before D1 1.82 before D1 1.4

27301 after D1 2.92 after D1 1.4Monzogranites 2700 3.1 1.9Tarkwaian conglomerates 27001 1.82 1.9


Banket series were deposited between 2104 ±6 Ma (age of the youngest detrital zircon; Milesiet al. 1989) and 2097 ± 2 Ma (age of the Bansogranitoid that is assumed to intrude the Tarkwa-ian deposits; Oberthür et al. 1998).

Constraints for erosion

The estimate of the erosion rate between theregional peak metamorphism (2100–2105 Ma)and the end of deposition of the TarkwaianBanket series (2097 Ma) is constrained by thepresence of mineralized pebbles of quartz in theconglomerate dated at 2105 Ma (Milesi et al.1989). Orogenic gold deposits related to anearlier Eburnean compressional event repre-sent one of the sources of gold for the aurifer-ous conglomerates (Milesi et al. 1989; Klemd etal. 1993; Oberthür et al. 1997; Pigois et al. 2003).These pebbles result from the reworking oforogenic deposits that are estimated to haveformed under the same conditions as the lode-gold deposits of the western margin of theAshanti belt (Klemd et al. 1993). The estimateddepth of formation for the deposits of theAshanti area is c. 4–15 km (c. 1–4 kbar)(Oberthür et al. 1997; Schmidt Mumm et al.1997; John et al. 1999; Yao et al. 2001). Thesedata suggest the erosion of 4–15 km of stratawithin these 10 Ma. For modelling purposes, weadopted a value of 10 km, which is in accordancewith the abundance of cataclastic and myloniticpebbles (Milesi pers. comm.).

Thermal modelling

Equations, boundary conditions,parameters and variables

The general form of the heat conductionequation to account for heat transfer in a solid,involving both internal heat production andmaterial advection in Cartesian coordinates, is:

∇(k∇T) – ρc

– ρf cfvf∇T+ A = 0

(3)

where T is temperature in K, k the thermalconductivity tensor depending on direction andtemperature (W m–1 K–1), A the heat-productionrate (W m–3), ρ the density (kg m–3), c thespecific heat (J kg–1 K–1), ρf cf the fluid heatcapacity (J K–1 m–3), and vf the Darcy velocityfield (m s–1). In our present study, fluid advec-tion is not taken into account, and vf = 0.

Moreover, k is considered to be isotropic,temperature independent and with the 2Dapproximation for the Ashanti belt, equation (3)reduces to:

k + A = 0 (4)

where x is the horizontal distance and z is depth.

Numerical modelling

The source code THERMIC, used to solve the2D conductive and advective heat-transferequation, was developed by Bonneville &Capolsini (1999). Based on a finite-elementmethod it can incorporate realistic geometries,heterogeneous material properties and variousboundary and initial conditions (e.g. Vasseur etal. 1993).

The original code was improved to take intoaccount phenomena encountered in Earthsciences, such as erosion, deposition and instan-taneous thrusting. Tests were performed tovalidate the modifications, and it turned out thatour results are in excellent agreement withthose of Nisbet & Fowler (1982a, b).

Description of the numerical model

The model is a rectangular grid 200 km long and150 km thick. This represents the crust and thelithospheric mantle. This grid is divided into9600 4-node isoparametric elements of 1 �1 km, down to 47 km depth. We used the thick-ness deduced for each rock unit to build theinitial model (Fig. 4a). Each 40 km-wide litho-logical unit corresponds to a specific material,whose thermal properties are defined by thethermal conductivity, the heat production rateand the heat capacity.

The solution of the thermal problem is afunction of the imposed boundary conditions,and the thermal properties of the different rocktypes. The chosen boundary conditions involvea constant mantle heat flow qm entering the baseof the grid, and a constant surface temperature(T = 0 °C). Various values of the mantle heatflow are used in order to adjust thermo-barometric data from the pre-Eburnean stage tothe different stages of Eburnean tectonism.

