Black shales and massive sulfide deposits: causal or casual relationships? Insights from Rammelsberg, Tharsis, and Draa Sfar

Abstract

Black shales and massive sulfides represent reduced lithofacies that require isolation from oxic environments to be preserved. This, together with the sedimentary affinity of both lithofacies, can explain their common concurrence in the geologic record. The present study is based on the comparison of Rammelsberg in Germany, Tharsis in Spain, and Draa Sfar in Morocco, three massive sulfide deposits closely associated with black shales that are distributed along the European and North African Variscan orogen. The study entails geochemical, biostratigraphic, and stratigraphic analyses of the black shale sequences hosting the three deposits and mineralogical and textural analyses of the sulfides. All three deposits were formed in immature, tectonically unstable basins within an active continental margin or continental magmatic arc. Their stratigraphic records consist of a sequence of black shales enclosing massive sulfides and variable proportions of bimodal volcanic and subvolcanic rocks. The major differences among the three deposits concern the size, composition, and mineralogy. Regarding age, they are diachronous and younger southward: Rammelsberg is middle Eifelian, Tharsis latest Famennian, and Draa Sfar late Viséan. The study of redox conditions of the paleoenvironment using organic and inorganic proxies highlights similarities and significant differences among the three ore-hosting basins during massive sulfide and black shale deposition. The black shales generally display low C_org and high S_tot contents. At Rammelsberg, the S_tot/C_tot ratios provide values typical for normal Middle Devonian marine environments, which suggests that the original reactive organic C is now fixed in carbonates. At Tharsis, most of the samples have C_org >1 and S_tot/C_org values equivalent to those of Devonian–Carboniferous normal marine sediments. However, some pyritic hanging-wall samples have C_org <1 and S_tot up to 5 wt.%, suggesting the epigenetic addition of HS⁻. The S_tot/C_org ratio for the Draa Sfar samples resembles that of Middle Carboniferous normal marine environments. Geochemical inorganic proxies used to define the environmental conditions include the enrichment factors of U (U_EF) and Mo (Mo_EF) together with V/Cr and V/(V + Ni) ratios. Footwall shales at Filón Norte (Tharsis) show positive and eventually elevated U_EF and Mo_EF values, which suggests anoxic conditions, whereas at Rammelsberg and Draa Sfar oxic bottom water is indicated. The relations V/Cr and V/(V + Ni) in all three cases point to a redox boundary near the sediment–water interface, although at Tharsis some samples indicate anoxic/euxinic conditions (i.e., V/(V + Ni) >0.9). Regarding the environmental conditions of the source areas, feldspar illitization and selective depletion in Na and Ca occurred at the three studies sites. Available sulfur isotopic data from the Rammelsberg and Tharsis sulfide ore indicate that biogenic reduction of marine sulfate was a major sulfur source during massive sulfide generation. Nevertheless, a hydrothermal sulfur source has also been detected. At Rammelsberg, this is indicated from the polymetallic sulfides that replace sedimentary and diagenetic pyrite. At Tharsis, the bacteriogenic sulfur signature is also restricted to sulfide with less evolved textures, whereas a hydrothermal source is more evident in sulfides showing evidence of recrystallization. Both geochemical and isotopic data suggest that the bacteriogenic reduction process was inhibited by rapid burial. The sedimentation rates calculated for Rammelsberg, Tharsis, and Draa Sfar were in the range 7–13, 8–14, and 19–27 cm/ka, respectively. Continuous sedimentation of black shale favored the isolation of the massive sulfides and organic material from bottom waters and hence favored their preservation. Accordingly, the relationships between black shales and massive sulfides are considered to be casual. Nevertheless, the tectono-sedimentary evolution of each basin controlled the deposition of both black shales and massive sulfides and the parameters that favored their coeval deposition.

Introduction

Black shales are a common lithofacies throughout the geologic record (e.g., Tourtelot 1979; Wignall 1994). They consist of original muddy deposits rich in organic matter usually with middle to high maturation grades. Black shales are also commonly rich in sulfur and, in many cases, elements having a euxinic (sulfidic) affinity. Economically, black shales are one of the main host rocks for sedimentary and volcanosedimentary exhalative ore deposits (Large 1992; Goodfellow et al. 1993; Barrie et al. 1999; Sáez et al. 1999; Franklin et al. 2005) and are also the foremost source rock for hydrocarbons (Sethi and Schieber 1998).

Organic-rich muds, including black shales, occur in multiple and varied sedimentary environments (Stow et al. 2001) generally associated with “critical palaeogeographic episodes” (Bitterli 1963), including superplume events and supercontinent breakup (Condie 2004). Locally, where these deposits are widespread, black shales are interpreted in terms of global anoxic events and are consequently used as stratigraphic markers (Eder and Franke 1982; Arthur and Sageman 1994; Walliser 1996; Racki 2005). In sequence stratigraphy, black shale levels are considered maximum marine flooding surfaces associated with high nutrient supplies and low sedimentation rates (Haq et al. 1987), but in many cases the conditions for black shale deposition are essentially the opposite, i.e., high sedimentation rates and low productivity (Müller and Suess 1979; Stein 1986; Canfield 1994). Between both extremes, all types of intermediate possibilities have been suggested (Arthur and Sageman 1994). In any case, oxygen depletion in organic-rich mud below the sediment–water interface seems to be the common factor for settings characterized by black shale sedimentation.

Massive sulfide deposits are stratabound and/or stratiform accumulations of sulfide minerals (Sangster and Scott 1976). They can occur in most supracrustal rock types but show preference for marine volcanic rocks and fine-grained sedimentary rocks (Franklin et al. 1981). Massive sulfide deposits are composed mostly of Fe sulfides (i.e., pyrite and pyrrhotite) and minor amounts of sphalerite, chalcopyrite, and/or galena as the main economic minerals. Conventionally, they have been classified in two major groups, volcanic-hosted massive sulfide (VMS) and sedimentary exhalative (SEDEX), according to the predominance of volcanic or sedimentary rocks, respectively. However, the boundary between these two groups has proved ambiguous. As a single class, massive sulfide deposits constitute a major source for Cu, Zn, Pb, and by-products such as Au, Ag, and other metals. The two types of major host successions may occur in different geologic settings; black shales are commonly the main host rock for the massive sulfides (Franklin et al. 1981; Large 1992; Goodfellow et al. 1993; Barrie et al. 1999; Sáez et al. 1999; Kettler 2000; Pratt and Warner 2000; Franklin et al. 2005; Leach et al. 2005; Goodfellow and Lydon 2007; Mosier et al. 2009).

Space/time concurrence of black shales and massive sulfides has occurred throughout geologic history, from Archean (e.g., Kid Creek—Walker et al. 1975) until the present times (e.g., Guaymas Basin, Middle Valley, and Red Sea; Simoneit 2000). The temporal distribution of this black shales–massive sulfides association is essentially heterogeneous, with peaks particularly fructiferous during Early and Middle Proterozoic, Silurian, Middle and Late Devonian, Mississippian, and, to a minor extent, during Jurassic (Goodfellow et al. 1993; Eastoe and Gustin 1996; Franklin et al. 2005; Leach et al. 2005; Lyons et al. 2006). At the global scale, peaks of maximum development coincide with oceanic anoxic events (Goodfellow 1987; Turner 1992; Eastoe and Gustin 1996). However, Earth history also contains strongly anoxic events, like that of Middle Cretaceous times, characterized by thick black shale sequences with few associated massive sulfides.

At the basin and deposit scale, the black shales–massive sulfide association has been variously explained by (1) ocean anoxic events (Goodfellow 1987; Turner 1992; Eastoe and Gustin 1996; Lyons et al. 2006), (2) generation of restricted and poorly oxygenated marine sub-basins within normal (oxygenated) oceans (Goodfellow and Peter 1996; Ozsoy and Unluata 1997), (3) local anoxia produced by accumulation of reducing, high-density fluids in “brine pools” (Ross 1972; Sato 1972; Solomon et al. 2002), or (4) mineralization below the sediment–water interface (Almodóvar et al. 1998; Moreno et al. 2008).

One of the problems in understanding the relationship between black shales and massive sulfides is the modification of the organic and inorganic geochemistry of the shales caused by fluid–rock interaction during hydrothermal alteration (Goodfellow and Blaise 1988; Al-Aasm and Blaise 1991; Kettler 2000; Simoneit and Gize 2000). In the present paper, we expand on such relations by studying three distinctive shale-hosted massive sulfide deposits located along the Variscan belt at Rammelsberg (Germany), Tharsis (Spain), and Draa Sfar (Morocco). These deposits exhibit significant differences in tonnage, grade, and mineral composition but show strong similarities in a geological setting. In all three deposits, the black shales occur above, below, and laterally adjacent to the massive sulfides, and the footwall sequences contain intercalated volcanic rocks of varying proportions (Large and Walcher 1999; Moreno et al. 2008; Tornos et al. 2008). In spite of such similarities, traditional classifications consider Rammelsberg to be a typical SEDEX deposit (Gustafson and Williams 1981; Goodfellow and Lydon 2007), Tharsis a siliciclastic felsic VMS (Franklin et al. 2005) or shale-hosted VMS (Tornos 2006), and Draa Sfar a typical VMS (Belkabir et al. 2008; Marcoux et al. 2008).

Methodology

The present study is based on geochemical, biostratigraphic, and stratigraphic analyses of the black shale sequences hosting the three massive sulfide deposits together with mineralogical and textural analyses of the sulfides.

