Introduction

We are glad that our recent papers (Höhn et al. 2020, 2021) sparked interest in resolving the controversy about the metallogenesis of the Aggeneys-Gamsberg ore district. The question to what extent the four base metal sulfide deposits there were subject to pre-Klondikean oxidation has the potential to be of significance also for our understanding of similar deposits elsewhere and, therefore, we are grateful for the opportunity to discuss this question further. The discussion by Cawood et al. (this issue) raises several points, some of which may be due to misunderstandings, others illustrate that we remain far from a definite genetic model that explains all data and observations accumulated over many years of research on this ore district. Specific issues addressed by Cawood et al. (this issue) are as follows.

Issue 1: Unconformity

We agree on the characteristics of the unconformity between the Hotson and the Koeris Formation as summarized by Cawood et al. (this issue) but disagree with their interpretation. The position of the unconformity in different stratigraphic levels and its channelized nature are not only in agreement with the oxidation-resulfidation hypothesis formulated by Höhn et al. (2020) but constitute the so far best possible explanation for variable intensities of oxidation and resulfidation in the Aggeneys-Gamsberg ore district (Höhn et al. 2021). Cawood et al. (this issue) claim that we stated “complete weathering at surface”, but such a statement was nowhere mentioned or even inferred by us.

We also agree with Cawood et al. (this issue) on the differences in thickness of rocks in the hanging wall of the various sulfide-enrichment zones but cannot follow their argument, why these should speak against the proposed oxidation at the time of erosion and peneplanation as indicated by the unconformity. For example, the metapelitic schists, reaching in some places > 10 m in thickness in the hanging wall of the Black Mountain deposit, mentioned by Cawood et al. (this issue), do not preclude the possibility of oxidation at the time of the hiatus. Recent examples demonstrate that alteration, weathering and oxidation can easily reach several tens of meters (e.g. Stiermann and Healy 1985; White and Buss 2014; Jin et al. 2010). Even under today´s arid climate conditions at Aggeneys, oxidation of the sulfides can, in many places (Swartberg, Big Syncline, Gamsberg North), be observed to depths of > 25 m. Currently at Swartberg, where the hanging wall above the ore unit consists of relatively permeable schists, complete oxidation of the sulfides persists down to 25 m below surface, with partial oxidation down to 60 m below surface. At Gamsberg North, a similar trend is observed, with (in extreme cases) oxidation down to a depth of  > 100 m. Apart from a simple vertical oxidation profile, significant lateral changes in rock composition due to oxidation along water pathways at the time of oxidation, as documented, for example, at the oxidized Skorpion deposit in Namibia (Borg et al. 2003), are also possible. In this context, we would greatly appreciate any kind of documentation of the synsedimentary “channel-structures” within these strongly deformed rocks referred to by Cawood et al. (this issue) as we have never observed such structures. While the presence of the unconformity between the Hotson and the Koeris formations remains uncontested, some kind of chemical weathering along this surface has to be assumed, bearing in mind that this unconformity reflects a hiatus of c. 80 million years (Cornell et al. 2022). It would be indeed difficult to envisage that such a long period of exposure (even with > 10 m of overlaying metapeltic schists) would leave a massive sulfide body underneath unaffected.

Issue 2: Local source of sulfur

We fully agree that the lack of an unequivocally identified sulfur source for the suggested resulfidation of previously oxidized sulfide bodies in the course of the Klondikean orogeny is a weakness of our hypothesis, and we have admitted this repeatedly (Höhn et al. 2020, 2021). The only indication of the S source is indirectly given by the available S isotope data: The source should have had a high δ34S, which is more characteristic of sulfates than sulfides. A possible sulfatic source could have been former evaporite beds in the succession – the most likely candidate being the Koeris Formation. Identification of such evaporites is, however, hampered by the high metamorphic grade in the area. It requires the specific investigation of indicator minerals (Warren 2016), which has only been attempted recently. Neither the reference to Rozendaal (1975), which is an unpublished 47 year old master thesis, nor that to Colliston and Schoch (2003), who did not address the possible presence of former evaporites at all, are helpful in this regard. Nevertheless, we concede that so far no unequivocal evidence of meta-evaporites could be found, although this is a subject of current research by us.