Transient evolution

We chose to represent the tectonic evolutionduring D1 stage using two different thrustingevents (Fig. 4). First, the westward compres-sion exerted by the Archaean craton caused

( )∂∂

+ ∂∂

− ∂∂

2

2

2

2

Tx

Tz

cTt

ρ

∂∂Tt



thrusting in which part of the monzogranites(10 km) and half of the sediments of the Birimbasin (3 km) were involved. The activation oftwo thrust faults located at the margin of theseblocks led to an intermediate stage where themonzogranites and the sediments were over-thrust onto the Kibi greenstone belt and theAshanti volcanic belt, respectively. The meanburial depth, estimated at 18 km (c. 5 kbar), atthe base of the sedimentary Birim basins couldonly have been met if they were covered by15 km of rocks. We thus assume that the unitoverthrust onto the Birim basin a sequencecomprising monzogranites (10 km) andvolcanics from the Kibi belt (5 km).

Numerical calculations are simplified bykeeping the Earth surface flat during thethrusting, which at the time scale of the modelcan be considered ‘instantaneous’. The choiceof instantaneous thrusting can be justified. Therate of burial does not influence the time lapsebetween the end of thrusting and maximumtemperature. It plays only a minor role indetermining the maximum temperature ofrocks during the later unroofing stage (Ruppel& Hodges 1994). Moreover, these authorsdemonstrate that the maximum temperaturedepth is linked only slightly to the rate of burialor to the amount of syntectonic heating. As weare interested in the thermal state after D1tectonism, the instantaneous approach isacceptable.

The period of thrusting was followed by astage of uplift and erosion, of which the Tark-waian conglomerates are the products. As aconsequence of isostatic response to burial, theBirim sedimentary basin, the most overthrustunit, will be the most affected by erosion anduplift. Based on geological observations of Tark-waian samples, we assumed a rate of erosion of1 km/Ma, so that 10 km were uplifted anderoded within the 10 Ma separating the end ofinstantaneous thrusting and the beginning of D2tectonism. In the model, erosion began immedi-ately after instantaneous burial, with a rate of1 km/Ma.

Results and discussion

Pre-thrusting

The thermobarometric conditions related tomonzogranites, c. 700 °C at c. 6 kbar, appear tobe reached only if the amount of heat at the baseof this unit is sufficient. The pre-thrusting stagecan be constrained using these data (Fig. 4a) byfitting temperature and pressure conditions withadequate model parameters. Those models

where the temperature at the base of the crustexceeded 900 °C were rejected.

Tests were initially performed using meanmeasured values of thermal conductivities andheat-production rates for the different litho-logical units (Table 4). Several experimentswere performed to evaluate mantle heat flowvalues that would be consistent with the ther-mobarometric conditions. Using averagethermal properties (Table 4), the minimummantle heat flow that leads to approximately700 °C at 22.5 km, is 60 mW m–2 (Fig. 5a). Forlower mantle heat flow values, the 700 °Cisotherm is located beneath the base of the unitwhereas for higher heat flow values, tempera-tures at the base of the crust unit exceed 900 °C.

In Figure 5a, the thermal anomaly thatappears at the base of the crust below the Birimbasin is due entirely to the thermal properties ofthe rocks, the basal heat flow being constant atthis depth. This figure also illustrates the effectof thermal refraction caused by a thermalconductivity contrast between the sediments (k= 1.8 W m–1 K–1) and the surroundings (k =2.9 W m–1 K–1). Figure 5c shows lateral heattransfer before the thrusting event. The crustalscale anomalies are focused towards the bound-aries between the different rock units, empha-sizing the role of lateral conductivity contrastsin the heat transfer pattern. Lateral contrasts inheat-production rates also contribute to lateralheat transfer anomalies, but with a much lesserimportance than conductivity contrasts.

Post-thrusting: Eburnean tectonism

The results need to fit the data recorded forregional peak metamorphism:

• in the greenstone belts, a temperature of500–650 °C is required at 4–6 kbar;

• in the sediments of the Birim basin, green-schist facies is reached

The stage before thrusting was calculatedwith mean thermal conductivities and a mantleheat flow of 60 mW m–2. Then, the differentunits were stacked instantaneously (Fig. 4b) andthe resulting final thermal field, after thrusting,uplift and erosion, is shown in Figure 5b.