A total of 95 samples were taken from the black shales hosting the ore deposits. Special attention was paid during sampling to avoid the effects of weathering, oxide crusts, or any other factor unrelated to the primary deposition of the shales. The Rammelsberg samples were collected at the Schiefermühle quarry, a black shale quarry contiguous to the mine. This quarry, used for backfilling stopes during the last period of mining activity, exposes a section stratigraphically equivalent to that hosting the ore deposit (Sperling 1986; Large and Walcher 1999). Therefore, it was possible to collect samples above, below, and lateral to the massive sulfides. At Tharsis, the black shale samples were taken from the Filón Norte open pit (see González et al. 2002 for sampling procedure) and from boreholes drilled by SEIEMSA. Samples were collected below and above the massive sulfides; however, intensely hydrothermally altered and highly tectonized rocks were avoided. At Draa Sfar, the sampling was carried out in the DS- borehole 125, which intersects the Sidi M’Barek massive sulfide horizon; the analytical results for these samples were reported in Moreno et al. (2008).

The black shale samples were pulverized in an agate mill and split into 20-g fractions at the University of Huelva laboratories. Pulverized samples were analyzed for major and trace elements together with carbon and sulfur in a fully certified commercial laboratory (ACME, Analytical Laboratories, Vancouver, Canada). Total abundances of major oxides and some minor and trace elements were determined by ICP emission spectrometry. Refractory trace elements were analyzed by ICP mass spectrometry. Both procedures followed Li-metaborate/tetraborate fusion and dilute nitric digestion of powders weighing 0.2 g. Sample splits of 0.5 g were analyzed for a selection of metals and semi-metals including Cu, Pb, and Zn after aqua regia digestion. Sulfur, C_tot, and C_org were determined by Leco analysis. Samples containing >1,000 ppm Zn, >200 ppm Cu, >200 ppm Pb, and >5,000 ppm Ba have been excluded from the data base in order to minimize the effects of the mineralizing processes on the original geochemistry of the black shales.

To interpret the analytical results and minimize the potential effects of weathering, the elements are generally normalized to a standard. The European Shale (ES) and the Post Archean Australian Shale (PAAS) (Taylor and McLennan 1985; Rollinson 1993) provide the best fit to the geochemical characteristics of the black shales associated with the massive sulfides at Rammelsberg, Tharsis, and Draa Sfar. In the present paper, analytical data have been normalized to PAAS because it best covers the analytical suite we used. For mineral identifications and characterization of textures, polished shale and sulfide samples were examined by reflected light microscopy and scanning electron microscopy.

The biostratigraphic data presented here derive from several sources. The black shales surrounding the orebodies at Tharsis and Draa Sfar were the subject of palynostratigraphic studies published by González et al. (2002), Moreno et al. (2008), and Playford et al. (2008). The extraction procedures followed the conventional palynological techniques described by Wood et al. (1996). The samples analyzed palynostratigraphically in these previous studies were in many cases the same ones used here for geochemical analysis. Palynological research carried out at Rammelsberg was unsuccessful, principally because of the poor state of preservation of the palynomorphs recovered. Therefore, the biostratigrafic information from Rammelsberg used in this study is based on the macro- and micropalaeontologic data reported by Buchholz and Luppold (2008) from the Schiefermühle quarry.

Geohistorical analysis of the stratigraphic record permits the application of one-dimensional decompaction models for determining the original sedimentary thickness (Van Hinte 1978). At Rammelsberg, Tharsis, and Draa Sfar, the original thickness of the black shale sequences hosting the ore deposits was affected by the total overlying sediment pile, but the thickness of the denuded sequences is unknown in each case. On the basis of available stratigraphical data (Strauss 1970; Bordonaro et al. 1979; Sperling 1986; Bernard et al. 1988; Large and Walcher 1999; González et al. 2002; Buchholz and Luppold 2008; Moreno et al. 2008; Tornos et al. 2008), we estimated the possible thicknesses of the original stratigraphic units above the massive sulfides. Maximum and minimum thickness values for each stratigraphic unit were selected from the various compaction scenarios investigated here. These values correspond to stratigraphic columns with and without overlying flysch sequences. In all cases, the compaction constants and initial porosity values are those recommended by Sclater and Christie (1980), Angevine et al. (1990), and Hölzel et al. (2008).

Sedimentation rates were conservatively estimated as minimum values because the time intervals used are uniformly the maximum permitted by biostratigraphic constraints. The numerical ages assigned to the biostratigraphically determined intervals were based on the Devonian correlation tables of Weddige (1996; 2003), the calibrated Devonian time scale of Kaufmann (2006), and the Carboniferous global stratigraphic scale of Menning et al. (2006).

Geologic features

The Variscan Belt crops out widely in the Paleozoic massifs of Central and Southwestern Europe and, to a lesser degree, in Northwest Africa (Fig. 1). Since the earliest physiographic descriptions of “Mesoeurope”, hypotheses about the origin of this orogenic belt have been constantly evolving. They can be classified into three groups: (1) those suggesting folding and/or contraction of static eugeosynclinal basins, (2) those considering the orogen as a symmetric folding chain derived from the collision of Gondwana and Laurussia, and (3) those, currently most accepted, which envisage a more complex collisional system, resembling a “caterpillar orogen”, built up by the collage-like accretion of terranes independently detached from Gondwana throughout the entire Paleozoic (e.g., Nicolas 1972; Lorenz 1976; Matte 1986, 1991; Piqué and Michard 1989; Martínez Catalán et al. 2002; Winchester et al. 2002; Simancas et al. 2005), revealing independent pre-collisional geological histories of different duration.

Fig. 1

The West European and North African Variscan belt showing the major exposed massifs and the location of the studied areas. HM Harz Massif, RM Rhenish Massif, BF Black Forest, BM Bohemian Massif, AM Armorican Massif, CM Central Massif, IM Iberian Massif, SPZ South Portuguese Zone, EM Eastern Meseta, WM Western Meseta, JM Jebilet Massif

Most Variscan geodynamic models assume correlation between the Central European, Iberian, and North African domains. In that sense, the Rhenohercynian, South Portuguese, and Moroccan Western Meseta zones are considered lateral equivalents located on the external margin of the orogen (e.g., Arthaud and Matte 1977; Ellenberg and Tamain 1980; Ziegler 1982; Piqué 1994; Martínez Catalán et al. 2002). These terranes experienced similar stratigraphic and tectonic evolutionary processes, although at different times during the Late Palaeozoic: during the Middle Devonian in the Rhenohercynian Zone, the Late Devonian in the South Portuguese Zone, and the Mississippian in the Moroccan Western Meseta. The sedimentary basins in which the studied massive sulfide deposits were generated and preserved, i.e., Harz Massif, Iberian Pyrite Belt, and Jebilet Massif, respectively, reveal similar diachrony. They represent paleogeographically immature, tectonically active, continentally influenced marine basins that originated during extensional episodes prior to the main Variscan phase in each of these zones. In each case, the stratigraphic logs of the host strata include thick sequences of black shales with massive sulfides and different proportions of bimodal volcanic and subvolcanic rocks. These sequences overlie sandy-lutitic units deposited on shallow siliciclastic shelves and are overlain by thick turbiditic successions composed of the Variscan flysch.

During the Late Mississippian, the main compressional Variscan phase affected the Harz, Iberian Pyrite Belt, and Jebilet Massif in a broadly similar manner, producing intense deformation characterized by thin-skinned fold and thrust systems with different regional orientations (Piqué and Michard 1989; Silva et al. 1990; Oncken et al. 1999) and also weak regional metamorphism below the greenschist facies in all three cases (Müller and Strauss 1985; Abad et al. 2001; Essaifi et al. 2001). Later, during the late-Variscan orogenic phase, the three zones were intruded by plutonic and subvolcanic rocks.

Rammelsberg

The Harz Massif, in the German part of the Rhenohercynian Zone (Fig. 1), consists of an uplifted NW–SE-trending Variscan block surrounded by Mesozoic rocks (Hinze et al. 1998; Franzke and Zerjadtke 1992). The uplifted basement includes Paleozoic and Pre-Paleozoic sedimentary and volcanic rocks that are divided into three major domains: the Upper, Middle, and Lower Harz (Fig. 2a). The Rammelsberg ore deposit is located in the Upper Harz, where the stratigraphic record is mostly comprised of Devonian–Carboniferous sedimentary (mainly detrital) and minor volcanic rocks. The shale unit hosting the massive sulfide deposit is known as the Wissenbach Shales (Fig. 3). It was deposited over the Calceola Shales and below the Banded Shales, which is the unit immediately beneath the local equivalent of the Kellwasser anoxic event horizon (Hannak 1981; Sperling 1986; Large and Walcher 1999). These three units constitute the main infill of the Goslar trough, a sub-basin that was initiated during the Middle Devonian (Krebs and Gwosdz 1985; Werner 1990; Large and Walcher 1999). The Rammelsberg ore deposit is located at the margin of the trough. As a result, one of the most striking stratigraphic features of the region is the conspicuous thickness (and also facies) variation observed between Rammelsberg and the trough depocenter. Toward the top, the Upper Devonian–Lower Mississippian (pre-Kulm) interval is represented by a mixed detrital-carbonate succession divided into different stratigraphic units. These units also display the regional pattern of condensed and expanded facies depending on the location within the basin.