With regard to the mineralogy of the supergene gossans, we have no reason to question the observations by Cawood et al. (this issue) but see little relevance of this for the discussion here. Naturally, any gossan mineralogy is a function of the physico-chemical conditions at the time and place of oxidation (e.g. Hitzman et al. 2003). As such, the presence or absence of sulfates becomes a function of local S-availability (see below). To assess the physico-chemical conditions on the mentioned unconformity and thus affecting the underlying sulfide bodies, it is worth looking at the climate at the time and the nature of the depositional setting for the lower Koeris Formation. A recent study on the geochemistry of the lower Koeris Formations (Höhn et al. 2022) indicates a small-scale, intra-montane basin at the very end of the Okiepian orogeny. In a small intracontinental basin, the ground water composition is mainly influenced by the intensity of silicate hydrolysis, atmospheric CO2-uptake, sulfates from the oxidation of sulfides, precipitation of neo-formed minerals and the chemical characteristics of the meteoric water (Rosen 1994). This multitude of factors increases the potential for fluid compositions deviating from the norm and explaining the minerals observed in the Aggeneys-Gamsberg ore district.

Regarding the S-availability, within a supergene environment, large amounts of Zn can bind a correspondingly large proportion of sulfur. Sulfuric acid generated by the oxidation of sulfides would immediately react with Zn to form ZnSO4 and thus retain the sulfur in the system, e.g. by the following reaction:

Zn + H2SO4 + 7 H2O → ZnSO4 • 7 H2O + H2.

Although entirely speculative, mineral phases such as goslarite, ZnSO4 · 7 H2O, zincmelanterite, (Zn,Cu,Fe)SO4·7 H2O, and gunningite, (Zn,Mn2+)SO4·H2O, might have formed if the mineralizing waters had been richer in sulfate than most supergene base metal deposits today. In short, the amount of sulfur lost due to oxidation could have been far less than suggested by Cawood et al. (this issue) but remains unquantified. If it was low, the problem of finding a suitable S source for the resulfidation during the Klondikean orogeny would fall away.

Issue 3: Sulfur-carrying capacity of metamorphic fluids

There is no doubt that fluids derived from metamorphism of average pelitic protoliths are typically not able to transport large amounts of base metals (for the lack of a high salinity) or sulfur (Barnes 1997). The reference given by Cawood et al. (this issue) in this context, that is, Zhong et al. (2015), is misleading. It refers to a gold-mineralized system of an entirely different setting, and no evidence exists of a similar metamorphic fluid composition in the Aggeneys-Gamsberg ore district. The metamorphic fluid composition in this ore district deviated, however, in all likelihood, from a typical, average crustal fluid composition as indicated by Zn- and S-isotope data from Gamsberg. If both elements came from the same source, the two isotope systems should be coupled. This is apparently not the case and different sources for the base metals and S are called for (Foulkes 2014). Unusually high base metal contents have been noted in the metasedimentary rocks and amphibolite of the Koeris Formation (Höhn et al. 2022), several tens of meters above the sulfidic ore of the Hotson Formation. The high Zn and Pb contents without a corresponding enrichment in Cu speak for metamorphic interaction of these rocks with the underlying Gamsberg base metal deposit at the time of the Klondikean orogeny. All of this clearly speaks against a fluid composition invoked by Cawood et al. (this issue) for the metamorphic overprint.

Issues 4/5: Spatial extent of the sulfidation

Nowhere in our papers on the Aggeneys-Gamsberg ore district (Höhn et al. 2020, 2021) did we infer or even hint at a complete oxidation of all deposits in this district. Instead, we suggested variable intensities of oxidation and sulfidation across the ore district. Thus, we see little point in addressing the points raised in the discussion pertaining to this specific issue.

The argument of a lack of structural pathways for channelized metamorphic fluids as presented by Cawood et al. (this issue) is invalid for several reasons. The Swartberg thrust as well as the “weak zone” at Broken Hill-Deeps are products of brittle deformation, which most probably took place towards the end of the Klondikean orogeny. The sulfidation, on the other hand, should have taken place on the prograde metamorphic path when devolatilization reactions led to pervasive fluid flow rather than the channeled flow typical of retrograde metamorphism. The observation that the most intensely sulfidized parts of the deposit are not spatially associated with the mentioned structural pathways is, thus, in agreement with our hypothesis.

In an area like the Aggeneys-Gamsberg ore district, which was affected by several deformation phases, timing is crucial. At the time of commencing Klondikean orogeny, the metasedimentary rocks of the Hotson Formation had already experienced medium- to high-grade metamorphism during the Okiepian orogeny. With the exception of the weathered uppermost parts of the Hotson Formation right below the unconformity, the permeability in the footwall of the unconformity should have been low. In contrast, the conglomerates and coarse-grained sediments of the Koeris Formation, still unmetamorphosed at that stage, should have had a high permeability and thus provided effective fluid pathways at least in the early, prograde stages of the Klondikean orogeny.

Therefore, it is possible that the unconformity, which had been shaped by erosion, peneplanation and oxidation, also provided the fluid pathway for a sulfidizing fluid during subsequent Klondikean metamorphism.