In order to validate the modelling, P–T pathsfrom various rocks within the Ashanti belt wereinvestigated. Figure 6 shows the paths for rockssituated in the Ashanti belt, in the middle of theunit at different initial depths (at the base of thegreenstones, in the middle and in between thesetwo levels). The rocks at an initial depth of15 km were first buried to 18 km during the




Fig. 5. Temperature field (a) before and (b) after the D1 tectonic phase, as inferred from thermobarometricdata and from a maximum temperature at the base of the crust of 900 °C. The undulated shape of theisotherms is due to the thermal properties of each lithological unit since mantle heat flow (60 mW m–2) staysconstant at the base of the model. This value allows the conditions for partial melting in the monzogranites tobe reached. After the D1 phase (b), the deposition of sediments and Tarkwaian conglomerates distorts theshallow isotherms and yields temperature gradients higher than 40 °C km–1 within the Ashanti belt. Lateralheat flow before the thrusting events (c) and after the entire D1 phase (d). Positive values correspond to awestward heat flow. Ore deposits have formed after the D1 phase and are mainly concentrated at the edges ofthe Ashanti belt, where shallow anomalous lateral heat flow is observed (see e).

05SP248_Harcouet (COL) (JB-S) 1/8/05 1:39 pm 3

instantaneous thrusting and then receiveddeposition of the Tarkwaian conglomerates.They reach P–T conditions that are in goodagreement with the amphibolite faciesconditions recorded in the greenstones. Athickness of around 12 km of greenstones(equivalent to an initial burial of 15 km) wouldalso enable the P–T domain defined by the peakmetamorphism conditions to be reached, inaccordance with the upper boundary estimatesof Sylvester & Attoh (1992).

The P–T paths of sediments located at thebase of the Birim basin were also investigatedand are reported in Figure 7. The rock, originallylocated at the base of the basin (18 km depth),follows a retrograde P–T path as it travelstowards the surface until it reaches its finallocation at 8 km depth. Computed values are ingood agreement with thermobarometric data,which suggest that these rocks reached green-schist facies conditions. The retrograde P–Tpath is representative of the response to burial,with an increase in temperature during the firstmillion years and then a decrease correspond-ing to the erosion phase. During the period ofunroofing to shallower crustal levels, the timelag observed between the end of the compres-sional interval and the reaching of peak temper-atures is in accordance with the theoreticalprediction of England & Richardson (1977).


Fig. 6. Results of numerical experiments (P–T paths) compared with thermobarometric data. Each P–T pathcorresponds to the burial of greenstones of the Ashanti belt, after deposition of the Tarkwaianconglomerates. The greenstones buried to 19 km record P–T paths that are in agreement with thethermobarometric data.

Fig. 7. Results of numerical experiments (P–T paths)for Birim sediments, compared with the greenschistfacies domain. Each P–T path corresponds to theerosion phase after thrusting, but before the end ofD1 tectonism (i.e. from 2105 to 2095 Ma). The fivecurves reflect different values for thermalparameters: (kmin and kmax) = minimum andmaximum values of thermal conductivity, (Amin andAmax) = minimum and maximum values of heatproduction rate, and (k–,A–) = mean values for thermalconductivity and heat production rate. The highesttemperatures (peak metamorphism conditions) arereached at 2101 Ma, as suggested by Hirdes et al.(1996) and Oberthür et al. (1998).


Moreover, in all cases, the peak temperatureis reached after 4 Ma of erosion, which corre-sponds to 2101 Ma ago. This correlates well withthe timing of the regional peak metamorphism,dated between 2092 ± 3 and 2105 ± 3 Ma byOberthür et al. (1998) and at 2100 ±3 Ma byHirdes et al. (1996).