Fig. 2

Geological schemes of the Harz Massif (a), Iberian Pyrite Belt (b), and Jebilet Massif (c) showing the location of the massive sulfides of Rammelsberg, Tharsis, and Draa Sfar, respectively. UH Upper Harz, MH Middle Harz, LH Lower Harz, Me Meggen, SWP Southwest Portuguese Domain, Nc Neves-Corvo, Rt Riotinto, WMSZ Western Meseta Shear Zone, MSZ Marrakech Shear Zone, Kt Kettara

Fig. 3

Local stratigraphic logs of the studied districts. Rammelsberg: CS Calceola Shales, BS Banded Shales, K Kellwasser horizon. Tharsis: PQG Phyllite–Quartzite Group. Draa Sfar: TF Teksim Formation. Question marks indicate stratigraphic units reported regionally but not locally as a consequence either of non-deposition or erosion. Asterisks indicate uncertainty between stratigraphic and tectonic contact relationships

The age of the Rammelsberg massive sulfide deposit has been estimated on the basis of macro- and micropaleontologic data for the Wissenbach Shales (Fig. 4). These data include an extensive and distinctive fauna comprising brachiopods, cephalopods, bivalves, ostracods, tentaculites, crinoids, and corals, indicating an Eifelian–Givetian (lower Middle Devonian) age. High-resolution biostratigraphic analysis reported by Buchholz and Luppold (2008) in the shaly sequence laterally equivalent to the mineralized horizon at the Schiefermühle quarry constrained the age of the ore host sequence to the middle Eifelian partitus to costatus conodont zones (Fig. 4).

Fig. 4

Summarized geochronologic (vertical black lines) and biostratigraphic (horizontal gray intervals) data from Rammelsberg, Tharsis, and Draa Sfar. 1, Buchholz and Luppold (2008); 2, González et al. (2002); 3, Mathur et al. (1999); 4, Nieto et al. (2000); 5, Playford et al. (2008); 6, Marcoux et al. (2008); 7, Mrini et al. (1992); 8, Essaifi et al. (2003)

Tharsis

The Iberian Pyrite Belt (IPB) is the central domain of the South Portuguese Zone (Fig. 1). It extends from south of Lisbon in Portugal to the outskirts of Seville in southern Spain (Fig. 2b). The IPB consists of Upper Paleozoic (Middle Devonian–Mississippian) rocks classically divided into three lithostratigraphic units (Schermerhorn 1971). Massive sulfide deposits within the IPB, including those of the Tharsis district, are included in the Volcanosedimentary Complex (Fig. 3). At Tharsis, the base of this stratigraphic unit has been termed the anoxic sequence (Moreno and González 2004), which consists of a black shale that hosts the sulfide orebodies and also some felsic volcaniclastic rocks. Here, the oldest episode of volcanic activity is expressed by a thin felsic pumice-rich tuff ca. 10 m below the sulfide ore (González et al. 2002). The anoxic sequence is structurally overlain by a thick sequence of felsic and mafic volcanic and subvolcanic rocks also intercalated within black shales (Tornos et al. 1998).

The age of the massive sulfide deposit at Tharsis (Fig. 4) was achieved directly by the palynostratigraphic analysis of the black shales enclosing the Filón Norte and San Guillermo orebodies in the Filón Norte open pit (González et al. 2002). They reported an abundant terrestrial and marine palynoflora, consisting mainly of miospores, acritarchs, and prasinophyte phycomata. The miospore assemblages were assigned to the West European Retispora lepidophyta–Verrucosisporites nitidus (LN) biozone of Clayton et al. (1977). This LN biozone is in the upper Famennian, and its upper limit approximately coincides with the Devonian–Carboniferous boundary. Accordingly, the age of the host shale sequence at Tharsis, and consequently the age of mineralization, is interpreted as Late Devonian (latest Famennian). This age assignment is consistent with the Re–Os ages of 353 ± 44 Ma and 348.6 ± 12.3 Ma by Mathur et al. (1999) and Nieto et al. (2000), respectively, for the massive sulfide and stockwork mineralization of the Filón Norte orebody (Fig. 4).

Draa Sfar

The Jebilet Massif is located in the Moroccan Western Meseta, the region displaying the largest number of Variscan outcrops in Africa (Fig. 1).The massif consists of an E–W trending belt to the north of Marrakech surrounded by Pliocene and Quaternary alluvial deposits (Fig. 2c). It comprises Paleozoic sedimentary, volcanic, and minor metamorphic rocks distributed in three paleogeographic domains: the eastern, central, and western massifs (Huvelin 1977; Beauchamp and Izart 1987; Houari and Hoepffner 2003; Moreno et al. 2008). The massive sulfides at Draa Sfar, those from Ketara, and other minor regional occurrences all occur within the central Jebilet Massif. There, the stratigraphic record consists mostly of detrital Carboniferous rocks of the Sarhlef Series (Huvelin 1977). This unit is made up of a thick succession of black shale intercalated with sandstone and subordinate felsic and mafic volcanics and displays numerous and abrupt facies and thickness variations (Bordonaro et al. 1979). The massive sulfide deposits occur either embedded within the black shale or overlain by black shales and underlain by volcanic rocks (Belkabir et al. 2008; Marcoux et al. 2008). The entire sequence was intruded by mafic sills related to intense syn- and post-sedimentary magmatic activity (Fig. 3).

The age of the massive sulfide deposits at Draa Sfar (Fig. 4) is considered to be Asbian, late Viséan, according to Moreno et al. (2008). This well-constrained age was established from the palynostratigraphic analysis reported by Playford et al. (2008) in the black shales hosting the Sidi M’Barek orebody and by faunal evidence from bioclastic limestone at the base of the overlying Teksim Formation as reported by Hollard et al. (1977). The biostratigraphic Asbian age assigned by Moreno et al. (2008) to the Draa Sfar ore deposits agrees with the Ar–Ar age of 331.7 ± 7.9 Ma by Marcoux et al. (2008) for hydrothermal sericite from rhyolitic volcanic rocks underlying the Draa Sfar deposit and with the U–Pb and Rb–Sr ages of 327 ± 4 and 330.5 + 0.68/−0.83 Ma by Mrini et al. (1992) and Essaifi et al. (2003), respectively, for cordierite-bearing granitoid and microgranite that intrude the Central Jebilet sequence (Fig. 4).

Paleogeographic model

The paleogeographic models for the Harz, IPB, and Jebilet Massif are very similar. The pre-orogenic history of the three basins was characterized by vast, stable marine platforms dominated by monotonous detrital deposition. Later, in response to extensional tectonics, crustal thinning, and magmatic activity associated with the beginning of the rifting phase of the Variscan orogeny, these platforms collapsed to form silled basins with subsidence rates that were highly variable in time and space (Werner 1988; Piqué and Michard 1989; Moreno et al. 1996; Moreno and González 2004).

Comparing the stratigraphy of the Rammelsberg, Tharsis, and Draa Sfar sites, several similarities and some differences are evident (Fig. 3). The basement does not crop out in any of these districts and the total thickness of the stratigraphic sequences is also unknown in each case. At Draa Sfar, the host unit constitutes nearly the entire local stratigraphy, whereas at Rammelsberg and Tharsis it extends far below the massive sulfides. In these two districts, the oldest stratigraphic units (Kahleberg Formation of early Middle Devonian age at Rammelsberg and PQ Group of middle Late Devonian age at Tharsis) represent deposits of the former marine platform. Both units are lithologically similar and display evidence of shallowing toward the top of the sequences.

The Wissenbach Shales, Volcanosedimentary Complex, and Sarhlef Series, the host sequences of the study sites, are all characterized by facies and thickness changes that produced important local variations in the local stratigraphy. Such facies and thickness changes are interpreted to be a direct consequence of basin compartmentalization. Syn-sedimentary tectonic activity manifests in the three districts as growth faults and associated sedimentary rocks with structures indicative of rapid sedimentation and loading (e.g., flame structures, mass flow deposits). Emplacement of the massive sulfide bodies may also have been related to the sub-basin margin faults as these likely provided the high cross-stratal permeability that focused the mineralizing hydrothermal fluids (Felenc et al. 1986; Sperling 1986; Beauchamp and Izart 1987; Moreno et al. 1996; Large and Walcher 1999; Houari and Hoepffner 2003; Moreno and González 2004; Sáez et al. 2005; Moreno et al. 2007, 2008). Bimodal magmatic activity predated and persisted after ore deposition, although the magmatic rocks are unevenly distributed among the given regions. At Rammelsberg, volcanic rocks are subordinate and occur homogeneously throughout the entire sequence, whereas at Tharsis and Draa Sfar they are more abundant and are concentrated above and below the massive sulfides, respectively (Fig. 5).

Fig. 5

Space–time outline showing the stratigraphic architecture of Rammelsberg, Tharsis, and Draa Sfar

Sedimentation rates

The sedimentation rates calculated here for Rammelsberg, Tharsis, and Draa Sfar, are in the range 7–13, 8–14, and 19–27 cm/ka, respectively. Such variations depend on whether or not one accounts for the Kulm sequence overlying the Wissemback shales, for the Culm Group above the Volcanosedimentary Complex, and of the Karrouba Flysh overlying the Sarhlef Series. A significant result is that the sedimentation rates calculated for Draa Sfar are twice those for Rammelsberg and Tharsis. Nonetheless, they all represent very high rates, especially considering the pelagic and hemipelagic nature of most of the sediments involved. These data also support the interpretations of immature character and tectonic instability for the three ore-hosting basins.

The calculated sedimentation rates refer to the entire stratigraphic columns because the available age data preclude further chronostratigraphic subdivision. Under such circumstances, sedimentation velocity and sedimentation rate are often considered equivalent, especially if the lithology is broadly homogeneous throughout the sequence. Nevertheless, the geochemical analysis presented here permits the resolution of substantial differences in sedimentation velocity within the same shale sequence.

Massive sulfide deposits

Rammelsberg, Tharsis, and Draa Sfar are all good examples of exhalative massive sulfides within black shales (Fig. 6) and subordinate volcanic rocks (Appendix 1). According to Goodfellow et al. (1993) and Franklin et al. (2005), the main features of exhalative deposits are: (1) occurrence of base metal sulfide deposits (Zn, Pb, Cu) with variable proportions of Fe-sulfides, (2) stratiform mineralization accompanied by intense hydrothermal alteration in the footwall, and (3) disseminated and/or stockwork mineralization below the main orebody that represents the feeder zones of submarine hydrothermal systems.