Furthermore, the mineral alabandite frequently occurs in and around the Gamsberg deposit as exsolution product of Mn-rich sphalerite and disseminated grains (Stalder 2004; Schouwstra et al. 2010). Its formation requires very high H2S activity, which can be achieved only in unusual environments (Cabral et al. 2019), and illustrates high H2S activity also beyond the ore bodies as inferred by Cawood et al. (this issue).

Cawood et al. (this issue) question the compatibility of our hypothesis with the findings by Hoffman (1993), who conducted his study at Broken Hill in the west of the ore district, where numerous oxides and silicates with a very high Fe/Mg ratio occur. This should speak against a high H2S activity. Apart from the fact that the Fe/Mg ratio depends not only on co-existing mineral phases but also on various other factors, such as bulk rock composition, fO2 and temperature (Hicks 1988), the stratigraphic position of the mineralized sulfide zone there is of critical importance. The ore bodies at Broken Hill are further below the unconformity than those in the eastern part of the district and, therefore, would have been less affected by the suggested oxidative pre-Klondikean weathering and, consequently, also less affected by subsequent re-sulfidation. In addition, it may be speculated that Fe (and base metals) in silicates are less prone to sulfidation than in other oxidized phases, such as oxides, carbonates or sulfates.

Issue 6: Biogenic/sedimentary S isotope signatures

We agree on the fact that sulfur provided by a metamorphic fluid should originally have quite uniform δ34S ratios, which would make them unsuitable to explain the pronounced sulfur isotopic gradient from the west to east of the Aggeneys-Gamsberg ore district. Unfortunately, the latter quote from Höhn et al. (2021) provided by Cawood et al. (this issue) is incomplete and taken out of context. The full quote should have been “the Gamsberg deposit probably experienced the most intense pre-Klondikean weathering-induced oxidation at the time of the hiatus (Höhn et al. 2020) and consequently the most intense resulfidation thereafter. In that case [italicized by the authors], the δ34S ratios of its sulfides should not describe the primary but rather the metamorphic sulfur source.” (Höhn et al. 2021, p. 725) The difference between the two quotes is essential as the quote given by Cawood et al. (this issue) conveys the impression that all four deposits of the Aggeneys-Gamsberg ore district were meant. This is not the case.

However, we discussed whether the strong δ34S-gradient observed within the Aggeneys-Gamsberg ore district could be the product of a spatially-controlled mixing of two different sulfur sources (Höhn et al. 2021): (i) one with sulfate-like high δ34S-ratios, atypical of Broken Hill-type deposits (Spry and Teale 2021), strongly marking the sulfur budget at Gamsberg where the unconformity is close to the ore; and (ii) one with lower δ34S-ratios dominating the western deposits (Black Mountain, Broken Hill) with > 10 m of shale between the ore and a possible unconformity.

Neither the broad internal range in δ34S ratios, nor their pronounced gradient or the positive mean δ34S of the deposits are similar to other Broken Hill-type deposits. For that reason, Spry and Teale (2021) excluded all deposits of the Aggeneys-Gamsberg ore district from their sulfur isotopic characterization of Broken Hill-type deposits elsewhere. The claim that the δ34S ratios observed at Aggeneys are just the product of “synsedimentary sulfide deposition” (Cawood at al. this issue) is, therefore, difficult to follow. Furthermore, the argument that these sulfur isotopic characteristics are a product of syn-sedimentary sulfide deposition in an east-west arranged cascade of sub-basins inevitably raises the question why this should only be the case at Aggeneys whereas the other Broken Hill-type deposits are explicitly characterized by narrow sulfur isotopic ranges centered around zero (Spry and Teale 2021). Obviously, the 34S characteristics of the ore district require a more complex explanation.

Issue 7: Mineral textures

Cawood et al. (this issue) argued that the mineralogical and textural observations described by Höhn et al. (2021) are rare and therefore no significant part of the big picture. We don´t see the point of this issue. As reported by Höhn et al. (2021), we performed 711 analyses on sulfides in and around garnet grains. We made it sufficiently clear that those observations and analyses were conducted on rock domains marked by low fluid-rock ratios (e.g. Frimmel et al. 1993; Stalder and Rozendaal 2005). “Although the long-term high-temperature history of the ore district would make complete diffusional homogenization of garnet likely, the observed zonation with respect to Fe points at very low fluid/rock ratios, which markedly decrease diffusion rates.” (Höhn et al. 2021: p. 719). The statement by Cawood et al. (this issue) that those garnets are “one of the exceptions rather than the rule” is not helpful at all because we never questioned the validity of previous studies on garnet from the Aggeneys-Gamsberg ore district but just added a hitherto unmentioned observation that has interesting bearings on the mineralization history.