The lateral heat flow calculated 11 Ma afterthrusting (corresponding to 2094 Ma ago) isillustrated in Fig. 5d. This shows the resultingshallow heat transfer mechanisms after D1tectonism and the associated erosion–deposi-tion phase. Compared to the initial lateral heatflow (Fig. 5c), it shows an important phase oflateral heat transfer (20–25 mW m–2) within theupper crust (maximum values are reachedwithin the upper 5 km, Fig. 5d). These thermalanomalies are more important at each beltboundary, including the eastern and westernboundaries of the Ashanti belt. As the thermalconductivity contrasts are the same as for theinitial stage, the anomalies cannot be attributedto heat refraction effects alone. Indeed, it is theparticular tectonic history of the Ashanti beltthat could explain these shallow thermal anom-alies: (a) before thrusting, isotherms wereuplifted beneath the surrounding sedimentarybasins (Afema, Birim) due to the insulatingeffect of fresh sediments; (b) thrusting of theseanomalously hot sediments disturbed theshallow temperature field, yielding the observedlateral heat flow pattern (Fig. 5d); and (e) thedeposition of insulating conglomeratesenhanced this effect.

Sensitivity to thermal parameters

The thermal regime of the crust is controlled byboundary conditions (i.e. mantle heat flow in thisstudy) and the thermal properties of representa-tive rocks. In order to constrain our results onmantle heat flow values, the effects of variations(and uncertainty) in thermal conductivity andheat production rates were also investigated.

We tested the sensitivity of the modelling toextreme values of thermal conductivity. If theminimum values of thermal conductivity for allrock types are taken (mean values minus thestandard deviation), then the temperature ishigher than in previous cases because of theenhanced insulating effect. In this case, a valueof 43 mW m–2 for the mantle heat flow would besufficient to satisfy the thermobarometricconditions. On the contrary, if the highest valuesof thermal conductivity are considered, then avalue of 70 mW m–2 for the basal heat flow isnecessary to comply with thermobarometricconditions before and after thrusting.

Tests were also performed to constrain theinfluence of heat production rates. Whenmaximum values of heat production rates areconsidered for all rock types, a value of42 mW m–2 for mantle heat flow is necessary tofit the thermobarometric conditions. This valueis much higher when minimum heat productionrates are taken, since qm must reach 75 mW m–2.Increasing the heat production rate of themonzogranites to its maximum (2.7 µW m–3)enables higher temperatures to be reachedwithin this unit, without changing the tempera-ture at the base of the crust dramatically.

When combining the effects of maximumheat production rates and minimum thermalconductivity, 25 mW m–2 is the minimum mantleheat flow value required to fit the thermobaro-metric data. This value is the lower validboundary for basal heat flow. It represents themost extreme case, which is quite unlikely.Figure 7, which also shows the case for theseparameters for heat production and thermalconductivity, demonstrates that the associatedP–T paths fall in the uppermost limit of thegreenschist facies domain. Tests using minimumheat production rates and maximum thermalconductivity present solutions that are not inaccordance with the thermobarometric data,and were thus rejected.

The lateral heat flow value also depends onthese thermal properties. Indeed, lateralthermal conductivity contrasts are the mostdeterminant parameters for the presence oflateral heat flow anomalies. The contribution oflateral contrasts in heat production rate wasestimated. First, by taking a thermal conductiv-ity of 2.9 W m–1 K–1 for all the rock types, theeffects of lateral thermal conductivity contrastsare cancelled. In this case, lateral heat flowvalues decrease dramatically compared to thestandard case, and a maximum value of only6 mW m–2 is observed, due to lateral contrasts inheat-production rates. Secondly, when uncer-tainties in heat production are considered, andcontrasts between units are maximized (e.g.heat production rate minimum for the sedi-ments and maximum for greenstone belts),lateral heat flow anomalies are modified by afactor of less than 25%.

Finally, the localization of the Tarkwaianconglomerates (here inferred from field obser-vations) also influences the pattern of lateralheat flow anomalies. For example, if depositiononly occurs within the surrounding units of theBirim basin, namely the belts of Ashanti andKibi, then the anomalies will be higher than inthe standard case and will be concentratedalong the margins of these units.