Fig. 6

Polished section photographs of hand samples representative of the mineralizations at Rammelsberg, Tharsis, and Draa Sfar. a Rammelsberg banded mineralization. Bright bands are mainly composed by crystalline pyrite and ankerite. Dark background corresponds to black shales. Note the synsedimentary faults affecting the banded structure. b Tharsis pyrite stockwork cross-cutting chloritized black shales. c Tharsis banded mineralization with grained pyrite showing graded bedding, millimetric layers, lenses of black shales, and polygenic breccia levels. The breccia consists of pyrite and black shale fragments in a matrix of detrital pyrite. Pyrite fragments are diverse in grain size. Black shale fragments occasionally include disseminated pyrite. d Draa Sfar banded mineralization with black shale layers and lenses partially replaced by pyrrhotite. BS black shales, Py pyrite, Ank ankerite, Po pyrrhotite

The original stratiform morphology of the deposits was intensely modified during Variscan deformation. All three have comparable lateral extents (ca. 1,500 m) but of different thicknesses. Consequently, their total tonnages vary greatly: Rammelsberg 30 Mt, Tharsis >100 Mt, and Draa Sfar >10 Mt (Appendix 1). Base metal contents are also a distinctive feature, being very high at Rammelsberg, moderate at Draa Sfar, and low at Tharsis (average Cu + Zn + Pb grades = 22%, 6.7%, and 3.8%, respectively). Mining at Tharsis was not focused on the recovery of base metals but mainly on the extraction of pyrite for sulfuric acid production.

The most striking difference among the three deposits is the occurrence of primary pyrrhotite as the major sulfide mineral at Draa Sfar as well as in other Variscan massive sulfide deposits in Morocco (Marcoux et al. 2008; Moreno et al. 2008). Rammelsberg and Tharsis both have pyrite as the principal Fe-sulfide. The particular nature of the Fe-sulfides at Draa Sfar implies differences in the physicochemical environment of formation compared to that which prevailed during the formation of most other massive sulfide deposits such as Rammelsberg and Tharsis.

At Rammelsberg, the footwall vein system is associated with intense silicification of a particular zone of the footwall shales locally named “kniest”. This zone is interpreted as the inner part of the hydrothermal system that was associated with massive sulfide formation (Krebs and Gwosdz 1985; Large and Walcher 1999). The associated proximal alteration includes chloritization, sericitization, and carbonatation, the outboard of which is a peripheral halo dominated by chloritization (Mueller 2008). Geometrically, this zone has an inverted cone shape, 35 m in thickness, and ca. 500 m in lateral extent. The most common sulfide minerals in the “kniest” are pyrite, galena, sphalerite, chalcopyrite, and arsenopyrite.

The ore textures and mineral paragenesis in the footwall vein system and in the banded ore were modified by Variscan and post-Variscan deformational processes (Muchez and Stassen 2006). The main massive sulfide body at Rammelsberg, composed of banded polymetallic sulfides (Fig. 7a), typically includes black shale fragments corroded and partially replaced by sulfides, which is indicative of epigenetic relations between the lithologies (Fig. 7a–c). Furthermore, framboidal pyrite occurs mainly within or immediately surrounding fragments of black shale. In this last case, they are commonly recrystallized or replaced by coarse-grain polymetallic sulfides. Recrystallized pyrite and marcasite are also linked to the black shale fragments (Fig. 7b, c). All these textures are attributed to the interaction of hydrothermal fluids with the shale and therefore an epigenetic origin for at least part of the massive sulfide deposit.

Fig. 7

Polished section photographs representative of the mineralizations at Rammelsberg, Tharsis, and Draa Sfar. a Rammelsberg sphalerite-rich massive mineralization with chalcopyrite, galena, and pyrite layers. Note the occurrence of thin lenses and disrupted layers of black shales. b Massive sulfide mineralization from Rammelsberg including pyrite and marcasite phenoblasts and a fragment of black shale including framboidal pyrite granules. Note the occurrence of framboidal pyrite also within sphalerite. c Detail of the relationship between black shales and the polymetallic massive sulfide mineralization from Rammelsberg. The black shale fragment includes framboidal pyrite and organic matter remains. The muscovite crystals at the border represent recrystallization of phyllosilicates from the black shale. d Massive pyrite mineralization from Tharsis including relicts of framboidal and colloform primary textures. e Carbonated ore from Tharsis showing breccia texture. It includes diverse pyrite crystals and aggregates, replacements of pyrite by chalcopyrite, interstitial chalcopyrite in pyrite, fragments of black shale, and sphalerite matrix. A late vein of siderite with chalcopyrite and galena is also present. f Massive mineralization with pyrrhotite, sphalerite, and black shale lenses from Draa Sfar. BS black shales, OM organic matter, Ccp chalcopyrite, Gn galena, Mrc marcasite, Ms muscovite, Po pyrrhotite, Py pyrite, Py _f framboidal pyrite, Sd siderite, Sp sphalerite

At Tharsis, the footwall vein (“feeder”) zone cross-cuts the black shale of the anoxic sequence (Fig. 6b) and the siliciclastic (sandy-shaly) strata of the underlying PQ Group. Hydrothermal alteration consists mainly of chloritization and subordinate carbonatation (siderite) and silicification (Tornos et al. 2008). Mineralogically, the veins are dominated by pyrite with varying proportions of quartz, Co–As–Fe sulfides (pyrrhotite, arsenopyrite, alloclasite, and cobaltite), sulfosalts of the joseite-tetradimite group, and native gold (Marcoux et al. 1996). Textural modification of the veins during Variscan deformation resulted in the remobilization of the most ductile phases, including chalcopyrite and gold (Marignac et al. 2003). Silicified black shale fragments, locally corroded, are commonly included in the massive sulfides and are interpreted here as evidence of epigenetic relationships.

Massive sulfide mineralization at Tharsis is mineralogically uniform and is characterized by banded textures defined by variations in the grain size of pyrite. This banded facies is commonly accompanied by framboidal textures (Fig. 7d) and by breccias (Fig. 7e) interpreted to reflect the collapse and fragmentation of seafloor chimneys and sulfide crusts (Tornos et al. 2008).

At Draa, Sfar the vein system of the feeder zone is poorly developed, consisting of sparse veins of pyrrhotite and chlorite (Belkabir et al. 2008; Moreno et al. 2008). Beneath Tazakourt, the southernmost orebody in the district, a chloritized feeder zone affects the footwall black shales and the underlying volcanic rocks and becomes more intense toward the immediate footwall of the massive sulfides (Belkabir et al. 2008). Chloritized and corroded black shale lenses devoid of original sedimentary or diagenetic textures form inclusions in the massive sulfides, producing pseudo-banded textures (Fig. 6d).

The evidence for syngenetic mineralization at Tharsis is mostly related to the occurrences of soft fragments of black shale included within banded massive sulfides. These typically consist of unsorted fragments of silicified shale and chert that occur in brecciated massive sulfide and banded shale–pyrite layers. The banded lithofacies generally contains angular fragments of shale showing sedimentary structures such as graded bedding and millimeter- to centimeter-scale synsedimentary faults (Fig. 6c). Its contact with the host black shales is characterized by a thin layer of siderite, a mineral also present as cement in brecciated banded ore. All of these structures suggest that the banded lithofacies represent synsedimentary reworking of shale and sulfides. In fact, the Tharsis banded ore is generally interpreted as a debris-flow deposit (Tornos et al. 2008), although an occurrence of the siderite layer indicates that hydrothermal activity persisted after sulfide deposition.

At Rammelsberg, the recrystallization of massive sulfides inhibited the preservation of primary textures and structures (Newhouse and Flaherty 1930; Ramdohr 1953; Mueller 2008). Nevertheless, banded lithofacies of alternating shale–pyrite layers occur at different scales at the margins of the massive sulfide bodies (Mueller 2008). Such lithofacies are typically disrupted by pervasive synsedimentary faults (Fig. 6a) that imply the syngenetic formation of at least parts of the orebody.

At Draa Sfar, inferred syngenetic relationships between black shale and massive sulfides are commonly obscured by intense tectonic deformation (Figs. 6d and 7f), although finely banded lithofacies of alternating shale–pyrrhotite layers occur locally, most commonly in the immediate hanging-wall of the massive sulfides (Moreno et al. 2008). Lenticular fragments of black shale are commonly included within the massive ore (Belkabir et al. 2008), but their relationships are difficult to discern. The host black shale is also devoid of textural or structural features indicative of sedimentary or diagenetic sulfides, with the exception of the presumably diagenetic nodular pyrrhotite and minor chalcopyrite, reported from the footwall shales (Belkabir et al. 2008; Marcoux et al. 2008).

Black shales

The black shales of the Sarhlef Series at Draa Sfar, the anoxic sequence of the Volcanosedimentary Complex at Tharsis, and the Wissenbach Shales at Rammelsberg are thought to have been deposited in young sub-basins characterized by tectonic, magmatic, and landscape instability.

The Rammelsberg and Draa Sfar black shales systematically alternate with the thin layers of fine-grained sandstone. Both the sand/lutite ratio < <1 and the sedimentary structures are indicative of diluted turbidity currents. The black shales at Tharsis are virtually devoid of intercalated sandstones.

Evidence of bioturbation has not been found at any of the study sites among the shales embracing the massive sulfides. Therefore, indices of bioturbation degree (Bottjer and Droser 1991; Sageman et al. 1991; Savrda et al. 1991) could not be employed. In all three deposits, structured organic matter includes marine and terrestrial particles indicative of high continental influence, presumably in the vicinity of emergent land masses. According to all the factors expressed above, the sedimentation can be considered of hemipelagic–pelagic nature (Stow et al. 2001).