The statement by Cawood et al. (this issue) that the larger inclusions (> 25 μm) occur in both the core and the rim and the reference to Höhn et al. (2021) are incorrect. The core zone is free of those large inclusions. Only at the contact between those two zones, large inclusions can be seen in the mentioned figure (Fig. 4 in Höhn et al. 2021) and all other investigated garnet grains. However, the slight overlap of mineral grains between two zones is inevitable and therefore common because a mineral grain is a three-dimensional object whereas a microphotograph is only two-dimensional.

Garnet is well-known for its tendency to develop euhedral to subhedral grains. The fact that all observed core zones display highly irregular shapes speaks in favor of diffusional zoning. On top of that, the observed zones also show a spatial association with late, most probably retrograde, metamorphic fractures filled with matrix minerals, which clearly demonstrates that these zones are not the product of crystal growth but the product of a process that is younger than the fracture.

Strictly speaking, the argument that all textures identified by Höhn et al. (2021) could have formed during prograde, peak and retrograde metamorphism is mostly correct. The presence of those textures within today´s Gamsberg deposits is, however, not only dependent on the possibility of their formation but mainly on the probability of their preservation. For example, it is effectively impossible that chlorite and muscovite associated with the vein-type sulfides are a product of prograde or amphibolite-facies peak metamorphic conditions. Other, more complex sulfide textures like sulfidic melts, which would be an alternative to the graphic-textured sulfide aggregates described by (Höhn et al. 2021), would re-equilibrate continuously down to very low temperatures and are usually absent for that reason (Frost et al. 2002).

Issue 8: Cu isotopes

We agree that Cu isotope data on SEDEX deposits are still rare but strongly warn against comparison with data from other genetically unrelated deposit types or laboratory studies conducted under entirely different conditions as listed by Cawood et al. (this issue). The only actually existing comparable data on chalcopyriteof sedimentary exhalative (SEDEX) deposits are from the Rammelsberg deposit (Germany), the type locality of SEDEX deposits, and indicate an even narrower δ65Cu range between − 1.03 and + 0.05 with a mean value of -0.54‰ (n = 7; Höhn et al. 2019). These values are distinct from the δ65Cu ratios obtained for the Broken Hill and Black Mountain deposits. Maybe a growing set of data will shed new light on this issue in the coming years, but currently, the data obtained on the Aggeneys-Gamsberg ore district cannot be explained by a pure SEDEX-model.

Issue 9: Aspects explained by a metamorphosed SEDEX model

The statement by Cawood et al. (this issue) that the observed high mineralogical variability can be explained simply by metamorphic overprint of a SEDEX deposit is not substantiated and lacks conclusive examples and references. Of interest in this context is the recent attempt to explain the mineralogical variability by Sangster (2020b), who ascribed the complex and varied nature of minerals and rock types in and around Broken Hill-type deposits to “compositionally unusual SEDEX-type chemical sediments” (Sangster 2020b, p. 1266). This, of course, avoids the more pertinent question as to why only the “unusual” ones seem to have experienced middle- to high-grade metamorphic overprints and thus became Broken Hill-type deposits.

The metal sources suggested by Cawood et al. (this issue) in order to explain the pronounced metal zonation in the ore district mostly refer to studies on VMS deposits. The only actual study of this kind on SEDEX deposits is by Sangster (2020a). The latter comes to the conclusion that lateral fluid flow in vent-distal SEDEX deposits may cause an enrichment in Zn compared to Cu and Pb. This is conclusive but cannot be harmonized with the idea of various sub-basins with strongly changing physico-chemical conditions – a notion that has been utilized to explain the gradient in sulfur isotope ratios, the presence and absence of barite and various other mineralogical anomalies within the Aggeneys-Gamsberg ore district (v. Gehlen et al. 1983; McClung et al. 2007; Stalder 2004).

Conclusion

A complete oxidation and subsequent resulfidation of all four deposits in the Aggeneys-Gamsberg ore district was neither suggested nor was such a suggestion intended by Höhn et al. (2020, 2021). Nevertheless, the stratigraphic situation with distances of several meters to an unconformity representing a hiatus of c. 80 My (Cornell et al. 2022)makes at least partial oxidation of the original base metal sulfide bodies in the uppermost Hotson Formation almost a geological necessity. We are fully aware that our suggested hypothesis leaves room for a number of open questions but re-emphasize that it is the currently best explanation for the wealth of observations and data on this ore district. We are confident that growing datasets on Cu isotopic characteristics of SEDEX and Broken Hill-type deposits and geochemical follow-up studies will make possible a more detailed assessment and characterization of the pre-Klondikean oxidation event at Aggeneys.