Conclusion

The thermal evolution of southern Ghana from2130 to 2095 Ma gives insight into the regionalthermal state of the crust before the mineraliz-ing events that led to the emplacement of thegiant gold deposits of the Ashanti belt.

Thermal modelling of tectonic and surfaceprocesses with adequate thermal parametershas allowed us to reproduce the thermobaro-metric conditions. Amphibolite- to greenschist-facies conditions are encountered for the rocksof the volcanic belts and the Birimian sediments,which is in agreement with the regional peakmetamorphism conditions indicated by previousstudies (Kleinschrot et al. 1993; Hirdes et al.1996; Oberthür et al. 1998; John et al. 1999;Klemd et al. 2002).

The obtained mantle heat flow values rangefrom 25 mW m–2 to 75 mW m–2, with a mostlikely value of 60 mW m–2, which is at least threetimes greater than the present-day value. Thisrelatively high value may be related to deepthermal perturbations. Indeed, the opening ofsedimentary basins during the first stage of D1tectonism, not taken into account in the model,may be related to a regional increase in mantleheat flow. Nevertheless, when present-daymantle heat flow values are considered, a valueof 60 mW m–2 does not represent a dramaticincrease in subcontinental heat flow, and the‘mantle overturning’ phenomenon (Condie2000; Goldfarb et al. 2001) may not be so cata-strophic. Indeed, even deep and large mantleplumes impinging at the base of the continentallithosphere does not necessarily induce sharpincreases in mantle heat flow (Lenardic et al.2000). If superevents in the mantle are relatedto anomalous deep processes such as super-plumes, then the associated thermal signatureswould not necessarily imply strong modifica-tions in mantle heat flow values. One must alsoindicate that a temperature-dependent thermalconductivity (not tested in this study) wouldtend to lower the mantle heat flow values sincethe decreasing conductivities would enhancecrustal temperatures. Finally, our results showthat thermal equilibrium of the crust can beeasily disturbed as soon as mantle heat flowincreases by only a few tens of mW m–2.

The results also indicate anomalies of lateralheat flow reaching 20–25 mW m–2, mainlyfocused on the western and eastern margins ofthe Ashanti belt and other belt boundaries. Suchanomalies, where associated with fractures, mayhave been of importance for the mineralizingevents, as they would have favoured fluidcirculation locally. Indeed, field studies have

emphasized the heterogeneous strain partition-ing and the highly focused fluid flow at shallowdepths, which could be triggered partly by suchlateral heat transfer anomalies. The correlationbetween the location of these anomalies on bothflanks of the Ashanti belt and the location of themajor ore deposits is an encouraging result.

A number of first-order simplifications weretaken in this study. On one hand, the choice ofa depth-independent thermal conductivity mayslightly overestimate the results since lateralcontrasts may decrease with increasingtemperature. Additional tests are required toquantify second-order effects, but the first-orderphysical mechanisms are expected to besufficiently important to affect local fluidcirculation.

Although our study does not take intoaccount fluid circulation, which is critical inunderstanding the details of the mineralizingprocesses, it is a first and necessary step in thecomprehension of the regional thermal evolu-tion of the area. A second stage of modelling,investigating the mineralizing processes at amore local scale and integrating fluid circu-lation, is the objective of an ongoing project. Itwill be based on the results of this regionalstudy.

V. H. was supported by a grant from BRGM (Bureaude Recherches Géologiques et Minières). The authorswish to thank sincerely Vincent Bouchot, Jean-LucLescuyer, Francis Lucazeau, and Jean-Pierre Milesifor their support and advice. We are grateful to thetwo reviewers, K. Gallagher and P. E. J. Pitfield, fortheir valuable comments and suggestions. We thankRowena Stead for proofreading the final text andEnglish editing. This is IPGP contribution #2919, andBRGM publication BRGM-CORP-03163.

ReferencesBARLEY, M.E., KRAPEZ, B., GROVES, D.I. & KERRICH,

R. 1998. The Late Archean bonanza : metallo-genic and environmental consequences of theinteraction between mantle plumes, lithospherictectonics and global cyclicity. PrecambrianResearch, 91, 65–90.