Geochemistry

Mudstone compositions depend on several interdependent factors related to processes that interact during the entire sedimentary cycle, i.e., weathering, transport, sedimentation, and burial (e.g., Johnson 1993). In geothermally anomalous environments, like those in which the massive sulfides formed, postsedimentary transformations related to water/rock interaction can noticeably change the primary geochemical signatures. Table 1 shows the compositional averages and standard deviations of the analyzed samples of the back shales studied here. The individual data are available as “Electronic Supplementary Material”.

Table 1 Compositional average and standard deviation of the analytical values from non-mineralized black shales at Rammelsberg, Tharsis, and Draa Sfar

Major elements

The black shales in all three sites studied show important variations in major oxide elements, especially in Fe, Mg, Ca, Na, Mn, and P. These variations can be related with processes common in sedimentary environments such as sorting of detrital particles and mixing with authigenic sediments, including carbonates, phosphates, biogenic silica, and organic matter. For comparison with the selected standard PAAS (Post Archean Australian Shale of Taylor and McLennan 1985) and between the three areas studied, the analytical data have been normalized to the “relatively immobile” Ti using molar element ratios (Fig. 8). This procedure minimizes the effects of sedimentary sorting. The average values of the most conservative elements (Si, Al, K) are close to one in each studied site. Elements that normally enter the crystal structure of carbonates (Ca, Mg, Fe, Mn) show more erratic behavior. At Tharsis, the average and range of normalized values for Ca, Mn, and, in minor degree, Mg are depleted relative to PAAS, whereas the normalized average for Fe is mostly near one. For Rammelsberg and Draa Sfar, normalized Ca values show strong positive and negative anomalies, whereas the average composition is slightly higher than that of PAAS standard. Data for Mn display a similar behavior although with a more limited absolute range for Draa Sfar but not for Rammelsberg that locally contains relatively high Mn concentrations. Fe and Mg are more conservative, with normalized averages generally close to one and limited ranges, although with small positive and negative anomalies for Mg in Rammelsberg and Tharsis, respectively.

Fig. 8

PAAS-normalized plots of major elements from the black shales samples of Rammelsberg, Tharsis, and Draa Sfar. Solid lines represent maximum and minimum values. Dashed lines indicate average values

The average Na content in the three cases is lower than that of PAAS, especially at Tharsis where all normalized values are clearly depleted. At Rammelsberg, normalized Na values are also less than one, but the average is on the same order of magnitude as PAAS. At Draa Sfar, normalized Na ranges from small positive to negative values; the normalized average is close to one. Positive values reflect unaltered albite-bearing black shale, whereas the negative ones could correspond to hydrothermally altered shales (Moreno et al. 2008). The normalized P average values at all three sites are broadly comparable to those of PAAS. In detail, Draa Sfar displays almost identical values, whereas Rammelsberg and Tharsis show lower average values and larger ranges.

The bulk compositions of the analyzed samples recalculated on a carbonate-free basis and plotted on an A–CN–K diagram (Nesbitt 2003) are indicative of shales moderately enriched in alumina and depleted in Ca and Na (Fig. 9a). K/Al ratios resemble those of dioctahedral phyllosilicates such as illite and muscovite. Only four samples stratigraphically far from the Draa Sfar mineralized horizon have values similar to PAAS.

Fig. 9

a Ternary diagram of carbonate-free black shale samples from Rammelsberg, Tharsis, and Draa Sfar in the A-CN-K space (adapted from Nesbitt 2003). The apices A (Al₂O₃), CN (CaO + Na₂O), and K (K₂O) are labeled on a molar basis. b A-CNK-FM ternary diagram (adapted from Nesbitt 2003) showing the main alteration trend of carbonate-free black shale data, involving the illite–muscovite and the chlorite poles. The apices A (Al₂O₃), CNK (CaO + Na₂O + K₂O), and FM (Fe₂O₃tot + MgO) are labeled on a molar basis. CC average continental crust, AS average shale, PAAS post-Archean Australian Shale, Pl plagioclase, Ka kaolinite, Chl chlorite, Ms muscovite, Bi biotite, FdK K-feldspar, Fd feldspars, Sm Smectite

Figure 9b shows the same data plotted on an A–CNK–FM diagram (Nesbitt 2003). The data for the Tharsis shale samples define a linear trend with extrapolated endmembers at theoretical compositions of illite/muscovite and chlorite, suggesting some Ca and Na depletion associated to hydrothermal alteration. The carbonate-free samples from Rammelsberg describe the same trend, although slightly displaced toward the CNK apex. Most of the Draa Sfar shales follow the Tharsis and Rammelsberg trends, but the same four weakly altered samples mentioned above also cluster about the PAAS composition on this diagram.

Carbon and sufur

The organic carbon content (C_org) of the analyzed shales is generally low (mostly <2 wt.%), although some samples from Tharsis have values close to 4 wt.%. The inorganic carbon content is very low at Tharsis and Draa Sfar, but the total carbon (C_tot), implicitly dominated by carbonates, is high for many samples from Rammelsberg. In terms of C_tot vs C_org (Fig. 10a), the Tharsis samples define a trend that passes through the origin, indicating that practically the entire carbon budget comes from organic sources. Among the Draa Sfar samples, all but one are aligned on the same trend, although typically at lower C_org values. In contrast, the Rammelsberg data display a trend subparallel to the y-axis, with the C_org contents of about 0.5 wt.% and C_tot values ranging from 0.5 to nearly 2.5 wt.%. On the C_tot vs S_tot diagram (Fig. 10b), these samples are positively correlated and broadly follow the trend of data for Middle Devonian normal marine shales proposed by Raiswell and Berner (1986). Petrographic analysis confirms carbonate cement in most of these shales that, according to Sperling (1986) and Sperling and Walcher (1990), is primarily Mn-rich ankerite.

Fig. 10

a Bivariate plot of the black shale samples from Rammelsberg, Tharsis, and Draa Sfar in the diagram C_tot vs C_org. b Bivariate plot of the black shale samples from Rammelsberg in the diagram C_tot vs S_tot

The ratio S_tot/C_org has been widely used as a proxy for the redox evaluation of depositional environments of modern and ancient black shales (Berner 1982; Dean and Arthur 1989). On a S_tot vs C_org plot (Fig. 11), data for the Rammelsberg black shales define a cluster with the low S_tot and C_org values displaying an ill-defined correlation. The regression trendline, steeper than that for normal marine environments at Middle Devonian times, probably reflects the conversion of organic C to carbonate C, from an original S_tot–C_org distribution that was similar to normal marine environments at that time. The Tharsis samples show a complex distribution: those with C_org close to 0.5 wt.% follow a trend parallel to the y-axis, whereas others plot close to the regression line for Upper Devonian–Lower Carboniferous normal, non-euxinic marine environments; the remainder are randomly scattered. All but two Draa Sfar samples follow Berner and Raiswell’s (1983) line for normal marine environments of the same age, although with the sulfur content slightly below the covariation line.

Fig. 11

Bivariate diagrams of S_tot vs C_org for black shale samples from Rammelsberg, Tharsis, and Draa Sfar. Gray fields indicate S_tot/C_org ratios for Middle Devonian, Upper Devonian–Lower Carboniferous, and Lower Carboniferous normal marine environments after Raiswell and Berner (1986) appropriate to the diagrams of Rammelsberg, Tharsis, and Draa Sfar, respectively. Dashed line represents the regression trendline for the Rammelsberg samples

Trace elements

Shale samples were analyzed for 32 trace elements according to the previously described method. In most of the samples, Se, Au, and Ag showed values below the detection limit. Figure 12 shows the Ti-normalized results of non-mineralized samples presented as ranges and averages relative to PAAS. The Draa Sfar black shales have average values similar to PAAS, whereas those from Rammelsberg and Tharsis show higher differences. The common features in all three cases are the distinct negative Cs anomaly and the high dispersion of V, Sr, Mo, Co, Ni, U, Pb, and Sn values. In general, the most conservative elements (i.e., Sc, Y, Th, Zr, Hf, Nb) show average values similar to PAAS. However, Sc, Y, and Th values have a slight positive anomaly in the three cases, and Zr average values depict a small negative anomaly in Rammelsberg and Tharsis black shales.

Fig. 12

PAAS-normalized spider-like diagrams for the Rammelsberg, Tharsis, and Draa Sfar samples. Solid lines represent maximum and minimum values. Dashed lines indicate average values

The concentration of large ion lithophile elements (Rb, Ba, Pb, Cs) is characterized by a generally strong negative Cs anomaly and patterns of Rb, Ba, and Pb being clearly different on each case. In Draa Sfar, the three elements have average values similar to PAAS, with ranges that are very wide for Pb and very narrow for Rb and Ba. The Tharsis black shales show a gently positive anomaly in Rb and Ba values that is more pronounced for Pb. The range is very wide for Pb and relatively narrow for Rb and Ba. In Rammelsberg, the Ba average value is similar to PAAS, whereas Rb and Pb show positive anomalies. As in Tharsis, the ranges are very wide for Pb and narrower for Rb and Ba. Although we have selected samples for this study with no evidence of hydrothermal alteration, the wide range shown by Pb in all three cases could be related to this alteration. Elements whose behavior is related in some way with the redox characteristics of the environment (i.e., V, Mo, Co, Ni, Cr, U, Sn) show values much different among the three districts.