BARRITT, S.D. & KUMA, J.S. 1998. Constrained gravitymodels of the Ashanti belt, southwest Ghana.Journal of African Earth Sciences, 26, 539–550.

BONNEVILLE, A. & CAPOLSINI, P. 1999. THERMIC: a2-D finite-element tool to solve conductive andadvective heat transfer problems in Earthsciences. Computers and Geosciences, 25,1137–1148.

CARSLAW, H.S. & JAEGER, J.C. 1984. Conduction ofheat in solids. Clarendon Press, Oxford, 2nd edn,510pp.

CLAUSER, C. & HUENGES, E. 1995. Thermal conduc-tivity of rocks and minerals. In: AHRENS,T.J. (ed.)



Rock Physics and Phase Relations : A Handbookof Physical Constants. AGU, Washington, D.C.,AGU Ref. Shelf, 3, 105–126.

CONDIE, K.C. 2000. Episodic continental growthmodels: afterthoughts and extensions. Tectono-physics, 322, 153–162.

CONDIE, K.C., DES MARAIS, D.J. & ABBOTT, D. 2000.Geologic evidence for a mantle superplumeevent at 1.9 Ga. Geochemistry, Geophysics, andGeosystems, 1, Paper number 2000GC000095.

DAVIS, D.W., HIRDES,W., SCHALTEGGER, U. & NUNOO,E.A. 1994. U–Pb age constraints on depositionand provenance of Birimian and gold bearingTarkwaian sediments in Ghana, West Africa.Precambrian Research, 67, 89–1997.

DE BOORDER, H., SPAKMAN, W., WHITE, S.H. &WORTEL, M.J.R. 1998. Late Cenozoic mineraliza-tion, orogenic collapse and slab detachment inthe European Alpine Belt. Earth and PlanetaryScience Letters, 164, 569–575.

ENGLAND, P.C. & RICHARDSON, S.W. 1977. The influ-ence of erosion upon mineral facies of rocks fromdifferent metamorphic environments. Journal ofthe Geological Society, London, 134, 201–213.

FEYBESSE, J.-L. & MILESI, J.-P. 1994. TheArchean/Proterozoic contact zone in West Africa:a mountain belt of decollement thrusting andfolding on a continental margin related to 2.1 Gaconvergence of Archaean cratons? PrecambrianResearch, 69, 199–227.

FEYBESSE, J.-L., BILLA, M., MILESI, J.,-P., LEROUGE, C.& LE GOFF, E. 2000. Relationships betweenmetamorphism-deformation-plutonism in theArchean-Paleoproterozoic contact zone of WestAfrica. 31st International Geological Congress,Rio, CD-ROM Abstract.

GALSON, D.A., WILSON, N.P., SCHÄRLI, U. & RYBACH,L. 1987. A comparison of the divided-bar andQTM methods of measuring thermal conductiv-ity. Geothermics, 16, 215–226.

GOLDFARB, R.J., GROVES, D.I. & GARDOLL, S. 2001.Orogenic gold and geologic time: a global synthe-sis. Ore Geology Reviews, 18, 1–75.

HIRDES, W., DAVIS, D.W., LÜDTKE, G. & KONAN, G.1996. Two generations of Birimian (Paleoprotero-zoic) volcanic belts in northeastern Côte d’Ivoire(West Africa): consequences for the ‘Birimiancontroversy’. Precambrian Research, 80, 173–191.

ISLEY, A.E. & ABBOTT, D.H. 1999. Plume-relatedmafic volcanism and the deposition of bandediron formation. Journal of Geophysical Research,104, 15 461–15 477.

JAUPART, C. & MARESCHAL, J.-C. 1999. The thermalstructure and thickness of continental roots.Lithos, 48, 93–114.

JAUPART, C. & PROVOST, A. 1985. Heat focussing,granite genesis and inverted metamorphic gradi-ents in collision zones. Earth and PlanetaryScience Letters, 73, 385–397.

JOHN, T., KLEMD, R., HIRDES, W. & LOH, G. 1999. Themetamorphic evolution of the Paleoproterozoic(Birimian) volcanic Ashanti belt (Ghana, WestAfrica). Precambrian Research, 98, 11–30.

KLEINSCHROT, D., KLEMD, R., BRÖCKER, M.,

OKRUSCH, M. & SCHMIDT, K. 1993. The Nsutamanganese deposit, Ghana: geological setting,ore-forming process and metamorphic evolution.Zeitschrift des Angewandtes Geologie, 39, 48–50.

KLEMD, R., HIRDES, W., OLESCH, M. & OBERTHÜR, T.1993. Fluid inclusions in quartz-pebbles of thegold-bearing Tarkwaian conglomerates of Ghanaas guides to their provenance area. MineraliumDeposita, 28, 334–343.

KLEMD, R., HÜNKEN, U. & OLESCH, M. 2002. Meta-morphism of the country rocks hosting gold-sulfide-bearing quartz veins in thePaleoproterozoic southern Kibi-Winneba belt(SE-Ghana). Journal of African Earth Sciences,35, 199–211.

KUKKONEN, I.T. 1998. Temperature and heat flowdensity in a thick cratonic lithosphere: the svekatransect, Central Fennoscandian shield. Journalof Geodynamics, 26, 111–136.

KUKKONEN, I.T. & PELTONEN, P. 1999. Xenolith-controlled geotherm for the central FennoscandianShield: implications for lithosphere–astenosphererelations. Tectonophysics, 304, 301–315.

LENARDIC, A., GUILLOU-FROTTIER, L., MARESCHAL,J.-C., JAUPART, C., MORESI, L.-N. & KAULA, W.M.2000. What the mantle sees: the effects of conti-nents on mantle heat flow. In: RICHARDS, M. ET

AL. (eds) The history and dynamics of global platemotions. AGU, Geophysical Monograph, 121,95–112.

MILESI, J.-P., FEYBESSE, J.L., LEDRU, P., ET AL. 1989.Mineralisations aurifères de l’Afrique de l’ouest,leurs relations avec l’évolution litho-structuraleau Paleoproterozoïque inférieur. Chroniques dela Recherche Minière, 497, 3–98.

MILESI, J.-P., LEDRU, P., ANKRAH, P., JOHAN, V.,MARCOUX, E. & VINCHON, C. 1991. The metallo-genic relationship between Birimian and tark-waian gold deposits in Ghana. MineraliumDeposita, 26, 228–238.

MILESI, J.-P., LEDRU, P., FEYBESSE, J.-L., DOMMANGET,A. & MARCOUX, E. 1992. Early Proterozoic oredeposits and tectonics of the Birimian orogenicbelt, West Africa. Precambrian Research, 58,305–344.

NISBET, E.G. & FOWLER, C.M.R. 1982a. The thermalbackground to metamorphism – 1. Simple one-dimensional conductive models. GeosciencesCanada, 9, 161–164.

NISBET, E.G. & FOWLER, C.M.R. 1982b. The thermalbackground to metamorphism – II. Simple two-dimensional conductive models. GeosciencesCanada, 9, 208–214.

OBERTHÜR, T., WEISER, T., AMANOR, J.A. & CHRYS-SOULIS, S.L. 1997. Mineralogical siting and distri-bution of gold in quartz veins and sulfide ores ofthe Ashanti mine and other deposits in theAshanti belt of Ghana: genetic implications.Mineralium Deposita, 32, 2–15.

OBERTHÜR, T., VETTER, U., DAVIS, D.W. & AMANOR,J.A. 1998. Age constraints on the gold mineraliza-tion and Paleoproterozoic crustal evolution in theAshanti belt of southern Ghana. PrecambrianResearch, 89, 129–143.



OPARE-ADDO, E., BROWING, P. & JOHN, B.E. 1993a.Pressure–temperature constraints on the evolu-tion of an early Proterozoic plutonic suite insouthern Ghana, West Africa. Journal of AfricanEarth Sciences, 17, 13–22.