Clay- and carbonate-hosted elements

Those elements commonly linked with the structure of clay phyllosilicates (i.e., Al, K, Ba, and Rb) are strongly conservative and show, in all three studied cases, limited ranges and average values close to PAAS (Figs. 8 and 12). An exception to this pattern is the behavior of Cs, which has strong negative anomalies in the black shales. Cesium, as well as Rb, tends to concentrate in clay material during sedimentary processes but, unlike Rb, the high ionic ratio of Cs severely precludes its inclusion in the structure of most phyllosolicates, with the exception of certain smectites (Liu et al. 2008 and references herein). A possible explanation for Cs anomalies could be that, during diagenetic maturation of the shales, the Cs partially escaped from the system, as dissolved species, due to its low ionization potential.

Strontium enters mainly the structure of carbonates and resembles the geochemical behavior of Ca, which has been previously discussed.

Redox-sensitive elements

Elements such as Ba, V, Mo, Co, Ni, Cr, Pb, U, and Re and their ratios have been used to evaluate redox conditions in marine waters, pore fluids, sediments, and sedimentary rocks (e.g., Bishop 1988; McManus et al. 1998; Hoffman et al. 1998; Algeo and Maynard 2004; Sageman and Lyons 2004; Rimmer 2004; Ross and Bustin 2009 and references therein). In the studied samples, the Ti-normalized values of some of these redox-sensitive elements are not anomalous relative to those of the standard, whereas Ba, U, Mo, Sn, V, and Pb show significant anomalies (Fig. 12). The Rammelsberg black shales have slightly positive anomalies in V and Ni and a strongly negative anomaly in Mo. The other elements (Co, Cr, U, Sn) show average values similar to PAAS and relatively narrow ranges. At Tharsis, those elements with a tendency to concentrate in oxygen-depleted sedimentary environments (V, Mo, and U) are characterized by strong positive anomalies. Co, Ni, and Cr values are very similar to PAAS, although Co shows a slightly negative anomaly. The black shales at Draa Sfar have average V and Mo values equivalent to the PAAS and small negative anomalies for the other redox-sensitive elements. Tin behaves differently in the three sites studied. It has a positive anomaly in Tharsis, negative in Draa Sfar, and values similar to PAAS in Rammelsberg.

Discussion

Sediment provenance

Sediment sources generally accord geochemically with their corresponding detrital deposits, especially if these are fine-grained. This correspondence can be demonstrated by using immobile trace elements (Bhatia and Crook 1986; McLennan 1989; McLennan et al. 1993). At Rammelsberg, Tharsis, and Draa Sfar, the concentrations of immobile trace elements are similar among the three areas and to PAAS. The entire data set, plotted on the Th–Sc–Zr provenance diagram of Bhatia and Crook (1986), defines a compact cluster between fields for active continental margins and continental magmatic arcs (Fig. 13). In detail, the Rammelsberg samples fall within the active margin field, whereas samples from Tharsis and Draa Sfar straddle the boundary between these two fields.

Fig. 13

Black shale samples from Rammelsberg, Tharsis, and Draa Sfar plotted on the Th–Sc–Zr provenance ternary diagram of Bhatia and Crook (1986)

The use of major elements in the discriminant function diagrams of Roser and Korsch (1988) suggests provenance of detrital material in the studied shales from sources dominated by magmatic and/or sedimentary rocks (Fig. 14). Samples from Rammelsberg probably derive from reworked sediments, whereas at Tharsis and Draa Sfar felsic and mafic volcanic rocks are also likely sources for the inorganic components of the black shales. Environmental conditions in the source areas seem to be partially obscured by overprinting weak hydrothermal alteration, which affected the black shales after deposition. Data plotted on the A–CN–K diagram of Nesbitt (2003) suggest significant feldspar illitization and selective depletion in Ca and Na, with Chemical Index of Alteration values close to 80 (Fig. 9a). Four Draa Sfar samples, compositionally equivalent to PAAS, are characterized by stable albite and very weak hydrothermal alteration. This behavior, also detected at Rammelsberg and Tharsis, could reflect weathering processes dominated by the moderate depletion in Na and Ca and preservation of Al and K as conservative elements. Nevertheless, the A–CN–K diagram is inadequate enough for discriminating whether the depletion in Na and Ca was related to leaching processes during weathering or to regional hydrothermal alteration.

Fig. 14

Black shale samples from Rammelsberg, Tharsis, and Draa Sfar plotted on the discriminant function diagrams of Roser and Korsch (1988) based on: a major elements and b major element ratios

On the A–CNK–FM diagram (Fig. 9b), all analyzed shales, except the aforementioned four Draa Sfar samples, follow the theoretical join between the compositions of illite/muscovite and chlorite. This trend suggests that the hydrothermal alteration produced, especially at Tharsis, minor geochemical variations. On this diagram, data for Tharsis are dispersed, whereas those for Rammelsberg and Draa Sfar are more clustered. The weak chloritization of Draa Sfar shales is interpreted to have occurred mostly in a closed system, and the Fe enrichment should be mainly associated with the precipitation of pyrrhotite, which has been interpreted as primary in Draa Sfar (Marcoux et al. 2008). In summary, data plotted in the A–CN–K and A–CNK–FM diagrams (Fig. 9a, b) suggest weathering effects dominated mainly by the loss of Na and Ca (and minor K), which is likely related to illitization of feldspar. At Tharsis, samples plot along a trend joining chlorite and muscovite (sericite) poles, suggesting minor modifications related to postsedimentary processes, including weak hydrothermal alteration.

Redox state of depositional environments

Evaluation of redox conditions during sedimentation and early diagenesis of ancient shales has generally been done by comparison with models for the behavior of redox-sensitive elements in modern sedimentary environments (Sageman and Lyons 2004 and references therein). For marine environments dominated by mud deposition, it can be assumed that the redox conditions of the water–sediment system are controlled by the availability and composition of organic matter and by the consumption of C_org during bacterial sulfate reduction. Sulfur reduced to S²⁻ reacts with Fe-rich detrital minerals to generate pyrite. These processes oxidize organic matter via bacterial metabolism, sulfur reduction from sulfate to sulfide, and, in places, iron reduction from Fe³⁺ (present in oxide minerals such as hematite or magnetite) to Fe²⁺. In marine basins with oxic bottom waters, the major limiting factor for diagenetic pyrite formation is the availability of organic matter because, generally, the high sulfate concentrations in Phanerozoic oceans did not restrict the sulfate supply for bacterial reduction (Sageman and Lyons 2004). Under these conditions, S_tot and C_org are typically covariant with the regression passing through the origin (Berner 1970, 1984).

The S_tot/C_org ratios of the studied shales are characterized by relatively high S_tot and low C_org contents (Fig. 11). According to Sageman and Lyons (2004), such distributions can be explained in three different ways: (1) hydrothermal or low-temperature sulfidation, (2) thermal or metamorphic C_org loss, and (3) euxinic (sulfidic) marine deposition. The euxinic scenario is classically associated with covariant C_org and S_tot with a positive S-intercept corresponding to sulfide additions from the water column (Raiswell and Berner 1985). The Tharsis black shales are grouped into two categories, most have C_org >1 wt.% and S_tot/C_org ratios equivalent to those of Devonian–Carboniferous normal marine sediments (Raiswell and Berner 1986). Samples with C_org <1 wt.%, chiefly represented by pyritic hanging-wall shales, are aligned parallel to the y-axis with S_tot contents of up to 5 wt.%. High-sulfide sulfur and low C_org characterize the late diagenetic or epigenetic addition of HS⁻ (Leventhal 1995).

Most of the Rammelsberg black shales have S_tot/C_org ratios higher than those common for Middle Devonian normal marine environments (Fig. 11). The broad positive correlation of S_tot and C_org values illustrated in this diagram suggests that organic matter oxidation was involved in pyrite generation during early diagenesis (Tourtelot 1979; Martin and Sayles 2004). However, the positive C_org intercept and negative S_tot intercept of the regression trendline suggest an unusual scenario. The Rammelsberg samples show C_tot/S_tot ratios similar to those for the Middle Devonian normal marine shales (Fig. 10b). This relationship suggests that the carbon in the carbonates was derived mainly from reactive organic matter and that the original C_org contents should correspond to the measured C_tot values. This involves that sulfate was reduced (and kerogen oxidized) via a thermochemical mechanism to form carbonate + sulfide. This interpretation is supported by the heavy S-isotopic data from sulfides at Rammelsberg (Anger et al. 1966). This process could occur in the water, in the unconsolidated sediment, or after lithification. Barite abundance in the massive sulfide ore suggests that base metals and Ba were transported by reduced and acidic mineralized brines (Lydon 1983; Cooke et al. 2000). Massive sulfide deposition might occur by dilution and temperature decrease caused by mixing with seawater and/or pore fluids and by interaction with bacteriogenic-reduced sulfur reservoirs, including pyrite-rich shales and H₂S within the pore fluids (Cooke et al. 2000).

All but one sample from Draa Sfar are aligned slightly below the line for normal Lower Carboniferous marine environments (Fig. 11). The relatively high C_org content suggests isolation due to the high sedimentation rate that characterizes the Draa Sfar basin.

Relationships between whole rock S and Fe are shown in Fig. 15. In general, correlation between these elements is poor or absent in the shales associated with all three deposits. Rammelsberg and Draa Sfar shales, very rich in Fe, plot far from the pyrite and pyrrhotite stoichiometric trends. In normal marine environments, the presence of non-sulfide iron can reflect the existence of non-reactive iron-bearing minerals and/or low availability of reduced sulfur during sedimentation and diagenesis. Such low S availability may inhibit the consumption of all available iron for pyrite formation. The occurrence of iron-bearing carbonates in the shales of both districts (Large and Walcher 1999; Marcoux et al. 2008) suggests that reactive iron was available, and therefore sulfide formation was likely limited by S availability. This relative deficiency of reduced sulfur may have arisen because of the original low availability of reactive organic matter or by isolation from sulfate-rich marine waters, e.g., during rapid sediment burial. At Tharsis, most shales have very high Fe concentrations, but five samples from the footwall of the Filón Norte orebody show values close to stoichiometric pyrite. This indicates that all of the Fe originally available is now incorporated in pyrite and that, at Filón Norte, environmental conditions were strictly anoxic, which is also supported by other geochemical proxies discussed below. The excess iron present in the other Tharsis samples can be explained by a rapid burial rate or by iron introduction by post-depositional hydrothermal processes, as appears to have occurred at Rammelsberg and Draa Sfar. The relatively low Fe content in the analyzed black shales can be explained in terms of isolation due to rapid burial.

Fig. 15

Black shales from Rammelsberg, Tharsis, and Draa Sfar plotted on the bivariate diagrams S_tot vs Fe (Leventhal 1979)

At Tharsis, the wide range of bulk C_org content and Fe sulfidation demands a more complex interpretation. The more C_org-rich samples have low Fe sulfidation, have an exceptional C–S systematics, and were probably deposited under normal marine conditions. However, most of the low-C_org samples have high degrees of sulfidation and were likely deposited during anoxic conditions. Their lower C_org content could reflect either the destruction of organic carbon during thermochemical sulfate reduction, hydrothermal alteration, or rapid sedimentation into the anoxic environment.

Other elements used for redox evaluation of sedimentary environments are those whose chemical behavior depends on their oxidation state and affinity for organic matter and/or insoluble sulfide species, i.e., Mo, U, V, Cr, Fe, Mn, Ni, Co, Pb, Cu, Zn, Cd (Calvert and Pedersen 1993; Arthur and Sageman 1994; Morford and Emerson 1999; Algeo and Maynard 2004; Sageman and Lyons 2004; Brumsack 2006; Tribovillard et al. 2006). Various threshold values for these elements and ratios have been used to distinguish the different depositional environments (Hatch and Leventhal 1992; Calvert and Pedersen 1993; Jones and Manning 1994; Hoffman et al. 1998; Sageman et al. 2003; Rimmer et al. 2004; Algeo and Tribovillard 2009; Lyons et al. 2009). In this study, we used Mo, U, V, Cr, and Ni as the ratios Mo/Cr, Mo/U, V/Cr, and V/(V + Ni) because they appear to have remained unaffected by hydrothermal processes.

U/Th is a commonly used proxy (Wignall and Myers 1988). In the sedimentary environment, Th is generally transported in the detrital fraction and resides mainly in monazite and zircon, whereas U shows affinity for organic matter (Wignall and Myers 1988; Klinkhammer and Palmer 1991). Consequently, the Th/U ratio in organic-rich muds is characteristically low (Anderson et al. 1989; Arthur and Sageman 1994). This geochemical behavior was used by Wignall and Myers (1988) to define the concept of authigenic uranium as the excess of uranium respective to common values in normal mudstones (U_auth = U−Th/3). This parameter is widely used to approximate the redox conditions of starved sedimentary environments but has some limitations when applied to basins with high sedimentation rates. In such basins, Th/U ratios can be anomalously high even under euxinic conditions (Arthur and Sageman 1994) as appear to be the case in the present study. On the U_auth/C_org diagram (Fig. 16), all Rammelsberg and Draa Sfar samples, and some samples from Tharsis, display negative U_auth values, but an important group of samples from the Filón Norte footwall, at Tharsis, have values >1 (locally close to eight), indicating anoxic conditions. The negative values can be reasonably explained by the high sedimentation rates we have calculated based on stratigraphic thickness and age constraints (8–14 cm/ka), whereas the high U_auth values for the Filón Norte footwall point to the development of anoxic conditions before sulfide mineralization commenced. We infer from the range of shale compositions and evidence of basin anoxia that the Tharsis deposit formed in a dynamic and restricted marine environment. However, in order to make this interpretation compatible with the high bulk sedimentation rates, it is necessary to take into account that sedimentation rate and velocity are different parameters. Sedimentation velocity during the deposition of the footwall to Tharsis must have been substantially lower than that during the deposition of the hanging-wall to the massive sulfides.

Fig. 16

Bivariate plot of the black shale samples from Rammelsberg, Tharsis, and Draa Sfar in the diagram U_auth vs C_org

The aqueous geochemistry of U and Mo provides new constraints to the assessment of the environmental redox conditions. Both elements tend to concentrate in strongly euxinic sediments (Morford and Emerson 1999; Tribovillard et al. 2006) but, in detail, their aqueous geochemistry shows some significant differences. Consequently, their relative abundance can be used for the evaluation of redox conditions of sedimentary environments (Algeo and Tribovillard 2009).

Under oxic–suboxic conditions in marine waters, U⁶⁺ behaves conservatively in the form of uranyl–carbonate complexes (Langmuir 1978; Calvert and Pedersen 1993; Algeo and Tribovillard 2009). Under anoxic conditions, U⁶⁺ is reduced to U⁴⁺ in the form of uranyl (UO²⁺) soluble species or less soluble uranous fluoride complexes (Algeo and Tribovillard 2009). The reduction from U⁶⁺ to U⁴⁺ occurs preferably below the sediment–water interface (Anderson et al. 1989; McManus et al. 2005; Algeo and Tribovillard 2009) and seems to be possibly favored by the catalytic action of enzymes produced by sulfate- and Fe-reducing bacteria (Morford et al. 2009). In the reduced form, U can be taken up by sediments as organometallic compounds or via precipitation as UO₂ (Klinkhammer and Palmer 1991; Zheng et al. 2002).

Molybdenum, like U, tends to concentrate in strongly euxinic sediments (Helz et al. 1996; Erickson and Helz 2000; Lyons et al. 2009) but in a somewhat different manner (Tribovillard et al. 2006). Under oxic conditions, Mo occurs in seawater as the largely conservative molybdate oxyanion (MoO ²⁻₄ ) (Morford and Emerson 1999; Algeo and Tribovillard 2009; Lyons et al. 2009). Under anoxic–euxinic conditions, Mo becomes activated at high H₂S concentrations, thus facilitating the transformation of the molybdate oxyanion into reactive oxy-thiomolybdate species (MoO _x S²⁻ _(4−x), x = 0 to 3) (Helz et al. 1996; Erickson and Helz 2000; Algeo and Tribovillard 2009; Lyons et al. 2009). Under these conditions, Mo becomes active and is fixed by sulfide species and/or organic matter (Helz et al. 1996; Tribovillard et al. 2004; Lyons et al. 2009). The strong Mo–C_org covariation detected in modern euxinic environments and many Phanerozoic black shales suggests that organic matter is the main agent for fixing Mo in euxinic sediments (Algeo and Lyons 2006; Lyons et al. 2009). Such fixation is more likely produced via diffusion from the seawater to the pore fluids of those sediments located near the sediment–water interface (Morford et al. 2009).

The Mo–U covariation has been recently proposed as proxy for redox conditions of sedimentary environments (Algeo and Tribovillard 2009). The effects of the detrital supply in the mining districts here analyzed can be minimized by double normalization to an essentially immobile element (e.g., Al or Ti) and to a standard (PAAS). This procedure permits the assessment of the “enrichment factors” (EF) of the amount of U and Mo that, as described above, has been extracted from seawater by means of oxidation–reduction processes. The enrichment factor is defined as the ratio between a given element and Al divided by the same ratio but using the standard X _EF = [(1/Al)_sample/(X/Al)_PAAS] (Algeo and Tribovillard 2009). Figure 17 illustrates the U_EF and Mo_EF covariation at Rammelsberg, Tharsis, and Draa Sfar. The black shales at Rammelsberg and Draa Sfar are distinguished by anomalously low EF values, generally below one, and the absence of major differences in EF between footwall and hanging-wall samples, with the exception of one sample from Draa Sfar with Mo_EF of 3.1. At Tharsis, by contrast, most of the footwall samples show Mo_EF and U_EF values greater than 1. Particularly significant are the extremely high Mo_EF values (up to 32) of some of the samples immediately below the Filón Norte orebody. According to the model proposed by Algeo and Tribovillard (2009) (Fig. 5), the low absolute values together with the absence of covariation between Mo_EF and U_EF suggest in Rammelsberg and Draa Sfar a depositional environment characterized by oxic bottom water, with the redox boundary below the sediment–water interface. At Tharsis, the Mo_EF and U_EF values suggest conditions that evolved from suboxic to strongly anoxic prior to massive sulfide deposition. After sulfide generation, the depositional environment recovered progressively to oxic conditions, which occasionally turned to anoxic, as indicated by Mo_EF values of up to 8 and U_EF values of up to 3 from the hanging-wall shales at Tharsis. This is well illustrated in Fig. 18, which shows the variation in Mo_EF and U_EF along one of the sampling profiles covering footwall and hanging-wall black shales in the Filón Norte open pit.

Fig. 17

Diagram U_EF vs Mo_EF for non-mineralized black shale samples from Rammelsberg, Tharsis, and Draa Sfar. Cluster at y-axis includes footwall and hanging-wall samples from Rammelsberg and Draa Sfar

Fig. 18

Sketch geological section for the eastern side of the Filón Norte open-pit (Tharsis) showing Mo and U enrichment factors

In general, anoxic environments are characterized by high V/Cr ratios and V/(V + Ni) values between 0.5 and 0.9, because of the disparate behavior of V, Ni, and Cr, during redox processes in marine environments characterized by fine-grained detritic sedimentation (Calvert and Pedersen 1993; Jones and Manning 1994; Hoffman et al. 1998; Algeo and Maynard 2004).