OPARE-ADDO, E., JOHN, B.E. & BROWING, P. 1993b.Field and geochronologic (U–Pb) constraints onthe age and generation of granitoids andmigmatites in southern Ghana, AGU Springmeeting, Baltimore. EOS Transactions AbstractsSupplement, 74, 301.

PIGOIS, J.P., GROVES, D.I., FLETCHER, I.R.,MCNAUGHTON, N.J. & SNEE, L.W. 2003. Ageconstraints on Tarkwaian paleoplacer and lode-gold formation in the Tarkwa–Damang district,SW Ghana. Mineralium Deposita, 38, 695–714.

PIRAJNO, F. 2001. Ore deposits and mantle plumes.Kluwer Academic Publishers, Dordrecht.

RICHTER, F.M. 1985. Models for the Archean thermalregime. Earth and Planetary Science Letters, 73,350–360.

ROLANDONE, F., JAUPART, C., MARESCHAL, J.-C.,GARIÉPY, C., BIENFAIT, G., CARBONNE, C. &LAPOINTE, R. 2002. Surface heat flow, crustaltemperatures and mantle heat flow in theProterozoic Trans-Hudson Orogen, Canadianshield. Journal of Geophysical Research, 107, B12,2341.

RUPPEL, C. & HODGES, K.V. 1994. Pressure–tempera-ture–time paths from two-dimensional thermalmodels: Prograde, retrograde, and inverted meta-morphism. Tectonics, 13, 17–44.

RUSSEL, J.K. & KOPYLOVA, M.G. 1999. A steady-stateconductive geotherm for the north central Slave,Canada: inversion of petrological data from theJericho kimberlite pipe. Journal of GeophysicalResearch, 104, 7089–7101.

SCHMIDT MUMM, A., OBERTHÜR, T., VETTER, U. &

BLENKINSOP,T.G. 1997. High CO2 content of fluidinclusions in gold mineralisations in the Ashantibelt, Ghana: a new category of ore formingfluids? Mineralium Deposita, 32, 107–118.

SERBAN, D.Z., NIELSEN, S.B. & DEMETRESCU, C. 2001.Transylvanian heat flow in presence of topogra-phy, paleoclimate, and groundwater flow.Tectonophysics, 335, 331–344.

SYLVESTER, P.J. & ATTOH, K. 1992. Lithostratigraphyand composition of 2.1 Ga greenstone belts ofthe West African craton and their bearing oncrustal evolution and the Archean-Proterozoicboundary. Journal of Geology, 100, 377–393.

TAVMAN, I.H. & TAVMAN, S. 1999. Measurement ofthermal conductivity of dairy products. Journal ofEngineering, 41, 109–114.

TAYLOR, P.N., MOORBATH, S., LEUBE, A. & HIRDES, W.1992. Early Proterozoic crustal evolution in theBirimian of Ghana: constraints from geochronol-ogy and isotope geochemistry. PrecambrianResearch, 56, 97–111.

TAYLOR, S.R. & MCLENNAN, S.M. 1995. The geochem-ical evolution of the continental crust. Reviews ofGeophysics, 33, 241–265.

TURCOTTE, D. & SCHUBERT, G. 2002. Geodynamics:Applications of Continuum Physics to GeologicalProblems. 2nd edn. Cambridge Univrsity Press,New York.

VASSEUR, G., DEMONGODIN, L. & BONNEVILLE, A.1993. Thermal effects arising from water circu-lation through thin inclined aquifers. GeophysicalJournal International, 112, 276–289.

YAO,Y., MURPHY, P.J. & ROBB, L.J. 2001. Fluid charac-teristics of granitoid-hosted gold deposits in theBirimian Terrane of Ghana: a fluid inclusionmicrothermometric and Raman spectroscopicstudy. Economic Geology, 96, 1611–1643.



Pre-mineralization thermal evolution of the Palaeoproterozoic gold-rich Ashanti belt, Ghana

Documents

Transcript of Pre-mineralization thermal evolution of the Palaeoproterozoic gold-rich Ashanti belt, Ghana