On the V/Cr vs V/(V + Ni) diagram (Fig. 19), most of the analyzed samples plot within the range of V/Cr ratios proposed by Jones and Manning (1994) for oxic bottom waters. According to these authors, V/Cr ratios close to one are indicative of a redox boundary near the sediment–water interface. However, all but one sample have V/Cr ratios characteristic of anoxic conditions in the sense of Hoffman et al. (1998). Regarding V/(V + Ni) relations, most samples are representative of anoxic environments (see Hatch and Leventhal 1992; Hoffman et al. 1998). A subset of the Tharsis black shales and a single Rammelsberg sample fall within the field of euxinic environments.

Fig. 19

Bivariate plot of the black shale samples from Rammelsberg, Tharsis, and Draa Sfar in the diagram V/Cr vs V/(V + Ni). Limits of environmental conditions after Jones and Manning (1994) and Hoffman et al. (1998)

Relationships between black shales and sulfides

Conceptually, the footwall rocks to the massive sulfide deposits reveal the environmental conditions that prevailed immediately prior to exhalative sulfide deposition, whereas hanging-wall rocks record the waning stages of mineralization and preservational conditions (Franklin et al. 2005). In the three cases studied, fine-grained sedimentation occurred before, during, and after sulfide generation. Coeval deposition of shales and sulfides is inferred because of the presence of intercalated shale and sulfide that is abundant at Rammelsberg (Mueller 2008), common at Tharsis (Tornos et al. 2008), and present but less frequent at Draa Sfar (Moreno et al. 2008). The presence of hydrothermal alteration in the hanging-wall shales at Tharsis and Draa Sfar is taken as evidence for the continuation of fine-grained sedimentation immediately after mineralization in these systems. At Draa Sfar, this is analogous to the hydrothermal activity that affected the shales below the massive sulfides, whereas at Tharsis, 1–5 m of the immediate hanging-wall black shales host a distinctive alteration facies characterized by partial chloritization with disseminated pyrite. The top of the post-mineral altered sequence at Tharsis is marked by a gray chert bed 0.2–5 m thick, continuous at the deposit scale, that also has disseminated pyrite.

Mineralogically, the shales above and below the studied deposits are characterized by variable proportions of C_org and S but have similar concentrations of most major and trace elements, which suggests basin stability.

Taking into consideration all of the geochemical data discussed above, together with textural and structural relations between the black shales and massive sulfides (Appendix 1 and references therein), it is possible to assert that some features of these sulfide deposits were substantially influenced by bacterial activity. At Rammelsberg and Tharsis, bacterial involvement is supported by available sulfur isotope data that also suggest contributions from a hydrothermal S source. In the case of Rammelsberg, different δ ³⁴S values of sulfide minerals depend on their mutual paragenesis (Anger et al. 1966; Nielsen 1985; Eldridge et al. 1988; Mueller 2008). The δ ³⁴S values of pyrite range from −15‰ to 20‰, indicating that S was derived at least in part from the biogenic reduction of seawater sulfate in a locally closed system (Large and Walcher 1999). In contrast, values measured in base metal sulfides yielded a narrower and isotopically heavier range, from 5‰ to 20‰, likely dominated by hydrothermal S sources (Nielsen 1985). At Tharsis, biogenically reduced marine sulfate, characterized by δ ³⁴S values between −33.2‰ and 4.1‰, has been suggested as the main S contribution in the massive sulfides (Mitsuno et al. 1986; Kase et al. 1990; Velasco et al. 1998; Tornos et al. 2008). Nevertheless, δ ³⁴S values for the stockwork sulfides range from −4.5‰ to 1.9‰ (Mitsuno et al. 1986; Tornos et al. 2008) and suggest a magmatic S source. This focused hydrothermal activity in the feeder zone apparently occurred following syngenetic mineralization, introducing isotopically heavier S into the system and producing replacement textures.

Conclusions

Rammelsberg, Tharsis, and Draa Sfar are three deposits located along the Variscan belt that exhibit comparable but diachronous paleogeographic features and tectonic setting.

The three deposits are located in young basins that originated due to crustal thinning during the early rifting stage of the Variscan Cycle in each of these zones. This tectonism favored the fragmentation of formerly stable platforms and the subsequent development of restricted basins characterized by high sedimentation rates and geothermal gradients. Deposition of massive sulfides was controlled by sub-basin growth faults that served as feeder structures for hydrothermal fluids. The stratigraphic records of the three basins, also comparable, consist of black shale sequences with minor volcanic and subvolcanic rocks, other sedimentary rocks, and massive sulfides. Many of these rocks display strong facies and thickness variations.

The most striking differences among the stratigraphic records of the three districts concern the ore deposits rather than the host rocks. These comprise tonnage, grade, metal content, and mineralogy. The main differences between the host successions are the amount and distribution of intercalated volcanic rocks. However, these rocks never played a critical role in the genesis of the massive sulfides and can be considered simply as products of contemporaneous magmatic activity. Their relationship with the sulfide deposits is casual because the thermal engine that drove the mineralization was the deeper magmatic system, not the volcanic rocks themselves. In contrast, the black shales represent the key lithology that appears to have influenced the formation and preservation of the massive sulfide deposits.

The geochemical data of the black shale host sequences together with the textural and mineralogical features of the massive sulfides suggest that all three districts inherited geochemical patterns characteristic of active continental margins or continental magmatic arcs, with source areas dominated by bimodal magmatic and continentally derived sedimentary rocks. The modification of rocks in the hinterlands of the host basins seems to have been dominated by the loss of Na and Ca (and minor K) and by illitization of feldspar; however, post-depositional hydrothermal alteration obscures these features in some cases.

The evaluation of redox conditions in the sedimentary sequences suggests that the measured C_org contents are, in three cases, only a fraction of the original amounts. At Rammelsberg, carbon produced during the oxidation of organic matter associated with thermochemical sulfate reduction was fixed by carbonates and, consequently, measured C_tot should correspond to primary C_org. The measured C_tot content resembles that proposed for Middle Devonian normal marine shales, and there is little evidence for bottom water anoxia during black shale deposition. Inorganic proxies (i.e., U, Mo, V, Cr, Ni) suggest that the redox boundary was located beneath the sea floor, near the sediment–water interface. On the other hand, available sulfur isotope data point to biogenically mediated sulfate reduction as a primary sulfur source mainly recorded in sedimentary-diagenetic pyrite showing framboidal and colloform textures. Heavier hydrothermal sulfur constitutes the main sulfur source for the polymetallic massive sulfides. The deposition of the Rammelsberg ore deposit was influenced by the presence of biogenic-reduced sulfur fixed to the shales in pyrite.

Black shale samples from the footwall of the Filón Norte massive sulfide body, at Tharsis, exhibit geochemical features common to sediments deposited in anoxic environments. The most sensitive redox indices (i.e., Mo_EF, U_EF, U/Th, V/(V + Ni)) suggest that the basin evolved into deeply anoxic (even euxinic) conditions prior to massive sulfide deposition. The beginning of sulfide deposition possibly occurred above the sediment–water interface as a direct consequence of pyrite precipitation during bacterial sulfate reduction. Similar to Rammelsberg, the sulfur isotope data imply that the bacterially reduced sulfur is an essential component of the sulfides, but at Tharsis the sulfides show less-evolved sedimentary/diagenetic features (i.e., framboidal, colloform, or/and banded features). We therefore consider that the Tharsis mineralization occurred under euxinic conditions and was controlled by an abundance of reactive organic matter.

Geochemical data for the Draa Sfar black shales suggest an oxic water column and an anoxic sediment pile, with the redox boundary situated below the sediment–water interface. A very high sedimentation rate (>19 cm/ky) favored the burial of organic matter and sulfate reduction of pore fluids. Nevertheless, sulfide formation appears to have been limited by low sulfur availability, which, paradoxically, was also controlled by rapid sediment burial. Despite the low C_org content, low sulfur and oxygen fugacities are indicated by the prevalence of pyrrhotite as the primary iron sulfide which strongly suggests anoxic conditions within the sedimentary pile. At Draa Sfar, the role played by the black shales was also apparently related to the availability of reduced sulfur, in particular, that dissolved in pore fluids or fixed in iron sulfides.

According to the present study, black shales play three different roles in the genesis and preservation of massive sulfide deposits (Fig. 20): (a) Black shales can deplete the benthic oxygen via redox processes that involve reactive organic matter oxidation and sulfate reduction. Both processes can result in the generation of euxinic bottom conditions and pore water characterized by H₂S as the main sulfur species. Such situation seems to have occurred at Tharsis. (b) In oxic environments like Rammelsberg, in which the redox boundary is located near the sediment–water interface, the black shales can first provide reactive organic matter for the generation of biogenic pyrite and serve later as barrier that isolates the hydrothermally derived massive sulfides from the oxic sea bottom water. (c) Black shales can also work as aquicludes and thermal isolators playing a physical role in concentrating hydrothermal fluids in permeable levels (Pratt and Warner 2000). In this case, formation of the massive sulfides occurs without contact or influence from the oxic bottom water. This could have been the case at Draa Sfar.

Fig. 20

Proposed models showing the disparate environmental conditions and the relationship between black shales and massive sulfides at Rammelsberg, Tharsis, and Draa Sfar. Gray and black colors represent anoxic conditions; light blue and yellow colors denote oxic conditions. Not to scale

However, the ultimate cause controlling the deposition of black shales and massive sulfides was the tectono-sedimentary evolution of each basin. This evolution promoted the generation of environmental conditions suitable for synchronous deposition and preservation of both lithologies.

Appendix 1

Table 2 Principal characteristics of the Rammelsberg, Tharsis, and Draa Sfar ore deposits